C.J.S.

PURDY
ON
THE ENDGAME

compiled and edited by

Ralph J. Tykod.i

with editorial assistance from

Frank P. Hutchings
GM Karsten Muller
John Purdy

Foreword by

GM Karsten Miiller

Thinkers' Press, Inc.

Davenport, IA
2003
 CJS Purdy On The Endgame

Copyright© 2003 by Nancy (Anne) Purdy and Frank P. Hutchings .

All rights reserved. No part of this work may be reproduced or transmitted

in any form or by any means, electronic or mechanical, including photocopy­
ing and recording, or by any information storage or retrieval system, except
as may be expressly permitted by the 1976 Copyright Act or in writing from
the publisher. These reserved rights also pertain to e-Books and similar
derivatives and the Internet.

January 2003

ISBN: 1-888710-03-9

Part of the C.J.S Purdy Library

compiled exclusively for
Thinkers' Press, Inc.
by
Dr. Ralph J. Tykodi

Requests for permissions and republication rights should be addressed in

writing to:

Thinkers' Press Inc.

Editor, Bob Long
P.O. Box 3037
Davenport IA 52808-3037 USA
e-mail: blong@thinkerspress.com

- ii
-
 OS Purdy On The Endgame

ACKNOWLEDGEMENTS

The publisher and the editor are indebted and grateful to Mrs. Anne Purdy,
owner of the copyright, for permission to reproduce from Australasian
Chess Review, Check, and Chess World the material presented here; and to
Frank Hutchings (F.P.H.) and John Purdy (J.S.P.) for their comments on,
suggestions for, and corrections to the working manuscript-Frank's work
on the MS was so extensive that in effect he served as an associate editor for
the book. See his SUMMING UP on page xi. We would also like to acknowl­
edge the valuable contribution by GM Karsten Muller for taking his time to
look over the manuscript and make pertinent additions and corrections.

C.J.S. PURDY and ENDGAME LORE

Cecil John Seddon (C.J.S.) Purdy (1906-1979) was four times Champion of
Australia, an International Master, and the first World Correspondence Chess
Champion. He published and edited what was arguably the premier chess
journal of its day: Australasian Chess Review (1929-1944), Check (1944-
1945), Chess World (1946-1967)-a continuing journal with an occasional
name change. Purdy's game annotations and his writings on all aspects of
chess have been widely praised. For some other books devoted to Purdy's
chess writings, see the items listed in The Purdy Library of Chess at the
end of this book.
C.J.S. Purdy had a deep and abiding interest in the endgame, covering
both the practical side and composed studies. Purdy remarked in an issue of
Chess World that he hoped to do a book on the endgame sometime. He never
got around to doing the book, but he did leave several structured series of
articles and a host of individual studies relating to the endgame scattered
throughout the pages of Australasian Chess Review, Check, and Chess
World. Collected here are most of Purdy's writings on the endgame (to­
gether with a few complementary articles by M.E. Goldstein and an article
by A.C. Harris). (There are also interesting and instructive remarks about
endgame themes in the notes to many of the games collected in the series of
books CJS Purdy's Fine Art of Chess Annotation and Other Thoughts­
see the description of The Purdy Library of Chess. )
The material for this book has been extracted from Purdy's monthly
magazines. It should be remembered that Purdy was subject to the same
time restrictions that confront all busy magazine editors; and in chess writ-

- iii
-
 CJS Purdy On The Endgame

ing under such conditions, it is impossible to avoid entirely, mistakes in

analysis. For this book, all the articles have been carefully reviewed; and
such errors as were found have been corrected. Our apologies to the reader
for any mistakes we have overlooked.
To aid the student in the systematic study of the endgame, we (Nathan
Long, Bob Long, Ralph Tykodi) have arranged the endgame material in the
following way: we have divided the material into two parts: How to Play
Endgames-The Short Course and All Facets of the Endgame. During the
years 1932-33, Purdy published in Australasian Chess Review detailed
discussions of the most commonly occurring endgames: Rook and pawn vs.
Rook endings, pawn endings, and Queen endings. In Part N of Guide to
Good Chess (first ed. 1950), Purdy listed and discussed such general rules
for playing endgames as he had found, fifteen in all. In the interval 1929-45,
Purdy published in his magazine a number of "practical" endgame studies,
mostly from actual play, in many of which players missed the right road,
resigning games that could have been drawn or conceding draws in games
that could have been won.
The Short Course consists of the material in Part N of Guide to Good
Chess, the detailed discussion in the 1932-33 articles of the most frequently
occurring endgames, and the "practical" endgame studies of 1929-45.
The material in All Facets of the Endgame deals, by and large, with such
specific endgame themes as caught Purdy's attention in the material that
passed over his desk as he edited his magazine. We have arranged the
material according to themes or topics. After each of your games that had an
endgame, we suggest that you go over all the articles in the book that have
relevance to the kind of endgame you played. In this way the amount of time
that you devote to the study of specific endgame themes or topics will be in
proportion to their occurrence in your own games.
Maurice E. Goldstein (1901-1966) was born in Eng1and and settled in
Australia in 1929; he was an Australian Master and was a co-editor of the
4th (1925) and 5th (1932) editions of Modern Chess Openings.

N.B.
Italicized material between square brackets is editorial commentary.

- iv-
 OS Purdy On The Endgame

CONTENTS
Acknowledgements ............................................................................................ iii
C.J.S. Purdy and Endgame Lore . .............................................. ...................... iii
Contents ............................................................................................................... v
Fbreword (GM Karsten Miiller) ................................... .............................. ....... ix
A Fascinating Battle Between Rook and Bishop (GM Karsten Milller) ........ x
Summing Up ....................................................................................................... xi

PART I. HOW TO PLAY ENDGAMES-The Short Course

Introductory ......... .................... ............... .......... ............... ........... ..... ............. 13
....

Chapter 1. Rook and Pawn vs. Rook ....... .... ...... .. .. .. . ... .. ..... .......... . 14
. .... .. . . . .

Type 1. The Defending King Commands the Queening Square 14 .............

Type 2. The Defending King Is Shut Out by the Opposing Rook 16 ............

Type 3. The Defending King Is Shut Out by One File

on the Wide Side 19
........................................... ..........................................

Type 4. The Defending King Is Shut Out by More than One File 22 ...........

Type 5. The Defending King is Shut Out by One or More Files 24 ..............

Type 6. The Kings Are Distant ............................. ...................................... 26

Chapter 2. Pawn Endings ...... .. ... ... .... ... . .. ... .... . ..... .. ... . ... . .... .. .. .. 30
. . . . . . . . .. . . . . . . .

The Square .
................ . .
............................................................. ................ ... 30
The Opposition ............................................................................................. 31
King and Pawn vs. King ........................................ ...................................... 31
The Capture o f Blocked Pawns .................................................................. 34
The "Trebuchet" .................................... ...................................................... 35
The Rule of the Rank .................................................... .............................. 35
Mutually Supporting Pawns .......... ............................................................. 36
Problems in Pawn Play ................................................. . ............................. 37
Fbrcing a Passed Pawn . ............ .
....................................... .......................... 37
Kill or Be Killed ............................................................................................ 38
The Distant Pawn ........................................................................................ 40
Two Pawns vs. One Pawn ......... .................................................................. 42
Two United Pawns vs. One Pawn ............................................................... 44
A Study in Doubled Pawns from the Australian Championship ............. 46

Chapter 3. Queen and Pawn Endings ........ ... .. .... .. ... . . . .... . . .. ..... .. 48
. .. . .. . . . ... ..

Queen vs. Pawn About to Queen ......................... . ......................... . ............ 48

Queen vs. an a- or h-Pawn .............................................. . ............ . .... .......... 49
Queen vs. the c- or f-Pawn .................... . ........................................ ............. 49
A Queen and Knight Ending . ........ .................................................... . ......... 50

-v-
 OS Purdy On The Endgame

Chapter 4. Practical Endgame Studies . . . ... . .. . . . . .... ...... ..... . . ...... ....... 52
... . .. . . .

"Sister Squares" ..................................................... . .................................... 59

PART II. ALL FACETS OF THE ENDGAME

Chapter 5. Road To The Endgame . . ..... ........ .. ....... . ... . .... .. . .. ...... 68
.. .. . . ... . . . . . ..

No. 1: The Pawn Center (Bogoljuboff-Reti, 1923) .................................. 68

No. 2: (Book-Petroff, 1937) ......................................................................... 71
No. 3: The Two Bishops (Stfiltlberg-Michel, 1942) ................................ 73
No. 4: Pawn Endings (Cohn-Rubinstein, 1909) ...................................... 75
No. 5: Rook Ending (Lasker-Levenfisch, 1925) ...................................... 78
No. 6: How to Wm with a Pawn More (Reshevsky-Rellstab, 1937) ...... 79
No. 7: Knight Versus Pawns (Hanks-Goldstein, NSW) . ................ ... ...... 80

Chapter 6. Endg8.Dle Principles ..... .. . . .. . . . . . .... . .. .. .. ... . . ... ...... ..... ......... 82
.. . . . . . . .

How To Play Balanced Endgame Positions . . . .... ......... ..................... ......... 82

Oh We Of Too Much Faith! . .. .
................................ ............................ .......... 84
Botvinnik on Errors . ..
............................ .
........... . ................................ .... 85
A Famous Ending .................................................................................... 85
Reconnaissance .
....................................................... ............... ............... 86
What Happened? . . ............ .. .... .
............... .
...................... ....................... .. 87
The Might-Have-Been ........................................................................... 89
The Correct Defense .
............................. ......................................... ....... 89
Fine's Diagram, Steiner's Analysis . ............. ........................................ 89
Znosko-Borovsky's Version . . .
..... ..
....................... . ... ............... . ............ .. 91
Summary of Analysis . ... .
....................................................................... . 91
We Celebrate the Day! ............................................................................ 91
The Mystery Endgame . .
...................... . .
......................... .................... ......... 91
That Teichmann-Blackburne Ending . .. .............. ...... . ............................... 94
An Overriding Principle Of Endgame Play . .
....................................... ...... 95
A Guiding Rule For Endgames ................................................................... 96

Chapter 7. Pawn Endings .. . .. ... . .... . .... ...

... .... . .... . . ....... ....... . ... . ..... . .... 99
. . . .. . . . . .

How To Make A Pawn Tell .......................................................................... 99

General Rules for Endgames with Pieces and Pawns 99 ........................... . .

Pawn Promotion . .... .

.................. . ................ 102
................................................

A Classic Pawn Ending ............................................................................. 103

The Zugzwanger Zugzwanged . . .
..................................... 105 ..... ........... ........

Three Pawn-Endings From The Melbourne Open . 106 ....... .........................

Pawn Ending .............................................................................................. 109

Chapter 8. Rook Endings ..... ... .. .... ..... ...... ... ....... ..... . .. . . ... ...... 111
. . ... . . . . . .. . ... . ..

The Rook In Endgames . . . ..

............... . . . . . . 111
....... . ...........................................

Rooks and Passed Pawns ......................................................................... 112

Tarrasch's Rule Amplified . .. .. . .
........... . .
. . ...................... .... ..................... ... 113
Bridge Building . .
...................................................... . . ...... ....... ................... 115
The Commonest Rook Ending .................................................................. 116
a- or h-Pawn With Its Own Rook In Front Of It . .................... ................. 116
Rook Endgames . . . ....... .
.. .... .............. ..
......................................... . .............. 119
No. 1 ....................................................................................................... 119

-vi-
 OS Purdy On The Endgame

No. 2 120
.......................................................................................................

No. 3 ....................................................................................................... 122

A Paradoxical Rook Ending . . . . . .
.......................... . 123 . .. .... .................. . ....... ....

Chapter 9. Knight Endings . ........ .................. ....................... ........ .... ........ 125
An Instructive Knight Ending ............ . . .. ..... . ........ . . .... .............................. 125
Knight and Pawn vs. Knight ................................ . . ... ................................ 128
Strategy In Knight Endings . .. ............. . ....................................... .. . . ........ . 130
.

Chapter 10. Bishop Endings . .. ............. ......... .. . ...... ..................... . ...... .. .... 133
Bishops Of Opposite Color ................................................................. . ... . .. 133
Bishops Of Opposite Color, With Rooks .... . .... . . . ............................ . ......... 136
Two Pawns Up And No W in . . ....... .......... ...... . ..
.. . ..... . ... . ........... . ................. 138
That "Bad" Bishop! . ........................................... . .............. . ....................... 140

Chapter 11. Bishops vs. Knights ..... ............ .. ........ ....... ... .. . ..... .............. . 142
Bishop vs. Knight . . . .... ... ....... . 142
............................................................... ......

Two Minor Piece Endgames From Maroubra 1952 ............................... 143

Chapter 12. Rook vs. Bishop/Knight ....... .. .. .......... ... ....... ......... ...... ..... ... 146
Endgame Value Of "The Exchange" . . . . ................ ............... .......... ...... ..... 146
Midgame Value Of "The Exchange" . . ................... ............... ..................... 151
Error InBCE .............................................................................................. 153
Purdy's Note . . .
............................ . . .
............. .... ...... ......... ........... ............. 155
I>evilish Rook . . . ...... . . . .
..................................... . . ....... .... ....... ....... ........... ..... 156

Chapter 13. Minor Pieces With Rooks . .. ........ ......... ........... ....... ............. 158
Endgame Strategy . . .. ........... ................. . ........ . .. .. . ...... . ............................... 158
A Finely Balanced Position ......... . ... . . . .. . ....... .. . ..... .............. . ...................... 161
Miracle! . . . . .
.... ............... .. ........................ . ............. . ................... . ............ .. . ... 161

Chapter 14. Queen Endings ... ...... .......... ....... ...................... .... .......... ..... .. 166
The Overrated Queen . .................. ............... . ........... . ............................ . .... 166
Two Bishops vs. Queen . .................. .................... . . .. ....... . ..................... . .... 168
Queen Endings .. . ......... . .. ........................ . ................. . ..... . ........................... 169
The Question Is Answered! . . ................. .. . ................... . ........... . ........... 172
Queen And b- Or g-Pawn vs. Queen ..... . . .
... ... ..... . ..... ..
. ...... ... . . ................ . 173
.

Chapter 15. Heavy Artillery ........ . .. ... .... .............. .. .. ... .. ......... .. ..... ....... ... .. 175
Queen vs. Rook . . ..... .. ..... . ...................... . ...... ............ . ....................... . ......... 175
Change Of Fortune . .. .. .. .... . ..................... .. .... .. ...... .. .. .. ......... . . .
... ... ..... . .... . ... 178

Chapter 16. Sure Draws (?) ....... ....... ..... ...... .. . .... .... ................... .. . ......... .. 180
Two Kelling Endings .. .
............. .
. .
............................. .................... ......... .... 180
A Missed Draw . . . . .. .. ... .. .... .
.......... ... ...................................... ...................... 182
Winning A "Drawn Game" . .. ......... .................................. ...... .................... 182
I>ead Draw! .. ...... . . .
. . . .
...................... .... .. ... ....................... ..... ..... ................. 184
Ragtime Band .. . ... ........ . . . ...
........ . . . .............. ....... ... . ... ...... ... ........... ............. 185
Attack Plus Three Pawns. No Wm! .......................................................... 186

- vii -
 OS Purdy On The Endgame

The Might-Have-Been ................................................ . ....................... . ..... 188

Is Endgame Theory Faulty? ................................................................. 188
Resign? ....................................................................................................... 190
The Critic On The Hearth ......................................................................... 191
A King Escapes . .................. ........ ................. . ............................................. 193
Endgame Suicide ....................................................................................... 194

Chapter 17. Find The Wm . . . .. ... .. ....... .. . ....... ... . . .. .. . ... . .. ... 195
. .. . . . . . . .. . . . .. .. . . . . .. .

Onlookers See Least Of The Grune . .... . ............................................. ....... 195

Wlillling A Drawn Endgame . ............... . ........................................... ......... 195
City Of Sydney Tourney .
............................................. .............................. 196
Premature Resignation .
.................................... ........................................ 197
Did You Try Diagrrun 5? ............................................................................ 198

Chapter 18. Interesting Endgames . .... ....... .. . .. ..... . .. .. ............... 201

. . . ... .. . . . . ..

The Goldstein-Crowl Ending . . ................................................................. 201

New Zealand Championship ................................. . ............................... .. 203 .

First Endgame .
............................ ........................................................ . 203
Second Endgame . ..... ............................. . . .
.. .. ............. .. . ........................ 205
Endgame .
............................. ....................................................................... 206
An Amazing Game . ......... ...................................................... ..................... 207
A Fighting Ending .
................................ ................................ . . . . . . . . ............. 212
Two Curious Endgames . . ............ ... .......... . ......... . ...................................... 214
First Endgame ...................................................................................... 214
Second Endgame . ......................... ..................... . ...................... ............ 215
A Lively Endgame ............... . ......................................................... . ......... . 217
..

Piquant Endgame .................................................... . ...................... ........... 218

The Confident Kibitzer . ........ ........................... .. . ........ . ...... . ...................... . 220

Chapter 19. Endgame Wizards . . . . . . ...... .. . ... . .. .. .. ... . . ...... . . .. . .. . . ...... .. . 223
. . . . . . . . .

Rubinstein's Black Magic . ... .

............................................. ....................... 223
Pillsbury Crashes Home . . .
.................... ......... ................. ..... . . . .................. 226
We're Off To Beat The Wizard .................................................................. 239
Two Watson Endgames . ........................ .............. ...................................... 230
First Endgame (Watson---Crakanthorp, 1922) ................................. 230
Second Endgame (Reti-Watson, 1922) ............................................ 231
Endzelins Again .
........................ .
........................................................ ....... 232
Grune Of A Lifetime .
............................................................................... ... 233

End of the Game . ...... . ...... . . . ... . ... .. ....... .....

. . . . . ... . .. .. . .. . . .... ...... . .. ...... . .. 237
... . . .. . ..

Rules for Playing Endgames . . ...................... ..... ....................................... 238

Explanations of Files, Ranks, Symbols, and the Queening Square ... . 243
.

Bibliography ............................................................................................... 244

The Thinkers' Press, Inc. Purdy Library Of Chess . .................... ........... 245
Colophon . .
................................ . .................. ..................... ........................... 247
SQUARES magazine . ....... ......................................................................... 248
Thinkers' Press Brief Catalog .................................................................. 249

-viii-
 CJS Purdy On The Endgame

FOREWORD
GM Karsten Muller

WWouldn't it just be outdated and only of interest for chess histori­

hy should anyone read a book on endings by C.J.S. Purdy?

ans? My answer is very simple: Not at all!

Endgame theory is developing much slower than opening theory and
so much of what Purdy had to say is still valid. In a few cases, where
later analysts or the computer has proven him wrong, this is pointed
out as a note to his analysis so that the reader can follow the search
for the truth directly in its historical context.
This collection of Purdy's articles is not a full encyclopedia of the
endgame, but the most important endings are covered in detail and
many rules are given to uncover the principles governing subsequent
play. The treatment is very instructive, particularly because his articles
on Rook endings will repay good dividends.
I am convinced that readers will enjoy Purdy's logical writing style
and the way he explores the mysteries of the endgame.

Karsten Muller
Hamburg
December 2002

- ix-
 CJS Purdy On The Endgame

A FASCINATING BATTLE BETWEEN

ROOK AND BISHOP
by Karsten Miiller
The following position (pg. 153ff) Kd7 Kb6 16. Rh5 Bd2 (16 ... Kc5 17.
derived from one of Purdy's games d6t+-) 17. Rh2 Bf4 18. Rc2 Be5 19.
is an original contribution to end- Kc8 Bf4 20. Rc6t Kb5 21. Rxc7+-.
game theory. Black (to move) can 5 ... Bg3 6. Re2 Be5 7. Kc5 Bd6 t 8.
draw by the skin of his teeth: Kc6 Kf4 9. Kd7 Kf5! (9 ... Be5? 10.
Ke6 Bd6 (10 ... Bc3 11. Rc2 Ba5 12.
Rc5 Bb6 13. Rc4t Ke3 14. Kd7+-)
11. Re1! Kf3 12. Kf5 Kf2 13. Re6
Bg3 14. Rc6+-) 10. Re1 Be5 11.
Rf1 t Ke4 12. Ke6 Bd6 13. Rd1 Bc5
14. Rh1

1 ... Bc5 2. Re8 Bd6 3. Ke3 Bg3 4.

Kd3. After 4. Rh8 Black should
avoid 4 ... Ke5?? as 5. Kf3! surpris­
ingly traps the black Bishop! 4...
Bh4 5. Kc4 5. Rh8 with the idea to
drive Black's King back is danger­
ous as well: 5... Be7?! (5... Bg3! is
= 14. . . Bd6 staying on the c7-h2 di­
safer) 6. Rh5t Kg6? 7. Re5 Kf6 (7... agonal is crucial as 14 ... Bb6? 15.
Kf7 8. Ke4 Ke8 9. Rh5 Kd7 10. Rh7 Rh4t Kd3 16. Kd7 Kc3 17. Rh6 Ba5
Kd6 11. Rh6t Kd7 12. Ke5+-) 18. Ra6+- shows. 15. Rh4t Kd3
8.Re6t Kf7 9. Kd4 Ke8 10. Ke5 Kd7 16. Kf5 Bc5 and it seems indeed,
11. Rh6 Bb4 (11... Bd6t 12. Rxd6t that White can't break through.
cxd6t 13.Kf6+-) 12. Rh7t Kc8 13. Purdy ends his analysis with:
Ke6: "Perhaps someone will test it prop­
A) 13 ... Bd6 14. Rh8t Kb7 15. erly now, thus contributing some­
Kd7 Kb6 (15... Be5 16. Rh5 Bg7 17. thing definitive to endgame the­
Rg5 Bd4 18. Rg4 Bc5 19. Rg3+-) ory."
16. Rh6 Kc5 17. Rxd6 cxd6 18. I disagree, this position is one
Ke6+-. of his original contributions to
B) 13 ... Ba5 14. Rh8t Kb7 15. endgame theory.

- x
-
 CJS Purdy On The Endgame

SUMMING UP
by Frank Hutchings

"l l Tbile Cecil Purdy's output of chess books was modest in number, his
V V tally of magazine articles was perhaps unequaled. Purdy's articles
cover all aspects of chess, with a good proportion on endings, though he
never produced a book dedicated to the endgame-had he done so, his
endgame articles would have been an excellent source of material.
Purdy frequently referred to comprehensive volumes of endgame theory;
mainly Reuben Fine's Basic Chess Endings; but he would not have tried to
emulate such a work. Rather, Purdy would have aimed, as he did in his
magazine, to instruct the less-experienced player-by presenting basic
rules and principles, combined with an appropriate measure of formal
theory. While containing first-rate instruction, a Purdy endgame book would
also have captured and maintained the reader's interest, as did his ar­
ticles-a feat difficult to achieve in a book intended mainly for reference.
Most of Purdy's endgame articles were based on his own thoughts,
research, and analysis. The principles and rules he expounded were often
original and of great practical value. Purdy selected illustrative examples
from games of the time, from existing chess literature, and from composed
studies. His examples were often a little offbeat, but they illustrated par­
ticularly instructive points; and they were frequently entertaining, or made
entertaining by Purdy's way of writing. His engaging style was an impor­
tant ingredient in his successful teaching formula.
Purdy emphasized that, for a chess student, time spent studying the
endgame was of more value than time spent on openings. Excellent open­
ing play, said Purdy, usually yields only a small advantage, which can be
easily thrown away by a small inaccuracy. Excellent endgame play, how­
ever, often yields extra half-points, or even full points, and has a much
more direct influence on the results of many games.
This holds true even in master play, though the relative importance of
the openings increases as the standard of chess rises. For players of all
strengths, however, the recent trend towards completing tournament games
in one playing session has emphasized the importance of endgame skill. If
there is no adjournment, the ending must be played unaided: no reference
books, computer data bases, or human assistants may be consulted.
Purdy often pointed out analytical errors in published endgames-not
implying criticism of the authors, but highlighting the difficulty of accurate
analysis in endgames. He realized that, as an author, he was as much sub-

- xi-
 CJS Purdy On The Endgame

ject to error as any, and would always acknowledge his own mistakes,
whether discovered by himself or by his readers. Purdy's instructional tech­
nique included drawing attention to the difficulties of endgame play, en­
couraging a critical attitude, and inspiring the reader to seek things out for
him/herself, rather than to accept an author's words and analysis without
question.
Since Purdy's time, endgame theory has made further progress. Recent
research has led to the reappraisal of some endgame positions, and com­
puters have been used to definitively analyze certain endings with very few
pieces on the board. Such investigations will continue, and inevitably er­
rors will be discovered in published analysis, including Purdy's. However,
as mentioned above, Purdy would regard that as normal. Such errors would
not detract from the instructive value of his writings or the soundness of
his advice.
The spirit of Purdy's endgame writing is, I believe, well captured in the
selection and presentation of the material in this volume. For me it has
been a privilege to check through the manuscript and, in collaboration with
John Purdy, to make a few editorial comments, which we hope will be found
useful. There is no doubt that all readers, including experienced players,
could hardly fail to pick up many valuable points from Purdy's endgame
writings. A thorough study of the material should instill sound principles
and should equip a player to conduct endgames with confidence, clarity of
thought, and success.
One point Karsten (Millier) made when he wrote to me was that he
thought Cecil's book would be a good complement to his (and Frank
Lamprecht's) own Fundamental Chess Endings. I think he's absolutely
right, especially for less experienced players, who would do well to have
Cecil's book plus a comprehensive reference work in their libraries.

- xii-
 The Short Course: Rook and Pawn vs Rook

PART I

HOW TO PLAY ENDGAMES

The Short Course

Introductory
As mentioned in the editorial introduction to this book on page iv, The
Short Course consists of the material in Part N of Purdy's Guide to Good
Chess and the material to be presented here in Chapters 1-4. Anyone who
studies the writings of C.J.S. Purdy should have access to his Guide to Good
Chess, preferably as a title in the student's personal library. Purdy continu­
ally refers to the Guide in his other writings, and he points out that there
are things in the Guide that you will find nowhere else in chess literature.
In that solid book are 15 italicized statements for fifteen general rules of
the endgame. See pages 137-140 of that book.

- 13 -
 CJS Purdy On The Endgame

Chapter I
Rook and Pa\\ln \IS. Rook

TYPE 1 onto that rank. Then, but not be­

The Defending King Com­ fore, immediatelyplace your Rook
mands the Queening Square on the attacker's back rank and
check forever (Cheron).

RULE1
The point is that when the pawn
-¢>
is on the rank next-but-one to the
T
he defense draws against
queening rank the attacker's King
any pawn that is not more
cannot shelter in front of the pawn.
than one rank beyond mid­
board (a white pawn on rank
5, e.g.) when the defender's
King stands on the queening
square and the defender's
Rook is on the rank next-but­
one to the queening rank (a
defending black Rook on rank
6, e.g.---Cheron). If the pawn is
on the rank next to or next­
but-one to the queening rank,
Either to move/Draw
the defender draws, provided
(Phllldor 1777)
he can safely give check-to do
that the defender should try to
Following our method, we have
get his Rook behind the pawn.
1. ... Ra6!
-¢>
2. e5 Rb6
3. Ra7 Rc6
4. e6 Rc1!
Method of drawing. The simplest 5. Kf6 Rf1t
rule is: Keep your Rook on the rank and Black draws. If the black Rook
next-but-one to the queening rank, could not check now, White would
preventing the advance of the en­ win.
emy King, until the pawn rrwves

-14-
 The Short Course: Rook and Pawn vs Rook

have been able to check again, and

the white King would have had to re­
2 turn to the defense of his pawn. This
observation shows that Black could
draw this way against a pawn on file
a, b, c, f, g, or h.
Continue now with Cheron's line:
2. Ke6 Kf8
3. RaSt Kg7
4. Kd6 Kf7!
Either to move/Draw And Black draws. The Rook and King
together prevent the advance of the
No. 2 is a position where, with pawn. If 4. Re8 (instead of 4. Kd6),
White to move (1. Kf6!), Black can­ then Black draws by 4... Ra1, threat­
not draw by the method of playing ening the "Rook on the rank next­
his Rook to the rank next-but-one but-one to the queening rank" busi­
to the queening rank; for if 1. Kf6! ness. Or if 4. Ra5 (say), then the black
Rb6t, then White plays 2. e6. Black King simply goes back to f8, getting
then has no time to get back for as near the queening square as pos­
checking because White's King is so sible.
far forward that White threatens Note that if in No. 2 the black Rook
Ra8t, leading to mate. were on the same rank as the white
Black can draw, however, as fol­ pawn the Rook would not be able to
lows: get behind the pawn and White, with
1. Kf6 the move, would win by 1. Kf6.
Of course, if 1. Ke6?, then 1... Rb6t; This refutes the idea of many
and Black draws as in No. 1. players that the command of the
1. ... Re1!! queening square by the defender's
(Cheron). So noted an authority King is alone suffic ient to draw.
as Berger gave 1... Rf1 t which loses In the by-play to No. 2 we saw that
against a center pawn, although it it made all the difference whether the
would draw against any other. Watch defending Rook had three files be­
carefully: 1... Rf1t? 2. Ke6 Kf8! The tween itself and the pawn or only two.
defender should always play his King That happens over and over again in
to the "narrow" side of the pawn; then Rook endings. Hence,
the attacking King must go to the
"wide" side where the defending " ... the black
Rook can check at a greater distance.
Remember: the Rook gains in King... gets as
strength with distance. 3. Ra8t Kg7 near to the
4. Ke7 Rf7t 5. Kd6 Rb7 6. e6 Rb6t 7.
Kd7 Rb7t 8. Kc6---and White wins. queening square
Had the black Rook been able to as possible."
get one file farther away, Black would
have drawn easily; for then he would

- 15
-
 OS Purdy On The Endgame

position owing to the attacking King

RULE2
having so little liberty on the a- or
�
h-file where he must go to aid his
T strongest when it has a
he defending Rook is
pawn.
Of course, in all this we assume
clear interval of three or more
that the attacking King is able to sup­
files or ranks between itself
port his pawn.
and the opposing pawn. In fu­
We shall first state the rules for
ture considerations, Rules 1
this type of ending and then proceed
and 2 will be used m a n y
to illustrate them.
times-they should be ab­
sorbed thoroughly.
Any Pawn Other Than
�
The a- or h-Pawn
a) One file between the pawn and
the defending King.
TYPE2 i) If the pawn is past the mid­
The Defending King Is Shut dle of the board, the attacker
Out By The Opposing Rook wins.
ii) If the pawn is not past the
We saw above that if the defen­ middle of the board, the de­
der's King could command the fender draws.
queening square then the defender ill) Important exception: If
could always draw, except in ex­ the defending King is on the
tremely rare cases. As this applies "narrow" side of the c-, d-, e-,
to any pawn whatever, it is the most or f-pawn and is at least as
important thing to know about Rook close to the queening rank as
endings. the pawn, then the defender,
The attacker, therefore, will natu­ with the move, regularly draws.
rally seek to shut out the defending
King. That can only be done with the b) More than one file between the
Rook. The attacker's first thought pawn and the defending King.
should be, Can I shut out his King i) If the pawn is past the
with my Rook, and what is the far­ middle of the board, the at­
thest out that I can shut him? tacker wins.
Whether the attacker can then win ii) If the pawn is not past the
depends on: middle of the board, the at­
1) how far advanced his pawn is, tacker wins always if the num­
2) by how many files the defend­ ber of ranks the pawn has ad­
ing King is shut out, and vanced from its home rank (0,
3) what kind of pawn the attacker 1, or 2) plus the number of files
has. between the pawn and the de­
For winning purposes, c-, d-, e-, fending King exceeds 3 in the
and f-pawns are best; then come the case of the c-, d-, e-, or f-pawn
b- and g-pawns, whereas a- and or exceeds 4 in the case of the
h-pawns seldom win in this type of b- or g-pawn. The defender can

-16-
 The Short Course: Rook and Pawn vs Rook

nearly always draw if the sum 4. Kc7 Ra7t

is less than 3 (or 4) and can The white King must go back to
frequently draw if the sum is defend his pawn, and thus allow per­
equal to 3 (or 4)-see later. petual check.

Rare exceptions and amplifica­ But move each of the chessmen

tions will be noted as we go along; other than the black Rook one file to
and the meticulous student, by writ­ the left.
ing out the rules afresh with all the
exceptions, may find it possible to
memorize the whole theory of the
ending.
It is easier to draw against an a­
or h-pawn than against the others,
but a handy covering rule is unavail­
able. The a- and h-pawns will have
some articles all to themselves later.
The strangest part of the rules
given above is the virtue inherent in Note that the defending King is
the defending King being on the "nar­ now on the ''wide" side of the pawn.
row" side of the pawn. Black cannot draw in this case, for
the method of horizontal checks fails:
1. ..
.
Ra8t
3 2. Kc7 Ra7t
3. Kc8 Ra8t
4. Kb7
The black Rook is too close-H
the Roo k could move one me to
the left, Black could draw.

Black draws with the move

Look at Diagram 3. Here the de­

fending King is cut off by only one
file, but as the pawn is on the rank
next to the queening rank one might
well expect White to win. The defend­
ing King, however, is on the "narrow"
side of the pawn; and Black, with the iabc de fgh
move, can draw by the method of dis­
tant horizontal checks. Thus, As it is, Black must all ow Kc7 and
1. R.aSt d8=Q.
2. Kd7 Ra7t The method of horizontal checks
3. Kd8 Ra8t is hopeless, therefore, if the defend-

-17-
 CJS Purdy On The Endgame

ing Rook is on the "narrow" side of A Surprising Resource

the pawn. Also, the defending Rook
must be on the opposite side to the
defending King; for if the Rook at­
tempted horizontal checks from the
same side, his own King would get in
the way. Thus, the defending King
must be on the "narrow" side; and
the defending Rook must be on the
''wide" side.
The defending Rook must be able
to check with an interval of three Black draws with the move
clear files between himseH and the
pawn-see Rule 2. In Diagram 3a Black draws by 1...
[Purdy considered here the case Ra7t, etc., as above. But suppose
where the defending Rook starts White then seeks to defeat the ends
one file closer: Diagram 3 with the of justice by 2. Kf6, evading check.
black Rook on bl instead of al. he Then,
then gave: 1. ... Ra7t

�
� �w � � � 2. Kf6 Kf8!

� � ��'<';: .� Commanding the queening square.

3. Rb2

::::�:;:
Threatening mate.
3. ... Rf7t!!

��� � ����
� �
By this surprise check, offering

�
� �� � � the Rook for stalemate, Black rends

�r� -----'�
the mating net and draws by com­
mand of the queening square.

1. Rb8t
2. Kd7 Rb7t Two Exceptional Positions
3. Kd8 Rb8t There are two exceptional posi­
4. Kc7 Ra8 tions which modify the rule about
and suggested th.at Black draws. "Defending King on the 'narrow'
However, as pQinted out by GM side." The rare occurrence of these
Karsten Muller, published theory two positions makes them of very
now shows th.at White still wins small practical importance in them­
here by 5. Ra2!, which Purdy had selves, but we give them because they
overlooked,-F.PH.] are useful practice for the ending in
general.
The modification is in the provi­
sion that the defending King, in or­
der to draw by being on the "nar­
row" side, must be at least as near
the queening rank as the pawn.

-18-
 The Short Course: Rook and Pawn vs Rook

In the case of the d- or e-pawn White wins.

(and not the c- or f-pawn) on the rank But now move all the chessmen in
next to the queening rank, the provi­ Diagram 3b one file to the right, so
sion is stricter: The defending King that the e-pawn becomes an f-pawn.
must be on the same rank as the When the play reaches the stage
pawn, not ahead of the pawn. In where White interposes his Rook on
Diagram 3, for example, the defend­ g6, Black can exchange Rooks be­
ing King must be on ft1, not g8. For cause the resulting position is stale­
an illustrative case, consider Dia­ mate. This shows that the pawn-on­
gram 3b. the-rank-next-to-the-queening­
rank exception applies only to the d­
����,�-� 3b
or e-pawn (and not to the c- or
f-pawn).
The other rare exception is with a

::::::::
pawn on its original square. Here,
to be sure of drawing, the defend­
ing King must be closer to the
-��� - ,� � queening rank than the pawn. Be­
,, � � � ing on the same rank will not do­
because the pawn can move two
Black to move/White wins squares on its first move.

White threatens both Rf8t, fol­

lowed by Kd7, and Rg'2t, driving the TYPE 3
black King another file away. Black's The Defending King Is Shut Out
only hope, as before, is the method By One File On The Wide Side
of horizontal checks. In this case,
however, Black does not guard f6; and This section is to cover all cases
White wins, in answer to checks, by of the defending King being shut out
Kd7-e6-e5(!)-f6-g5-g6 and then inter­ with only one file separating him
position of the white Rook at f6. Thus, from the pawn, except for the case
1. R.a8t discussed in section 2 (with the King
2. Kd7 Ra7t on the "narrow" side of the c-, d-, e-,
3. Ke6 Ra6t or f-pawn--of course, the King can­
4. Ke5! not be shut out hy one file and still
If 4. Kf5?, then 4.. Kf7.
. be on the "narrow" side of the a-, b-,
4. ... Ra5t g-, or h-pawn, there just isn't room
Or ...Ra8. enough on the board for that).
5. Kf6 Ra6t The rule covering the present
6. Kg5 Ra5t case is that the pawn wins if it is
Or .. Ra8.
. past mid-board and only draws if
7. Kg6 Ra6t it is not. Very simple and handy!
8. Rf6 Diagram 4 is the final stage nor­
Black cannot exchange and must mally reached where the defending
retire to a8, allowing Rd6-d8t, and King is shut out by one file on the

-19-
 OS Purdy On The Endgame

"wide" side. gram 4. (This concept is called the

Lucena Position.)

Either to play/White wins

Either to move/White wins
White threatens 2. Re7t Kf8 (if 1...
Kf6, then 2. Ke8, etc.) 3. Re8t Kf7 4. Thus, in Diagram 5,
Kc7, winning. The only reasonable 1. Kc5 RcSt
initial move for Black, therefore, is 2. Kb6 Rd8
1... Rel, preventing the white King's 3. Kc6 RcSt
egress; for we know that the method Once again the black Rook is too
of horizontal checks is hopeless since close to the white King to draw.
the black Rook cannot put three clear 4. Kd7
files between himseH and the pawn. followed by d6, and so on, winning.
So,
1. .. . Rc1 !
then, Exceptional Position
2. R�t Kg7 But with a pawn on the fifth rank
3. Rf4! Rd1 (i.e. with a pawn that requires three
4. Ke7 moves to reach the queening rank)
or4. Kc7. there is one exception. In Diagram 5,
4. Re1 t put the black King on f7 instead of
5. Kd6 Rdt t on ffi.-and you will have the excep­
6. Ke6 Re1 t tional position. Black, with the move,
7. Kd5 Rdtt draws by 1... Re8! The white Rook
8. Rd4! must move off the e-file, and the black
And White wins. King then comes across and draws
This method, interposition on the easily by command of the queening
fourth rank, is of frequent use. The square.
defending King must not be close White, with the move, could win
enough to draw. by playing his Rook to e5, the one
Move the pawn back to the fifth or safe square on the e-file. If thereupon
sixth rank and put the white King 1... Re8, then 2. d6!, winning. White
anywhere supporting the pawn. You could also win by 1. Kc5.
should find it quite easy to play the
pawn to the seventh rank, thus bring­
ing about the winning position in Dia-

-20-
 The Short Course: Rook and Pawn vs Rook

Pawn On The Fourth Rank

But put the pawn on the fourth
rank or farther back and Black
draws. For now Black can once more
employ the method of distant checks,
not horizontal this time but frontal­
Black can put an interval of three
clear ranks (not files this ti.me) be­
tween his Rook and the pawn. Thus,
Rule 2 is once again invoked.
White to move/Win
Black to move/Draw
6
Black, with the move, draws by
playing 1... Kf6--he is then in the po­
sition of Diagram 6. White, with the
move, starts out as in the play for
Diagram 6:
1 . Kc4 Rc8t
2. Kb5 Rd8
Either to move/Draw 3. Kc5 Rc8t
4. Kb6 Rd8
In Diagram 6, for example, Black Now White continues:
draws: 5. Re4! Kf6
1 . Kc4 RcSt (had the black King stood on f6 in
2. Kb5 Rd8! the initial position-as in Diagram
3. Kc5 RcSt 6--Black would here play ...Kf5 and
4. Kb6 Rd8 draw)
White must allow perpetual check 6. Kc7 Rd5
or lose his pawn or defend the pawn 7. Kc6
with his Rook, letting in the black followed by d5, giving an instance of
King-the result in each case is a the "pawn past mid-board" rule and
draw. making it a win for White.
If the pawn is on the second or
third rank, the draw is easy.
Exception Again, it is very easy to draw
In Diagram 6, had the black King against the b- or a g-pawn in such
been on f7 or f8 instead of on f6, White positions (pawn not past mid-board)
could win . Consider Diagram 6a. because the white King has less
freedom of movement. But with the
pawn on the fifth rank that does not
matter.

- 21 -
 as Purdy On The Endgame

TYPE 4 In Diagram 7, we have the d-pawn

The Defending King Is Shut with a sum of 4 (2 plus 2), so the
Out By More Than One File position should be a sure win for
White:
Let us recapitulate the rule gov­ 1. Kc4
erning this type from section 2: It is no advantage for Black to
b) More than one file between the have the first move; for if 1... Rc8 to
pawn and the defending King. stop the text, then simply 2. d5.
i) If the pawn is past the 1. ... Rc8t
middle of the board, the at­ 2. Kb5
tacker wins. If the pawn is the b- or g-pawn,
ii) If the pawn is not past the this maneuver (getting two files away
middle of the board, the at­ from the pawn on the "narrow" side)
tacker wins always if the num­ is impossible; and that is why a b- or
ber of ranks the pawn has ad­ g-pawn, if only on the fourth rank,
vanced from its home rank (0, can only draw-unless the defend­
1, or 2) plus the number of files ing King is more than two files away.
between the pawn and the de­ 2. ... Rd8
fending King exceeds 3 in the If 2. .. Rb8t, then 3. Kc6!
case of a c-, d-, e-, or f-pawn or 3. Kc5 Rc8t
exceeds4 in the case of a b- or 4. Kb6! Rd8
g-pawn. The defender can Now note the winning device.
nearly always draw if the sum White can afford to use his Rook be­
is less than 3 (or 4) and can cause of the distance of the black
frequently draw if the sum is King. That is not necessary if the
equal to 3 (or4). pawn is closer to the queening
rank-see Diagram 5.
The win for a pawn on the fourth 5. Rd1! Kf7
rank is shown in Diagrams 7 and 8. 6. Kc7 Rd5
If the pawn is on the fifth rank, the If 6... Ra8, then 7. Rei! Ra5 (or 7...
win is quite easy; for we showed­ Ra7t 8. Kb6 Rd7 9. Kc5 and White
Diagrams 4 and 5---that the pawn wins for the same reason as given in
then wins even if the interval is only the main variation) 8. Kc6 Ra6t 9.
one file. Kb5 and Black can no longer prevent
the advance of the pawn to the fifth
rank-then White wins, as was
7 shown in the preceding section.
7. Kc6 Rd8
If Black makes any other move,
then 8. Rei; and White wins as in the
variations given above.
8. d5
And White wins. If the black King
attempts to approach closer to the
Either to move/White wins key squares of the d-file (... Ke7 or

-22-
 The Short Course: Rook and Pawn vs Rook

. . . Ke8), he is driven back by Re1t­ move, wins by:

the black King is barred from d8 by 1. Kf4
his own Rook. Or 1. Kd4)
But now look at Diagram 8, a very 1. ... rust
pretty study. 2. Ke5
Threatening Kd6 followed by ad­
vance of the pawn.
8 2. . .
.
Re8t
3. Kf6!
Were the black King on h7 or h8,
White would win by 3. Kf5 Rf8t 4.
Ke6 Re8t 5. Kf7! threatening mate
as in the text.
3. ... Rf8t
4. Ke7
Either to move/White wins followed by advance of the pawn to
the fifth rank, winning.
Here White's Rook cannot afford d) Now suppose it is Black's move
to leave the cutting-off file, as he does in the position diagrammed. 1... Kh4
on move 5 in the play to Diagram 7, and 1. .. Kh6 lose as shown in b) and
because Black would then draw by c) . Again, if 1... Re7 to prevent ad­
the "King on the 'narrow' side" rule. vance of the pawn, then Black re­
White, however, has a special way duces the interval between his Rook
of winning because Black's King is and the pawn to less than three
on the edge of the board and subject ranks-the old, old story-and White
to threats of mate. wins by
1. Re7
2. Kd4 Rd7t
Demonstration Of Diagram 8 3. Kc5 Re7
a) The first point to note is that 4. Kd5 Rd7t
once the pawn can get to the fifth 5. Ke6
rank its farther advance is quite easy, followed by e5. White again wins by
the r e a s o n b e i n g that fro n t a l getting the pawn to the fifth rank.
checks by Black are then inef­ e) Again, suppose
fective since he cannot put three 1. ... Ra8
clear ranks between his Rook 2. e5 Ra4
and the pawn-see section 3 and By thus retarding White's King,
also good old Rule 2. Black hopes to keep the pawn from
b) Suppose Black's King were at advancing farther. But White impu­
h4 instead of h5. Then White, with dently plays 3. e6! !
the move, wins by 1. e5! ; for if 1.. . 3. e6! ! Kh6
Rxe5t, t h e n 2. Kf4-and White If 3 ... Ra6, then 4 . Kf4! Rxe6 5.
mates or wins the Rook. Kf5; and White mates or wins the
c) Suppose Black's King were at Rook.
h6 instead of at h5. White, with the 4. e7! ! Ra8

- 23 -
 OS Purdy On The Endgame

5. Kf4 Re8 the beginning of this section-and

6. Kf5! Rxe7 here is another:
7. Kf6 The defending King is usually best
And White mates or wins the placed just two ranks ahead of the
Rook. pawn-and never less than one
We have now shown that any move ahead, nor more than three.
by Black loses-Black is in Zug­ Furthermore, the student should
zwang. have gathered by this time the gen­
f) Now suppose it is White to move eral aim of each side. The defender
in Diagram 8. White wins simply by will strive at all costs to prevent the
bringing about the same position with pawn from advancing, relying mainly
Black to move. Thus, on distant frontal checks�r hori­
1. Rg3! Kh4 zontal checks if his King is on the
If Black moves the Rook, he loses "narrow" side of the pawn. The at­
as before; and 1... Kh6 loses as in c) . tacker, in order to advance the pawn,
2 . Rg2 Kh5(g) may have to move his Rook away and
3. Rg1! ! let the defending King approach a file
And the deed is done. nearer.

g) If
1. Rg3! Kh4 TYPE 5
2. Rg2 Kh3 The Defending King Is Shut
then Out By One Or More Files
3. Rg5!
threatening e5. Special Characteristics of the
3. ... Kh4 a- and h-Pawns
4. Rg2! Kh3
If 4 . . . Kh5, then 5. Rg1 ! ! as above. We shall deal with the same types
5. Rg1 ! Kh2 of positions as in sections 2, 3, and 4 ,
If 5 . . . Kh4, then 6. e5! ! but we shall concentrate this time o n
6. Rg5! Kh3 the special features o f the a- and
7. e5 Kh4 h-pawns. Drawing chances are con­
8. Rg1! ! siderably greater against an a- or an
And White wins; for, as we know, h-pawn, partly because:
Black cannot take the pawn. 1) the attacking King can move
only on one side of the pawn and
partly because,
Second And Third Rank 2) the defender can frequently ex­
We do not propose to discuss these change Rooks and bring about the
cases in detail, for their ramifications usual draw in the pawn ending.
are bewildering-to go into those
now would delay discussions of much
greater practical importance.
However, we have given the chief
rule that applies to these cases-see

-2�
 The Short Course: Rook and Pawn vs Rook

9 10

Either to move/Draw Either to move/Draw

Diagram 9 illustrates feature 1) . Diagram 10 illustrates feature 2-

The black King is shut out by three very important.
clear files, and the white pawn is on We know that any kind of pawn
the seventh rank; but White cannot on the fifth rank, other than an a- or
win . White's King has no move, and h-pawn, wins even if the interval be­
White's only hope is to play his Rook tween the pawn and the defending
to b7 or b8, thus making an escape King is only one file (two in the case
for his King. But if 1. Rh2, then 1... of "Defending King on the 'Narrow'
Kd7 2. Rh8 Kc7; and now the white Side" -see section 2).
King can never escape. But here Black draws with an in­
If White's Rook were on the c-file terval of three files!
or if Black's King were at e6, say, Black does it by threateningto ex­
instead of at e7 and if it were White's change Rooks. Thus, White to move,
move, then the white Rook could get 1. Kb5! Rd8!
to b8 in two moves and thereby force White cannot afford to exchange:
the win. For example, Diagram 9 with 2. Rxd8? Kxd8 3. Kb6 Kc8 and Black
the white Rook at c2: draws. So White must leave the d-file
1. Rc8 Kd6! and let Black's King approach a file
To prevent later Kb7-a6-b6-c5! nearer.
2. Rb8 Rh1 2. Rc4
3. Kb7 Rbtt If 2. Ra4, then 2 ... Kd7.
4. Kc8 Rctt 2. ... Rb8t
5. Kd8 Rh1! An important interpolation.
6. Rb6t Kc5 The white King cannot advance;
7. Rc6t! Kb5 for if 3. Kc6, then Black exchanges
If 7 . . . Kd5, then 8. Ra6. Rooks and draws. If 3. Ka6, then the
8. Rc8 white King is hemmed in in front of
And White wins. his own pawn; and Black draws eas­
ily. White can get his pawn to the sev­
enth rank, but only in a drawn posi­
tion such as that in Diagram 9. So
White must retire.
3. Ka4 Ra8!

- 25 -
 CJS Purdy On The Endgame

This is our own move, giving a

very much shorter solution than Che­
ron's more obvious . . . Kd7 at once
which allows White to push his pawn
to the sixth rank. For now Black
threatens ... Kd7; and White has noth­
ing better than 4. Rd4, bringing about
the original position. If 4. Kb5, then
4 . . . Rb8t as before. The method of
drawing by threatening to exchange
Rooks was also used in an excep­ Black to move/White wins
tional position with the d- or e-pawn
in section 3. But there, of course, only By way of example, consider Dia-
one file separated the pawn from the gram 10a. If now 1. .. Rc8, then
defending King. 2. Rb4 Kc7
3. a7 Ra8
Rule (for an a- or h-pawn on the 4. Ka6
2nd, 3rd, 4th, or 5th rank in positions and so on-White wins.
such as that of Diagram 10): There are other variations, but
White wins in all.
RULE3
¢-
TYPE6
T
h e defender draws by
threatening to exchange The Kings Are Distant
Rooks, provided that his King
is no farther from the queen­ So far, we have been considering
ing square than the attacker's only cases where the attacker's King
King. The defender then draws assists the pawn and the Rook is
with or without the move. used to shut out the defender's King;
<¢- but sometimes the attacker's King is
too far away to protect his pawn.
That gives rise to an entirely distinct
In Diagram 10, for example, the and very important class of ending
Kings are equidistant from (the same in which the Rooks attack and de­
number of moves from) a8. fend the pawn.
A pawn on the sixth rank does not Nearly always it is the a- or h­
obey this rule. pawn that is involved-sometimes
the b- or g-pawn, but seldom others
(c-, d-, e-, or f-pawn), simply because
the Kings can so easily reach those.
Consequently, the a- and h-pawns will
be our heroes in the diagrams.
Two main considerations govern
this type of ending.

-26-
 The Short Course: Rook and Pawn vs Rook

1) The Positions Of The Kings

If the defender's King can safely
approach the pawn or the queening 11
square and if he is closer to either of
these (approachable) targets than his
rival, then the draw is, of course, cer­
tain. If the attacker's King is nearer
the queening square than the rival
King, the attacker will win, in gen­
eral, if the pawn is not an a- or
h-pawn-and in most cases the at­ Either to move/Draw
tacker will win even if the pawn is an
a- or h-pawn. Diagram 11 illustrates the final po­
sition when the attacker's Rook is in
2) The positions Of The Rooks front of his pawn.
The great rule is, whether you are The position of the white King is
the attacker or the defender: immaterial-it can never shift the
If you can, have your Rook behind defender's Rook from the a-file, and
the pawn. if it goes near the pawn, it has to
For with every step the pawn ad­ suffer checks till it goes back.
vances, the Rook behind it gets more Everything depends on the black
room while the Rook in front of it King. With White to move, the draw­
gets more confined. When the pawn ing squares for the black King are
finally gets to the rank next to the b7, a6, b6, c7, c6 (all obvious) and g7
queening rank, the Rook in front of and h7!
it usually has not a single move on Why? Well, clearly the black King
the board. must not be in front of the seventh
So if you are the attacker and you rank unless he can command the
get your Rook behind your pawn, you queening square if checked. But why
should win, other things being equal, not anywhere on the seventh rank?
provided that the pawn is past mid­ Because if the black King is at d7,
board. If you are the defender and e7, or f7 White wins by 1. Rh8! Rxa7
you get your Rook behind the pawn, (forced) 2 Rh7t, picking up the black
you should draw, other things being Rook (skewer). But if the black King
equal, even though by being behind is at g1 or h7, he stops that maneu­
the pawn you allow it to reach with ver. Diagram 11, therefore, is the
impunity the rank next to the queen­ stock drawing position.
ing rank. Note that if the pawn is at a6 in­
This rule greatly increases in im­ stead of at a7 then White can win
portance when each side has pawns sometimes through being able to
on the wing opposite to that of the shelter his King at a7.
candidate Queen-which is about the Now transpose the Rooks. White,
most commonly occurring type of with the move, wins because his King
ending in chess. can reach b7 and dislodge the black
Rook.

- 27 -
 CJS Purdy On The Endgame

Pawns On The Opposite Wing giving check.

This type is so common that ev- If White's f-pawn were at f2 in-
ery chess player can be practically stead of at f4, White would draw eas­
certain of striking it at least once in ily; for White's King could protect his
his career. f-pawn from g2.
Knowing that he must lose if he
Sir George Thomas allowed the c-pawn to canter along
to the second rank, White should
have tried his luck with a royal tour
12 to the queen.side. White would lose
one of his kingside pawns of course,
but his King might be close enough
to play into a drawn position. That is
so because Black's candidate Queen
is a c-pawn. In the case of the b- or
an a-pawn, the journey would take
the white King too far for White to
Kostich have any chance of drawing-that
White to play can be taken as a general rule.
Kostich prefers merely to set a
In Diagram 12, Kostich-Thomas, trap. Thus,
Nice 1930, the pawn ( candidate 1. g5
Queen) is a c-pawn and not the White hopes for 1... hxg5, which
a-pawn as in Diagram 1 1 . With dis­ would enable him to eliminate his
tant Kings, a-, b-, and c-pawns (also fatal f-pawn and to draw. After 2.
f-, g-, and h-pawns) are much of a fxg5, Black could play the c-pawn
muchness. down to the second rank and then
In general, this type of position is maneuver his King to f5 to attack
a draw, just as much as when there the White g-pawn . But then White
are no kingside pawns at all; but the could check with his Rook (Rc5t)
example given in Diagram 12 is an and, after ... Kf4 (the white King then
exception-White is probably lost. being on g2 or h2), simply move his
We must visualize the position af­ King to h2 or g2.
ter Black has got his pawn down to 1. . . . Rc3t
c2, with his Rook on c1. After that, 2. Kg2
we know that the white King's only Giving better chances was 2. Kf2,
safe squares are g2 and h2. Conse­ to be nearer to the queen.side. For if
quently; at that stage of the play, thereupon 2 ... R:xh3, then 3. Rxc5;
White's f-pawn will be quite defense­ and White should draw.
less; and it will be impossible for 2. c4
White to prevent the black King from 3. h4 Rc1
working around to e4 and winning 4. Kh2
the white f-pawn-Black's kingside The royal tour would now be too
pawns are inviolate, the white Rook late: 4. Kf2 c3 5. Ke2 c2 6. Kd2 Rf1 ,
being tied to the c-file or needed for etc.

-28-
 The Short Course: Rook and Pawn vs Rook

4. c3 although behind its pawn, has little

5. Kg2 h5 mobility since it has only three ranks
6. Kh2 c2 to move on, compared to four ranks
7. Kg2 f5 for the white Rook. But with the
White, having tamely walked into Rooks still transposed, put the
a clear loss, now resigned. For when c-pawn on the fourth rank. Now the
the black King gets to e4-and he black Rook has four ranks to move
cannot be stopped-the lateral check on, and the white Rook has only
is in vain. Upon . . . Ke3 (after the three. This extra mobility generally
check), the white King must go to h2. allows a forced win, especially if the
Then .. Kf3, the white Rook moves,
. candidate Queen is a b- or an a-pawn
and the black King grabs the pawn (or a g- or an h-pawn); for then the
on f4. Now we see why the weak defender's King can never afford to
f-pawn is fatal for White, whereas a cross over to the other wing-after
weak g- or h-pawn would not matter. the exchange of Rooks when the can­
Thus we see that this type of end­ didate Queen is captured, the at­
ing should nearly always be drawn­ tacker's King walks in and the at­
it was just the accident of the ad­ tacker wins the pawn ending. With a
vanced f-pawn that led to White's c-pawn (or f-pawn) as candidate
loss. Queen, the crossover is a possibility
Now transpose the Rooks in Dia­ to be considered.
gram 12. Still a draw, in general, for The student should experiment
the c-pawn stands blockaded on the with the positions we suggest and
fifth rank. The black Rook, therefore, work them out to wins or draws.

- 29 -
 CJS Purdy On The Endgame

Chapler Z
Pa\\11\ El\dil\qs

I have seen a strong player make a

C Rook, is concluded. We are tak­
hapter 1 , on Rook and pawn vs.
vital oversight in an important tour­
ing pawn endings next, in order of nament, thereby missing a prize, be­
usefulness. Perhaps pawn endings do cause-incredible as it may seem­
not occur as frequently as some other he had never heard of this rule.
endings, but they are more amenable
to clear-cut rules than any other The rule is:
type-consequently it is with pawn If the enemy King is "in" the
endings that the value of book study square of which that line is the di­
is the most apparent. agonal, or if the enemy King can get
into that square before the pawn ad­
vances, then the enemy King can stop
Rule 4 the pawn from queening-otherwise
"The Square" the pawn can march in under its own
The first thing to learn is how to steam.
tell at a glance whether a passed Some players look at the "square
pawn can queen unaided, i.e. whether of the pawn.. itself-that is harder to
it can get to the queening square two do. It is much easier to picture the
moves ahead of the enemy King. In "square of the pawn" by looking at
these circumstances, a single pawn the diagonal than by looking at a side.
may win though the enemy has a vast
superiority in numbers (of pawns) .
You tell by the "rule of the square."

RULE 4 (The Square)

�

L
ook at the pawn, and run
your eye along the diago­
nal that leads from the pawn to
the queening rank.
� White to move wins
Black to move draws

-30-
 The Short Course: Pawn Endings

Look at Diagram 1, and run your Kings, we have "distant opposition."

eye along the diagonal from the pawn "Distant opposition" is advantageous
to c8. You see at once that the black because it turns into close opposi­
King is one move outside the square tion as the Kings approach each
of the pawn. So if it is White's move, other.
he can queen his pawn; and if it is Also, there is "diagonal opposi­
Black's move, he can stop the pawn tion." When the Kings are on the
from queening. same diagonal with one square be­
tween them, they are in immediate
diagonal opposition. The advantage
RULE5 of the diagonal opposition is that the
� King moving first must allow his ri­
'l l Th en the Kings are on the val to take up either "real" opposi­
V V same rank or file with tion, in rank or file, or else to re­
one square between them, as sume again diagonal opposition-but
in Diagram 1a, we notice that in this case, the first-moving King at
each one prevents the other least has the option of which kind of
one from advancing, either opposition he will allow. Diagonal
straight ahead or diagonally. opposition is, therefore, the lesser of
Therefore the first King that the two kinds of opposition.
has to move is at a disadvan­ The chief things to know about the
tage; for wherever he goes, he opposition are:
gives the opposing King the op­ i) that one King always has it, in
tion of either advancing or of some form,
again "taking the opposition," ii) that if ever one King renounces
provided there is no obstruc­ it the other King can take it, unless a
tion. pawn is in the way, and
iii) that, with the same proviso,
the King that has it can never lose it
if he chooses to keep it.
Rule 5 That is enough of "the opposition"
The Opposition to go on with-let us now consider
When the Kings are so placed, some examples of its value.
they are said to be "in opposition."
The player who moves into such a
position "takes the opposition;" and King And Pawn vs. King
when he has got there, he "has the When the pawn cannot queen un­
opposition "- because the onus of aided, it needs the assistance of its
moving is on the opponent. King. It is extremely helpful to know
The number of squares between the winning and drawing positions
the Kings need not be one-it can be in this ending so that you can tell at
any odd number. In other words, the a giance whether it will pay you to
squares the Kings stand on must be exchange Queens or Rooks-Qr other
of the same color. When there are pawns-in a certain position.
three or five squares between the

31
- -
 CJS Purdy On The Endgame

Any Pawn Other Than Rule 6D is included in the other

The a- or An h-Pawn rules, but it's helpful to remember it.
The winning and drawing posi­
tions are covered by the following
four rules:

RULES6A-6D
-¢'

K
ing in front of his pawn and
on the rank next-but-one
to the queening rank: Always a
win (6A)
.

Either to play/White wins

V ing in front of his pawn and
.l�ot on the rank next-but­
Diagram ta will illustrate all four
one to the queening rank: A win
rules. The white King is in front of
only if the King has the opposi­
his pawn and is on the rank next­
tion or if he can advance or if his
but-one to the queening rank, so the
pawn is more than one rank be­
position is a win for White (Rule 6A) .
hind him. (6B)
If White has the opposition, the win
V ing not in front of his pawn: is easy; for Black must move aside
.lu win only if the King is on and let the white King move into com­
the rank next-but-one to the mand of the queening square-then
queening rank and his pawn can the pawn walks in.
go to the rank next to the queen­ It is important to advance the King
ing rank without checking. (6C) at every opportunity, provided that
he defending King can move you do not thereby allow the enemy
T onto the square immedi­ to take the pawn.
ately in front of the pawn: Al­ If it is White's move, we have:
w ays a draw. (6D) 1. Ke6 Ke8
-¢' 2 . d6!
This resource is only available on
the rank next-but-one to the queen­
Every chess player worthy of the ing rank. We now have the excep­
name knows these rules, though per­ tional position of Rule 6C.
haps not in this simple and concise 2. .
.. Kd8
form. Of course the advantage of hav­ 3. d'7
ing your pawn more than one rank And White wins-the pawn mov­
behind your King (Rule 6B) is that ing to the rank next to the queening
you can move the pawn to gain the rank without giving check. Had
opposition and still have the pawn Black's King been in opposition at
behind your King. To be able to ad­ e8, then after d7t the black King
vance your King enables you to put would have been able to go to d8, the
another rank between your pawn and square immediately in front of the
your King-and so is just as good. pawn (Rule 6D) , and draw since Kd6

-32-
 The Short Course: Pawn Endings

(forced) would give stalemate. If Black retires, White simply ad-

As it is, the black King must move vances his King to the 5th rank and
aside ( ... Kc7) and allow Ke7. then to the 6th rank, winning easily.
Now imagine that Black still has For example, 4 . . . Kd7 5. Kd5 Ke7 6.
another rank to retire to-for the Kc6 Ke6 7. d4, etc. Not 7. Kc7 as that
moment call it rank 9. Then after 3. loses the pawn.
d7, Black simply plays 3 . . . Kd9 ! , 5. Kc5
whereupon 4. Ke7 Ke9. Black now Always advance, if possible, rath-
has the opposition and draws, for er than take the opposition; for after
White must advance his pawn with all, the opposition is only a stepping
check. stone to advancement, not an end in
This proves that if we push all itself.
the chessmen in Diagram 1a back a 5. . .. Ke7
rank then White could not win-ii- Black takes the diagonal opposi-
lustrating Rules 6B and 6C. tion. The lateral opposition by 5 ...
Ke5 would be useless (6. d4t Ke6 7.
Kc6) .
2 6. Kc6 Ke8
If 6 ... Ke6, then 7. d4 Ke7 8. d5
Kd8 9. Kd6; and we are back at Dia­
gram 1a.
7. d4! Ke7
8. d5! Kd8
It is quite safe to advance the
pawn once the white King is on the
White to play and win 6th rank, so long as the pawn re­
Black to play and draw mains behind the King.
9. Kd6!
Now we take Diagram 2 and give Now we have the position of Dia­
the whole play as a fuller illustra­ gram 1a, and White wins . Not 9. d6?
tion. White, with the move, wins. because then 9. . . Kc8, and the pawn
1. Ke2 Ke8! cannot advance without checking­
White advances his King o f for example, 10. Kc5 Kd7 1 1 . Kd5
course, and Black takes the opposi­ Kd8! 12. Ke6 Ke8, etc.
tion. The move 1 .. . Ke7? would let In Diagram 2 let Black move first.
White take the opposition next move, Then Black draws thus:
and 1... Kd7 2. Ke3 Ke7! would com� 1. ... Kd7!
to the same thing as in the text. 2. Ke2 Ke6
2. Ke3 Ke7! 3. Ke3 Ke5
3. Kd4! Kd6! White can never get the opposi­
4. d3! tion without advancing his pawn to
White must move his pawn to get the level of his King. The continua­
the opposition; but the pawn is still tion is:
behind his King, so all is well. 4. d4t
4. ... Ke6 No better is 4. d3 Kd5, etc.

- 33 -
 CJS Purdy On The Endgame

4. Kd5
5. Kd3 Kd6
6. Ke4 Ke6 3
7. d5t Kd6
8. Kd4 Kd7
9. Kc5 Kc7
10. d6t Kd7
1 1. Kd5 Kd8!
The critical stage.
12. Ke6 Ke8
And Black draws, as we have White plays and wins
seen. Black plays and draws
The analysis shows that at least
White can force Black back right to In Diagram 3 we see that in four
Black's back rank, and that is often moves White captures the a-pawn
helpful if there are other pawns on and Black gets to the d-file. So if it is
the board. Black's move in the diagram, then
after White captures the a-pawn
Black gets to the c-file-the white
An a- or An h-Pawn King is shut in, and Black draws. If it
is White's move in the diagram, White
RULE 7 can escape from the a-file after cap­
<} turing the pawn and thus win:
J\ gain.st an a- or an h-pawn, 1. Kd7 Kg6
fithe defense always draws 2. Kc7 Kf6
if the defender can either reach 3. Kb7 Ke6
the queening square or if he 4. Kxa6 Kd7
can keep the opposing King 5. Kb7
confined to the file of the pawn. And White wins. Not 5. Kb6? be­
<} cause of 5 . .. Kc8.
All very simple, but most impor­
tant to know.
In Diagram 1a, move all the chess­
men three ranks to the left. Dead
draw! The black King can oscillate The Capture Of Blocked Pawns
between a8 and b8 interminably­ The two rules relating to the cap­
else stalemate. .ture of blocked pawns are of funda­
mental importance. But although
they are very easy, not many players
have them at their fingertips; and the
second is known by very few.
Rules 8 and 9 refer to the forma­
tion of pawns shown in Diagrams 4
and 5, i.e. to two opposing pawns mu­
tually blockading each other.

-34-
 The Short Course: Pawn Endings

�� 8 ��9
''The Trebuchet" The Rule of the Rank
[See also THE RULE OF LIMITS
in Guide to Good Chess, Part IV.]
4

Whoever moves wins

See Diagram 4. Here the natural Black to move/White wins

tendency of the beginner is to play 1.
Kf7, as there he both protects his own In Diagram 5, the Kings are in op­
pawn and attacks the enemy pawn. position; and Black loses if he has to
But then Black plays likewise ( 1 . . . move. Thus,
Kd6) , and White i s in Zugzwang. 1. ... Kb7
This little affair is known as the "tre­ 2. Kb5 Kc7
buchet" (bird-trap) . The first player's 3. Kc5 Kd7
right course is, by attacking from 4. Kb6!
behind, to force his opponent to be White has reached the rank of the
the first to take the fatal square. For enemy pawn, which must now fall­
example, the opposition no longer matters.
1 . Kf8! Kd6 4. . .
. Kd8 (say)
Fbrced. 5. Kc6 Ke8
2. Kf7 6. Kd6 Kf7
and White wins.
Thu s , in Diagram 4, whoever RULE 9
moves first can win the enemy pawn. -¢-
Move each King a file farther away l 'I Then the Kings are on the
and the same thing happens, Rule 8. VV same side of the pawns'
file, as in Diagram 5, if one King
RULE S can reach the rank occupied by
-¢- the enemy pawn then he wins
l 'I Then the Kings are on op­ the enemy pawn and saves his
VV p o s i t e s i d e s o f t h e own. The King in question can
pawns' file, the King who can always do what is required if
first attack the enemy pawn he can reach, with the opposi­
from the rear wins the enemy tion, the rank occupied by his
pawn and saves his own. own pawn.
-¢-

35
- -
 CJS Purdy On The Endgame

7. Kd7 3. Kc5
And White wins the pawn. This And White wins, as was shown for
means that White wins the game; for Diagram 5.
after capturing the pawn, the white
King will be on the 6th rank and in
front of his pawn (Rule 6A). Mutually Supporting Pawns
But move all the chessmen one or Every beginner knows that two
more ranks toward White's side of united passed pawns defend each
the board; and although White wins other from the enemy King; for if the
the pawn just the same, he does not enemy King captures the rear pawn,
win the game (if there are no other then the other one queens. But the
pawns on the board) because the same beginner doesn't always know
black King will so maneuver that that two separated passed pawns do
when White captures the pawn Black that just as well.
can take the opposition in front of
White and draw by Rule 6B.
7

Either to play/White wins

White to play and win Diagram 7 shows that two passed

pawns with one file between them
After the discussion above, Dia­ always protect each other, so that
gram 6 becomes simplicity itself. The their King can just come up at his
only move White can hope to win by leisure and shepherd them through
is 1. e5, as otherwise Black will sim­ to the queening rank. For as soon as
ply play ... e5; then when White cap­ the enemy King moves up to attack
tures the pawn, Black plays ...Ke7 one pawn, the other pawn advanc­
and draws (Rule 6B). es-if the enemy King then captures
By playing 1. e5 White does win, the back pawn, he falls outside the
for the white King gets onto the 6th square of the forward pawn and can­
rank and wins as in Diagram 5 . For not keep that pawn from queening
example, (Rule 4) .
1. e5 Kd7 In Diagram 7, move the left pawn
2. Kb5! a file more to the left-the result is
Taking the (diagonal) opposition the same; move it two files over, and
while getting on the rank of his pawn the result is the same again.
(Rule 9). There is one exception. When the
2. ... Kc7 two pawns have two files between

-36-
 The Short Course: Pawn Endings

them, have moved off their home with the play of the Kings.
rank, and are not beyond mid-board, But there are also problems in the
the enemy King, if properly posi- play of the pawns. Here are examples
tioned, can keep both pawns from of the two main departments of pure
queening. pawn play, to wit, forcing a passed
pawn and maneuvering for Zug­
zwang. The first is quite easy, but
8 the second is quite knotty.

RULE 10
-¢-
T protect each other from
wo passed pawns always

capture unless they are not be­

yond mid-board, away from
Either to play/White wins their home rank, and have two
files between them.
Thus in Diagram 8, neither Black -¢-
pawn can advance; and as soon as it
is White's move, he plays 1. Kh4.
Then
1. Kh4 e4 (forced)
2. Kg3 e3
3. Kf3 h4
4. Kxe3 h3
5. Kf3
The white King is in the square of
the remaining pawn-all is lost.
Had the pawns been a rank far­
ther on, the white King would not
have been in the square at the end.
Again, had the black pawns been on Look at Diagram 9. White's job is
their home rank, then as soon as one to fo rce a passed pawn through. It is
was captured the other one could a job one is faced with very frequently
move two s quares and bolt. in pawn endings, so one may as well
So we have: know how to set about it. The ques­
tion is, Which pawn is it best to ad­
vance? The answer, which is almost
Rule 10 never wrong, is:
Mutually Supporting Pawns

Problems In Pawn Play Rule 11

In pawn endings, the nicest prob­ Advancing Pawns
lems that beset us-some of the most Thus in Diagram 9, we play, not 1 .
beautiful in the game-have to do b4-for then Black blockades by 1...

-37-
 CJS Purdy On The Endgame

b5, holding up all three of our pawns The position diagrammed oc­
with one (!) of his-and not 1 . a4 (for curred in a tournament for the Aus­
then Black blockades by 1 . . . a5-in tralian championship. The game was
this instance the blockade could be thrown away by a blunder-showing
shifted), but 1. c4! that these simple-looking things do
If Black replies 1... a5, what then? not altogether play themselves.
We must play for b4; but if a3? at The position is one which must be
once, then Black blockades by a4!­ either lost or won-a draw is ruled
and two of our pawns are held by out of court altogether. For whoever
one of his. No, we again advance the runs short of pawn moves first, and
pawn that cannot be blockaded: 2. someone must, has to move his King,
b3! That enables us to play a3 safely. thereby losing his passed pawn and
So the play goes (we assume that the game.
Black has moves available on the Seeing that the player who can
other side of the board, and we indi­ make the last pawn move wins the
cate such a move by ...any): game, one would think that the player
1 . c4 a5 who had to move first would lose; but,
2. b3 any on the contrary, in this position the
3. a3 any player with the move wins!
4. b4 axb4 The rules that govern this type of
5. axb4 any position are:
6. c5 Exception: Two pawns abreast on
And White gets a passed pawn. their home rank against two pawns
on the rank next to their home rank
RULE 11 win, with or without the move.
-¢-
J\dvance the pawn that can­
.1"\..n o t be blockaded by an RULE 12
enemy pawn.
-¢-
I tions, the player who has to
n symmetrical pawn posi­
-¢-
move first will run out of moves
first-unless he can blockade
both the enemy pawns with his
Kill Or Be Killed
first move.

10
The reason for Rule 13 is that an
unmoved pawn's option of moving
one or two squares is a good thing to
keep.
For an illustration of all the points
in Rule 12, imagine that the kingside
pawns in Diagram 10 are off the
Whoever moves wins board. In the position that remains,

- 38
-
 The Short Course: Pawn Endings

Then, reacting to the one- or two­

RULE 13 square move by the second player,
'¢-
the first player moves his unmoved
Dor purposes of gaining
pawn in such a way as to produce
.I1 moves with pawns, always symmetry. For example (White to
move a pawn that has already
move), 1. g4 g6 2. h4. Now symmetry
moved rather than an unmoved
is produced; and Black, having to
pawn.
move first, obviously loses-e.g. 2...
g5 3. h5.
Now we take the position in Dia­
White's position is bad because gram 10 as a whole. Will the move
White has forfeited the option of mov­ win, as on the kingside, or will
ing two squares with one of his Black's better queenside determine
pawns; and Black wins with or with­ the game in his favor with or without
out the move. the move? First of all, it is obvious
For if it is Black's move, Black im­ that Black, with the move, wins eas­
mediately, by 1... b6, produces a sym­ ily by 1... h5; for we saw that this won
metrical position with White to move even without the queenside pawns,
and lose-for whatever White does, which are placed in Black's favor.
Black just goes on copying until Black actually threw away the
White is blockaded. game by the howling blunder 1. . .
If it is White's move, White can g5??-the very worst move on the
produce symmetry by 1. a3; but then board. White won, of course, by 2.
we have the exceptional position of g4; for then he had a spare move on
Rule 12, where Black wins because the kingside which he could use at
he has two unmoved pawns and any time-he had only to wait until
White none. Black wins by any move the queenside was blocked up. An­
on the board, e.g. 1. a3 a6 2. a4. Now other illustration of symmetry that
we have symmetry again, and again does not win-by 1... g5 Black pro­
with Black to move, but alas! Black duced symmetry on the kingside, but
can blockade both White pawns im­ White could immediately blockade
mediately by 2 .. a5. If White plays 1.
. both of Black's pawns.
b4 instead of 1. a3, then Black imme­ It is interesting to note that Black
diately produces symmetry by 1... b5. would lose by any first move other
For an illustration of Rule 13, than 1... h5 or 1... b5. The student
imagine, in Diagram 10, that the should work out how all the other
kingside pawns are on and the queen­ moves lose, just as an exercise. In
side pawns are off. This position is particular, it seems mysterious that
not symmetrical, the two sets of g­ 1.. . b5 wins but 1... a5 loses-the rea-
and h-pawns are not placed in mir­ son is that 1. .. b5 threatens ... b4,
ror-image fashion, and the move whereas 1... a5 does not threaten ... a4
wins: whoever has the move moves since the a4 square is protected by
his already-moved pawn (Rule 13) White.
and thus, this is the point, forces the The most interesting point arises
opponent to move his unmoved pawn. if White has the move. According to

-39-
 CJS Purdy On The Endgame

Rule 13, White has two moves to The Distant Pawn

consider, 1 . b4 and 1. g4. We should No one can play pawn endings un-
expect 1 . g4 to be the winning move, less he knows about the "distant
for it was so with the queenside passed pawn," the phrase occurs
pawns off. But in fact 1. g4 loses; for over and over again in annotations.
although it enables White to move
last on the kingside, he still has a Blaekbnrne
queenside that loses with or without
the move. Black's reply to 1 . g4 is
1 . . . b6! 11
The winning move for White, how­
ever, is 1 . b4! The point is that 1 . b4
threatens b5! which must win the
game if Black allows it (the student
should work out why) , so that Black
is forced to reply 1... b5. That imme­
diately produces symmetry on the
queenside. Black's good queenside Teichmann
position has deteriorated to the point Black to play (his 57th)
where if Black has now to move first
on the queenside then he loses there. Faced with the diagrammed posi­
But White can force just such a con­ tion (Teichmann-Blackburne,
tingency by playing 2. g4 which wins Berlin 1897), any strong player
for him on the kingside. would say at once, "Winnin g chances
This all sounds complicated for Black-distant passed pawn and
enough, perhaps, even when worked all that." That is because when he
out purely by principles; but try to do sees a pawn formation like that on
it by calculating all the variations the king-wing he immediately pic­
possible on either side, and you will tures the result of an exchange of
need to swathe your head in a wet pawns after Black's ... h4 or White's
towel. And this though only pawn f4. Then each King will have to cap­
moves are possible ! Yet Capablanca ture the enemy passed pawn; and af­
said chess was getting too easy! ter that White's Kingwill be two files
farther from the queen-wing than
Black's, and White can throw in the
towel. In short, it is almost invari­
ably a winning advantage to have the
more distant pawn, where the other
wing is static.
Here Black cannot force such an
exchange, but the threat of it keeps
the white King hovering around g3.
Can Black win? Obviously the first
move we consider is 1... h4, for the
reasons discussed; then 2. Kh3 [Ed.:

-40-
 The Short Course: Pawn Endings

K.M. & FL. give 2. Kg2! =] hxg3 [Ed.: Kf2?, and similar play followed with­
2... hxg3 isn't forced as Black can out Black having to use his spare
win with 2 ... Keo! (l(Jtording to KM/ move at all.
FL.] 3. Kxg'3. Now Black must use 8. Ke2
his spare move to get to f4, and we If 8. Kg2, then Black goes after
have 3... c5 4. Kg2! Kf4 5. Kf2 g4 6. the white c-pawn (...Ke3-d3xc3) and
fxg4 Kxg4 7. Kg2; and White, having we find that he queens in four more
the opposition, draws (see Rule 9). moves-meanwhile White, after four
If only Black could make the cap­ moves, will only have his pawn at f6
ture on g3 when his King was not on and will lose. Here note the following
·
f5, he would win easily; for then he rule, which is the most important of
could play . . Kf5. The white King
. all in pawn endings.
would have to retire, and Black could 8. l{g3
then play . . . Kf 4 with the spare move 9. Ke3
still in hand to recover the opposi­
tion. Can that be wang1ed? RULE 14
¢-
0 nee the Kings can no
The game went:
1. ... Kf6
2. Kg2 Kg6 longer interfere with each
Not even yetwill ... h4 win because other's business, i.e. the game
of 3. f4; and if thereupon 3... gxf4, has become a race, never cal­
then it is White who gets the distant culate from move to move in the
passed pawn by 4. gxh4. ordinary way. Simply count up
3. Kh2? the moves the first queenerwill
The fatal square. A strange move, queen in, and then see how far
as Kh3 is so obvious and seems to the opponent's queening pawn
give Black no chance of winning. Did will get in the same number of
Teichmann think he could draw any moves.
way?
3. ... h4!
4. Kh3 In the "turning'' maneuver, Black
Now 4. f4 gxf4 5. gxh4 will not has had to lose the opposition. What
serve, for Black wins the white pawn will he do about it? If 9 ... Kg2, simply
and saves his own by 5... Kh5. And of 10. Ke2, etc.
course 4. gxh4 gives Black the dis­ 9. .. . Kh3 ! !
tant passed pawn. White cannot now take the oppo­
4. . .
.
hxg3 sition because both d3 and f3 are
5. Kxg3 Kf5 barred. Black must get the opposi­
Now follows a King duel well wor­ tion back again next move.
thy of study by the learner. 10. Kd2 Kh2 !
6. Kg2! Kf4 11. Ke3 l{g3
7. Kf2 c5 12. Ke2 Kg2
White has made Black use his 13. Ke3 Kf1!
spare move to regain the "opposi­ Black is gradually forcing the
tion. " Teichmann actually played 6. white King in front of the f-pawn,

-4 1 -
 OS Purdy On The Endgame

where the white King has only one

defending square, e4.
14. Ke4 Kf2
15. Kf5
Now both Kings capture, and
Black wins because his pawn was the
more distant. In this duel, the pawns
were not "passed," so the term "dis­
tant passed pawn" does not apply­
" distant pawn" is more exact.
[The Teichmann-Blackburne end ­ Either to move, White wins
ing has been used many times in
the chess literature as a fine illus­ White makes a passed pawn by
tratio n of the "distant pawn " forcing an exchange of Black's single
theme. Though this article is cer­ pawn. White then goes across to the
tainly inst ructive, Purd y reveals queenwing, he can get three clear
so me thirteen years lat er that files away and can capture all the
things are not as simple as they pawns he wants to. Thus White wins
seem. For an "in-d epth" analysis in the same way as with the "distant
o f this ending, refer to Purd y's ex­ passed pawn," except that he does
traordinary articles OH WE OF not even have to capture a pawn on
TOO MUCH FAITH (Chapter 6) and his way over. All absurdly easy.
the t wo f ollo wing ones.- FP.H.J But when there are pawns on only
one side of the board, or in the cen­
ter, it is often hard to win with a pawn
Two Pawns vs. One Pawn plus; and sometimes impossible. You
With a pawn plus, barring doub­ have to keep pegging away in the one
led pawns, you can generally make a place, and when the final exchange
passed pawn; the way to do that was of pawns comes, to make a passed
shown in a previous section. A pawn­ pawn, the lone King may be able to
plus advantage is always enough to play into one of the drawn positions
win if there are pawns on another in the ending King and pawn vs. King.
part of the board. F or then you can
simply make the passed pawn and
clear out-while the enemy spends
time in capturing the passed pawn,
you can be capturing one or more
pawns elsewhere.
In any part of the game, it is al­
ways a great advantage for the win­
ning side to have more than one
sphere of action.
F or instance, in Diagrams 12 or Either to move/Dra�
13, imagine that both sides have a But White wins If bis pawns
pawn or pawns on the queenwing. are on the second rank

-42-
 The Short Course: Pawn Endings

There is a sharp difference be­ the pawn gives check on moving to

tween Diagrams 12 and 13 with which the 7th rank-see Rule 6C.
not many players are familiar, de­ But , in Diagram 1 3 , suppose
spite the common occurrence of such White's pawns are at home instead
positions. Against the g-pawn, White of on the fourth rank. Then White
wins; against the h-pawn, he only wins because he has the option of
draws. We shall see why that is so in moving a pawn one or two squares
a moment. on its first step, according to Black's
In Diagram 12, play play. As White has the opposition, let
1 . Kf5 Kf8 him have the move so as to make
If 1... g'6t, then 2. Ke5 Ke7 3. g5 ! ; things harder for him.
and White must win Black's blocked 1. g3
pawn by Rule 6. White has to give away one op­
2. Kg6 Kg8 tion right at the start to gain the op­
3. h5 Kh8 position, but he still retains one­
4. g5 Kg8 and that is enough. Equally good
Now simply would be 1. g4, but it is more logical
5. h6 gxh6 to keep as many spare moves as pos­
The student should work out the sible.
variation following on 5... Kh8. 1. ... Kh8
6. K:xh6 Or 1... Kg8, whereupon White sim­
And White wins-the white King ply plays h4 on move 5. Or 1 . . . h6t 2.
is on the 6th rank and ahead of his Kf5 Kf7 3. g4 Kg7 4. h4 ! (not 4. h3)
pawn. 4 ... Kf7 5. h5; and White, having the
But now let us play into a similar opposition, wins the blocked pawn
position in Diagram 13, giving Black by Rule 9.
the move as that helps White who 2. Kh6 Kg8
wants to get his King to h6. We have: 3. g4 Kh8
1. . .. Kh8! 4. g5 Kg8
If 1 . . . Kg8?, we have the same play 5. h3 !
as in the text, except that the white And White wins. Not 5. h4. Any­
pawn goes to the 7th rank without one ought to be able to tell why at a
checking and White wins. glance, and it is very useful to know
2. Kh6 Kg8 how. If White could play g7 at once,
3. g5 Kh8 he would win because the pawn
And now the maneuver corre­ would go to the 7th rank without
sponding to that for Diagram 12, checking. Consequently, if White can
namely g6, would be absurd since play g7 on his third move from now
White would be left with an h-pawn. or on his fifth move from now or on
And after his seventh move from now, he will
4. h5 Kg8 win because every second move finds
5. g6 hxg6 the black King on the same square.
6. bxg6 Kh8 So all White has to do is to play a
White's King is not ahead of his pawn to g7 in an odd number of
pawn, and White cannot win because moves, neglecting the p awn ex-

-43-
 CJ S Purdy On The Endgame

change which cancels out a move on require careful play.

both sides. If White plays h4, h5, g6, Let us say that White's pawns
gl, he will take an even number of have already moved, as in Diagrams
moves to attain the goal-therefore 14 and 15; for unmoved pawns make
White knows he must first play h3. things easier for White because of
If Black's King were at h8, then their one-or-two squares-on-the­
h4 would be the move. We can make first-move options.
that into a general rule. Where the united pawns are c-and
d-pawns or d- and e-pawns or e-and
RULE 15 f-pawns, there are no drawing posi­
¢- tions. Diagram 14 will illustrate the
winning process.
I
n a position where it is your
move and where the oppo­
nent must arrive at his present
position every alternate move,
imagine that you have already
played two, four, six, or eight
additional moves (neglecting
pawn exchanges) and that you
then have to make your move.
Then ask, Do I win? If the an­
swer is no, ask, Can I lose or
gain a move somewhere? Either to move/White wins
-¢-
First of all, note the little trap: 1 .
Kc4?? d5t 2. exd5t Kd6 3 . K moves
If you are the underdog, you fol­ Kxd5, and Black draws-Rule 6D.
low a converse process. So White must play more circui­
tously.
1. Kb4 Kb6
Two United Pawns 2. d5
vs. One Pawn Unfortunately necessary to regain
Two united pawns against one the opposition.
pawn regularly win. In the preced­ 2. ... Ka6!
ing section, we dealt with the g- and And now White has to go around
the h-pawn (the same procedure to the kingwing-the only way to win.
would hold for the a- and the b-pawn). 3. Kc4 Kb6
Now we take up the case of any other 4. Kd4 Kc7
two united pawns, i.e. two united 5. Ke3 Kd'7
pawns neither of which is the a- or 6 . Kf4 Ke'7
the h-pawn. This is better for White, 7. Kg5!
for the a- or h-pawn is usually not a By his last move, White takes the
good winning pawn. However, as we opposition (here diagonal) on the
shall see, there is one drawn posi­ same rank as his own blocking pawn
tion; and the winning positions often and wins-Rule 9.

--44-
 The Short Course: Pawn Endings

sition mentioned above. But without

RULE 16 knowing about that position, you
�
should see at a glance that 2. Kd4 is
I
n pawn endings, when you .
right. Why? Because 2. Ke4 allows
have a chance of taking the
Black to take the opposition, whereas
opposition or of preventing the
2 . Kd4 prevents Black from doing
opponent from taking the op­
so-the black pawn being in the way.
position, do not neglect to do
so without good reason.
Now continue with the analysis.
�
2. ... Ke6
3. Kc5 Kd7
If we shifted the whole position If 3 .. .f5, then 4. g5 wins easily,
two ranks to the right, White would Black's King being hampered by
not have this resource on move 7; White's threat of queening.
that explains why White cannot win 4. Kd5 Ke7
in Diagram 15 with Black's King at 5. Kc6
e6, the one drawing position apart White soon follows with f5. Then
from Diagram 13 in the preceding White has his King on the same rank
section. as the blocked black pawn, and White
wins-Rule 9.
Take now the exceptional posi­
15 tion, in Diagram 13, put the black
King on e6 instead of on g6. Black,
with the move, naturally thinks of ad­
vancing by 1... Kd5; then White wins
by taking the opposition with 2. Kd3.
If thereupon 2 ... Ke6, then 3. Kc4, etc.
Black's only move to draw is 1... Ke7! ,
whereby Black takes the opposition
Either to move/White wins. (distant) himself. We mention this to
But with Black King at e6/Draw illustrate the following general
maxim.
With Black' s King unfavorably
placed, White wins even with moved RULE 17
�
w.hen in doub t , retreat
f- and g- pawns (or b-and c-pawns ) ­

his worst formation apart from the

g- and h-pawns (or a- and b-pawns). your King and gain dis­
Thus, in Diagram 26: tant opposition rather than ad­
1.•.. Kf7 vance and lose the opposition.
White wins easily with the move �
by 1. Kd4, but of course not by 1 .
Ke4??, a s we have seen. The elements of the opposition
2. Kd4! were discussed in an earlier section.
Not 2 . Ke4 because of 2... Ke6, The study of pawn endings begins
which produces the exceptional po- and ends with the opposition.

-45-
 CJS Purdy On The Endgame

A Study In Doubled Pawns queenside pawns to force an entry

From The for his King. That must be done, evi-
Australian Championship dently, by b4 and ultimately a5, which
In the eleventh round of the 1932- will force an entry at c5.
33 Sydney congress, a featureless But let us beware of playing a5
opening and middlegame led to this too soon, for it involves undoubling
classical pawn ending-the nicest the Black b-pawns. Here is how the
piece of chess in the whole tourney. game went.
35. ... f6
A.E. Nield Or 35. . . Kd6 36. Kc4 (or Kd4),
merely transposing. It will soon be
seen how the pawn at f6 strengthens
16 Black's resistance.
36. Kc4
Kd4 could be played at once, but
the text has the merit of tempting
...Ke5.
36. ... Kd6
37. b4 Kd7
If 37... Ke5?, then 38. a5 bxa5 39.
W.S. Viner bxa5 Kd6 (forced) 40. f4; and White
Black to play his 35th move can successively force entry at c5 and
b6, winnin g. For example, 40... g5 41.
White has two advantages: his f5 Kd7 42. Kc5 Kc7 43. c4 Kd7 44. Kb6
King is a rank up on Black's, and he Kc8 45. c5 Kb8 46. a6 bxa6 47. Kxa6,
has only one set of doubled pawns and White wins the blocked pawn by
whereas Black has two. Viner dem­ our well-known Rule 9 (King on the
onstrated that these advantages same rank as the blocked pawn).
were enough to force a win . Black must be ready to meet a5
What will be White's plan? Clearly, with ... Kc7!
to make use of White's extra mobil­ 38. f4!
ity on the queenwing. There Black is While Black retains command of
quite immobile-if the b6 pawn c7, the move a5 fails. For example,
moves, Black gets a set of isolated, 38. a5 Kc7! 39. axb6t (else Black
doubled b-pawns. If Black's c-pawn plays ... bxa5 and ... b6!) Kxb6 40. f4
moves, Black's doubled b-pawns be­ Kc7 41. Kc5 b6t, etc.
come "backward"; and White could Therefore White must try fo r an­
easily force an entry at b5. From other point of entry, and he can get
there White could then fo rce the cap­ one by dislodging Black's f-pawn
ture of the b6 pawn since he has a from f6.
sufficient number of spare pawn 38. ... Kc7
moves with his b- and c-pawns. 39. Kd4
So White can count on Black's The white King now takes his
queenside pawns staying put. White proper post, commanding both of the
simply has to advance his own future points of entry, c5 and e5.

- 46
-
 The Short Course: Pawn Endings

39. .. . Kd6 47. c4!

40. g5 Kc7 The point. White has no time to
If 40... c5t, then 41. Kc4 Kc6 42. use his King now, but he can break
gxf6 gxf6 43. g4 Kd6 (if 43... cxb4, through by sheer weight of pawns, in
then 44. Kxb4; and the isolated, spite of one pair being doubled.
doubled b-pawns are easy prey) 44. 47. ... Kf5
bxc5t bxc5 45. Kb5; and White wins. 48. b5 c5t
41. gxf6 gxf6 If 48 ... cxb5, then 49. c5! followed
42. g4 Kd6 by c6, winnin g easily.
43. g5 fxg5 49. Kxc5 g5
44 . f:xg5 Ke6 If 49 . . . Ke5, then 50. Kb4 Kd6
At last White has his point of en- (forced-if 50... g5, then 51. c5 wins,
try at e5--to defend e5, Black must White queening with check) 51. c5t
now give up his command of c7. How- Kd7 52. Kc4; and White captures the
ever, there is now an entrance for g-pawn and wins.
Black also; and White had to calcu- And if 49 ... Ke6, then 50. Kd4; and
late deeply to see that he could give White wins.
up the g-pawn. 50. Kd5 g4
45. a5 ! Kf5 51. c5 g3
If 45... bxa5, then 46. bxa5 Kd6 (if 52. c6 bxc6t
46... Kf5?, then 47. Kc5, etc. , win- 53. bxc6 g2
Ding) 47. Kc4; and White forces en- 54. c7 g1 = Q
try at b6, winning as in the note to 55. c8 = Qt Kf4
move 37. 56. Qf8t Resigns.
46. axb6 Kxg5 The exchange of Queens is forced.

-4 7-
 CJS Purdy On The Endgame

[hapler J
Oueen and P a\UI\ Endinqs

lntrodudory as possible by checks;

C
ertain Queen and pawn endings 2) force the enemy King in front
are always cropping up and must of his pawn for one move; then
be known. They usually arise out of 3) move your King a square nearer;
pawn endings, hence our sequence. 4) repeat the process;
Chess endings are what you can 5) when your King is near enough,
remember about them; and where mate.
memory work of some sort is com­ The only part that gives beginners
pulsory, the true art of the teacher diffi culty is 2), so we show how to do
lies in stating rules in memorable it in Diagram 1. No key position is
form. Endgame books draw wiggly given-too easy.
cordons all over the board to show
when a Queen can win against the c­
or f-pawn. But those cordons are not
memorable. What you really remem­
ber is the final mating position, the
tricks, and when such an ending oc­
curs in your own play, you just see if
your King is close enough to get into
one of the positions.
We shall, therefore, diagram just
those key positions. White wins easily

1. Qa2
Queen vs. Pawn Or 1. Qe8t-check or pin, it does-
About To Queen n't matter.
The b-, d-, e-, or g-Pawn 1. Ket
2. Qe6t Krl
Against the b-, d-, e-, or g-pawn, 3. Qb3
the Queen always wins, and easily. Or Qd5 or Qf5t-any old thing.
All you have to do is: 3. Ke2
1) Bring your Queen up as close 4. Qc2 Ket

-48-
 The Short Course: Queen and Pawn Endings

5. Qe4t Kf2 In Diagram 3, Black can queen;

6. Qd3 Ket but the new Queen only serves to ob­
Now we have him. struct the black King, and White
7. Q e3t Kd1 (forced) mates next move.
8. Kb6 Kc2 These are the two winning posi­
The Queen keeps moving until we tions. All you have to do is to see if
can check at c3; then the King moves you can play into one of them. For
up again. And so on. example, with White's King at a5 or
d5 in Diagram 2-but with Black's
pawn still at a2, one move from
Queen vs. An a- or h-Pawn queening-simply 1 . Kb4 (or Kc4)
In Diagram 1 , move Black's men at = Q 2. Kb3; and White wins. How­
three files to the left. Then White can­ ever, if the white King starts on b6
not win because he can 't leave (say), then no go-Black draws. If
Black's King in front of the pawn the white King starts on e4 or e2
without stalemating. Try it. (say) , then 1 . Kd3 (or Kd1 ) a 1 = Q 2.
But White can win if his King is Qc2#.
close enough. See Diagrams 2 and 3. Your Queen must first get to d2,
by checks (and/or other threats).
Bring your Queen as close as pos­
sible before starting to think at all .

Queen vs. The c- or f-Pawn

Risking Stalemate
In Diagram 1, move Black's men
only one file to the left, making it the
c-pawn. Again White cannot win; for
Winning position the black King, instead of going in
front of the pawn, hides at a1, letting
In Diagram 2 , Black has queened; the Queen take the pawn-stale­
and White has then calmly played mate!
Kb3. Black can move his queen but
canno t escape mate in two or three.
4

Winni ng position

Winning position But again White can win if close

-49-
 OS Purdy On The Endgame

enough. Diagrams 4 and 5 give the 4. Kc4 Kb2

mating positions. Notice that Black 5. Qd2 Ka1
can be allowed to queen-but not to or Kb1)
move the new Queen, as in the case 6. Kb3
of an a- or h-pawn. Thus, in Diagram and White wins.
4, 1... c1=Q 2. Qa2#.

A Queen And Knight Ending

5 A fine Queen and Knight ending
won by F.A. Crowl from an old oppo­
nent, S.F. James. The play is worthy
of the closest study.

Crowl

Winning position 7

In Diagram 5, the other method of

winning is illustrated-catching the
pawn with the King.
If the black King has not yet got
to the narrow side of the pawn, as he
is in Diagrams 4 and 5, White's King
can afford to be at least one square James
farther away since White gains a Black to play
tempo due to Black's King having to
block the pawn for one move as the 48. Nf8!
King journeys to a1 (or to h1). The ending is a striking example
of the cooperative power of Queen
and Knight. Black's advantage lies
6 chiefly in that, rather than in the fa-
vorable pawn formation.
49. Kg2 Qd7
50. Qxd3 Qxa4
51. Qf5 Qd7
52. Qc2
[52. Qh5, posting the Queen ag-
gressively, looks stronger.-F.PH.]
White to play and win 52. . .. Qe8
53 . .Kr2 a4
In Diagram 6, for example, 54. Kft a3
t. Qd4t Ket 55. Qb3 Qa8
2. Qc3t Kd1 56. Qa2 Qa4
3. Qd3t Ket 5'7. Ke1 Ng6

-50-
 The Short Course: Queen and Pawn Endings

58. Be3 b5! 70. Qxd6t Kh5

59. cxb5 Qe4 71. Qe6 Qxc3
60. Ke2 Qf3t 72. d6 Nd3
61. Kd3 Qxg3 The final blow. With other moves
62. Qxa3 Nxf4t in reply, too, White runs out of checks
63. Kd4 Ne2t or is met with a fatal cross check or
64. Kd3 Qf3 the Knight gets south in time to stop
65. QaSt Kh7 any nonsense.
66. Qc6! 73. Qe8t Kxh4
A painful surprise which nearly 74. d7 Qctt
upset the attack. Black's plan right 75. Ke2 Nf4t
from his 57th move had been predi- 76. Kf2 Qd2t
cated on Qb7 here, whereupon ...Nf4t 77. KU Qd1t
wins the Bishop since if in response Here a move or two could be saved
Kd4 then ...Ne6t followed by ...Nc5t by the crude thrust ...Kg3, but it
wins the Queen instead. seemed fitting that the Knight should
66. .. . Nc1t have the reward of his gallantry by
67. Kc2! finalizing the matter.
White decides, probably correctly, 78. Kf2 g3t
that the sacrifice of the piece offers 79. Ke3 Nd5t
him the best chance. 80. Ke4 Nf6t
67. Qe4t! 81. Resigns.
68. Kd1 Qxe3
69. Qxc7t Kg6 -Notes by F.A. Crowl.

- 51 -
 CJS Purdy On The Endgame

(l\apler 4
Practical El\dqal\\e Studies
fPu,rdy recognized the importance, in practical endgame play, of con­
stantly being on the lookout for tactical points and surprise rrwves
arising from themes specific to endgames. Such possibilities can arise
even in seemingly harmless positions and are easily overlooked�en
by strong players. In the following collection, the "studies" -almost all
from actual play - demonstrate in each case an overlooked artistic
win or draw.]

An Endgame for Poe 0. Buras

[Before taking up the theme of this
section as outlined in the editorial
commen t above, let's look at a com­
posed study by Duras that Pu,rdy
used to illustrate the desiderata of
a good endgame study.]

I
n The Murders in the Rue Morgue,
Edgar Allan Poe accuses chess of
"elaborate frivolity:" he says, "In
chess, where the pieces have differ­ White to play and win
ent and bizarre motions, with vari­
ous and variable values, what is only Confronted with this position in
complex is mistaken (a not unusual actual play, a player who merely tried
error) for what is profound." Some out different variations, like a fly in a
time we may discuss this philosophi­ flytrap, would be almost sure to miss
cally, but here we merely point out the win. Being told that a win is there
that some of the hardest problems makes it easier, but even so we ad­
that arise in play occur in endgames vise the reader not to adopt the trial
with very few pieces and with none and error method but to solve the
of that complexity of which Poe problem by a logical train of reason­
speaks. A particularly good example ing-in practical play the latter is the
of profundity without complexity is only way to success.
the following endgame study by 0. The correct procedure is first to
Duras, the Bohemian master. inquire into Black's possibility of

-52-
 The Short Course: Practical Endgame Studies

drawing. For the benefit of the be­ ing one with the Bishop to free the
ginne r, we point out that the answer pawn. After four moves, the black
is: Black draws: King may be at c6, c7, or c8. There­
i) if he captures the white pawn, fore, the first move of the Bishop must
or take him, if possible, to a square from
ii) if he attains command of the which--0n whichever of the indicated
queening square before the pawn squares the black King rests-he can
queens. shut out the black King in one move.
For once having attained that 3) First suppose that the black
command, he can never be ousted King, after four moves, is at c7-we
from it; and so the pawn can never take this first because it looks the
queen without being taken. Against most dangerous for White: from c7
a Bishop and the a- or h-pawn, where the King threatens to go to either b6
the Bishop is not of the color of or b8. To hold both those squares,
the queening square, the King can the Bishop would have to go to a7.
always draw if he can command that But then ...Kc6, threatening to win
square-nothing worse can befall the pawn; and the Bishop must re­
him than stalemate. We leave the rest tire so as to unblock the pawn. There­
to the reader. The solution is given upon ... Kc7 and we are headed for a
below. draw by repetition.
Probably most readers found Dia­ This proves that Black draws if at
gram 1 too difficult, and they may be any stage he gets his King to c7.
interested in seeing how it can be Therefore, on its first move, the Bish­
solved by a simple chain of reason­ op must not give up command of d6;
ing-without giving a singie line of for if Black had that square, he could
moves. get to c7 in four moves.
We have already seen that White 4) Now suppose that the black
can win only if he can shut off the King, after four moves, is at c6. From
black King from a8. How is that to be there, he threatens to go to b6 or c7.
done? To protect both those squares, the
1) We cannot use the white King; Bishop must go to either a5 or d8.
o
f r in the three moves that he would His first move, therefore, must be ei­
take to get to e4, Black could play ther Bb4 or Be7. It is easy to show
his King to e6 and move his pawn, that Be7 won't do because the black
thereby allowing himself a safe path King can chase the Bishop and Black
to a8. Therefore the shutting off of will have time to make a path for his
the black King from the square a8 King by moving his pawn if neces­
must be the work of the white Bishop sary. Therefore, the only solution, if
and pawn, acting alone. any, is the unexpected 1. Bb4. We
2) What positions should the already know that this move wins if
Bishop and pawn take up? For the the black King goes to c6, for then
pawn, the answer is easy: he must Ba5-and the Bishop and white pawn
guard the light square b7, so he must completely shut out the black King.
post himself on a6. To get the pawn The white King then only needs to
to a6 will require four moves, includ- come and finish things off.

- 53 -
 CJS Purdy On The Endgame

5) But is Bb4 a solution if the black Actual play

King, after four moves, is at c8? It is,
for then the Bishop could go to d6-
and again the black King is shut out.
(N.B.: The Bishop must go to d6 as
soon as the black King goes to d8.)
The key move, therefore, is 1. Bb4.
The above makes the rest easy.

We consider this little gem of

Duras' to be an example of the ideal
type of endgame problem because it White to play and win.
is profound without being in the least
complex. Difficulty with simplicity is, 1. Qb6t Kc4
in our view, the summum bonum in 2. Qd4t Kb3
every kind of puzzle or problem. 3. Qd1t Ka3
4. Qa1t Kb3
Sam Loyd 5. Nxa5t Kb4
From an actual game, 6. Qd4t! Ka3
abandoned as a draw The King can't capture the Knight.
7. Nc4t Kb3
8. Nd2t Ka3
9. Qa1t Kb4
10. Qb2t Kc5
11. Qb6t Kd5
12. Qd6#
A pretty mate in midboard.
This ending illustrates the power
of "'1 + � since "'1 + 1it can never
do tricks like this.
White to play and win
Demonstrated by G. Koshnitsky
This occurred in a game and was
abandoned as a draw, owing to
Black's threat of stalemate. Sam 4
Loyd saw a win as soon as he looked
at the position.
1. Rc8t Rxc8
2. Na7

White to play and draw

Occurring in a game between two

-54-
 The Short Course: Practical Endgame Studies

members of the Metropolitan Social In Diagram 5 we have a magnifi­

Chess Club, Sydney, the game was cent piece of work of great practical
agreed a draw, but G. Koshnitsky set value. The position occurred in a
several other members a puzzle: find tournament at Munich, and Nimzo­
a draw for White. No one was able to vich and Bernstein worked out a
do that at the time. subtle win for White. The main varia­
White must get his King to g2, thus tion is under ten moves.
establishing the blockade. The idea is to force the black
1. c5 Kf6 pawns to advance by Zugzwang.
2. Kc4! Kxf5 Then the square c6 will become avail­
3. Kd3 Kg4! able.
If 3... Kf4, then 4. Ke2 Bd4 5. c6 Note that if 1. Rh1, then 1...Rxh1
Bb6 6. c7 Bxc7 7. Kf2; and the white 2. d8=Qt Ka7; and Black has set up
King gets to g2 easily. an impregnable fortress. Black can­
4. Ke3! not be prevented from locating his
Not 4. Ke2, as Black then compels Rook on one of the protected squares
the pawn sacrifice too soon by 4 .. . a5, c5, or c6, after which White's King
Bd4 5. c6 Bb6 6. c7 (if 6. Kf1, then 6... and Queen are easily kept at bay.
Kf3) Bxc7 7. Kf2 (not 7. Kf1 because 1. Kf6! b5!
of 7 ... Kf3) Kh3 8. Kf3 Bb6; and Black The black Rook must stay where
wins. he is-to check if Re1-e8. For ex­
4. ... Kh3 ample, if 1... Rg8? then 2. Rei Kc7 (if
5. Kf3 2 ... Rd8, then 3. Ke7) 3. Re8; and
Threatening Nf2t. White wins.
5. Bd4 2. Rc2!! b6
6. c6 Bb6 If 2... b4, then 3. Rh2! Rxh2 4.
7. c7 Bxc7 d8=Qt Ka7 5. Qa5t Kb8 6. Qe5t; and
8. Nf2t Kh4 White wins the black Rook.
9. Kg2 3. Ke7 Rh7t
And White draws. 4. Kd6 Rh6t
5. Kd5 Rh5t
Actual pla� Munich 1905 6. Kc6 Rh8
Demonstrated by If 6... Rh6t, then 7. Kxb5 Rh5t 8.
Nlmzovicb and Bernstein Kxb6 Rh6t 9. Rc6.
7. Rd2!
Not 7. Kxb6 because of 7... b4, and
5 not 7. Kxb5 because of 7... Kb7.
7. ... Rd8
If 7... Rh6t, then 8. Rd6.
8. Re2
And White wins.

White to play and win

-
55-
 CJS Purdy On The Endgame

Actual play Demonstrated by F.J. Prokop

Demonstrated by C.J.S. Purdy

If Black plays 1 RM!,

•.•

White to play and win then White to play and win

This occurred after Black's 39th The diagrammed position o c-

move in a game that decided the curred in a grune Lasker-Spielmann.
championship of New Zealand-it is 1. ... Rb4!
a fine practical exercise. The gen­ 2. Kf4 Rxd4t
eral idea is not hard to find, but there 3. Kg5
is a modification that gives a clear­ And White wins.
cut win beyond cavil in eight moves [Black will soon have to rrwve his
in the longest variation. If you don't Rook to try to stop the h-pawn.
find it, you will hardly be able to dem­ Then White plays Rxa4, and his
onstrate a forced win at all neatly. Rook and King combine to shep ­
1. Bh7! Nb2 herd the h-pawn through. E.g. 3...
If 1... Nd2, then 2. Bd3!! (the only Rdl 4. Rxa4, and the white Rook
neat way) Nxf3 3. Bxb5t Kc7 (the will soon be able to get behind the
King must head off the passed pawn) h-pawn-or else it will be able to
4. Ke6! Nxh2 5. Be2. This shuts out shelter the King from checks on the
the Knight, and Black has no move. g-file. -F.PH.]
2. Bf5t!
Driving the black King to one side Actual play
or the other.
2. ... Kc7
If 2... Ke7, then 3. Bc2! Na4 4. Bd3
Nc3t 5. Kc5; and White wins on the
other side.
3. Bc2!
Shutting out the Knight.
3. ... Na4
4. Ke6 Nc3
5. Kxf6 Na2
6. Kg6! Nxb4 White to play/What result?
7. Be4
And White wins easily. Black wins! After 1. Ng4! Bxg4 2.

-56-
 The Short Course: Practical Endgame Studies

Kf1, the fly in the ointment is 2... 9. Rg8t Bb8

Be2t! The Bishop is now forced off the
[A1wther way to win after 2. Kf1 i,s critical diagonal.
2... e3 3. fxe3 BrJ 4. Kgl Bxd5 5. Kh2 10. Kb5 Kb7
Bg2, and Black will advance hi,s 11. Rg7t Bc7
King and the g-pawn. -F. P.H.] If 11... Ka8, then 12. Kb6 Bh2 13.
The original position was given as a4 Bf4 14. Ra7t Kb8 15. Re7 Ka8 16.
a win for White in the London Ob- a5; and this process is repeated, of
server. preventing a check, till the pawn gets
to a7.
Actual play 12. a4 Kb8
13. Ka6 Bf4
14. Rb7t Ka8
8 If 14... Kc8, see the first note.
15. Rf7 B-moves
Then the Rook attacks the Bishop
until safe for a ("far away") check
on the eighth rank, the Bishop inter­
posing. Then
16. Kb5 Kb7
17. R t o 7th rank t
White to play and win 17. . .
. Bc7
If 17... Kc8, see the first note; and
This position occurred in a New if 17... Ka8, see the fifth note.
York Chess League match. It needs 18. a5 Kb8
fine thinking, but the Goldstein­ 19. R to 8th rank t
Crowl ending, together with C.G. 19. ... Kb7
Watson's game, should help-see 20. a6t
chapter 18. And White wins.
1. Kc7 Bt2
2. Ra.'Jt Ba7 Actual play
3. Kc6 Kb8 Demonstrated by K. Richter
4. Rb3t Ka8
If 4 ... Kc8, then the pawn marches
on to a7, supported by the Rook at
b7--one of the positions Black must
avoid.
5. Kb5 Kb7
Stopping Ka6.
6. Ka5t Ka8
7. Ka6 Bg1
8. Rg3
Wherever the Bishop goes in this White to play and win
direction, it would be attacked.
8. . . . Bh2 T h i s o c c u r r e d i n the game

- 57 -
 CJS Purdy On The Endgame

Schallop-Blackburne, Frankfort 3. AB a first step, White can easily

1887. It was given up as a draw, get his King to d4 and his Knight to
Schallop seeing no way of escaping e5 (simply by Kc5 followed by Nd6
perpetual check or stalemate. The and Nc4t, etc.).
German master K Richter, however, 4. Imagine that White has got his
proved that White can win. King to d4 and his Knight to e5, that
1. Kxe1 Rf1t Black has moved last, and that it is
2. Ke2 Hett now White's move. Both of Black's
3. Kf3 Re3t pawns are under fire, so Black must
4. Kg4 Re4t be defending both.
5. Kh5 Rh4t 5. If the black Bishop is defending
6. Kg5 Rh5t both pawns, then it must be at f3,
7. Kg4! Rh4t whereupon White wins by exchang­
8. Kf3! Rf4t ing pieces, or at f5, whereupon White
9. Ke2-d.2-c2 wins the e-pawn by Nc4-d2. There­
And White wins. fore, the King must be defending at
least one pawn.
6. Whenever White plays Nc4 or
Nf7, Black must play ...Ke6 to stop
Nd6 (proposition 2).
7. By propositions 5 and 6, Black's
King must be at f5. Thus the only
variable factor is Black's Bishop.
8. The Bishop must be protecting
the e-pawn, to give the pawn double
protection, else White wins the
e-pawn by 1. Nc4 KE!a (forced) 2. Kxe4
Either plays/White wins (or 2. Nd2 first).
9. The Bishop can protect the
In this case it is possible to give e-pawn from the Northwest, South­
an absolute or mathematical demon­ west, or Southeast (but not from the
stration, much more pleasing than Northeast since the black KinWf5).
the ordinary empirical method popu­ 10. If the Bishop protects from the
larly termed "analysis." We proceed Northwest or Southwest, then White
by a chain of propositions. wins by 1. Nf7 Ke6 (forced) 2. Nh6!
1. It is clear that Black can never since one pawn must fall.
afford to exchange the minor pieces, 11. If the Bishop protects from the
as the black pawns would fall rap­ Southeast, the only squares are g2
idly. and h1 since f3 is ruled out by propo­
2. The obvious winning position sition 5. White now forces his Knight
for White is: King at d4, Knight at d6. to d6 by Kd5! White will then bring
If White's pieces ever attain these po­ his King back to d4 (if necessary) and
sitions, White automatically wins the win by proposition 2.
e-pawn since Black's King could not Propositions 8, 9, 10, and 11 prove
then be at either e5 or f5. that White wins in any possible posi-

- 58 -
 The Short Course: Practical Endgame Studies

tion-Q.E.D.

12
11

White to play
Either plays/White wins What outcome?

The trick in this ending is that the Draw!

black King is cut off from the pawn. If 1. Kb5, then 1... Re5t 2. Ka6 (or
If his King were across the fourth Kc6) Re6, followed by ...Rxb6!! with
rank, White would have a book win: stalemate if the Rook be taken.
Any pawn but an a- or h-pawn, If 1. Kb3, then 1... Re3t 2. Kc2 Re2t
beyond the mldline and supported 3. Kc3 Re3t 4. Kd4 Rb3 5. Ra6 (if 5.
by its King, wins if separated by Kc5, then 5... Rxb6) Rb1; and Black
more than one file from the enemy draws simply by keeping his Rook
King. on the b-file (if 6. Kc5, Black doesn't
even need to check).
See Chapter 1. If the black Rook stands on d4 in­
stead of e4, White wins by 1. Kb5
Here the first few moves are the since 1... Rd5t is useless (2. Kc6);
obvious ones, b nt the final winning and if 1... Rd7, the white Rook takes
maneuver is pretty. an open file, with an easy win. A Rook
1. Kd3 Rb4 likes to check at a distance.
Or 1... Kh7.
2. e6 Rb6
3. Ret Rb8 "Sister Squares"
4. e7 Re8 [The studies di,scussed in Purdy's
5. Kd4 Kg'7 (1937) article below were composed
6. Kd5 Kf7 in the late nineteenth and early
7. Kd6 Ra8 twentieth centuries and have sub­
8. Ritt Kg'7 sequently appeared in a number
9. Rat! of publications. It is not clear
And White wins. whether the label "s'ister squares"
was coined by Purdy, but hi,s ex­
planation of the concept and hi,s
commentary on the examples ap­
pear to be original.
Znosko-Borovsky, in his "How
to Play Chess Endgames" (1940)

-59-
 CJS Purdy On The Endgame

uses the label "related squares," C.D. Locoek

which is also mentioned by Euwe
and Hoaper in '!4. Guide to Chess
Endings (1959)." Averbach and 15
Maizelis in "Pawn Endings" (En­
glish translation 1974) use the la­
bel "corresponding squares," and
they credit Grigoriev (1922) with
first presenting a theoretical ba­
sis, later further develaped by oth­
ers-Averbach andMaizelis devote
a chapter of 40 pages to the devel­ White wins
opment of the theory of coordinate
squares. All three of these studies can be
Clearly the tapic is complex, but solved in the same way-that is, by
Purdy's e:rplanation should more the method of "sister squares," an
than satisfy most readers. -F.PH.] important idea for certain kinds of
pawn endings.
F.Sackmann A pair of sister squares is a pair
of squares such that if the attacker's
King moves to one square of the pair
then the defender must move his King
to the sister of that first square or
lose the game. A corollary is that if
the attacker can move his King to a
square that is the sister of the square
occupied by the defender's King then
the attacker wins.
In Diagram 13, an obvious pair of
White wins sister squares is d6 & b6. For if White
plays Kd6, then there is only one
square that the black King can go to
Lasker and Reichhelm if Black is to save the game, b6.
Another obvious pair is d7 & b7. A
third pair is d8 & b8. For if the white
14 King ever gets to d8 and if the black
King thereupon goes to any other
square but b8, then White can rap­
idly force an entry at c7 (or c8 first)
and can win easily.
A fourth pair is e7 & a7. For e7
commands d6, d7, and d8. There are
two squares that command b6, b7,
White wins and b8, namely a7 and c7. But a very
little calculation shows that if the

-60---
 The Short Course: Practical Endgame Studies

white King ever gets to e7 and if the would say-but Black can't. Here is
black King thereupon goes to c7 then the proof.
White will rapidly win by playing his 1) When the white King goes to
King to e6; therefore, a7 is the only c4, the black King must go to b�
sister square of e7. the square a6 will not do because the
A fifth pair is e8 & a8. For the white white King would thereupon run
King at e8 commands d7, dB, and e7; around to g5.
so the sister square must be the one 2) When the white King goes to
that commands b7, b8, and a7. h4, the black King must go to g'&­
Yet another pair is f7 & b7 (by sim­ again not to h6.
ilar calculation). True, if Kf7, then 3) Therefore c4 & b6 and h4 & g6
Black would draw if his King could are pairs of sister squares.
go to d7; but we assume that the 4) At d3, the white King can get to
black King cannot carry out such a c4 in one move and to h4 in four
maneuver, since the white King could moves; so when the white King goes
go to f7 via e7 or e8---and obviously to d3, the black King must go to c7
the black King could not get from a7 (to be able to get to b6 in one move
or a8 to d 7 in one move. and to g6 in four moves). It follows,
Still another pair is f6 & b6. For therefore, that d3 & c7 are a pair of
the white King on f6 commands e6, sister squares.
e7, and f7. 5) By similar reasoning, we can
Finally, we arrive at f5 and­ find in succession the following pairs:
what? From f5 the white King com­ c3 & b7, d2 & c8, c2 & b8, b2 & a8 (or
mands e6 and f6; and if Kf5, then the c8), bl & a7 (or c7)-and more if we
black King must go to a square com­ like.
manding both a6 and b6. But if 1. 6) The last-mentioned pair tells us
Kf5! , then Black cannot do as re­ that White must win by
quired. 1. Kb1!
Therefore, without the slightest since the black King cannot go to ei­
room for argument, White wins by ther a7 or c7 on his next move!
1. Kf5! The demonstration is absolute.
Furthermore, 1. Kf5 is the only
solution; for any other move enables
Black to move his King to a sister
square and to repeat the operation In practical play, it would be al­
for each of White's subsequent most impossible to keep all the pairs
moves. in one's head; but if we look at the
various pairs on the board, we find a
geometrical relationship common to
every pair.
An amazing affair is Diagram 14 Thus, in Diagram 14 the sister
by Emanuel Lasker and Reichhelm. squares are always of different col­
Clearly, White can win if he can ever ors and on adjacent files. In other
get his King to b5, g5, or h5. Surely words, all relationships are simply
Black can keep White out, anyone extensions of the key relationships,

- 61 -
 CJS Purdy On The Endgame

namely c4 & b6 and h4 & g6. For con­ threatening Kc4. So

venience, let us call the relationships 5. .. . Kc7
"Knight's-move" opposition. is forced, and now we can take
Just as in "opposition studies" "Knight's-move" opposition again by
such as Diagram 13, we have the ex­ 6. Kd3!
tension of close direct opposition to but on the third rank-a consider­
distant direct opposition. For in Dia­ able advance. Then
gram 13 the sister squares are al­ 6. ... Kb7
ways of the same color and on the 7. Ke3!
same rank or file. White soon reaches h5 and wins.
The reason that the "opposition"
solves nearly every study of blocked
pawns is that the key relationship is
nearly always a form of close oppo­
sition (as in Diagram 13).
Our theory is that you simply have
to find the key relationship and then
extend it, like a telescope, and that
there is no need to keep on working
out new pairs of sister squares­
once you have worked out enough to
"see" whatever geometrical system
governs the situation.
Using our theory to solve Diagram White wins
14, we have only to keep Black out of
"Knight's-move" opposition. We can Now we come to Locock's famous
do that by simply assuming it our­ study, Diagram 15. Clearly, e3 & g5
selves, but we must be on the look­ and d4 & f6 are pairs of sister
out for other ways-for we must squares. From this, another pair is
ADVANCE whenever we can. obviously d3 & g6.
For example, after Note that on the two pairs of dark
1. Kb1! Kb7 squares in question simple diagonal
2. Ket! Kc7 opposition holds, but with the light
3. Kd1! (suppose) squares we have a different relation­
3. Kd7 ship because of an intervening pawn.
Then we can afford This relationship is: squares of the
4. Kc2! same color, but with an even number
Eureka! An advance! True, Black of rows intervening. Let us call this
can take "Knight's-move" opposition "even-number diagonal opposition."
by retiring to d8: According to our theory, these re­
4. ... Kd8 lationships should hold all over the
(actually, the only sister square to board. That is, sister squares will al­
c2 is b8), but then we can again ad­ ways be of the same color; but in the
vance by case of dark squares we expect to
5. Kc3! find that the relationship is always

- 62 -
 The Short Course: Practical Endgame Studies

one of diagonal opposition (odd imm­ with an odd number of intervening

ber of squares intervening), whereas squares. In diagonal and oblique op­
in the case of ligilt squares we ex­ position, the Kings are not on the
pect to find always a relationship of same rank or file; and there is not
"even-number diagonal opposition." only an odd number of intervening
For example, we expect the sister ranks, but also an odd number of in­
square of c3 to be g1 and of c2 to be tervening files. Thus, there are three
h7. Other pairs, we should expect, forms of real opposition: direct, di­
are c4 & f7, d2 & h6, e2 & h5, b3 & g'S, agonal, and oblique.)
b2 & h8, b4 & f8, a2 & f7, a1 & gl, bi So, let Black try
& g6. 6. ... Kf7
That is what we do find if we take 7. Kd2!
the trouble to puzzle out the sister Now Black cannot take diagonal
squares as we did for Diagram 14, it opposition (at h6), so the best he can
is just a little mathematical problem. try is oblique opposition by ...Kf6.
On our theory, the solution now 7. . .
. Kf6
plays itself. Black, with the move, 8. Ke2!
would draw easily by 1... Kg7! On the Threatening Ke3-f4, so Black must
other hand, White, with the move, play
wins by 1. Kb1 ! White must just keep 8. . .
. Kg6
on going either to a sister square or, 9. Kd3!
failing that, to a square that does not And now White wins easily.
permit the opponent to take a sister If you work out the sister squares,
square. you will find that where the limits of
The play migilt go: the board do not permit of distant
1. Kb1 Kh8 diagonal opposition, oblique opposi­
2. Kb2! Kg8 tion takes its place-e .g. di & g6 or
3. Kb3! c1 & gl.
An advance, preserving at the Locock's study is a beautiful example
same time "even-number diagonal of profundity and simplicity.
opposition." •

3. Kf8 Actual play

4. Kb4 Kg8 Berger--Chigorin
5. Kb3 Kf8
Now we puzzle hard. We tried Kb4
before, and it has led to repetition. Is
the whole thing a frost? No-try
6. Kc2!
If now 6... Kg8 or 6... Ke8, Black
will have real (odd number) opposi­
tion on a ligilt square-which is no
good to him.
(This is an example of "oblique
opposition." In direct opposition, the White to play and win
Kings stand on the same rank or file,

- 63 -
 CJS Purdy On The Endgame

1. Bf4 Nd5 3. a4 Kxg4

Not 1... Nxh5 because of 2. Be5. 4. b4 axb4
2. Bg3 Nb4 5. Kd3 h5
If 2. . Ne7, then 3. Kd7.
. 6. a5
3. Be1 And Whit e wins.
And White wins. •

•
Actual Play
Actual play (White resigned!)

White to play and wtn White to play/Draw

1. Bb3! a4 This position occurred in a corre­

2. Kc4 axb3 spondence game in which White re­
3. axb3 signed. The late T.D. Clarke of Grass­
And White wins. dale, Victoria, sent the position to us
•
many years ago, pointing out a draw
Actual play for White-pretty enough to have
Berger-Bauer 1889-1891 made a composed study by Gregor-
iev.
1. a4! Kxa4
2. Kc2 Ka3
3. c4!
And Whit e draws.
•

Actual play

White to play and wtn

This was the finish to a con-espon­

dence game by the great endgame
writer Johann Nepomuk Berger
(1841>-1933).
1. c4! bxc3 e.p.
2. Ke3 Kg5 White to play and win

-64-
 The Short Course: Practical Endgame Studies

I n Diagram 20, Bogoljuboff 2. Kc4 Ra4t

(White) agreed to a draw against 3. Kb3 Raat
Thomas, Hastings 1922. Bogoljuboff 4. Kc2 Rc3t
could have won. How? 5. Kb2
1. Kd5 Ra5t And White wins.

- 65 -
 PART II

ALL FACETS OF THE ENDGAME

 CJS Purdy On The Endgame

Chapter�
Road To The E1\d4an\e

Playing For The Endgame who could fail to be enchanted by

M.E. Goldstein Reti's Knight maneuver Ne3-d1-c3-b1
in the very heart of the enemy camp?
A series of articlesdevoted to the The opening concerns us only indi­
.ft.transition from the middlegame rectly, as showing how surrender of
to the endgame. the center may sometimes be tanta­
mount to resignation. The game was
played at Mahrisch-Ostrau 1923,
No. 1 Reti then being at the height of his
The Pawn Center powers. The following year at New
Domination of the center is the York, Reti defeated Capablanca-the
real aim of any sound opening. The latter's first tournament loss in eight
player with a strong pawn center can years.
maneuver freely behind his pawns,
switching his pieces from one wing Bogoljuboff-Reti
to the other; and his opponent, Fren ch Defense
cramped for room, is at a serious dis­ (Gledhill Attack)
advantage. The following game is an 1. e4 e6
admirable example of how a strong 2. d4 d5
pawn center by itself ensures a fa­ 3. Nc3 Nf6
vorable middlegame. When first play­ 4. e5 Nfd7
ing over the score for use in a lec­ 5. Qg4
ture at the Hampstead Chess Club, I Introduced by W. Gledhill of Leeds
was immensely impressed. Replay­ half a century ago. It is hardly sound,
ing the game twenty years after­ as Black's g-Knight has already been
wards, I am still more fascinated at driven to a good square for the coun­
this peerless example of Richard terattack.
Reti's genius. 5. . .. c5
Tarrasch once sagely wrote that 6. Nb5
"Chess, like music, like love, has the In conjunction with his eighth
power to make men happy"-and move, this is Bogoljuboff's idea. It is,

-68-
 All Facets Of The Endgame: Road to the Endgame

of course, against sound principles

to move a developed piece a second
time so early in the opening. The 2
move 6. Nf3, played by Marshall
against Maroczy at Lake Hopat­
cong, USA, 1926, is also inadequate.
6. ... cxd4
7. Nf3 Nc6
Not 7 ... Bb4t; for after 8. c3!, the
black Bishop has to return home with
the loss of a tempo. Study this position carefully. The
8. Nd6t Bxd6 amateur counts noses, finds equal
9. Qxg7 pawns and Bishops of opposite color,
and dismisses the game as an obvi­
ous draw. The master reasons dif­
ferently-he sees a strong pawn cen­
ter depriving the white minor pieces
of good squares and sees open b- and
g-files for the black Rooks. That was
the price paid by White for the re­
covery of his gambit pawn.
16. ... c5
17. Ne2 Kf7
White expects the reply .. .Rf8, 18. f3?
whereupon exd6 would give him a Fatally weakening e3, upon which
fine game. square the black Knight soon settles.
9. ... Bxe5! But how can White develop his
Lasker's deperado motif. As this Bishop? If the Bishop is played to the
Bishop is doomed anyhow, he sells Kingside, it is promptly harried by
his life as dearly as possible. the advancing pawn phalanx, where­
10. Nxe5 Qf6! as b3 weakens the square c3. If 18.
The keystone of Black's defense. Ng3, then 18... Nd6!
The Queens are forced off, leaving 18. . .
. Nd6
Black's pawns in possession of the 19. b3
center. Now c3 will soon be occupied by
11. Qxf6 Nxf6 Black (see Black's 27th move).
12. Bb5 Bd7 19. e5
13. Nf3 Ne4! 20. Ba3 Rac8
Freeing the f-pawn. 21. Rad1 d4!
14. 0-0 f6
15. Bxc6 bxc6!
16. Nxd4

- 69 -
 CJS Purdy On The Endgame

The first step forward, securing

outposts for the black Knight at c3
and e3. Also, . c4!, opening a file for
..

the black Rook, is now imminent.

22. Nc 1 Nf5 easily.
23. M Ne3 30 • .•• Bxb3
24. Re1 c4! 31. a.xb3
25. b4 Still the black Knight cannot be
Forced, as the c-file must be kept captured, as ... Bxa2 and ...cxb2 would
closed at all costs. The threat was follow.
... cxb3 and ... Nc2. Now White's Bish­ 3t. ... Nd2!
op shares the lack of mobility of his 32. Ret Rhd8
other pieces, and the black Bishop 33. Bet d3!
intrudes with a threat. The entry of the d-Rook is quickly
25. ... Ba4 decisive.
26. Ree2 34. cxd3
The Laughing Cavalier is still safe;
for if 34. Bxd2, then 34... cxd2 35.
4 Rxd2 dxc2; and the passed pawn
costs a Rook.
34. ... Rxd3
35. Bxd2 Rxd2
36. Rat Ke6
37. Kf1 Rxr.!t
38. Kxf2 c2
39. Rct Kd5
26 . ... Nd1! 40. Ke3 Rc3t
Skillful maneuvering of his minor 41. Kd2 Kd4
pieces is characteristic of the mod­ 42. h4
ern master. The pawn ending after 42. Rxc2
27. Rf1 Nc3 is hopeless. Black would exchange
28. Rer.! Nb1! Rooks and then play .. Ke3, ... f5!, and
.

29. Bb2 c3! ... e4.

42 . ... Rd3t!
43. Resigns.

-70-
 All Facets Of The Endgame: Road to the Endgame

The response 43. Kxc2 is forced At first glance the position ap­
(else ... Kc3!), whereupon ... Rc3t, pears quite equal. Black's Knights
...Rxc1, and ...Ke3 (or ...Kc3) wins, have good mobility, whereas White's
as in the previous note. Bishop is rather restricted. That sug­
The heart of a true artist such as gests that White may try to exchange
Reti would be uplifted by this mas­ Bishop for Knight at c4 or e4 at some
terpiece. later time.
Black's doubled g-pawns may
No. 2 prove a handicap in a Knight ending
Book-Petroff, Kemeri 1937 or a simple pawn ending because of
Playing for a win in an equal posi­ the lack of mobility of the rearmost
tion, while keeping the draw in hand, pawn. In compensation, the pawns
is one of the hardest problems in are doubled toward the center, con­
chess. Transformation of the seeds trolling important central squares.
of minute potential advantages into In summary, Black has the initia­
the flowering tree of victory requires tive-that is all. The win is a conse­
cool judgment, acquired only by quence of superior endgame tech­
years of practical experience. nique plus planless moves by White.
The analogous task, drawing a The first step forward is to create
game in which the opponent has a chances on the queenside. So,
slight advantage, is even more diffi­ 27. ... a4!
cult. Euwe himself blamed his loss in Threatening ...a3! at a judicious
the return world championship moment (see Black's 38th move). If
match against Alekhine on his lack then bxa3, White's a-pawn becomes
of a drawing technique. isolated and weak; and if b3 (or b4),
In such cases passive defense is the a-pawn is weaker still, being
usually futile, as the opponent can threatened with capture (see Black's
build up his position step by step. 54th move).
Consider the typical position that 28. Ne2 Ne6!
resulted after White's 27th move in Black is playing for a win; hence
Book-Petroff, Kemeri 1937. he avoids simplifying exchanges.
Also, Black can now centralize his
Petroff King for the ultimate breakthrough.
29. Net Ke'i
30. Nd3 Kd6
31. Bd1 b5
32. Bg4
Loss of time, merely provoking
Black into playing a good move.
32. ... f5
33. Be2 g5!
The despised doublet advances
with line-opening threats, also con­
Book trolling the important f4 square.
Black to move 34. Bd1

-71-
 CJS Purdy On The Endgame

Moving aimlessly, White obviously passed e-pawn. Moreover, Black's

assumes that a draw is his birthright. strong e- and f-pawns control the im­
34. ... Nf4! portant squares d3, e3, f3, and g'3.
Removing White's best posted 41. Kd2 Kd5
piece, his Knight. Now 35. Nxf4 leads 42. Kc2 Kc4
by transposition into the actual game. 43. Rd1 Rd8
35. Bc2 g6 Else follows 44. d5! cxd5?? 45.
36. Nxf4 gxf4 Rd4#!
37. Kf1 Nf6 44. Rd2 g5!
38. Ke2 45. Rd1 Rd7
46. Re1 Re7
Black prepares the final liquida­
7 tion by ...e3!-when he has overcome
his clock-trouble.
47. Rd1 Kd5
Directed against 48. d5!
48. Rf1 Re6
49. Kd2 Kc4
50. Re1

38. ... a3!

39. b4 8
Compare the note to Black's 27th
move. Here 39. b3 looks preferable,
as it retains control of c4, a square
soon to be occupied by the black
King. After 39. b3 Black's task would
be much harder. Black could open
files for his Rook by ... Rc8 and ...c5!,
or even by ...Rg8 and the advance of 50. ... e3t!
the g-pawn. That, in conjunction with The winning move, requiring ex­
the superiority of his Knight over act calculation. After the forced ex­
White's Bishop, would give Black change of Rooks, White queens his
good winning chances. h-pawn at the same time that Black
39. ... Ne4! queens his a-pawn. But closer inves­
Here 39... Re8t is plausible but in- tigation reveals that the resulting
effective. For example, 40. Kd2 Rxe1 Queen ending is hopeless for White-­
41. Kxe1 Ne4 42. f3! Nxc3 43. Bb3 his remaining pawns are weak and
any move 44. Kd2 Na4 45. Bxa4, with his King is seriously endangered.
a safe draw for White. Before the student can expect
40. Bxe4 dxe4! to play endgames well, he must
Black has greatly improved his be able to count correctly. Arith­
prospects. His three pawns hold metic is as important in chess as
White's four queenside p awns, in our everyday work.
whereas he can at any time force a 51. fxe3 Rxe3

- 72 -
 All Facets Of The Endgame: Road to the Endgame

52. Rxe3 fxe3t from the tournament at Mar del

53. Kxe3 Plata, 1942. On his 9th move, Stahl­
Of course 53. Kc2 is much worse. berg decides to obtain the Bishop
For example, 53. Kc2 e2 54. Kd2 Pair. He then systematically ex­
e1= Qt, and then the black a-pawn changes off as many pieces as pos­
will queen with check. sible so that his Bishops may preach
53. Kxc3 more eloquently. Playing as Black is
54. g'J! Kb2 the German master, P. Michel.
55. h4 gxh4
56. gxh4 Kxa2 StAhlberg-Michel
57. h5 Kb3 Queen's Gambit Declined
58. h6 a2 (Slav Defense)
59. h7 a1=Q 1. d4 d5
60. h8=Q Qg1t 2. c4 c6
A few checks to drive White into a 3. Nf3 Nf6
mating net. 4. Nc3 dxc4
61. Kd3 Qftt 5. a4 Bf5
62. Kd2 Qf2t 6. e3 e6
63. Kd3 f4! 7. Bxc4 Bb4
Black calculated as far as this on 8. 0--0 0-0
his 50th move. Now two separate 9. Nh4
mates are threatened, at c2 and e3. Preparing the way for his Two
64. Resigns. Bishops. The text, introduced by
For if 64. Ke4, then 64... Qe3t; and Stahlberg, is a welcome change from
the f-pawn queens. Or if 64. Qg8t , the usual 9. Qe2.
then 64 ... Kb 2! 9. Ne4
Played with superior skill through­ 10. Nxf5 exf5
out by the youthful Petroff who tied 11. Nxe4 fxe4
for first place with Reshevsky and 12. Qg4! Qe7
Flohr, ahead of Alekhine and Fine. 13. f3!
Removing the black e-pawn which
hinders the formation of a strong
No. 3 White pawn center.
The Two Bishops 13. ... exf3
Since the days of Steinitz, the su­ Not 13... Re8? because of 14. fxe4,
periority of two Bishops over Bishop winning.
and Knight has been well recognized. 14. e4! Qd7
Masters such as Janowsky and Ru­ Black spots the trap: 14 ... fxg2?
binstein were famous for their Two 15. Rxf7! and 16. Qc8t. [Ed.: If 16. ..
Bishop games three decades ago. Qf8, then 17.Qxb? Nd? followed by
Nowadays the technique of exploit­ 18. Qxd?.]
ing the Bishops is an essential weap­ 15. Qxd7 Nxd7?
on in the armory of every well-tr_ained Would you expect this natural
master. move to be Black's undoing? Correct
I have chosen a typical example is 15 ... f2t, depriving White's e-pawn

-73-
 CJS Purdy On The Endgame

of support by the f-pawn. The pawn­ move seems safest. Compare Black's
move would permit counterplay on 27th move in No. 2: Book-Petroff.
the open e-file. If hereabouts White 18. Rd1 Rad8
plays e5, Black gets a good post for 19. Be3 Rfe8
his Knight at d5 and can later on at­ 20. Rd3
tack the backward d-pawn. Stopping ...Nd5.
16. gxf3 20. ... Rd'7
21. Rb3! Bf8

Study this position carefully.

White's opening strategy has been Black has taken up his strongest
crowned with success-White re­ defensive position and confidently
tains his Two Bishops, together with awaits developments. We see that
an impregnable Pawn Center. Behind White can at any moment force a
his center pawns, White can regroup passed pawn (d5!). This pawn will
his forces at will-switching them prove especially dangerous, as every
from one flank to the other-seeking square on its way to the eighth rank
a favorable opportunity for a break­ is controlled by the Bishop pair.
through. But there is no need to hasten this
16. ... Nf6 particular move, which is momen­
Unfortunately, Black cannot main­ tarily hindered by the tactical threat
tain his well-posted Bishop at b4. If to White's loose c-Bishop. StAhlberg
16... a5!?, then 17. Be3 Nb6 18. Bb3, utilizes a second property of the Two
threateningd5! Black must therefore Bishops, that of cramping the enemy
withdraw his Knight, enabling White, position by reducing the mobility of
after suitable preparation, to force a the Knight. First the c-Bishop is
passed pawn by d5. Note that if brought onto the more effective diag­
White's b-pawn were already at b3, onal, c8--h3.
that square would not be available 22. Bf1! Red8
for the White pieces (see move 21). 23. Bh3 Re'7
17. a5! a6 24. Rd1 h6
As the late Spencer Crakanthorp 25. Kf1!
used to say, Black is "on the horns of Cannily avoidinga possible check
a quagmire." Shall we allow a6, or is on the long dark diagonal. When you
...a6 necessary, leaving a backward have built up a winning position stra­
b-pawn and a new weakness at b6? tegically, that is the time to guard
All things considered, the text against tactical surprises.

-74--
 All Facets Of The Endgame: Road to the Endgame

25. ... Ne8 tin.nation for the benefit of students

who have not previously met such
endings: 36. b4 Rc1t 37. Kg2 Rb1 38.
11 b5 axb5 39. a6, followed by a7 and
Rb8t.
34. Rxb7 g5
Although material is still equal,
White has the "Seventh Rank Abso­
lute" (Nimzovich).
35. Rg7t Kf8
36. Rh7 Rd8
Black is virtually in Zugzwang, 37. Rxh6 Rd1t
as his Rook at e7 is tied down to the 38. Kg2 Bd2
defense of his backward b-pawn. To save the threatened a-pawn.
Black hopes to post his Knight at b5 39. Rxe6 Rat
(via d6), shielding the weak pawn 40. Kh3 a5
from frontal attack. White now 41. b4 Ra3
breaks through into the hostile camp. 42. Rf6t
26. d5! cxd5 And White ultimately won.
27. Bb6! Ra8 A masterly game by a master of
28. Rxd5 Nc7 the endgame.
29. Rd7! Ne6
30. Bxe6!
Parting with one Bishop, to occupy No. 4
the seventh rank with a Rook. Pawn Endings
30. fxe6 Ak:iba Rubinstein is considered by
31. Rxe7 Bxe7 Reuben Fine to be the greatest end­
32. Bc7 Bd8 game player of all time. In his golden
33. Be5!! age, 1907-1914, the phrase "Rook
Much stronger than the obvious ending" was synonymous with Rubin­
pawn win: 33. Rxb7 Bxc7 34. Rxc7 stein, his artistry in this branch of
Rb8!, with counterchances for Black. the game being outstanding.
Endgames with Rooks only are very The first diagram is taken from
tricky, two pawns extra being some­ the game Erich Cohn-Akiba Rub­
times insufficient to force a win . With instein, St. Petersburg 1909. Black
the forced loss of the b-pawn, Black's has just exchanged the last remain­
game becomes untenable. ing minor pieces at c4, and it is time
33. ... Bxa5 to take stock.
The alternatives, ... Ra7 or ... Bf6,
are no better. If 33... Ra7, then 34.
Bb8 Ra8 35. Rxb7 Bxa5 36. Be5!,
transposing into the actual game. Or
if 33... Bf6, then 34. Bxf6 gxf6 35.
Rxb7 Rc8; and White has a quick win
by pushing the b-pawn. I give the con-

- 75 -
 CJS Purdy On The Endgame

Rubinstein Leading to an endgame study,

"Black to play and win."
26. ... Kg5
27. Ke2
Simple counting shows that if
White plays his King to c7 then
Black's h-pawn queens too soon.
27. Kh4
28. Kf1 Kh3
29. Kg1 e5
30. Kh1
Cohn When exchanging Rooks, White
After Black's !3rd move presumably had foreseen this posi­
tion. By oscillating between g1 and
Analysis shows that material is h1, White vainly hopes to keep Black
level, Black having a slight positional at bay.
advantage owing to White's doubled 30. ... b5!
pawns and undeveloped Rook. Always a good move in such cases.
Wherein lies the weakness of White's Before commencing the decisive
kingside pawns? In the first place, breakthrough, Black fixes the queen­
the h-pawn is isolated and open to side and retains a tempo (a6) for
attack by the black King penetrating emergencies.
to h3 (if White still had a pawn at g2, 31. Kg1 f5
guarding the square h3, the position Black's King can make no further
would be absolutely equal). Secondly, progress until White's defending
White cannot advance his center pawns are exchanged-the more fa­
pawns without further weakening his vorable position of the black King will
position. If White plays e4, then the then settle matters. Thus the win­
square f4 becomes an entry point for ning plan is to force g4 at a mo­
Black; and the move f4 weakens, to a ment when White's King is on ht,
lesser extent, the squares e4 and g4. as far away from the queenside as
Since White has the slightly infe­ possible.
rior position, he should retain his 32. Kh1 g5
Rook. Emanuel Lasker points out in 33. Kg1 h5
the tournament book that 24. f4! (pre­ 34. Kh1
venting the mobile black Rook from
playing to h4) Kf6 25. Kd3 b5 26. Rg1!
would still draw, as Black can make
no headway.
Confident that he could draw the
pawn ending, White instead offered
to exchange Rooks by
24. Rc1? Rxc1!
25. Kxc1 Kf6
26. Kd2

-76-
 All Facets Of The Endgame: Road to the Endgame

34. ... g4! 42. Kfl Ke3 43. Kel Kd3 44. a4!!;
35. e4 and if Black takes, White's King
Or 35. f:xg4 hxg4 [Ed.: K.M. points reaches the corner-after which,
out that 35... fxg4 wins more eas- possessing a-pawns only, Black
ily.] 36. Kg1 f4 37. exf4 exf4 38. Kh1 cannot win.If 44... a6, to maintain
g3! [Ed.: Best move says Purdy, but the b-pawn, then 45. axb5 axb5 46.
a big mistake says K.M. Instead Kdl; and when Black captures on
38 ... f3! wins-see Nunn's "Tacti- b4, White'sKing will be able to take
cal Chess Endings," pgs. 166-167.] the opposition on b2, with a book
This characteristic maneuver should draw. Note that if Black's a-pawn
be memorized. Black occupies the were already on a6 he would not
third rank and then switches his King have to waste a move protecting
over to the other flank-against his b-pawn, and White's King
which White has no defense: 39. fxg3 would then arrive too late to take
fxg3 40. hxg3 Kxg3, etc. the opposition.
[F.PH./J.S.P.: Purdy assumes the As K.M. indicates above, Black
position reached is won for Black, had opportunities to win differ­
as have other commentators, see ently, at moves 35 and 38, but the
e.g. "Encyclopedia of Endgames" instructive point is that the flaw
(1982), and his comment after 38... in the simple plan given by Purdy
g3 would indeed be valid in almost and others was obscure enough to
all similar positions. So although deceive commentators for many
it is not too difficult to work out decades. Who would suspect that
that 38... f3! wins, why is 38 ... g3 Black wins with his pawn on a6,
an error? It appears to win also, but only draws with it on a7,
but when first looking at this where it appears to give Black a
manuscript, J.S.P. immediately normally-useful spare move.It all
recognized the ending as one in depends on seeing the tactical
which he had noticed an analyti- point 44. a4! !-a point arising
cal flaw some decades previously. only through a quirk of pawn end­
Can you spot the draw for White ings.
after 40 ... Kxg3 without reading It seems 35. fxg4 would have
further? been a better practical chance than
the move played, which also loses
quickly. Had the situation arisen,
we do not suggest that the great
endgame expert Rubinstein would
have been deceived, but many
lesser players might well have
been, and equally, many players
of White might have resigned af­
ter or before 40 ... Kxg3.]
35. .. . fxe4
By pulling a rabbit from the hat, 36. fxe4
White draws as follows : 41.Kgl Kf3 No better is 36. fxg4 hxg4 37. Kg1

- 77 -
 OS Purdy On The Endgame

e3!! 38. fxe3 e4! 39. Kh1 g3! Black has an easy win by ...Kb5, ex­
36. ... h4 pelling the blockadingRook, followed
37. Kg1 g3 by the rapid advance of the a-pawn.
38. hxg3 hxg3 In such cases, passive resistance
39. Resigns. is hopeless. White must at all costs
For if 39. f4, then 39 ... exf4 40. e5 create counterchances, even by sac­
g2 41. e6 Kg3 42. e7 ra 43. e8=Q f2#. rificing one or more pawns. The fa­
A fine example of how to play such vorable position of White's King per­
endings. mits this, but the draw hangs on a
hair. As usual, the slightest inaccu­
racy is fatal; and here we witness
No. 5 the unusual sight of Lasker making
Rook Ending the second-best move in an endgame.
More blunders are made in Rook White's only way to secure a draw
endings than in any other phase of is to set up a passed h-pawn by the
the game. Even the greatest masters ingenious sacrifice 1. f5!! Obviously
fall prey to the seeming simplicity of 1... gxf5 2. h5! is uncomfortable for
such endings, which in fact are full Black. Therefore, 1... exf5. Then fol­
of hidden subtleties. lows the characteristic maneuver 2.
Consider the diagr amme d posi­ e6! fxe6t 3. Kxg6 Kb5 4. Ra1 f4 5. h5
tion, Emanuel Lasker-Levenfisch,
Moscow 1925.

Levenftseh 15

14

After 5. h5 (analysis)

5... e5. The dangerous alternative 5 . .

.

Lasker ra requires great exactness in de­

White to play fense. White just draws by 5... ra 6.
Rf1 a4 7. Rxffi a3 8. Rf1! a2 9. Ra1
Levenfisch, a pawn ahead, prop­ Kc4 10. h6 Kb3 11. h7 Ra8 12. Re1!
erly posted his Rook behind his Kb2 13. Re2 Kb3 14. Rei.
passed pawn, following Tarrasch's With the move 5... e5, Black in­
dictum: "The Rook's place is behind tends to sacrifice his Rook when
the passed pawn: behind the enemy White's pawn reaches h7-the three
one to hold up its advance and be­ passed pawns will then win against
hind one's own to support its ad­ the white Rook.
vance." At first sight, it seems that 6. Re1! The only move to draw. If

-78-
 All Facets Of The Endgame: Road to the Endgame

now 6... Re7, then White chases the 8. Rxf3 a2 and ... al.
Rook off the e-file by Kf6-f7-hence 7. ... e5!
Black is forced to play 6... a4 7. Rxe5t 8. Ret!
Kb4 8. Re4t Kb3 9. Rxf4 a3. The Naturally, the black pawns win
ending is now a clear draw, each after 8. h7 Rxh7!
passed pawn costing a Rook. 8. ... a3!
How did Lasker meet the problem 9. Rxe5t Kc4
in the original position? Naturally he 10. Ret a2
visualized the breakthrough f5!!, but 11. h7 Ra8
thought he might as well first place 12. Kg7 f2!
his King on f6. Precisely that move 13. Rat
cost Lasker one tempo and half a Or 13. Rf1 a1= Qt! 14. Rxa1 Rxa1
point. The game continuation is easy 15. h8= Q Rg1t!, and Black either
to follow, and the student should com­ wins the Queen or queens with check.
pare Lasker' s actual play with the 13. ... Kb3
analysis given earlier. 14. Rf1 at=Qt!
t. Kf6 Kb5 t5. Rxa1 Rxat
2. Rat a4 16. Resigns.
3. f5! exf5
4. e6 fxe6 No. 6
5. Kxg6 f4! How to Win with a Pawn More
6. h5 f3! Most experienced players realize
7. h6 the great difficulty of forcing a win
Not now 7. Rfl because of 7... a3 when only one pawn ahead. Gener­
ally the win of the pawn involves
some temporary disadvantage which
"How did Lasker may be exploited by an alert oppo­
nent. The following diagram from the
meet the problem
Kemeri tournament of 1937 typifies
in the original po­ the difference between theory and
s ition? Naturally practice.

he visualized the L. Rellstab

breakthro u gh f5! ! ,
b u t t houg ht h e
might as well first
place his King on
f6. Prec is ely that
move cost Lasker
one te mp o and
hall a point." S. Reshevsk:y
White to play

-79-
 CJS Purdy On The Endgame

A glance shows that Black's 46. Kd5 Kg5

Queen is very well centralized and 47. Kc6 Nd8t
that most of White's pawns are weak. 48. Kd7
Therefore an ordinary mortal would
be satisfied with a draw, readily
forced by 37. Nf6t! gxf6 38. Qxf7t 17
Kh8 39. exf6; and Black must take a
perpetual check.
Not so Reshevsky. Despite his
usual clock trouble, Reshevsky ex­
pects more from the position than is
justified. There follows:
37. Qxf7? Qd4t
38. Kh1 Qxe5 48. ... Kh4!
39. Kg1 Qxb2 With White's King decoyed from
40. Qxe6 Qx g2t! the kingside, this sacrifice is quite
Overlooked in White's previous justified. Black obtains passed pawns
calculations. on both wings, against which the lone
41. Kxg2 Nf4t Knight is helpless-the white King
42. Kf2 Nxe6 being too far away.
The ensuing Knight ending is very 49. Kxd8 Kxh3
tricky. Although White is a pawn 50. Ke7 h5
down, he has excellent drawing 5 1. Kf7 h4
chances: 52. Kxg7 Kg3
1) White's King is better central­ 53. Kf6 h3
ized; 54. Nf1t Kf2
2) Black's queenside pawns are 55. Nh2 a4!
weak and liable to joint attack by 56. bxa4 b3
King and Knight; 57. f4 b2
3) owing to the reduced material And after deep thought even the
left on the board, White may be able cheery optimist Reshevsky resigned.
to exchange most of the pawns and
then sacrifice his Knight for White's
last remaining pawn. No. 7
43. Ke3 Kg6 Knight Versus Pawns
44. Nd2 The short-stepping Knight is least
The threat of Nc4 forces Black's effective when trying to stop two
reply of ... Nc5, whereupon White ex­ passed pawns on opposite sides of
pects to drive away the defending the board. Study the following posi­
Knight and to pick up the queenside tion which occurred in a New Sauth
pawns with his King. Black, however, Wa l e s c h a m p ion s h i p g ame,
has foreseen this danger and coun­ Hanks-Goldstein.
ters with a surprising sacrifice.
44. Nc5
45. Kd4 Nb7

-80-
 Al l Facets Of The Endgame: Road to the Endgame

M.E. Goldstein The student should work out the

slower and less logical win by 2... Nh5
3. a4 Kc5 4. Kc3 Nf4!
18 3. a4 Kc4!
The key move. If Black checks at
e3, White retires to c1-and Black
cannot check again with the Knight
for three moves. Therefore a square
must be found from which the white
King, on reaching c1, can be checked
in one move. From d4 the Knight can
J.N. Hanks reach e2 (or from b4, a2); hence the
Black to play key square, to be vacated by the King,
is d4.
White, having defended very care­ 4. a5
fully for seven hours, has just sacri­ If 4. h5, then Black can take the
ficed his Knight for a pawn, expect­ h-pawn and reach b6 with his Knight
ing to obtain an easy draw. White's just in time to stop the a-pawn from
h-pawn restrains the hostile Knight; queening.
and at a suitable moment, White will 4. l"Jf5
push his a-pawn, forcing Black to 5. a6 Nd4t
chase it with his King. But Black can­ 6. Ket Kd3
not very easily capture the a-pawn Not 6... Kb3 because 7. a7 Ne2t 8.
while keeping his own pawn pro­ Kd1 c2t 9. Kxe2, and White draws.
tected by the King. 7. a7 Ne2t
At the second adjournment, Black Black rejects the flamboyant
sealed ...Nb3t-although he wins the hos­
1. ... Kd4 tile Queen just the same, he loses his
The threat of ... Kd3 forces Knight in the process: 7... Nb3t 8.
2. Kc2 Kb1 c2t 9. Ka2c1= Q 10. a8= Q Qa1t.
First I shall demonstrate two un­ 8. Kb1
successful attempts-to show up the If 8. Kd1, then mate in four.
nature of Black's difficulties. Let's try 8. c2t
2... Nf5 3. h5 Nh6 (if 3... Ne3t, then 4. 9. Ka2 c1=Q
Kc1 Kd3? 5. h6!; and White actually 10. a8=Q N c3t
wins) 4. a4 Nf5 5. a5 Kc5 6. Kc3, and And Black wins the Queen by a
White draws by a6 later on. couple of checks.
Or, alternatively, 2... Ne4 3. h5 Ke5 This game, which lasted 102
4. h6 Kf6 5. a4; and the Knight must moves, is the longest I have played
soon release his control of c3. for 17 years-since Reti beat me in
Therefore, since Black has such 95 moves at the Tunbridge Wells
difficulty in restraining the two tournament of 1927.
pawns, he must try to queen his own.
The gume continued:
2. ... c3!

-81-
 OS Purdy On The Endgame

Chapter�
Endqal\\e Principles

sum.ably on the assumption that an

exchange of Queens would make the
game drawish.
Of course White should have re­
captured with the Queen, preserving
a completely balanced position. If
Black then exchanged Queens, it
would make the position less drawish
rather than more drawish. With mo­
bile Queens on the board, the prob­
ability of a draw is usually greater
than in a minor pieces' ending be­

T
he diagrammed position oc­ cause of the possibility of the weaker
curred in the current tourney for side playing for a perpetual at some
the women's championship of New stage. A Knight versus Knight end­
South Wales. The subsequent play il­ ing may look drawish, but a skillful
lustrates the kind of errors that natu­ endgame player will always win it
rally enterprising players are bound against a player whose endgame lore
to make if they lack experience in is scant.
handling such positions-positions In any case, one should never sub­
that are neither wins, losses, nor mit to an inferior position just to
clear draws, and that do not offer avoid exchanges-a drawish posi­
any indication for a set plan. tion, after all, is better than an infe­
It is Black's move; and after 1... rior one.
Qd5t she has an evident advantage, Black's first move is obvi o u s
White's Queen being temporarily out enough.
of play and obstructed by the white 1. ... Qd5t
Knight. It is interesting to note that 2. Kg1
White got into this awkward situa­ This leaves the light squares
tion through "playing to win ." Black weak. However, to plug them with f3
had exchanged Rooks on d3, and would weaken White's King by leav­
White retook with her Knight, pre- ing him open at some time to a check

-82-
 All Facets Of The Endgame: Endgame Principles

on the second rank; and that seems was better to make some quite use­
at least as great an evil. less one with a piece rather than a
Now, after 2. Kg1, let us take weakening advance such as ...g5.
stock. Black's Queen is very well Another way of stopping the
placed-if she can stay at d5, White threat was by 2... e5, but that is also
will have to expend a move on safe­ slightly weakening.
guarding the a-pawn so as to free But in any case, the first reaction
her Queen. This thought turns our to a threat should be: Can I ignore
eyes to White's threat of Nf4, driving it? Visualize the move made (Nf4).
the black Queen away and at the What then? Black can reply simply
same time freeing White's pieces. with ...Qd2!, a position just as effec­
Seeing that, Black played tive for the Queen as d5, as the Queen
2. ... g5? still bears on a pawn, tying up
That certainly stops Nf4, but it is White's Queen.
bad because it disorganizes Black's Therefore, why not utilize the
pawns and weakens squares in her move to improve the position of the
camp. Black was deluded by the fact black Knight, instead of stopping an
that White has no immediate way of imaginary threat?
exploiting the weakness. A weak­ It is now time to seek out a strate­
ness, once created, cannot but harm gic aim, so as to form a general plan.
one, even if it is not directly ex­ White's weakness lies in the light
ploited. One always has to keep an squares around her King. Clearly
eye on it, and one will find that cer­ Black should try to post her Knight
tain undertakings will be barred be­ on one of those squares. The ideal
cause of the counterchances the en­ square would be f3, but a more fea­
emy would have. sible goal is e4. Thus, suppose, after
If one insists on learning the dis­ 1... Qd5t 2. Kg1, that the play goes:
advantages of such moves by experi­ 2... Nf5 3. Nf4 Qd2 4. b3 g'6! To play
ence, the cost is heavy. It is better to ... Nd6 at once would give White a
avoid them and to let your opponent chance to free herself with Qd3. Black
make them; then try to profit from therefore plays to threaten ...e5, driv­
them-learn from your opponent's ing White's Knight back first so that
mistakes instead of your own. Qd3 will not be on. Of course 4... g5?
Clearly a better move than 2... g5 is too weakening (it allows Nh5) 5.
was 2... Ng'6 since not only does it Qa1 Qc2! Again Black prepares the
avoid creating any weakness, but it way. If at once ... Nd6, then Qe5. In
develops a piece. Black discarded it reply to the text, if 6. Qf6, then 6...
because she thought (rightly) that Qxa2-as White cannot afford to sac­
after 3. Nf4 Nxf4 4. exf4 the position rifice the Knight.
is very drawish. The black Queen re­ Black now has a clear advantage,
tains her commanding position, but as she threatens ...Nd6, followed by
has no piece to combine with. As al­ Ne4; and it is hard to see a good line
ready stated, Queen vs. Queen end­ for White.
ings do tend to be drawish. Doubtless White can make a bet­
But there are other moves, and it ter defense, but the article is only

-83-
 CJS Purdy On The Endgame

designed to show the general way to that chess books should be read
proceed in nearly balanced positions. critically, thus enhancing their in­
structional value enormously; and
RULE 18 here he demonstrates plenty of in­
-¢-
seek out the enemy's weak­
structional value hidden beneath
the surface.
nesses and if possible in­ It is indeed extraordinary that,
duce him to create more, but do in following up some analysis
not try to force him by playing from a "Chess World" subscriber,
weakening moves o f your Purdy (with a vital contribution
own-that is, until you have from Lajos Steiner) should dis­
obtained a crushing advantage cover that in the first post-game
and can afford to launch out. half century three highly reputable
All the time, your first care authors published flawed analy­
must be to avoid giving coun­ sis of this apparently not-too-dif­
terchances (note Black's 4 ... g'6 ficult ending (J. Berger, in "Theorie
and 5.. . Qc2 by way of illustra­ und Praxis der Endspiele, " 1922,
tion). was a fourth). It is curious also,
noted Purdy (citing a book on
Blackburne's games as his author­
ity), that none of these authors
As the game actually went, Black used as the starting position for
first played 2... g5? and then, to pre­ their analysis the position th at ac­
vent also Nb4, played 3. .. a5, creat­ tually occurred in the game. (For
ing further weaknesses on the other a similar superficial treatment by
wing. Subsequently, to straighten out Purdy himself, refer to pages 40-
her pawns, Black advanced more of 42).
them, while White prudently did noth­ Perhaps even stranger is that,
ing at all. Finally, the black pawns another half century on, despite
fell like ripe fruit. new commentaries by other au­
The position struck us as being thors, Purdy's 1946-47 series of
one that very few players would know "Chess World" articles was prob­
just how to handle. ably the most authoritative treat­
ment in print of this e n d i ng.
Steiner's crucial piece of analysis
Oh We Of Too Much Faith! was repeated in the 1950s by
[The following remarkable series B o n c h-Os m o l o v s k y a n d Ter ­
of three articles deals "in-depth" Pogosov, who were credited by
with the Teichmann-Blackburne Averbach and Maizelis (in "Pawn
pawn ending. The pawn formation Endings," 1974) and by the ''Ency­
in this famous ending is clearly clopedia of Chess Endings" (1982);
ideal for illustrating the "distant but if Purdy was correct, these emi­
pawn" theme, so it is not su:rpris­ nent autlwrities missed other vi­
ing that many authors have made tal points. Euwe and Hooper, in '.ti
use of it. Purdy makes the point Guide to Chess Endings" (1959),

-84-
 All Facets Of The Endgame: Endgame Principles

were the only authors of those men­ This is a good and valuable
tioned to start with a correct dia­ book by our esteemed and
gram-they cover this ending well, merited master, but even in
but do not identify the line that this book there are quite a
Purdy gives as the best defense. lot of err ors ....
"

Was Purdy right? He makes no

claim that his analysis is error­ Botvinnik seems to imply that
free, but it seems probable that master players could run through a
Purdy and Steiner produced the composer's work and pick up most
most accurate and comprehensive of the errors without much difficulty.
analysis of this ending in the 100 We, on the other hand, venture to
years that have elapsed since think that a painstaking critical
Teichmann and Blackburne strug­ analysis of endgame literature of
gled with it over the board. which masters themselves were the
As you will see below, Muller & authors would show many errors.
Lamprecht have definitively au­
topsied this game position in their A Famous Ending
"Secrets of Pawn Endings." The full Now for a staggering case in point!
truth? Black wins.-F.PHJ The pawn ending, Teichmann­
Blackburne, Berlin 1897, has been
made so famous by three popular
Botvinnik on Errors authors: Edward Lasker in Chess
First some remarks by Botvinnik Strategy, Znosko-Borovsky in How
culled by us from the new American to Play the Chess Endings, and Fine
fortnightly, Chess News from Rus­ in Basic Chess Endings that there
-

sia. must be very few strong players who

have not come across it. All three
"... And another thing: the masters, as well as Blackburne him­
majority of our composers self, have given the position as a win
are not too strong in the for Black. And yet, as will be seen by
technique of the endgame. the method adopted here, there is no
Consequently the studies win. [Ed.: This is Purdy's claim
la ck s o l u t i o n s , ha v e which KM/FL slww to be an error.
'cooks,' duals, etc. H there They analyze a win for Black. See
cou l d be e stablishe d a pgs. 231-232 in "Secrets of Pawn
clo ser contact between Endings."]
endgame studies and mas­ Strangely enough, Edward Las­
ters, it would be possible to ker, Znosko-Borovsky, and Fine all
find means which, if they vary from the actual play in the early
could not e liminate al l stages; and they all vary in different
these defects, would at ways-then they all glide into the
least diminish them consid­ correct text after a few moves. But
erably ... Not long ago, I granting each his own version of the
looked over the collection of play, each is wrong about it!
studies by A.A. Tr oitzky. Our remarks are not made in dis-

-85-
 CJS Purdy On The Endgame

paragement of these eminent writ­ Blaekburne

ers, whom we greatly admire. Error
is of the very life-blood of chess; and
just as error is inevitable in actual
games, so it is inevitable-though of
course much less frequent-in chess
books dealing with numerous diffi­
cult positions. The aim of this article
is to induce players to read chess
books more critically, with ten times
the benefit to their own play-but let
them not imagine that everything Teichmann
that they find obscure is a mistake Blaek to play his 56th
by the author!
The strange thing is not that the The moves immediately preceding
play in this ending was unsound, but the diagram were 55. hxg5 fxg5 56.
that the unsoundness should have K(h3)g'2, reaching the diagram with
been missed by tens of thousands Black to move-see Diagram 2.
who played through the ending. It The note inBlacklJUrne's Games
seems that the vast majority of us says that everybody thought the end­
prefer faith to reason. ing a draw with the exception of
At last a Sydney amateur, J.A Blackburne-who then proceeded to
Phillips, read his Znosko-Borovsky win. Blackburne subsequently dem­
critically enough to discover the er­ onstrated a much prettier win against
ror-he thereupon sent his claim to a different defense. The victory plus
us. Not then having time to examine the demonstration led to the prema­
it fully, we sent it to Lajos Steiner. ture conclusion that Blackburne was
Steiner found an extremely interest­ right in his assessment of the posi­
ing flaw in the analysis as submit­ tion in Diagram 2 (Black wins).
ted-the flaw itself made an entirely
new endgame study-but we then Reconnaissance
found an earlier improvement which Let us first of all discuss the end­
upheld Phillips's claim. ing methodically. That will give us a
To start from firm ground, we go better understanding of it than pro­
back to the original game in Mr. ceeding move by move from the out­
BlacklJUrne's Games of Chess. set. The following are thoughts that
might flash through the mind of a

good endgame player:

1) Black is a pawn up, but the only
uses for the rear member of the set
of doubled pawns can be
a) to gain a move at a critical stage
or
b) to hold a square.
From a) it follows that ...c5 should

-86-
 All Facets Of The Endgame: Endgame Principles

be held in reserve as long as pos­ answer to an immediate ...h4, namely

sible-because you cannot both eat f4t! After Black replies with ...gxf4,
a cake and have it. White himself acquires a remote
2) Black has two other small ad­ passed pawn with gxh4!-and then
vantages. Black's King is an aggres­ White must either win or draw.
sor, White's a defender. Above all, 7) We conclude that Black must
Black's h-pawn is a potential "remote seek to play ...h4 when the Kings are
passed pawn," after ...h4. Because of so situated that White cannot draw
his three small advantages, one must by f4 in reply.
conclude that Black has at any rate 8) Now for White's possible pawn
excellent winning chances. sorties. Obviously f4 is ruled out even
3) The ending is not of the com­ with Black's King some way back, as
pletely blocked type. There are move­ it at once gives Black a remote
able pawns on both sides; so before passed pawn with ...gxf4. But f4 will
we can profitably consider the ma­ often come in as a reply to ...h4-see
neuvering of the Kings--0n which the 7) above.
result always ultimately depends­ 9) Also g4 is impractical in the dia­
we must examine the results of all gram because of either ...hxg4 or
the possible pawn moves for both ...h4. But if White first plays his King
sides. to h3, then if ever Black's King hap­
4) First of all, we can rule out ...g4 pens to be back on the sixth rank
for Black at any stage since then instead of on the fifth White has a
White would draw easily by f4. possible drawing combination with
5) But ...h4 is very important­ 1. g4 h4 2. f4! gxf4 3. Kxh4.
clearly, Black can win, if at all, only
by a well-timed ...h4. Envisage that What Happened?
move made in Diagram 2. White can­ One could carry this reconnais­
not exchange pawns as that would sance further, but it may interest the
give Black a remote passed pawn and reader to see how the original game
an easy win (each side captures the actually went.
enemy passed pawn and then the From Diagram 2:
King nearer the other wing just 56. ... Ke6
comes across and wins). Nor can Edward Lasker makes Black play
White move his King because then 56... Kf5, probably because it looks
. .. hxg3, Kxg'3, ...Kf5; thereafter the obvious and takes less explaining.
white King must retreat, letting the But Blackburne sheers off from the
black King into f4. From that point, square f5 deliberately-he wants to
the ending at least looks like a win hop in there at the moment when
for Black (King four ranks ahead of White's King comes to g3, after ...h4
its home rank vs. King one rank and ...hxg3. But if Black is there al­
ahead of its home rank). So for the ready, he can't go there; and to stay
moment, we will assume that this there he would have to use up his
position is to be avoided by White, valuable spare move ( ... c5).
leaving the actual game to prove it. 57. Kh2?
6) White, however, has an easy White should have played 57. Kh3.

- 87 -
 CJS Purdy On The Endgame

If in response 57... Ke5, then 58. Kg2!; 59. Kh2?

and if 57... Kf5, then 58. Kh2!! This Seeing that this lost, Phillips
will be clearer later. [Ed. : After 58. . . asked himself, Why not to h3? Why
Kf6 59. Kh3 Ke6 60. g4 h4 61. f4 gxf4 did Edward Lasker and Znosko­
62. Kxh4 Ke5 White loses.] Borovsky ignore this obvious move?
57. ... Kf6? We return to that later. [Ed.: It
Blackburne, whose handling of doesn't matter if 59. Kh3 After 59...
.

the ending has been regarded as a Kf6 White still loses on 60. g4, 60.
choice bit of wizardry, actually Kg2, on 60. Kh2. -K.M.]
missed a forced win here by 57... h4! 59. ... h4!
at once. That threatens ...hxg3; and Noticing White's strange propen­
as soon as White is forced to recap­ sity to move his King only to h2 when
ture on g3 with his King, then ...Kf5 unable to go to g2, Blackburne has
wins as in the game. And upon 58. cleverly played into a position where
f 4, the "drawing" move, then 58... Kh2 is particularly bad. For if now
gxf4 59. gxh4 Kf5 60. Kh3 (forced) 60. f4, then 60... gxf4 6 1. gxh4; and
Ke4!!; and we arrive at a position Black wins easily by 61... Kh5! 62.
proved a win for Black by Lajos Kh3 c5.
Steiner-see further on. The square g6 has no merit as a
58. Kg2? goal apart from this trap. We men­
Teichmann, one of the greatest tion this in case students have been
masters, again finds the wrong move, mystified when seeing this strange
showing how beset with pitfall and journey of the black King passed over
with gin are these simple-looking in books without any explanation.
pawn endings. Again White should 60. Kh3 hxg3
have played Kh 3!, with the same ef­ 61. Kxg3 Kf5
fect as on the previous move. [Ed. : The ideal position for which Black
Not so, as 58. . . Ke6 once again wins has been angling-which he could
as KM & FL pointed out.] never have forced against correct
58. ... Kg6? play, though the books all tacitly as­
And again we find that play hith­ sume that he could have. Black has
erto regarded as extremely ingenious got his King to f5 just when White's
and subtle is just a missed opportu­ King is on g3-without using the
nity. Black again had a forced win­ move ...c5. White must now retreat.
this time by 58... Ke5! Then, after ei­ It was useless for White to delay
ther 59. Kh3 or 59. Kh2, Black wins Kxg3. For example, 6 1. Kg2 Kf6 62.
by 59... h4! We shall prove this later, Kh 3 Ke5; and now White must re­
showing that White must always capture. Black has carefully walked
meet ...Ke5 with Kg2. [Ed. : Except round f5, not setting foot there till
in this instance as White's King is White captured.
already on g2.] Therefore, White Teichmann now played 62. Krl?,
must never move to g2 when Black allowing Black to go to f4 with "the
goes to any square adjacent to e5 opposition." Play proceeded:
(that is, of course, while the pawns 62. Kf4
remain as in the original diagram). 63. Ke2 Kg3
-88-
 All Facets Of The Endgame: Endgame Principles

64. Ke3 c5
With the vital spare move, Black
regains "the opposition." 4
65. Ke2 Kg2
66. Ke3 Kf1
67. Resigns.
Black wins the f-pawn and loses
his own g-pawn, but he is then clo­
ser to the queenside and thus wins
easily.
Here let us try the obvious move
59. Kh3!
The Might-Have-Been This threatens to draw by 60. g4
The nicer finish was the following h4 61. f4 gxf4 62. Kxh4, etc. See 9)
one as demonstrated by Blackburne. above.
62. Kg2! Kf4 59. ... Kf5
63. Kf2 c5 Here Phillips' analysis continued
64. Ke2 Kg3 with 60. Kg2, producing the impor­
65. Ke3 tant Diagram 5. (pg. 89).

�
� �
� � � � �
� ,� � 5
��
� ��� T
��
� ��
��-�

:�1�=��--�
-'
�� "� �ji,> �
�'� ····-'� �� �
���� � 1
· · ··

� � � �� �
Black to play and win After 60. Kg2 (Philllps)
Black to play and win
And now comes the pretty move
that is the main reason for the end­
ing having been put in so many Fine's Diagram,
books. Do you see it? The solution is Steiner's Analysis
at the end of the article. {See pg. 91) By a coincidence, Diagram 5 hap­
pens to be the diagram that Fine
The Corred Defense gives as the original position though
Now we return to the position af­ it never actually occurred in the
ter Black's 58 ... Kg6-Diagram (4). game-Fine takes several liberties
with the text. Fine gives the follow­
ing moves to show a win for Black:
1... Kf6? 2. Kh3 Kg6 3. Kh2?, and

.
White loses as in the game. Fine says
that if 3. g4, then 3. h4-he fails to

-89-
 CJS Purdy On The Endgame

see that White then draws easily by Let us return to Diagram 4.

4. f4 (pointed out originally by Lajos 59. Kh3 Kf5
Steiner). Now for 60. Kg2? we substitute
But in Fine's diagram, Black has 60. Kh2!
a win by the immediate 1... h4, which
Fine does not consider. The demon­
stration is by Lajos Steiner: 1... h4!
2. Kh3 Ke5! (threatening the good old
... hxg'iJ, Kxg'3, .. . Kf5, with the same
win as in the game) 3. f4t gxf4 4.
gxh4 Ke4!
Here we have an interesting ex­
ception to the general rule that a re­
mote passed pawn wins. Note that
had White played 2. Kh2 then Black We thus arrive at a position
would interpolate 2 ... Ke5! , "forcing" wrongly given by Edward Lasker as
Kh3 [Ed. : As 3. f4f loses. ] Black occurring in the actual game [Ed. :
would thereupon follow with ...Ke5 Instead, this position is in a note
as in the text. to move 57. Kh2? on page 87.]
There are now two variations. Lasker says that if now 60... h4 then
a) 5. h5 f3 6. h6 Ke3 7. h7 f2 8. White would play 61. Kh3. That, how­
h8 =Q f1 =Qt, and Black must win as ever, would lose after 61... Ke5! (see
he can assail White's pawns with Steiner's analysis). Actually, White
both of his pieces, whereas White's would answer 60... h4 with 61. Kg2!,
King is out of play. An illustrative with a forced draw. If thereupon 61...
line is: 9. Kh2 Qf4t 10. Kg2 Qe4t 11. K moves, then 62. f4 at least draws.
Kg3 Kd2 12. Qd8t Qd3t, and Black And if 61.. . hxg'iJ, then 62. Kxg'3; and
wins. If this check were not on, the Black finds himself on the square to
win would still be easy by ...Kc2, etc. which he wants to move. Black,
b) 5. Kg2 Kd3! (threatening ...Ke2! therefore, has to use up his spare
and thus gaining a vital tempo) 6. pawn move, whereupon White draws
Kf3 (or Kf2) Kxc3. White queens first, easily.
but Black wins easily on material. Instead of ...h4 in reply to 60. Kh2 ,

Fine says the position was adjudi­ suppose Black plays 60... Ke5. Then
cated a draw, but that position never 61. Kg2 gets us back to Diagram 2.
arose-and the actual game was not Or if 60... Ke6, then 61. Kh3. And if
adjudicated. There seems to have 60... Kg6, then 61. Kh3 (back again)
been some hoodoo over this ending, or 61. Kg2 [Ed. : Purdy likes Kg2 in
forcing every writer to blunder about a situation where i t isn't war­
it until the spell was broken by ranted. After 61 . . . Kf7! 62. Kh2 Ke6
Phillips. I must add that Fine's Basic White is in trouble.] Finally, if 60 ...
Chess Endings is a monumental g4, then 61. Kg2.
work, one of the most complete and
authoritative on endgames in any
language.

-90-
 All Facets Of The Endgame: Endgame Principles

Znosko-Borovsky's Version We will go further. There may be

To obtain Znosko-Borovsk:y's dia- no forced win by any method, but
gram, which-like Fine's-never oc- Black has one line that forces a clear
curred at any stage, place the White advantage with excellent winning
King on f2 in Diagram 4. Z-B contin- chances.
ues with 1... Kf6 2. Kg2, thenceforth Can you find that strongest line
as in the game. Apart from the wrong for Black?
diagram, Z-B's errors are only errors
of omission, i.e. failure to comment Solution For Diagram 3
on incorrect moves. Black wins by 65... Kh3!! This is
the only move that deprives White of
"the opposition," White being unable
Summary Of Analysis to move either to f3 (occupied) or to
With the pawns remaining as they d3 (guarded). If 66. Kf2, then 66 .. .

are, White must always meet ...Ke5 Kh2; and Black must soon get to f1
with Kg2; and he must always meet and win as in the game. For example,
...Ke6 or ...Kf6 with Kh3 (threatening 67. Ke1 Kg1 (or Kg3) 68. Ke2 Kg2 69.
g4). But White can meet ... Kd6 or Ke3 Kf1.
...Kd5 or ...Kf5 (King on a square ad­
jacent to e5) with either Kh3 or Kb.2,
and he can meet ...Kg6 with either The Mystery Endgame
Kh3 or Kg2 (or if he is already on h3, A lthough we know that many
with g4 of course). readers tried their hand at the Teich­
The reason that ... Ke6 and ...Kf6 mann-Blackburne ending, we have
are so compelling is that the King's received no demonstrations of the
square then adjoins both e5 and f5. strongest line for Black. It is easy to
White cannot reply with Kg2 because find the preliminary moves. See Dia­
of ... Ke5, nor with Kh2 because of gram 2 .
... h4, threatening ...hxg3, Kxg3, ...Kf5.
If Black goes to e6 and then to f6 or Blaekburne
vice versa, then White plays g4! See
9) above (pg. 87)and also our com­
ment to 3. Kb.2? under Diagram 5.

We Celebrate The Day!

We haven't given all the secrets
of this ending yet! Did we give the
impression it was a draw? But if you
reread the article, you will find no
claim that a draw can be forced from Teichmann
the original position. The only claim Black to play his 56th
was that the published play was un­
sound and would not force a win. Analysis
In the preceding article, we found

-91-
 CJS Purdy On The Endgame

that White must answer either ... Ke6 endings. White has queened first and
or ...Kf6 with Kh3. Therefore, Black can check first, yet he should prob­
should play 1... Ke6 and 2 .. Kf6 (or
. ably lose-for Black can attack the
vice versa) since White cannot an­ white pawns with both his pieces,
swer Kb.3 to both. So we have: whereas White cannot do the same
56. . . . Ke6! to the Black pawns. A tempting idea
The very move Blackburne played. now for White is 65. Qg6t Ke3 66.
Teichmann replied 57. Kh2?, where­ Qxc6; but Black plays ...Qf4 t and then
upon Blackburne could have won by ... Qf3t, forcing the exchange of
57... h4! Queens. So that's off.
57. Kh3 Kf6! Straightforward play would be:
58. g4! 65. Qe6t Kd3
As shown in the preceding article, 66. Qxc6 Kxc3
58. Kh2 would lose to 58... h4; and 67. Qxa4
58. Kg2 would lose to 58... Ke5! We Here, it is true, White very likely
found that this g4 draws easily if the has a draw-even without the aid of
black King is at g6, but it is different his pawn.
now. In Basic Chess Endings, Fine
58. h4 gives the rule:
59. f4 gxf4
60. Kxh4 Ke5 RULE 19
61. g5 Ke4 �

Q
62. g6 f3 u een and pawn versus
So far it is all forced, but now lone Queen is normally a
White has a choice between g7 at win only with the c-, d-, e-, or
once and Kg-3 first-to be followed, f-pawn that is on the rank next
after ...Ke3, by g7. If White does the to the queening rank.
latter, we shall reach a position like �
that in Diagram 8 but with the white
King on g3 and the black King on e3. As to the possibility of getting a
63. g7 � pawn onto the rank next to the
64. gS=Q f1 =Q queening rank, Fine says:

"In the first place, we must

;:+:;:�
8 note that in general it is im­
possible to advan ce the
pawn very far. The de­

�-" ��--/"����
fender keeps on checking;
and when he runs out of
��"--������
. v,

checks, he pins the pawn.

�����=�� In view of the huge number
of possible positions of the
We have now left the realm of ab­ Kings and the pawn, all of
solute demonstration and are in the which require individual
relatively uncharted ocean of Queen attention, it is out of the

-92-
 All Facets Of The Endgame: Endgame Principles

question to subject this 67. Kg5! Qd2t!

statement to precise analy­ Where is White to go? Not to f6,
sis. At any rate, in actual losing his Queen. Nor to f5 or to g6,
practice it is usually found as then ...Qc2 threatens discovered
that the pawn can be held check. And the squares g4 or h5 give
back." the same light-square trouble as be­
fore. Therefore 68. Kb.4. Now comes
[Ed.: In thi s case, however, 68... Qc2 again. Then, however, if
Black can win! 67. Qxa4 Qf2t 68. White proceeds with 69. Qd6t Kxc3
Kh3 Qc5! 69. Kg2 Kb2 70. Qa6 c3 71. 70. Qb4t Kd3 71. Qd6t, the black King
a4 c2. Now Rule 19 kicks in and can this time go to e3. It will be im­
Black will win.] possible for White to prevent the
However, on move 66, instead of black King from ultimately getting to
. .Kxc3, try:
. shelter on the queenside. Once there,
66. ... Qf2t! the black King can also attack the
67. Kg4 Qc2! white a-pawn. So, again we have a
We feel sure that Black should win win for Black.
from there. Black threatens ...Kxc3; White can make sure of winning
and if that is stopped, then ...Qb3 fol­ Black's a-pawn by the following
lowed by a stealthy creep with the play-start again from Diagram 8:
King round to the a-pawn. 65. Qe6t Kd3
White can nearly get a perpetual 66. Qd6t Kxc3
check. Thus, 67. Qb4t Kd3
68. Qd6t Kxc3 68. Qxa4
69. Qb4t Kd3 But then, after judicious checks,
70. Qd6t Ke2 Black can get his front pawn to the
71. Qb2t Kd1 second rank. Even if Black gives up
72. Qg1t Kd2 his rear pawn, he should then win .
73. Qd4t Qd3 The white pawn is not an asset to
74. Qb2t Ket White-it would support a possible
Now Black would decisively inter­ check on b4, but, on the other hand,
pose his Queen against all checks but it would prevent White from check­
one. If now 75. Qb4t, then 75... c3 ing on the a-file.
wins. White is helpless-he cannot We do not propose examining
even take the a-pawn because of here the results of interpolating 63.
...Qd1 t, with exchange of Queens. Kg3 before g7. The interpolation,
however, ought not to be good for the
As the presence of his King on g4 following reason:
is embarrassing to White, suppose
he tries 67. Kg5! instead. A smart
ruse, but Black can circumvent it by
playing 67 ... Qd2t before moving the
Queen to c2. Thus, from Diagram 8:
65. Qe6t Kd3
66. Qxc6 Qt?t

- 93-
 CJS Purdy On The Endgame

dure. See the following diagram.

RULE20
'¢- Blaekburne
ry1he rule for the defending
.l player in Queen endings is
to have his King either very
close to the scene of action (so
that he can afford to exchange
Queens), or, failing that, as far
away as possible (so as not to
limit his choice of checks). The
middle way is fatal here.
'¢-
Teichmann
An absolute demonstration might Black to play his 56tb
need a book, but we are sure that
Black should win . This tallies with Analysts
the opinion any strong endgame play­ Briefly, the best play for both
er would form on seeing the original sides is:
position (Diagram 2). On one's gen­ 56. Ke6!
eral experience of pawn endings, 57. Kh3! Kf6!
Black's three tiny advantages look 58. g4! h4
as if they should add up to a win. 59. f4 gxf4
The main features of interest, 60. .Kxh4 Ke5
however, were that there was no win 61. g5 Ke4
by the method actually adopted and 62. g6 f3
subsequently enshrined in books and And the position reached is be­
that play treated as a model for al­ yond reasonable doubt a win for
most half a century was full of de­ Black, though absolute demonstra­
monstrable errors on both sides. tion is difficult.
Now Brian Reilly, the well-known
British player, makes a tentative
That Teichmann­ claim to a draw, varying on move 61.
Blackburne Ending Reilly gives:
In the first of the two preceding 61. Kh3 Ke4
articles, we showed that everything 62. Kg2 f3t
previously published on this ending 63. Kr.!
was wrong. Far from being a mag­ He then goes on to show a draw,
nificent piece of endgame play by mentioning that if 62 ... Ke3 then 63.
Blackburne, the actual play was a Kf1!
succession of errors on both sides; Black, however, actually makes
and there was no win by the method the surprise move
adopted. In the previous article, how­ 62. . .. Kd3!
ever, we proceeded to show that This threatens ...Ke2 just the same
there was a win by a different proce- as ... Ke3 would, so White must ap-

-94-
 All Facets Of The Endgame: Endgame Principles

proach with his King. But then 63... have come about in the game We­
Kxc3 and 64... Kb2, and Black wins ber-Sarapu had Weber played, in
the subsequent Queen ending by the diagrammed position, 27. b4!
sheer weight of material. which he suggests in place of the pas-
This ...Kd3 embodies the impor- sive 27. Kf1 (probably a mistake in­
tant principle that a King can ignore duced by fatigue).
Euclid. A diagonal route or even a
zigzag one, as here (Ke4-d3-e2), is to Sarapu
him just as short as a straight line.
To many new readers, Reti's fa­
mous little ending will still be new. 11

10

Weber
After 26... Rxd!

After 27. b4, perhaps as many as

99% of the chess players in the world
Black to play/White draws! would automatically reply 27... c4. We
all know the value of a supported
At first glance it seems that even passed pawn in general. Besides, the
if White had the move he could not only alternative would be 27... cxb4,
draw-he cannot catch the Black producing a weak backward a-pawn
pawn and he is too far away to save in Black's camp.
his own. And yet White does draw And yet 27... c4? would be a posi­
even if he gives Black the move. Thus: tional blunder, allowing White to
1. h5 equalize with ease, whereas 27 ...
2. Kg7 h4 cxb4! retains winning chances.
3. Kf6! Kb6 The reason is that the advantage
ElseKe7. for Black lies in his powerfully posted
4. Ke5!! Rook, and the Rook can maintain its
The miracle has been accom­ post only if Black can always meet
plished-by forgetting Euclid and Kf1 with ...Bc4 t.
remembering the chess King's own After 27. b4 cxb4! 28. axb4 f5 (pre­
special geometry. venting Be4), White is very cramped.
If 29. e4, then 29 .. .f 4; and Black
threatens to develop his King strong­
An Overriding Principle ly. If 30. e5 to stop that, then 30...
Of Endgame Play Re2 gets the pawn. If 29. Bd1, then
A very instructive position would 29... Rb2 should win as White's back
,

-95-
 OS Purdy On The Endgame

rank is too vulnerable for the white A Guiding Rule

Rook to leave it. And after 29. h3 Kf6, For Endgames
it is hard to find an equalizing line Although as a general work on
for White. White's pawns are clearly endgames there is nothing to equal
better than Black's, but White's Euwe and Hooper's Guide to the
pieces are very much worse, espe- Chess Endings, students ought not
cially the white Bishop. to dispense with the Endgame sec-
tion of Guide to Good Chess, as it is
RULE21 the only book, as far as I know, that
-¢> gives the general principles of

I
n a middlegame, one feeble endgame play.
piece need not be a serious Reuben Fine, in his Chess the
handicap; for it is only one unit Easy Way, gave "ten rules for the
among many. But with very few ending;" but some of the rules are
pieces on the board, the immo­ rather vague, and important ones are
bility of one means a propor­ left out. However, Fine's 6th rule, "Do
tionately bigger burden to not place your pawns on the same
carry for the others. color as your Bishop," is an impor­
-¢> tant one and is now included in
Guide to Good Chess, 11th and 12th
This is really the overriding prin­ printings (omitted inadvertently in
ciple of endgame play, yet how few printings 1-10).
players appreciate it! The first and guiding rule for
There is another principle in­ endgames-and the one this article
volved here, also insufficiently appre­ is about-is stated nowhere but in
ciated. Really, it is included in the Guide to Good Chess and in occa­
bigger principle just given. sional articles in my magazine down
the years. You will find that the mas­
RULE22 ters of endgame play all follow it; but
-¢> they must do it intuitively, i.e. on the

T
o get the best out of a basis of unconsciously formulated
Bishop, avoid clogging his experience. For those who are not
diagonals with pawns. great masters and who have no such
-¢> intuition, the rule can be extremely
helpful; and the only reason that it
That is what 27... c4? would do. does not appear in textbooks in gen­
Doing that is what makes a Bishop eral is that the writers themselves
"bad." Yet many players are continu­ have never thought of it and have
ally tempted to do it because it seems probably not read the Guide, per­
to make the pawns safer to have the haps thinking that too elementary.
Bishop protecting them. The rule stated in the Guide is:

-96-
 All Facets Of The Endgame: Endgame Principles

in placing a black pawn on h5 instead

RULE23 of on h7. In any case, Hay actually
-¢- resigned two moves earlier.

B think of making a passed

efore ever beginning to If confronted with the diagram, I
am sure that the great majority of
pawn, get all your pieces into players below master class would
as good positions as possible. quickly play 1. Rc8t Kh7 2. Ra8-
-¢- and even some masters might care­
lessly play that. True, it wins the
This presupposes that you do not a-pawn; but it takes a very well
already own a passed pawn. If you placed Rook momentarily out of play
do, the rule could have many excep­ at a8, and Black can start good
tions-it might be advisable to push counterplay with 2... Rd3 3. R8xa7
the passed pawn quickly, even Rxb3. If thereupon 4. Rb4, then 4...
though your pieces have not yet at­ Rd3 5. Bb2 Nd5 6. Re4 Rb3; and the
tained their best positions. road to victory still stretches for
Where you are likely to get side­ miles.
tracked in endgames is in excursions If White applies Rule 23, the win
for winning material. Material is cer­ is not startlingly easy; but it makes
tainly important in endgames, and just that little difference that is all
yet... I was reminded what a good the difference.
rule this Rule 23 is when looking at How to apply the rule? White's
the end of my game with Hay in Perth. Rooks are already well situated, but
In the position where Hay resigned his Bishop would be better on the
because he would lose the Exchange, long diagonal. Then again, Black has
there was a possibility of a long and a threat, namely, ...Nd4.
tedious struggle if White made what That suggests, of course,
most players would consider the ob- 1. Bb2!
vious moves. We quickly see that this leaves
Black with no progressive reply. If
the black Rook leaves the seventh
Rank, then Bxf6 and Rxa7 threatens
mate. Or if 1... Nd5, then 2. Be5 places
White in so dominating a position
that the win is not in doubt. If Black
plays a developing sort of move, 1...
Kf7, it is at least clear that 2. Rca1
saves a move as compared to the
originally obvious move 1. Rc8t, as
White to play the a-pawn is won with only two Rook
moves instead of three.
I am under no obligation to use Again, if 1... Kh7 (to protect the
the Purdy-Hay position; and to King from check), then White could
make it slightly harder for White to afford to simplify by
win, I have made a slight alteration 1. ..
. Kh7

-97-
 CJS Purdy On The Endgame

2. Bxf6 gxf6 be overanxious to "do something."

3. Rca1 Kg6 The win won't run away. See first if
4. Rxa7 Rd5 you can apply Rule 23, i.e. improve
5. Rb7 Rxb5 your position without giving the op­
6. Ra6 Rxb3 ponent a chance to improve his in
7. Rbxb6 any substantial way. Of course, if the
This forces off the black Rook and opponent has threats, they must be
gives White a slow but easy win. looked after. But if the situation is
The analysis is not very impor­ "quiet," remember that a winning
tant-it is given only to prove what position is a winning position-and
is fairly obvious. Since 1. Bb2 im­ you can keep it so. Don't jeopardize
proves White's position without giv­ the win by hurrying to take weak
ing Black a chance to improve his, it pawns-a weak pawn can usually
must be a good move-whereas 1. wait.
Rc8-a8 only amounts to an exchange The reason for Rule 23 is that in
of pawns. the endgame there are very few
And, as Fine succinctly observes: pieces, and the fewer the pieces the
more important it is to get the most
RULE24 out of each one.
{> (That means, in general, getting

T
o win , exchange off pieces; all the pieces into attacking types of
to d r a w, ex c h a n g e off positions. The King is often an ex­
pawns. ception: he is frequently better kept
back to protect pawns that might oth­
erwise fall prey to a marauding Rook.
In the example where we ex­ It is mainly in minor-piece endings
changed off the black Rook, we did it that the advice to rush the King out
at the cost of swapping off all the is relevant.)
pawns on the queenside-but it was
well worth it. It does not matter if a
win is slow if it is easy. With pawns
all on one wing, a pawn up is usually One final word. Why did I place
not enough to win; but the Exchange the black pawn on h5 instead of on
up does win. The final winning h7? Had the pawn been on h7, 1. Bb2
coup ls almost always to give up would still have been the best move;
Rook for Piece and pawn, thereby but the difference between that and
producing a winning pawn-ending. the obvious play would have been less
In a winning endgame-if it is not marked. The rule, however, would
a question of racing as yet-do not have held just the same.

- 98
-
 All Facets Of The Endgame: Pawn Endings

[hapler 7
Pa\Un Endinqs

How To Make A Pawn Tell if the enemy has two or more minor

T
his article is not about interro­ pieces, especially Bishops, a Rook
gating a pawn about its alleged must be careful about leaving his own
misdeeds. Rather, our object is to camp and frequently must be con­
assist the worthy chess practitioner tent with commanding an open file.
who has slithered into an endgame A Rook is more vulnerable than a
with a pawn plus and then finds him­ Queen because he has at best only
self in a mental fog. one line of retreat, whereas the
Queen has three.
Knights: a square in the center
General Rules For Endgames or in the enemy camp, pawn-sup­
With Pieces And Pawns ported and not attackable by an en­
emy pawn.
Bishops: central diagonals. Pawn­
RULE25
support is not essential, but certainly
�
A void p a w n-moves while
desirable.
King: central, provided that he is
.fiyou are following Rule 23
not there dangerously exposed. If the
because pawn moves create
enemy has a Rook, the King is often
lasting weaknesses and thus
best posted at or adjacent to d2 (d7)
make your task harder.
or e2 (e7)-fairly near the center, but
�
still preventing the enemy Rook from
seizing the King's home rank or the
Before ever beginning to think of rank next to it.
making a passed pawn, get all your
pieces into as good positions as pos­ RULE26
sible. �

T
ry to free your position
Note from weaknesses; and, if
Broadly, ideal positions are: possible, make it hard for the
Queen: central. opponent to do likewise.
Rooks: seventh (second) rank. But �

-99-
 OS Purdy On The Endgame
If you are a pawn up, follow these piece from its good position.
rules just the same. Don't try to do
anything with your odd pawn until
the position is as hotsy-totsy as you Illustration
can make it. The odd pawn will be Illustrative of the foregoing four
working for you all the time as a rules is the following ending from a
moral force-by threatening ex­ New Zealand championship tourney.
changes, you can often force the en­
emy to give ground. Dunlop
Rules 25, 26 apply whether you
are a pawn up or not. If you are a
pawn up, add the following.

RULE27
-¢-

E
xchange pieces (not
pawns) wherever you can
do so without losing in position.
Exception: do not rush an ex­
Wade
change that will produce Bish­
White to play
ops of opposite color (though if
such an exchange is pressed on
White is a pawn up; and as the
you, evasion may be worse
position is otherwise about evenly
than acceptance). Corollary:
balanced, White should win.
refrain from forcing an ex­
28. Bxf6t?
change that will give your op­
A fiendishly tempting exchange­
ponent two Bishops against
because it forces the further ex­
Bishop and Knight.
change of Rooks. In spite of the big
-¢-
simplification, however, the move is
bad, as will be shown. We shall see
Note to corollary: Two mobile Bish­ here an instance of the high impor­
ops are nasty foes at any time, but tance of the Corollary to Rule 27.
particularly so against Bishop and The big point to remember in such
Knight when the latter are trying to endings is that there is actually no
win-the Knight has to shy away hurry to simplify. White has a win­
from certain squares for fear of al­ ning position as long as he preserves
lowing an exchange leading to "Bish­ equilibrium between the fighting
ops of opposite color." forces. By conceding Black the two
Of course, always think well be­ Bishops, White destroys that equi­
fore exchanging a well-placed piece librium.
for a badly-placed piece. At the same Looking at Rule 23, we see that
time, remember that to evade such White's f-Bishop is already in an ex­
an exchange is usually the greater cellent position (28. Bc2!? is qui te a
evil if it means withdrawing your tempting move, but on principle it is

-100-
 All Facets Of The Endgame: Pawn Endings

not good to withdraw pieces from be pinning the black Knight. The time
good posts). White's King cannot, at lost by the preliminary move Be3
the moment, improve his position. would be returned by Black in the
Neither can White's Rook-see the process of unpinning.
note to Rule 23. If, after 28. Be3!, Black forestalls
But both the white Knight (which the pin by 28... Ne4, then White giee­
can make no forward move) and the fully exchanges Knights, thereby
white c-Bishop (which is not on a cen­ eliminating his weakest piece. Thus,
tral diagonal) could be better placed. even at g3, the white Knight has some
Having decided on that, let us, in effect.
accordance with Rule 26, examine Or 28. Be3! Be7 29. Bd4 Kf8 (...b4
the position for weaknesses. White would be futile now, as the Bishop
has a g1aring weakness at g2. An ob­ can recapture on c3) 30. Be5; and
vious way of eliminating that weak­ now White threatens to exchange
ness would be by 28. f3!? But that Bishops at d6, after which the white
would break Rule 25-the dark Rook would threaten to emerge.
squares would be badly weakened. That is enough calculation to
It seems, then, that this problem may verify that 28. Be3 is good-there is
be better postponed until White is in no need for long range planning. The
a position to challenge the long light moral of all this is CENTRALIZE.
diagonal with his own Bishop. Before leaving Diagram 1, con­
The Knight's ideal post is d4 (cen­ sider 28. Bc2!?-with a view to chal­
tral and pawn-supported). But the lenging the long lig1It diagonal on e4:
pawn-support is not permanent-it 28. Bc2 h6 29. Bxf6t Bxf6 30. Rd7.
is easily undermined by ... b4. For ex­ Yes, but now, because of the anti­
ample, 28. Ne2!? Be7 29. Nd4 b4! Now, commonsense withdrawal of the
to avoid an isolated pawn, White Bishop, the exchange of Rooks is not
must exchange pawns. Thereupon, forced-and cannot be forced until
the white Knight will be "loose"; and the Bishop returns to b3. For ex­
the black Bishops will have gained ample, 30... Bc6 31. Ra7 (or 31. Rc7
in mobility, due to the opening of Bd5 or 31. Rd6 Bb7) b4! and White is
lines. unable to carry out his plan of chal­
White cannot consider a3? to stop lenging on e4. All this goes to show
... b4. That would defy Rule 25 and the wisdom, in over-the-board play,
would get nowhere, as Black could of not wasting time in considering
easily renew his threat by ... a5- moves such as 28. Bc2 that involve
Black would then have the new op­ breaking simple principles.
tion of ... a4, whereupon White's three 28. Bxf6
queenside pawns would be held by 29. Rd7 Re7
the two black ones. 30. Rxe7 Bxe7
What of the c-Bishop? At g5 the
Bishop is doing nothing, once we ad­
mit that exchanging at f6 is impru­
dent. At d4, however, the Bishop
would be centrally posted and would

-101-
 CJS Purdy On The Endgame
more possibilities of complicating on
both wings.
31. .. . b4
32. f3 a5
33. cxb4
And after 33... Bxb4 the draw is
"dead," as White's two queenside
pawns are held by one.
Black actually suffered a mental
aberration, forgot to recapture the
Diagram 2, even with best play on pawn, and played 33... Ba6t??? That
both sides, is very likely only a draw! is irrelevant. The main lesson of the
The fighting forces are no longer game is the rapidity with which
equal (see the Corollary to rule 27). White, after outplaying his opponent
31. a4? in the opening and the middlegame,
This not only breaks Rule 25, but dissipated his advantage in the end­
seems to lack rhyme or reason. A game.
more justifiable breakage of Rule 25 Such things are done mainly be­
would have been 31. f3!?, covering cause of the dearth of literature on
the weakness on gil--but, of course, general endgame strategy. It is that
that would seriously weaken the dark want which we hope our articles will
squares. A deeper idea would be to help fill.
establish the Knight on d4 and then
to challenge the long light diagonal
by Bd1-f3. But that is very slow-as Pawn Promotion
we have seen, Black would mean­ M.E. Goldstein
while undermine the Knight's sup­ Spain is famous more for its bull­
port by ...b4. fights, cloistered senoritas, and mafi­
Note the idea behind 31... b4! (e.g. ana than for the skill of its chess play­
31. Ne2 b4!)-it illustrates the gen­ ers. Yet here is a real gem from a
eral technique with Bishops against game Ortueda-Sanz, Madrid 1934.
Knights: Open the game so as to de­
prive the Knights of pawn-supported Sanz
central posts.
More feasible is the challenge on
d5. Thus, 31. Ne2 b4 32. cxb4 Bxb4 3
33. Nc3 f5 34. Bd5. But that permits a
terribly drawish ending with Bishops
of opposite color after 34... Bxc3 35.
Bxb7 Bxb2 36. Bxa6. White had bet­
ter alter his plan and occupy the di­
agonal with the Knight. Thus, 34.
Nd5! Then after 34 ... Bxd5 35. Bxd5,
the Bishops-of-opposite-color end­ Ortueda
ing is not quite so drawish. White has Black to move

- 102 -
 All Facets Of The Endgame: Pawn Endings

If all the pieces were exchanged

off, the resulting simple pawn end­
ing would be an obvious win for 4
White. But Black, to move, can seize
the second rank while his Bishop
maintains a masked attack on the
white King who is endangered on the
long a7-g1 diagonal. The thematic
solution is a Rook sacrifice to facili­
tate the advance of the doubled
c-pawns. Hence, The Rook is driven from the key
1. .. . Rd2 square b4-to which it can no longer
2. Na4 Rxb2!! return after 6. Rxc4 cxb2.
The first surprise, infusing new 6. Resigns.
life into the doublets. For if 6. Rb7, then 6 ... c2 7. a4
3. Nxb2 c3 c1 =Qt; and the second doubled pawn
Now the Knight dare not move; for advances triumphantly.
if 4. Nd1 or 4. Nc4, then 4... c2-and This ending is the most brilliant
the pawn queens. Or if 4. Nd3, then promotion combination I have ever
4... c4 t 5. Rxb6 cxd3! Two united seen.
passed pawns on the third rank win
against a Rook except in rare cases
where mating threats allow a draw A Classic Pawn Ending
(Keres-Eliskases, Noordwijk This next endgame study, taken
1938). directly from the Dixon-Goldstein
4. Rxb6 game in the New South Wales
Expecting 4 ... axb6 5. Nd3, where­ Championship tourney, is a really
upon White wins easily. remarkable pawn ending-one can
4. ... c4!! hardly believe that it is not a com­
The second shock. Control of the posed study. In order to win Black's
,

square d3 is far more important than King must journey via devious by­
a mere Rook. With 5. Nd3 again pre­ ways to the very unlikely looking
vented, ...c2 becomes an acute threat. square h1.
Nor can White save the day by offer­ Yes, Black wins. The clearest
ing to return the Rook. For example, analysis that we have seen is by
5. Rd6 c2 6. Rd1 c3! !; and the two Dixon himself, though he claimed a
white pieces are curiously helpless draw through not carrying out his
against the doubled pawns. analysis far enough. It was Goldstein
5. Rb4 who originally drew our attention to
Apparently a knockout, as ...cxb2 the position; he showed a win in most
or ...c2 loses. But Black still has a variations-though in the trunk line
Joker up his sleeve. he broke off at the stage where we
5. . .
. a5!! give Diagram 6.
The final position deserves a dia­
gram .

- 103 -
 CJS Purdy On The Endgame

Endgame Study however, threaten to win the pawn.

Now let us proceed.
1. ... f5!
5 2. Ke2 Kc5!
Not 2... Kd5, giving White "the op-
position."
3. Ke3 Kc4
4. Ke2 Kd4
5. Kf3 Kd3
6. Kf2 Ke4
The inevitable has happened.
Black to play and win White has his King on its best pos­
sible square. White's next move is
The first move, of course, is 1... obvious.
f5! 7. g3
That creates a blocked pawn po­ If now 7 .. hxg3t, then 8. Kxg3 Ke3
.

sition. By a certain rule, we know at 9. Kh4 Kxf4-stalemate! This is the

once that White cannot save his main point of the study.
f-pawn. The rule is (see Rule 14 in
Part IV of Guide to Good Chess): 7. ... Kd4!

RULE28
� 6

F same side of the file of the

or the case of Kings on the

two blocked pawns (and each

King on its own pawn's side),
the King who occupies a limit
of the enemy pawn wins that
pawn and holds his own pawn.
� 8. Ke2!

White, too, must avoid disturbing

A pawn's "limits" are the three the pawns; for if 8. gxh4, then the
squares on each side of it on its own loss of the f-pawn is fatal. Thus, 8 ..
.

rank. Thus, in Diagram 5, the white Ke4 9. K(.3 Ke3 10. h5 gxh5 11. Kh4
f-pawn's limits are c4, d4, e4, g4, and Kxf4 12. Kxh5 Kg3 13. h4 f 4, and
h4. If the black King, then, can reach Black wins.
either c4 or d4, he must win the white White must also avoid playing Kf3
f-pawn. We can easily see that the at any stage. For example, 8. Kf3 Kd3
black King can reach c4 (e.g. 2. Ke3 9. Kf2 Ke4 10. Kg2 (if 10. Ke2, then
Kc5 3. Kd3 Kd5 4. Ke3 Kc4), so we 10... hxg3 wins) Ke3, and Black com­
know that the white f-pawn must fall. mands entry at f3, forcing White to
But here is the catch: if Black wins play gxh4 and lose.
the pawn, he only draws! Black must But after the text (8. Ke2!), how

104
- -
 Al l Facets Of The Endgame: Pawn Endings

can Black hope to win? Black's only finish seems to me far more beauti­
chance is to attack from the rear­ ful than many a spectacular Queen
to maneuver into a position where sacrifice. It is miraculous that such
he can play ... hxg3 and then, after effects can be obtained with so little
Kxg3, play ... Kf1! We may not see at material. Surely the inventor of chess
once that this should win, but we can was inspired by supernatural forces!
see that it is a chance. So let's work 16. h4 Kgt
for it. Now if White moves his King, one
8. ... Kc4! pawn or the other falls fatally.
9. Kd2 17. h5 gxh5
It seems now that Black has shot 18. Kh4 Kf2
his bolt; for if 9... Kb4, then 10. Kd3- 19. Kxh5
and Black gets nowhere. If 19. Kg5, then 19... K"3 wins.
We return to 9. Kf2 at the end. 19. . .. Ke3
9. . . . hxg3 ! 20. Kg5 Ke4
Now is the moment. And Black wins the pawn and the
10. Ke3 g2! game.
11. Kf2 NOTE: If 9. Kf2 (instead of 9. Kd2),
And now Black shoots down to f1. then 9... Kd3 10. Kf3 Kd2 11. Kf2 Kd1
11. Kd3 12. Kf3 (if 12. Ke3, then 12... hxg3;
12. Kxg2 Ke2 and Black wins as in the trunk line)
13. Kg3 Kit!! 13. Ke3 Kf1 14. Kf3 Kg1; and Black
We know that 13... Ke3? wins the wins (if 15. gxh4, then 15... Kh2).
f-pawn, but White just gets his old
stalemate. At the very beginning, if 2. g3, then
14. Kf3 2... Kd5, with a similar win. If at any
If 14. Kh4, then 14 ... Kg2 wins. If stage White plays g4, then the loss of
14. h4, then 14 ... Ke2 wins the f-pawn his f-pawn is obviously decisive.
and the game, the stalemate setup The whole study seems to us a
no longer being there. classic.
14. ..
. Kg1
15. Kg3 Kh1!!
The Zugzwanger Zugzwanged
S. Nikolenko of the Perth Chess
Club contributes the following inter­
esting pawn ending from a Perth
Chess Club championship.

Black wins by taking his King into

its position of minimum mobility­
nearly stalemating himself! Such a

- 105 -
 CJS Purdy On The Endgame
I? Salkauskas trying for a "remote passed pawn."
Were the position a rank higher up,
White would win-but now he loses:
8 36... f4! 37. h5 (if 37. gxf4?, then 37...
gxh4) fxg3. After both sides queen,
Black's Queen wins the white Queen
in three moves (exercise for stu­
dents).
Nik.olenko pointed out that White
should have played 33. h3! White can
afford to let Black move last on the
G. Lindley kingside, provided a block results; for
After 3% Kd6
... he still has a5 in reserve. If Black
forestalls that with 33... a5, Nik.olenko
From the diagram, Lindley lost. gives
Nik.olenko pointed out the correct 33. h3! a5
continuation, leading to a win-in his 34. Kb5 Kxd5
opinion. 35. Kxa5 g5
Lindley's error is an example of 36. Kb6! gxf4
the danger of applying a general rule 37. gxf4 e5
without a proper examination of the 38. a5
position. Clearly, whoever has to White wins since he will queen
move his King first loses. On the with check, thus giving Black no ti.me
k:ingside, the pawn position is sym­ to queen at all.
metrical; and the general rule is: At one stage in this analysis, how­
Whoever moves first in a symmetri­ ever, our correspondent made Black
cal pawn position yields to his oppo­ play a wrong move; and Black can
nent the privilege of moving last. Cor­ play a different move which should
ollary: If you are playing for Zug­ draw. Can you find it? See solution.
zwang, you should try not to be the
first to disturb such a setup.
This led Lindley to play 33. a5. Solution
There then followed 33... h6 34. h3 In Nik.olenko's analysis, Black
gf}! could improve with 35... e5! H 36.
Now White cannot continue sym­ fxe5, then 36... gf}. Black will queen
metrically. If White does continue first; and although White will queen
symmetrically, the rule about mov­ with check, White cannot force a win.
ing first being a disadvantage breaks
down. This means that each sym­
metrical pawn position must be ex­ Three Pawn Endings From
amined on its merits. If now 33. a5 The Melbourne Open
h6 34. h3 g5! 35. h4 gxh4 36. gxh4 h5, M.E. Goldstein
and Black gets the last move and The transition from an endgame
therefore wins. So White tried 33. a5 with Rooks or minor pieces and
h6 34. h3 gf}! 35. fxg5 hxg5 36. h4, pawns to a simpler type with pawns

-106-
 All Facets Of The Endgame: Pawn Endings

requires a flexibility of judgment

which even the most experienced
players may lack. 10
In endings with pieces, the sec­
ond best move is often sufficient; but
in pawn endings, the first error is
usually fatal. And the recent Mel­
bourne Open Tournament proved no
exception.

Looks all right, as White will se-

1: Watson Pulls Through cure a distant passed pawn on the
kingside. Then after both Kings have
G. Karoly mopped up on the kingside, White
will be nearer to the queenside.
Yet this natural move should lead
9 to a forced loss. White can still have
a draw by 41. fxe5t Kxe5 42. a5! bxa5
43. bxa5 Kd6.
41. . .
. Kc6?
Loses a precious move. As usual
in pawn endings, one tempo makes
all the difference.
A forced win for Black was subse­
C.G. Watson quently demonstrated by the surpris­
After 38. f4 ing 41... b5t! ! , forcing an entry for
the black King at d5.
Watson has just played 38. f4. The There are two main lines after 42.
game continued: axb5 axb5t. First White can take the
38. .•. Nxd4 b-pawn, play Ka6, and queen his b­
39. exd4 pawn. In the meantime, Black plays
Intending to establish a pawn ma­ Kd5 and queens his e-pawn. Black
jority on the kingside after Black's then checks on the a- and b-files, ex­
anticipated ... e5, Watson is playing changes Queens, and wins the king­
for a win; he therefore disdains the side pawns-White's King arriving
draw that should result from 39. on the scene too late.
Kxd4 with balanced pawn positions. Secondly, White can refuse the
39. e5 pawn after 41... b5t!! 42. axb5 axb5t
40. dxe5t fxe5 by 43. Kc3 [Black will be able to
41. f5 force pawn exchanges leading to a
drawn K + P vs. K ending. Try it
as an exerci.se. FPH.J Kd5 44. Kd3
e4t 45. Ke3 Ke5 46. g4 h5!, splitting
the white kingside pawns. The only
attempt at a "wriggle," viz. 47. f6 gxf6

-107-
 CJS Purdy On The Endgame
48. gxh5, fails against ...f5---Black is II: A Fifty-Minute Move
in the square of the h-pawn. My second example is from a
In the actual game, Karoly played game between two rivals of a quarter
41... Kc6?; and Watson won thus: century's tournament chess, Frank
42. b5t axb5t Crowl and myself.
43. axb5t Kd6
44. g4 F.A. Crowl
Comparing this position with that
in the previous note, we see that
Black is in Zugzwang and must lose 11
his e-pawn. If 44... Ke7, then 45. Kd5
Kf6 46. Ke4.
Black can still "play for traps" by
44... e4! 45. Kd4 e3 46. Kxe3 Kc5 47.
h4 Kd5 48. g5 Ke5, but White wins by
49. f6! gxf6 50. gxh6-Black's own
pawn at f6 prevents him from catch­
ing the passed pawn. M.E. Goldstein
The game concluded After 33 Rxb4
•••

44. h5
45. g5 e4 Black has just recaptured a pawn
46. Kd4 e3 on b4 with his Rook. The opening and
47. Kxe3 Ke5 middlegame had been played with
48. f6! gxf6 frivolous speed, the first 33 moves
49. g6 Ke6 taking barely half an hour.
50. Kf4 Resigns. I now took 50 minutes over my
next move, realizing that I was drift­
ing into an inferior position. I rejected
34. Rxb4 cxb4 35. Kd2 Kd6 because
Black would obtain a superior Queen
and pawn ending after White secures
his passed h-pawn (by h4! and g4).
The game continued:
34. Rc1 Kd6
35. Rc4 Kd5
The experts pointed out 35... Rb6!
36. Ra4 [Ed.: 36. Rh4!?] Rb4! 37. Ra2
Kc6, with advantage to Black.
36. h4! Rxc4?
Black should refuse the exchange
of Rooks and play ...Rb6.
37. dxc4t Kxc4
Loses off hand. But even after 37...
Kd6, there follows 37... Kd6 (if 37...
Kc6, then 38. Kd3!) 38. Ke4 Kc6 39.

- 1 08 -
 All Facets Of The Endgame: Pawn Endings

g4! Kd6! 40. gxh5! Ke7 41. Kd5; and offset the passed pawn that White
\Vhite queens either the c-pawn or will create on the queenside. So Black
the h-pawn. should refuse the exchange of Rooks
38. g4! hxg4 and should play 31... Rc7. If White
If 38 .. Kb4, then 39. g'5 and gxf6--
. then transfers his King to the queen-
queening on the f-file with check. side, Black secures counterchances
If 38... Kb3, then 39. g5-queening by ... e5!
on the g-file with check. And if 38... The pawn ending seems a forced
Kb5, then 39. gxh5 as Black is no win for White.
longer threatening to queen his c- 32. Kxd4 Kf5?
pawn. Loss of time, as the King has to go
39. h5 g3 back. 32... Ke7 leads to positions simi­
40. h6 � lar to the actual game, and 32... g5 is
41. Kf2 Resigns. refuted by 33. fxg'5t, 34. Ke5!, and
35. b4.
33. g3 Kf6
Ill: ''Playing To Lose" Realizing that 33... h4 34. b4 gives
My third example decided the win­ White an easy win.
ner of the tournament. Lazare, trail­ 34. b4 Ke7
ing Ozols, decided that he must play 35. c5 b5
for a win at all costs. Had Lazare This attempt to keep the position
drawn, he would have tied for first closed fails owing to the strength of
with Ozols. This game was played in White's supported passed pawn. No
the penultimate round. better is 35... Kd7 36. cxb6 axb6 37.
a4 f6 (to keep White's King from e5).
S. Lazare White then simply creates a passed
pawn by a5, forcing Black's King to
the b-file. That allows White's King
to penetrate on the kingside and to
gobble up the kingside pawns.
36. h3 f6
37. g4 h4
38. g5! e5t
39. Ke4 fxg5
40. fxg5 Ke6
Now that the pawn position has
K. Ozols been stabilized on the kingside,
Black to play his 30th White forces the decisive entry of his
King by sacrificing his passed pawn.
Play continued: 41. c6 Kd6
30. ... Nxd4 42. c7 Kxc7
31. Rxd4 Rxd4? 43. Kxe5 Kc6
A serious error of positional judg­ 44. Ke6
ment. Black has no real chances on And White ultimately won.
the kingside in the pawn ending to

- 109 -
 CJS Purdy On The Endgame

Pawn Ending But Shmeleff refused to believe that

Here is the pawn ending from the he could lose the pawn ending with
game John Purdy-Shmeleff. A three pawns to three, all united and
pawn ending, provided it has some all on one wing. And so, without
quirk or other, is always "box-office." checking up, he played:
Shmeleff 53. ... Ra7?
54. Ke5 Rxd7
Probably too late for . Ra2 now,
. .

as f4! works.
55. Rxd7t Kxd7
56. Kf6 Ke8
57. f4 Kf8
58. f5 gxf5
It is clear enough now that White
wins.
59. Kxf5 Kg7
John Purdy 60. Kg5 f6t
Black to play 61. Kxh5 Kh7
62. Kg4 Kg6
It is not easy to tell whether Black 63. Kf4 Kg7
could draw this ending with best play. 64. Kf5 Kf7
But one thing is certain: the only 65. g4 Kg7
chance is to keep the Rooks on-for 66. h5 Kf7
example, 53 ... Ra2 54. f3 for a start. 67. h6 Resigns.

-110-
 All Facets Of The Endgame: Rook Endings

[l\apler i
Rook Entlinqs

The Rook In Endgames ing more than one unprotected pawn;

R
ook endings are much com­ for if one pawn moves, the Rook can
moner than any other sort; and take another.
without book knowledge, they are the This explains the Rook's well­
hardest. Book knowledge is more known love for the seventh (second)
useful in Rook endings than in any rank. For on that rank there are
other department of chess. nearly always pawns, even far into
Here, I shall not try to discuss any the endgame; and none of them can
special type of Rook ending, but shall ever have a pawn supporting it.
just explain the Rook's peculiarities, If the enemy King is on his back
idiosyncrasies, and possibilities. rank, there is another advantage in
The Rook differs from all the other having a Rook on the seventh (sec­
pieces in having no special hanker­ ond). The Rook makes it harder for
ing for the center. The Bishop com­ the King to develop, especially to­
mands 13 squares from the center, ward the center. Often the King has
but only 7 from the edge; the Bishop, to creep out via h7 (h2), going away
therefore, is very keen on central from the center before he can get
posts. The Queen, being partly a back to it.
Bishop, is also keen on the center. Sometimes the King cannot get
The Knight's command, too, varies away from the back rank at all be­
from 8 in the center to 4 on the edge. cause he has no pawn to shelter him;
But a Rook, unobstructed, always then the Rook has the "seventh (sec­
commands 14 squares. ond) rank absolute," which is terrific.
So, as far as mobility is concerned, Diagram 1 shows an example of
all the Rook asks for is to be unob­ the "seventh rank absolute." If Black
structed. had a pawn on g7, he could play ... h6
and could then emerge by ...Kh7, etc.
But as the position now stands, the
Seventh (Second) Rank black King's confinement is "abso­
The Rook is particularly venom­ lute."
ous in attack along a rank contain-

- 111 -
 OS Purdy On The Endgame

Avoid Immobile Defense mobility is decreasing as fast as your

A Rook loves mobility (plenty of own Rook's mobility is increasing.
squares) and he loves to attack. A
Rook hates immobility and he hates
defense. It is often permissible to 1
place a Rook where he is mobile and
attacking nothing. It is sometimes
permissible to place a Rook where
he is immobile but attacking some­
thing. But above all , avoid putting a
Rook in a position of IMMOBILE DE­
FENSE. For instance, the enemy at­
tacks your unmoved pawn on a7 (a2), In Diagram 1 that state of affairs
and you can defend the pawn only by has gone as far as nature will let it.
... Ra8 (Ra1). Always prefer to give The enemy pawn is on the rank next
up the pawn, and use your Rook for to the queening rank, and the pro­
attack in some way. tecting Rook in front of it is a blob.
The advanced passed pawn, which
would make the position a win for
Rooks And Passed Pawns Black if the Rooks were to change
The posting of a Rook in relation places, is useless except that it can
to a passed pawn is a matter of vital be sacrificed to gain a tempo for
importance. Tarrasch's Rule is: Black's King to escape. If Black does
not do that-and tries to preserve
RULE29 his odd pawn till the crack of doom­
� then Black can easily lose, as the
white King can come marching down
P pawn.
lace your Rook behind the
with fearful effect. White's pawns act
� as a shelter for him against a check
by the black Rook-which would, of
course, ruin everything.
It doesn't matter whose Rook or
whose pawn-you can read the rule
either way. If the pawn is your own, Distance Lends ...
you are gaining doubly with every Another thing you must remem­
move it makes: the pawn is getting ber about the Rook is that he is a
nearer to queening, and your Rook stand-offish fellow-he likes to shoot
is gaining in mobility. If it is an en­ at long range, out of reach of the en­
emy pawn, you can regard each fresh emy. For instance, if the Rook is
square it gives your Rook (through checking a King, then he likes to be
its advance) as some compensation as far from the King as he can be-­
for the sad fact that it is advancing so that the King cannot easily get
at all. If the enemy's Rook is in front close enough to stop the checks.
of his passed pawn, to protect it, you For instance, in Diagram 5 it is
have the further joy that his Rook's Black's move; and Black sees that

-112-
 Al l Facets Of The Endgame: Rook Endings

White is going to check and then However, if the black Rook ceases to
play Kf7, threatening to queen. So shut out the white King, that King
Black wants to be ready to check the can approach nearer to the queen­
white King and to keep on checking. ing square; and White should be able
Black might attempt that by playing to draw.
... Rh5 or ... Rh4 or ... Rh3 or .... A be­ By the way, note here how very
ginner might wonder which square important it is for White's Rook to
to go to, or else might push his Rook be as far from the black King as he
back and back and suddenly get lazy is. Were the white Rook at b2 instead
and stop halfway. But a good player of bi, his first check would be his
doesn't think at all -he plays ... Rh2 last-Black would then answer with
automaticall y, and if White's Rook ... Kb3 and thereafter push up his
weren't stopping him he would go to pawn with an easy win.
h1. Black would even like to get right Now suppose we put the white
off the board and check from the Rook behind the passed pawn, ex­
table. Don't forget that. pecting to draw all the more easily
because of the Tarrasch Rule. All is
lost, we find, because our checks only
Tarrasch's Rule Amplified serve to drive the black King IN
Now it is time to get back to the FRONT of his pawn-from whence
Tarrasch Rule, which is not nearly he can help the pawn forward.
as good as it is cracked up to be. If now we start with Black's pawn
already on the fourth rank (at b4),
we find that we are lost anyway,
2 whether the Rook is behind or in
front of the pawn.
All this leads to the correct sup­
position that the Tarrasch Rule ap­
plies only if the passed pawn has
crossed (or can cross) the midline. If
the pawn is still on its own half of the
board, it is better for the enemy to
White to play and draw have his Rook in front of the pawn.

In Diagram 2, White's Rook is in

front of the passed pawn; but White 3
can draw by incessantly checking
until the black King is forced back
behind his pawn to b6. Then the white
Rook hops back to bi, preventing the
pawn from advancing, and stands
ready to resume checking if the black
King pokes his head out. Black, there­
fore, needs the help of his Rook if he To verify that, look at Diagram 3.
hopes to push his pawn forward. Here the white Rook is in front of the

- 113 -
 CJS Purdy On The Endgame

pawn; but because the pawn has not Lateral Adion

crossed the middle of the board, the When the passed pawn is your
white Rook has more mobility than own, your Rook is often quite well
the black one. In particular, the white posted i f protecting the pawn
Rook is shutting out the black King LATERALLY, provided thatthe pawn
from the fourth rank; and, therefore, has crossed the middle-line. For then
the white King can win the h-pawn. your Rook is in the enemy camp and
In Diagram 3, the logical play, with is playing an aggressive as well as a
White to move, is: defensive role.
1. Kh2 Kg5 In Diagram 1, for example, place
2. Kxh3 Kf5 the black Rook on the second rank
3. Kg3 Ke5 instead of in front of his pawn; and
4. Kf3 Kd5 we find a big difference. The black
5. Ke3 Kc5 King now has the possibility of help-
6. Kd3 Kb5 ing his pawn to queen by marching,
7. Rat Rc8 via b8, to b1, after which a check
Then, to all intents and purposes, would be answered by interposing
we are back to Diagram 2. White the black Rook. Furthermo r e ,
draws by checking the black King White's own pawns are vulnerable if
until that King goes behind his pawn. the white King leaves them.
In Diagram 3, transpose the With respect to lateral action by
Rooks. You will thereupon find that the Rook, here is an important point.
the black Rook can protect both of If you have an advanced passed pawn
the pawns by going to a3 when nee- and if you want your Rook to help
essary. Black wins. the pawn queen, the Rook should be
We may restate the Tarrasch Rule on the same side of the pawn as the
thus: enemy King. Otherwise, the enemy
King can use your pawn as a shel­
RULE30 ter-see Diagram 4, with irrelevant
� units omitted.

P
ut your Rook BEHIND a
passed pawn if the pawn
has crossed or can cross the
middleline. Put your Rook IN
FRONT of a passed pawn if the
pawn is held in its own half of
the board.

As we should expect, it all de­ White to move

pends on whether the pawn has more
squares in front or behind. In an actual game, White, receiv­
ing odds, played 1. c7?, allowing
Black to stop the pawn by 1... Kd7.

-114-
 Al l Facets Of The Endgame: Rook Endings

White should first have put his Rook So,

on the same side of the pawn as the 1. .. . Rh2
enemy King by 1. Rh7, whereupon 2. Re1t Kd7
the black King could no longer have 3. Re4!
stopped the pawn. This motif occurs Not 3. Re5 because then 3... Kd6
over and over again. hits the Rook away.
3. Kd6
4. Kf7 Rt?t
Bridge Building 5. Kg6 Rg2t
One of the most important devices 6. Kf6 Rf2t
in Rook endings is known as "bridge­ 7. Kg5 Rg2t
building." You have a pawn wanting 8. Rg4!
to queen; the enemy Rook is behind The bridge.
the pawn; you find some way of in­ 8. . .
. Rxg4t
terposing your Rook (supported of 9. Kxg4
course) between the pawn and the And the pawn queens. If the black
enemy Rook. Then the enemy Rook King were one square nearer (i.e.,
must either move away or exchange. K/e7 or K/e6) the pawn, the game
In either case, the enemy Rook is no would obviously be drawn.
longer there; and the pawn queens­
provided that the enemy King can't Conclusion
stop it. Diagram 5 was concocted by Merely reading over this material
L.R. de Lucena in 1495. will not help the tyro much. But if he
will read it and reread it, remember­
ing to come back to it again some
5 time, until he has thoroughly ab­
sorbed its contents into his system,
he will have acquired a valuable
grounding in the most useful depart­
ment of chess theory.
Good players, too, may find that
getting down to fundamentals has
removed confusion.
Black moves/White wins
RULE31
Black starts with 1... Rh2, as men­ ¢-
tioned previously; and then White has J\ void putting a Rook in a
to find the win. Naturally White starts .l"\..position of immobile de­
with 2. Re1 t, driving the black King fense. For an attacked, un­
away (2 ... Kd7). But what then? If 3. moved a- or h- pawn, usually
Kf7, then Black just checks until the
prefer to give up the pawn
white King has to go back to g8. What rather than to tie up a Rook for
White must do is to post his Rook its defense-instead, use your
where it can interpose at the right Rook for attack some way.
time.

-115-
 CJS Purdy On The Endgame

The Commonest a- or h-Pawn With Its Own

Rook Ending Rook In Front Of It
It is easy to see why the aggres­
sor Rook should so often find itself
6 in this awkward place, i.e. in front of
the pawn. The Rook has come down
and grabbed some pawns, and then
it is forced to defend its own passed
pawn. If that pawn is attacked later­
ally, the Rook has no alternative but
to go in front of it. If the pawn is at­
tacked from the rear, the Rook may
White King anywhere be able to defend it laterally, but
Black to move/Draw there is no advantage in that unless
the pawn has crossed the middle of
the board. By going in front of the
pawn, the Rook enables it to advance.
7 It is when the pawn has advanced
to the penultimate rank that the Rook
really feels the awkwardness of its
position-its mobility is reduced to
a minimum.
Diagram 6 illustrates that state of
affairs. If the Rook had a safe check,
White would win at once if he had
Either moves, White wins the move. Also, if Black's King were
at d7, e7, or f7, then White, with the
move, would win by 1. Rh8! Rxa7 2.
Rh7t, winning the enemy's Rook.
8 But as long as Black's King re-
mains at gl or h7 White can do noth­
ing. As soon as White's King plays
up to b6 or b7, defending the pawn,
Black simply checks until the white
King abandons the pawn; the black
Rook then resumes his vigil on the
a-file.
White King anywhere In such endings, a beginner is of-
Blaek to move/Draw ten amazed to see an expert scut­
tling his King away for dear life, tak-
The three positions given above ing it out of play instead of into play.
are vital to the understanding of the The expert wants to reach the safe
most commonly occurring type of squares.
endgame:

-116-
 All Facets Of The Endgame: Rook Endings

Evidently, if the defender's King If the pawn were not an a-pawn,

can reach the safe squares, we the King could find shelter by get­
should bring our pawn to a halt on ting to leeward of the pawn.
the 6th rank. That enables our King But the a-pawn has no lee side.
to shelter at a7 from vertical checks, So, in Diagram 8, you can put the
defending the pawn at the same time. white King anywhere at all; and
The aggressor Rook then becomes Black will either check it (if neces­
mobile. sary) or take some other suitable
Diagram 7 tells when we can win square on the 6th rank.
in that position. This time Black's For example, if the white King is
King would like to be close up, but at b5, then Black must start check­
we must assume that he has been ing immediately. And here note that
forced to stay at g7 until the move a7 it is of vital importance for the black
was no longer a threat. So the black Rook to be at safe checking distance.
King has been able to get no closer If the Rook were at d6 instead of far­
than f7. Black's last move was obvi­ ther along and if the white King could
ously ... Rb1 t, driving the white King go to b5, all would be over-for ex­
in front of its pawn. If Black had let ample, 1... Rd5t 2. Kc6, and there are
the white King stay at b7, then White no more checks.
could have freed his Rook without Would e6 do for the Rook? Yes,
having to block the pawn with his 1... Re5t 2. Kc6 Re6t 3. Kd7, and
King. Black has time for 3... Rf6!, getting
It is now Black's move again, and into the diagrammed position. But 3. ..

Black finds that his King is just one Rg6 or 3. . Rh6 would not do! The
.

square too far away. Thus, white King could not then be checked
1. Ke7 on the 8th rank. The f-file is the per­
2. Rb8 R moves fect one.
3. Kb7 Rb1t In Diagram 8, suppose the white
4. Ka8! R moves King is at d5. Then it would be fatal
5. a7 to check because of Ke6. But the
And White wins easily; for he can black Rook could safely play to b6;
force his King out either by Kb7 or, if and if Kc5, then back again to f6. Or
the black Rook stops that by taking Black could preserve the status quo
the seventh rank, by Rh8 and Kb8. with ... Kh8.
We thus see that Tarrasch's rule
that either the attacker's or the
defender's Rook is always best
But in Diagram 8 we see that placed behind the passed pawn is
Black, by some crafty device, has subject to at least one important ex­
transferred his Rook from the a-file ception. Sometimes "on the same
to the 6th rank. Black is still attack­ rank" is better. Note that if White
ing the passed pawn, and he can give plays Ra8t and a7 in Diagram 8 then
lateral checks instead of vertical the black Rook can resume his vigil
checks. Now our King can find no on the a-file. That might be impor­
shelter at all. tant.

-117-
 OS Purdy On The Endgame

Now For Another Endgame! is not at safe checking distance. The

With Diagrams 6, 7, and 8 to help black Rook must go back behind the
us, we can very easily fathom this pawn and abandon any attempt to
extraordinarily subtle endgame take the 6th rank.
which baffled all our solvers. We re- The only problem that remains is,
print the diagram. Can White produce Diagram 7? He
can.
There are two ways that White
can get his King to a7. The obvious
one is to play Ra8t and then simply
bring the King up.
If the white King is distant, this
method should not be used unless it
becomes necessary, as it involves giv­
ing up the seventh rank.
It does become necessary if Black
Black plays/Draw plays his King to f8; for Black then
White plays/Win threatens ...Ke8!, whereupon Ra8t
and a7 would come too late-the
With the move, Black can imme- black King would get right up to the
diately obtain Diagram 8. passed pawn.
1. Ra5! Example:
2. Ke3 Re5t! 1. Ke2!! Ra3
3. Kd4 Re6! 2. Kd2 Kg8
Easy draw. 3. Kc2 Kf8
Note that the black Rook had to 4. Kb2 Ra5
reach the 6th rank by means of a 5. RaSt (now forced)
check. If the Rook leaves the a-file 5. ... Kg7
without a check, then White wins (not ...Ke7??, as explained under Dia­
with ridiculous ease by Rb7! and a7 gram 1)
(the seventh rank absolute!). Inciden­ 6. Kb3 Rf5!
tally, this shows the merit of keeping And now White has to win by
our Rook on a7 as long as possible 7. Rb8! Ra5
rather than on a8. Do not unneces­ 8. Rb6!
sarily give up the seventh rank abso­ The white King can just get
lute. around before the black King can
Now the hard part, White to move! head him off.
Not 1. Ke3?, as Black again obtains The less obvious method, which
Diagram 3 by 1... Re1t and 2... Re6. should be used where possible since
Rather it avoids tricks, is to play the King
1. Ke2!! down to b8. There the King is safe
Now Black's only way to reach the from a vertical check since his Rook
6th rank with a check is 1... Ra5 2. would interpose at b7.
Kd3 Rd5t 3. Kc4 Rd6. But then 4. After 1. Ke2!!, it is so easy to force
Kb5, and we have seen that the Rook the passage of the white King to b8

- 118 -
 All Facets Of The Endgame: Rook Endings

(as long as Black does not play his 6. Kb5 Kf7

King to f8) that we need not give all 7. Re4
the play. Just imagine that White has And White wins easily by the
transferred his King from f2 to b6 well-known device of interposing the
and that it is Black's move. Let the Rook and then playing it behind the
black King be at g8, say. pawn.
Playing from that situation, we
have:
1. •.• Rbtt Rook Endgames
2. Kc7 Rctt M.E. Goldstein
If 2... Ra1, we use the Ra8 method, A series of articles on endgames
as White's King is no longer distant. with Rooks and pawns.
For example, 3. Ra8t Kf7 4. Kb7. If
now Black tries to steal a march on
us with his King by 4... Ke7, then we No. 1
win by 5. a7! The black King is out of Endgames with Rooks and pawns
his crease (i.e. away from the safe occur more frequently than any oth­
squares gland h7); and if the white ers, yet they are seldom played well.
King is checked, he simply retreats My example is taken from a recent
until the checks end. "A" Grade match between two promi­
3. Kb8! Rat nent Sydney players.
Now K i ng and Rook change
places.
4. Ra8 Kf7 9
If 4. . Rb1 t, then 5. Ka7t; and if
.

4 ... R elsewhere, then 5. Kb7t-and

the same thing comes about. Com­
mon sense tells us that we should
not play the King in front of the pawn
until driven to.
5. Kb7! Rb1t!
6. Ka7 White to move
And we have Diagram 7.
By good middlegame strategy,
White has won a pawn; and a simple
endgame win is now in sight. A giance
There is some interesting play if at the diagram reveals that all the
Black forces us to play Ra8 immedi­ black pawns are weak-to win White
,

ately. Starting from the diagram, we need merely force an entry with his
have: King. The two squares indicated are
1. Ke2!! Kg8 b5 and d5, the latter for preference.
2. Kd3 Kf8 Hence White should continue: 1. Rf3!,
3. Ra8t! Kg7 threatening Rf6t, 1. .. Ke7 2. Kd5 Re6
4. Kc4 Rf1!? 3. Rf6!! Evidently 3. Re3 would win a
5. Re8! Rat second pawn, but the text leads to

-119-
 CJS Purdy On The Endgame

an amusing finish. 3... Rxf6 4. gxf6t 11. Kf5? Kg'3! wins for Black).
Kxf6 5. Kd6! The point of the pawn­ 8. ... fxg6
sacrifice, forcing a Queen. Accurate 9. h5 g5!
counting is necessary, as Black also On his 8th move, White had reck­
queens. 5... e4 6. c6 e3 7. c7 e2 8. oned on 9... gxh5-after which White
c8=Q e1=Q 9. Qh8t!-and White can get to h1 and draw.
mates in two moves by Qg8t and 1 0. Resigns.
Qg5#.
If, instead, Black pushes 1... e4 in
response to 1. Rf3!, then 2. Rf6t Kd7 No. 2
3. Rxf7t Ke8 4. Rf4 Re6 5. Kd5; and Few players thoroughly master
White wins. the correct method of exploiting an
White chose the alternative break­ extra pawn in Rook endings (see No.
through b5, which leads to a more 1 of this series). Reuben Fine, in his
difficult win. monumental workBasic Chess End­
As a general rule, the player who ings, gives the following rules for the
is ahead by a pawn should not ex­ typical case where the pawn posi­
change pawns in such endings-for tion is balanced and the black King
the fewer the pawns remaining on is in front of White's passed pawn:
the board, the greater is the danger a) Weaken the black pawns by
of a draw. compelling their advance, whereupon
The actual game continued: they may be blockaded.
1. b5 axb5 b) Tie up the black Rook by at­
2. Kxb5 e4 tacking the weakened pawns.
3. Re3? c) Advance the white King and
Throws away the win, which was passed pawn as far as practicable-­
still there by 3. Rd4!-if thereupon usually the pawn reaches the fifth
3... Ke5, then 4. Rxe4t! The student rank.
should note that the text move per­ d) Put Black in Zugzwang by
mits Black to play ...Kd5, a square threatening to enter with the King or
denied him after 3. Rd4! to capture more material.
3. Kd5 e) Reduce to an elementary book
4. Rxe4 Rxc5t! win by favorable exchanges or reduc­
5. KM Rb5t! tion to a stock winning position.
6. Kxb5 Kxe4
7. Kc5 Kf4 Many are familiar with the score
8. g6?? of Rubinstein's brilliant win against
Dejected by his previous errors Emanuel Lasker at St. Petersburg
and convinced that all his pawns are 1909. By a subtle combination, Ru­
lost, White now hurls the draw after binstein won a pawn in the mid­
the win! Actually, White can draw dlegame; but the actual win offered
very easily by 8. Kd4! Kxg4 9. Ke4 technical difficulties. I shall show
Kxh4 10. Kf4!, and White has the op­ how Rubinstein's genius overcame
position. For example, 10... Kh5 11. the difficulties, giving a classical
Kf5! Or 10... Kh3 11. Kf3! (note that demonstration of endgame artistry.

-120-
 All Facets Of The Endgame: Rook Endings

Lasker since Black cannot permit Kg'6.

8. e5 Rb7
9. Rd6 Kf8
10 10. Rc6 Kf7
11. a3! Resigns.
Why resign?, you ask. Black is
now hopelessly Zugzwanged, his
only possibilities being:
a) 11... Kf8 12. Kg6 Rb3 (the point
ofWhite's 11th move was to bar Black
from the square b4) 13. Rc8t Ke7 14.
Rubinstein Rc7t Ke6 15. Rxg7 Rxa3 16. Kxh6,
White to play and White's two united passed pawns
win easily.
The first two steps in Fine's gen­ b) 11... Re7 12. e6t Kg8 13. Kg6!
eralization have already been carried Re8 14. e7! Kh8 15. Rd6 Kg8 16. Rd8,
out, as Black's Rook is tied to the and White mates.
defense of Black's weakened a-pawn.
Play proceeded: Now let's take up the variant on
1. Ra6 Black's 6th move (see Diagram 11).
First confine the hostile King be- 6. ... Kf7
fore following rule c). Passive resistance, postponing de-
1. Kf8 feat for a few more moves.
2. e4 Rc7 7. Kf5 Ke7
3. h4! Kf7 8. g5 Kf7
4. g4 Kf8 9. e5
5. Kf4 Ke7 Postponing g6 until Black vacates
6. h5 f7.
9. Ke7
10. g6! h6
11. Re6t! Kd7
Or 11... Kf8, whereupon 12. Rd6
wins more slowly, as Black cannot
prevent the white Rook from reach­
ing the 8th rank with the deadly
threat of Rg8.
12. Rf6!
Naturally the Rook is taboo-if he
6. ... h6 is captured, then the passed pawn
This weakening move is virtually wins off hand after 13. Kxf6!
forced, as Black dare not let White's 12. ... Ke8
g-pawn reach g6 (see the variation 13. Rf7!
below). The villain still pursues him . The
7. Kf5 Kf7 resulting pawn ending is a simple win
Rule d) had now been obeyed, for White.

-121-
 CJS Purdy On The Endgame

13. ... Rxf7 with his King.

14. gxf7t Kxf7 47. ... Ra8
15. e6t The Rook is best placed as far
Now work out the win for your­ away from the King as possible, in
self. The black King is to be stale­ order to check repeatedly on the file.
mated on e8, forcing Black ultimately 48. Kb2 Kf4
to give up his g-pawn. 49. Rc3! Rb8t
50. Ka2 Ra8
Or 50.. Rb7 51. a4 and Ka3 .
.

No. 3 51. Kb3 Rb8t

The master must excel in end­ 52. Kc4! Rc8t
game technique if he is to reap the Tarrasch's Rule, .. the Rook be­
reward of his middlegame tactics, longs behind the hostile passed
and the once New South Wales cham­ pawn," does not apply here. For if
pion, Lajos Steiner, is no exception. 52... Rbt, then 53. a4 (threatening
On the 46th move of my State Cham­ Ra3) Rat 54. Kb4; and the King es­
pionship game with him, the follow­ corts the passed pawn to the queen-
ing position arose. ing square.
53. Kb4 Rb8t
M.E. Goldstein 54. Ka4 Ra8t
55. Kb5 Rb8t
56. Ka6 Ra8t
12 57. Kb6! Ra4
Zugzwa ng! One more check
would be useless; for if 57... Rb8t,
then White's winning line is 58. Ka7!
Rb5 59. Ka6!; and the advance a4 can
no longer be prevented.
58. Kb5 Ra8
59. a4 Rb8t
L. Steiner 60. Kc4 Rc8t
Black to play 61. Kd3 Ra8
62. Ra3
Black sealed The winning position, with Rook
46. ... Ke5 behind the passed pawn. Contrast
Not, of course, 46... Kf4 because the note to Black's 52nd move.
of 47. Rb4t!, forcing exchange of 62. .. . Rd8t
Rooks and queening one move be­ 63. Kc4!
fore Black. Emerging for the final sally. A
47. Kc2 plausible try is 63. Ke2; for if 63 . . .

Although White is a pawn ahead, ReSt, then 64. Kf2 Ra8 65. a5!, fol­
the win offers technical difficulties. lowed by Ra4t and Ke3. However,
White must try to advance his Black throws a spanne r in the works
a-pawn, supported by his King, while by 64... Re3!! 65. Rxe3 stalemate, or
preventing Black from entering 65. Rat Rxh3-and Black draws.

- 1 22
-
 All Facets Of The Endgame: Rook Endings

63. . .. RcSt erful. Actually, Black's position is

The stalemate threat 63... Rd4t! very bad-can anyone find a draw
is averted by 64. Kc5 Rd5t! 65. Kc6 for him?
Ra5 66. Kb6, and the a-pawn marches One can look at it this way. Sup­
on. pose Black's two passed pawns were
Black now gave a few more checks off the board. Then White must win
and, having reached the three score with his five pawns to three. But if
moves and ten, belatedly gave up the Black never utilizes his two passed
unequal struggle. pawns, those pawns are as good as
off the board. How can Black utilize
those pawns?:
A Paradoxical 1) By bringing his King around to
Rook Ending them. But a very brief analysis will
Diagram 13 shows the final posi­ show that by that time White will him­
tion in the second of the two games self acquire two passed pawns (start­
between Purdy and Koshnitsky in an ing with Kf3, g4, etc.), well advanced
invitation tourney in Sydney. Kosh­ and supported by their King. Black
nitsky, as Black, had deliberately will be doomed. In other words, the
played into it, lured by the thought of plan is too slow.
two united passed pawns. However, 2) By giving up one pawn in order
having arrived at the position, Kosh­ to advance the other, thus driving
nitsk:y quickly realized his mistake White's Rook into limbo. The trouble
and offered a draw. As a draw made with this is that White not only gets
Purdy certain of first place and since one pawn, but ultimately the other
he was fairly short of time, Purdy ac­ as well. For example,
cepted the offer. Purdy's last move 1. ... b5
was K(h2)g'J. 2. Kf3! b4
If 2... Kg8, then 3. Ke2; and Black
KoshnJtsky cannot play . ..Kf7 yet because of the
fork.
3. Ke2 Kg8
If ...b3, then Rc3; and the advanced
pawn falls.
4. Rc4 c5
5. Rxc5 Rb7
6. Rc2 Kf7
7. Kd3 Ke6
No help is ...b3.
8. Kc4
White wins the black pawn with­
out losing any of his own; or if Black
forces the exchange of Rooks, White
At first sight, most players would wins the pawn ending.
like Black's position-two united 3) To try to attack some weak­
passed pawns are proverbially pow- ness in White's position, abandoning

- 1 23 -
 OS Purdy On The Endgame

the two passed pawns in order to re­ 3. g4 hxg4t

duce White's majority on the king­ .
4 hxg4 Kg7
side. But that will not work. For ex­ 5. gxf5 gxf5
ample, 6. Ke2!
1. ... h5 Useless here is Rf6.
2. Kh4! Rd2 6. ... Kf8
The only chance. Losing still is ... b5, White playing
3. g3 Rd3 Rc5 at once this time.
4. Kxh5 Rxe3 7. e4! fxe4
5. Kh4 b5 8. Ke3
6. Rxc7 b4 The advanced passed pawns win.
7. Rb7 b3 Or
But now 8. g4 wins. White still has 1. ...
Kg8
his two pawns up on the kingside. 2. Kf3 Kf8
Or The black King can't go to fl be­
1. g6 cause of the fork.
2. Kf3 h5 3. g4 fxg4t
3. g4 h4 4. hxg4
To fix White's pawn on h3. White's pawns aided by their King
4. gxf5 gxf5 will win.
But now White can afford What is the chief factor in Black's
5. Rf6! b5 defeat? Is it the remoteness of
6. Rxf5 Black's King, or is it the confinement
For example, if 6... c5, then of Black's Rook to a defensive task?
7. e6 Rc7 Certainly the latter; for if we start
8. e7 Rxe7 with Black's King a move nearer, at
9. Rxc5 Rb7 g8 instead of at h7, we shall still find
10. Rc3 the game extremely difficult for
followed by Rb3; and White should Black-even if you put the black King
win. on f8, the advantage is still with
4) To play defensively for a while, White. But with the black Rook on
waiting for a better opportunity to b7, or even on c8, Black should win.
sacrifice on the queenside. For ex­ We conclude, correctly, that the
ample, worst thing in a Rook ending is to
1. g6 have a Rook confined to a purely de­
2. Kf3 h5 fensive role.

-124-
 Al l Facets Of The Endgame: Knight Endings

[hapler?
Kniqhl Enllinqs

An Instructive Knight Ending us, criticizing as we go.

J\ study of Knight endings-a spe­ A point of interest is that claims
fi.cialty of the late Aaron Nim­ have been made that the ending is
zovich, is attractive as well as use­ very difficult to win perhaps even im­
,

ful. possible!
The play went as follows:
Crowl 49. g3
Here we have a point of technique.
In such positions, as a general rule,
what boots it to move a pawn when
one's aim is to develop the King? If
the white King particularly wants to
go to f3, it takes no longer to make
three King moves than to make one
pawn move and two King moves. And
if you want to go to some other
square, moving the King gains a
Coultas tempo. So moving the King must pay
White to play bis 49th move in any case, unless there is some ob­
vious objection. Here White had a
A case in point. The diagrammed straightforward win by 49. Kh2, aim­
positio n o c curred in the game ing at Kg3 and ultimately at the cap­
Coul t as-Crowl in a Victorian ture of Black's g-pawn when Black
Championship tournament, and goes for the queenside.
Black (Crowl) was allowed to draw 49. ... Nd3
it. Coultas seemed to have been in 50. b3
one of his slack moods and played Here White had a more difficult
the remainder of the ending at about problem. This way, the white Knight
the level of the average club player. is tied to the defense of the b-pawn
That being so, it should be instruc­ after Black's next move, a serious
tive for the more inexperienced of our handicap. On the other hand, if 50.
readers to play over the moves with b4, then the black King can enter at

-125-
 OS Purdy On The Endgame

c4 via e5 and d5. A little calculation, then automatically cast out from your
however, would show that that inva­ mind all side issues such as pawn­
sion need not be feared, as it enables grabbing. Everything will be subor­
White to play his King to f3 and to dinated to the main purpose.
defend the c-pawn by Ne2, leaving a Here it is obvious that if there is a
comfortable win without complica­ chance of winning it can only be by
tions. playing up the King to the assistance
50. Nc1 of the passed pawn. Therefore, un­
51. a3 Ke5 less there is some obvious reason for
52. Kg2 Ke4 doing something else first, we natu­
Had White played 50. b4, this rally consider 60. Kg3 at once. If there
would fail because of 53. Nb5. is any objection to that, it could only
53. h4! be 60 ... a6 or 60 . a5; for otherwise
..

Owing to his King being kept back, the a-pawn can wait. If 60 ... a6, then
White is compelled to try a pawn­ 61. Nd6 Nh8 62. Kg4!! (again not the
race; and the win is no longer the obvious pawn grab) Kxa3 63. Kg5!;
armchair affair that it was. Rapid and it is obvious that White cannot
King development is always prefer­ lose--it is not difficult to prove that
able where feasible. White wins against any defense, by
53 . ... Kd3 Kf6-g'7, etc. Don't fail to verify this.
54. g4 Kxc3 And if 60 ... a5, then 61. Kg4 Ka4
55. Nb5t Kxb3 62. Kf5!, winning easily by sacrific­
56. h5 gxh5 ing the Knight.
57. gxh5 Nd3 If 60... Ka4, then White wins ei­
58. h6 Ne5 ther by 61. Nxa7 (having gained a
59. h7 Nf7 tempo) or by simply sacrificing the
Psychologically better than 59... Knight.
Ng6, as White would then have been Thus 60. Kg3! wins in all varia­
more likely to see that he could win tions. So it would have against 59...
by 60. Kg3!, etc. If thereupon 60... Ng6 also, as the student is advised to
Kc4, then simply 61. Kg4!, offering work out.
the Knight. 60. ... Kxa3
60. Nxa7? 61 . Nb5t
Losing a vital tempo. Players will The straight-out King tour is now
greatly simplify their endgame tech­ one move too late; for after the white
nique if they will always keep in mind King captures the Knight on h8, he is
our golden rule: then shut in by the black King-a
position that every player should
RULE32 know.
-¢-
;n if you can.
61. Kb4

Q
ueen a pa 62. Nd6 Nh8
63. Kf3?
This is the only really bad mis­
take that White makes. If the King
If a pawn looks queenable, you will tour was of no use on move 61, still

- 126 -
 Al l Facets Of The Endgame: Knight Endings

less can it be of any use here. By 63. does not decide the case. The truth
�xb7, White would probably have is that if, in the diagram, Black plays
won still; for the saving defense is so ... Kf7 (the move nearly everyone
subtle that the odds would be in fa­ would make in actual play as a mat­
vor of its being overlooked over the ter of course) then White wins. How?
board-even by a player of Crowl's On the other hand, if 1... Ke7!, then
strength. After 63. Nxb7! Kc4 64. Kf3 White can only draw. Why? We leave
Kd5 65. Kf4 Ke6 66. Kg5, we arrive at these two little puzzles for our stu­
the position in Diagram 2. dents-most players will find them
quite entertaining.
The key to such positions is that
you cannot gain or lose a move with
a Knight.
The rest of the game went as fol-
lows:
63. ... b5
64. Ke4 Kc5
65. Ke5 b4
66. Ne4t Kc6
Black to play and draw 67. Kf6 Kd7
68. Nc5t Ke8
The endgame Knight and pawn vs. Drawn by agreement.
Knight is governed by the following The white King eats the Knight
rule: on h8, but then he can never get out.

RULE33
� Solution
The solution of the ending dia­
I
n the endgame Knight and
pawn vs. Knight if the pawn grammed above is as follows.
is on the rank next to the The natural move is 1... Kf7; that
queening rank and if it can be loses, however, because after 2. Kh6
supported by its King on that White can maneuver with his Knight
same rank/Always a win, un­ so as to force the black King to lose
less the defender's King can command of g7-then White plays
occupy the queening square Kg; and wins.
and cannot be ousted there­ 1 . ... Kf7?
from. 2. Kh6 Kf6
3. Nd6 Ng6
4. Ne4t Kf7
Otherwise the outcome is uncer­ 5. Nc5! Nh8
tain. 6. Nd7
The a-, b-, g-, and h-pawns are the The key square. Here the Knight
most winnish. prevents ... Kf6, so that Black must
Here the white King can be pre­ play either 6 ... Ke7 (allowing 7. Kg;)
vented from reaching g;, so the rule or 6... Ng6 (allowing 7. Ne5t, etc.).

-127-
 CJS Purdy On The Endgame

Black cannot evade his fate by A freak exception is the case

moving his Knight because a Knight where the defending Knight is in an
cannot gain or lose a move. If in initial position from whence he can
the above, 2 ... Ng13, then 3. Nc5!, etc., safely fork the King and new Queen,
wins just the same. Try it. if the pawn queens immediately.
But if 1... Ke7! 2. Kh6 (forced) Kf7!,
then White can never play Nd7 when
Black's King is on f7, again because
a Knight cannot gain or lose a
move. For example, 3. Nc5 Kf6 4.
Nd7t Kf7 5. Ne5t Kf6, and so on ad
infinitum.
Therefore, 1... Kf7? loses, and 1...
Ke7! draws.
Andre CMron, in his Traite Com­
plet d'Echecs, states that in the end­
ing Knight and pawn vs. Knight, the
pawn always wins if on the rank next The method of winning is simplic­
to the queening rank and "supported ity itself if only you work out what
by its King." However, we see from you want to do. First make Black's
the example above that there is an Knight move by putting the white
exception in the case of an a- or Knight en prise to it. The black
h-pawn. Knight has only one flight square
(here f5) from whence he threatens
a fork if we queen. Then we move
Knight And Pawn vs. Knight our King to b8 to avoid the fork-the
The general rule for the ending black Knight must then take up a new
Knight and pawn vs. Knight is: post, here d6. The Knight is more re­
• Pawn requiring one move to stricted on d6 than he was on e7 be­
queen: A win. cause now he has no flight square-­
• Pawn requiring two moves to so one more smite from our Knight
queen: Often a win if the attacker's will settle things.
pieces are relatively favorably Now let us start. We shall play the
placed. best moves for Black, leaving the stu­
• Pawn requiring three moves to dent to work out wins against oth-
queen: To win requires that the ers.
attacker's King and Knight both be 1. Nc4 Kf5
well placed and that the enemy King 2. Ne3t Ke5
be very distant. 3. Ng4t Kd6
The variation 3 .. Ke6 4. Nh6 Ke5
.

The a-, b-, g-, or h-pawn has the (best) 5. Ng8 gives the text line.
best chance of winning because the 4. Nf6! Ke5
defending King has less scope. 5. Ng8 Nf5
The vital thing is to advance the 6. Kb8 Nd6
pawn as far as ever you can in safety. The hard part is over.

- 128 -
 All Facets Of The Endgame: Knight Endings

7. Ne7 Ke6 rank by ...Ne7, afterwards winning

8. Nc6! the h-pawn. Knight and pawn against
And White wins, for his Knight Knight, with the pawn held to the fifth
must get either to b7 or to c4. rank, could never be a win with one
of the attacker's pieces so distant.
Black's 54 ... Kg2 looked like los­
Correspondence Endgame ing time, as the King eventually had
For the position that follows, we to go back to g4. The move, however,
are indebted to R.D. McKenzie who forced White to lose time in moving
arrived at it in a correspondence his doomed h-pawn twice.
game. The players devoted much Let us now see if White could have
analysis to the position and came to improved anywhere. On his first
the correct conclusion: the game is move, White had to do something to
drawn! White's men are all unfortu­ free his Knight for its journey to the
nately placed, but it comes as a dis­ kingside. For if 1. Kb5? Kf5 2. c4, then
tinct shock to find that White cannot simply 2... Nxc4, drawing.
win with two pawns up. So the only alternative to 1. Kb3
was 1. c3. That has the advantage of
WE. Roberts not retreating the King, who is want­
ed to help the c-pawn when it gets
near to queening. And if the game
4 continued as in the text, then White
would win comfortably by 58. c5, as
he soon produces the book win (P/c7,
K/b7). [Ed.: Unfortunately for our
purposes, the Nalirrwv Tablebases
shows it to be drawn after 58. . .
Kxh5 as KM mentioned to us.]
Still, cannot Black punish White
R.D. McKenzie for keeping his King on the edge of
White to play; but cannot win! the board? Yes, on move 54 Black
would play 54 ... Kg4!, threatening
The play went: ...Kh3, etc., and Black would draw,
51. Kb3 Kf5 for 55. Nf2t is only repetition, and
52. Nd3 Kg4 other variations are easily worked
53. Nf2t Kf3 out.
54. Nh1 ! That raises the question, Won't
The only chance to win . 54... Kg4 also draw in the actual
54. Kg2! game? No, and the following analy­
55. h4 Kh3 sis shows why (see Diagram No. 5).
56. h5 Nf5
57. c4 Kg4
Dead draw! If the c-pawn ad­
vances, Black forks it; otherwise
Black holds up the c-pawn on its fifth

-129-
 CJS Purdy On The Endgame

Strategy In
Knight Endings
5 Our illustration for this article is
a classic ending won by Frank Mar­
shall against Leonhardt. It has never
previously been thoroughly dis­
sected, as far as we know, and, like
many "classic" endgames, it is re­
markable for the number of errors
on the part of the loser that a search­
1. c4 Kh3 ing analysis reveals.
If 1... Nxc4?, then 2. Nf2t! wins.
2. c5 Kxh2 Marshall
If 2 ... Nf5, then 3. Ng3! Ne7 4. Nf1
Kg2 5. h4; and White wins.
3. c6 Nd5! 6
If 3... Nf5, then White soon gets
the book win.
4. Nrl Kg3
5. Nd3 Kg4!
If 5... Kf3, then 6. Kc4 Nc7 7. Kd4!
Ke2 (if 7... Kg4, then 8. Ke5 wins). 8.
Nb4 Kd2 is drawn.
6. Kc4 Nc7 Leonhardt
7. Kc5 Kf5
8. Kd6 Nb5 t The diagrammed position, with
If 8... Na8, then 9. Kd7 Ke4 10. White to move, would not appear to
Kc8! Kd5 11. Kb7 Kd6 12. Ne5!; and the average player to be at all disad­
White wins. vantageous to White!
9. Kd7 Ke4 Indeed, one untrained in endgame
10. Net Kd5 lore would probably prefer White's
[Ed.: While 10. Nel looks like a position. White's King is a square
winner, the NT indicate it's drawn. nearer the center, and Black has an
Purdy gives "10 ... K-rrwves" fol­ isolated pawn. And is not an isolated
lowed by 11. Nc2 K-rrwves 12. Na3 pawn generally a more serious weak­
N-rrwves and then 13. c7. But K.M. ness in the endgame than in the
points out th.at 10 ... Kd5 11. Nc2 middlegame?
Ke5! 12. Na3 is spoiled by 12 .. . This estimate would be quite cor­
Nxa3! After 13. c7 Black plays 13. . . rect with a slight alteration of the
Nc4! and now White is stopped position, e.g. with Black's Knight on
from queening without forfeiting e6 instead of on f6, or if any of White's
the new piece on c8.j kingside pawns were still unmoved.
White's weakness lies in his king­
side pawns-all have moved. Moved
pawns are weak, in general, because

-130-
 All Facets Of The Endgame: Knight Endings

they are nearer to the enemy; and would still prevent the King from de­
that applies particularly where the veloping.
enemy is a King or a Knight, to whom Only a spirit of overconfidence, a
(unlike a Bishop or a Rook) distance feeling that it was impossible to lose,
is a serious matter. could have made White choose this
In this instance, Black threatens line on his second move; and it is this
at once to win a pawn by ...Nh5. To spirit that loses many endings.
meet that, White must either advance 3. c3?
his pawns still farther, by 1. g4 Nd5 As the King cannot develop via d3,
2. f5-weakening more squares-or this does not really get White any­
else make a defensive maneuver with where; and it creates a new weak­
his Knight. ness by gratuitously advancing a
1. Nd4? pawn. The only chance was 3. Kd1!,
Not realizing how easily the game offering the h-pawn, which it would
can be lost, White makes an error not be good for Black to grab (the
right at the start. The text move per­ student should verify this). The
mits ...Ne4, hindering the develop­ threat of the grab is stronger than
ment of White's King. Correct was 1. the execution; for even after going to
�d2!, permitting White's King to de­ d1, the King is still unable to de­
velop via d 1, as 1... Nh5 is met ad­ velop-his own Knight blocks him!
equately by 2. Nf1 and 1... Nd5 by 2. And so Black would preserve good
Ne4. winning chances by simply develop­
Thus 1. Nd2 d5 2. Kd1! Ne4 3. Nf1 ing, with 3... Ke7 or perhaps 3... h5.
Kf2t? 4. Ke2 Nxh3 5. Kf3 g5 6. Kg2 g4 White would have to continue with 4.
7. Ke3, regaining the pawn. Ke1 and would still have difficulty in
1. ... Ne4 developing his King. For even then
2. Ne2? White is not threatening Kf1, because
A second error. White refuses to of ...Nd2t and ...Nc4!-winning either
take the risk of giving himself two the b-pawn or the a-pawn. Note how
isolated pawns with 2. Nf5 g6 3. the a-pawn, because advanced, is a
Nxh6t Kg7 4. Ng4 Nxg3 5. Kd2. weakness, even though protected!
\:Vhite's King could then have devel­ This line (3. Kd1), however, does
oped, and Black's winning chances give drawing chances. The feeble and
would have been less. planless text move, the one most
2. ... Kf8 players would make mechanically,
We now have a position of a type compromises White's game beyond
that always presents winning chanc­ repair. Leonhardt was a master, but
es in Knight endings: the white King evidently not of endgames.
cannot develop safely because of the 3. Ke7
weakness of one of his pawns. The 4. Kc2 Ke6
King cannot go to d3 (after c3 and 5. b3
Kc2). Moreover, to push h4 would White has no defense-he can
weaken White fatally-if he then only wait for Black to carry through
tried to fill up the hole at g4 by play­ the following decisive maneuver.
ing g4, the new weak pawn at g4 5. ... Nf2

- 131 -
 CJS Purdy On The Endgame

6. Nd4t 1 4. Resigns.
Preventing ... Kf5 or ... Kd5 (be­ Note that the isolated d-pawn
cause of Nf5). never came into question. It would
6.••• Kf6 have if it had been a square farther
7. h4 h5 up. In the middlegame, an isolated
The fixing of weaknesses, where d-pawn or e-pawn is weaker if one
possible, is a particularly vital fea­ rank away from home than if two
ture of Knight strategy-much more ranks away from hom�that is so
so than with other pieces. because it is more obstructive to its
8. c4 own pieces when only one rank away
The poor old white King is still from home. Naturally, the same thing
imprisoned, now in reverse fashion. does not apply in Knight endings.
The King's di and d3 are barred di­ There are fewer pieces to be ob­
rectly, and his d2 is barred because structed; and, besides, the disadvan­
of his g-pawn. tage of being nearer to the enemy is
8. Ne4 greater.
9. Ne2 Kf 5 The decisive factor in this ending,
10. Kd3 Nc5t hindering the enemy King from de­
11. Kc3 Kg4 veloping, is one that should be
12. b4 Ne4t watched for in every Knight ending
13. Kd4 Nxg3 as soon as it arises.

-132-
 [l\apler ID
Bishop Entlinqs
Bishops Of Opposite Color
AC. Harris

[The first of our 1940 series on practical endgame play is contributed at

our request by A.G. Harris of Western Australia. His epic correspon­
dence battle against Koshnitsky, which should become one of the his­
toric Bishops-of-opposite-color endings, gave him more insight into
this special branch of chess lore than most of us gain in a Zifetime.
Students may now benefit from his hard-won experience. -C.J. S. Purdy]

F ing with "Bishops of opposite

or many chess players, an end­
RULE34
color" appears to involve a draw so -¢-

A
automatic as to render the game un­ s far as possible, place
worthy of further play. Although this your pawns o n the
strong tendency to a draw does exist squares of opposite color to
(and in some cases not even two that of your own Bishop.
pawns plus will enable the possessor -¢-
to do more than draw), there are cir-
cumstances where a win can be
forced even with equal pawns. Posted that way, your pawns re­
The drawing tendency of the Bish­ strict the action of the opposing
ops of opposite color is due chiefly to Bishop and, in cooperation with their
the lack of control over squares not own Bishop, the activity of the op­
of the same color as one's Bishop, so posing King. Your pawns will of
that pawns cannot safely advance course, need their King's protection
across such squares unless aided by if they are liable to attack.
their own King-the pawns thus can
be easily restrained or blockaded. An ending of this type occurred in
A few general rules will assist the a critical game in the current Aus­
student in formulating a plan of op­ tralian Correspondence Champion­
eration-particularly when he is try­ ship.
ing to win .
 CJS Purdy On The Endgame

Seeing his error, White tries to

RULES35-37
-¢'- induce Black to recreate the old po­
use your King to command sition after 41... Be8. But-
41. ... Ng5t!!
squares of the color not
The student should see why this
controlled by your own Bishop.
and the following line could not be
played in the diagrammed position.
J\ ttempt to conduct maneu­
42. Bxg5 Qxe5
.1'"1.vers on the wing not occu­
pied by the enemy King.
43. Qxh6t Kg8
44. Bf6! Qe3t!!
45. Qxe3 f4t
E
ndeavor to secure a passed
pawn, or at least compli­
cations, on both wings simul­
taneously. fl!fllfl�= 2

�����-·��
��,�����
��V/4,�li
�,� W@� v,��
[This was the Inaugural Austra­
lian Correspondence Chess Cham­ rr�
�������-
--

pionship, 1937-1941, wonbyC.J.S.

� � � �
....�

Purdy, with Koshnitsky coming

second-see "Ho w Purdy Won," White is now faced with a tricky
Chapter 2.] ending, though the chances of a win
are strong, despite the material
Koshnitsky equality, because:
a) White can stop both of Black's
passed pawns by Bc3, at the same
time protecting his b-pawn and main­
taining command of the long diago­
nal and the queening square of his
h-pawn.
b) The white King can cooperate
at once with the two united passed
pawns.
White's plan should be as follows:
Harris 1. Advance the King to g5.
Mter White's 39th; Black to pla� 2. Bring up the pawns to h5 and
g4; then protect the h-pawn by play­
Black played ing Kh4 or Kh6 as required by cir­
39. ... Be8 cumstances. Advance the g-pawn to
White can win straightforwardly g6, and return the King to g5, where­
by 40. Qxf5t Qxf5 41. R:xf5, etc., but upon the h-pawn canno t be stopped
he over-finesses. unless Black sacrifices his Bishop for
40. Qe2? Bd7! the two pawns.
41. Qh5 3. In view of that possible sacri-

-134-
 All Facets Of The Endgame: Bishop Endings

fice, White must maintain a pawn on The real mistake, as it loses a

the b-file-so operations on the tempo in the King's necessary pro­
queenside are at present ruled out. gress to g5. Although the aim of the
That is so because if White plays text move is to allow the rapid ad­
bxa5, then Black later sacrifices his vance of the h-pawn, that is of less
Bishop as above; and the black King urgency than restraining the black
then gets over to a8, whereupon a King. Actually, 47. Kh4! with 48. Kg5!
draw is secured because White's still wins on lines already indicated­
Bishop does not control the queen­ an exercise for the student.
ing square of the a-pawn. The book 47. ... Kf7
draw of King vs. a- or h-pawn and 48. Bc3 Ke6!!
Bishop of the wrong color is one that Black's King now has the com­
every chess player should keep in manding position and threatens Kd5-
mind in such endings. e4-d3 if allowed. Passive defense by
Let us see how this plan would 48... Kg6 would lose.
work. 46. Kh4! fxe3 47. Kg5! 49. h4 e2
Then either 47... Kf7 48. h4, fol­ 50. h5
lowed by 49. h5, etc., winning, or 47... The white pawns assume their
e2 48. Bc3 Ba4! 49. h4 Bc2 50. g4 and best position (see Rule 34). If 50. Kf3,
either: then 50... Kd5 draws.
i) 50... Kf7 51. h5 Bd1 (this stops 50. ... Be8
52. Kh6 because thereupon 52 ... If now 51. h6, then 51. .. Bg6; and
e1=Q 53. Bxe1 Bxg4 draws) 52. Kh4 the white pawns are permanently im­
Bc2 53. g5 Bd3 54. g6t Bxg6 55. hxg6t mobilized.
Kxg6 56. Kg4 Kf7 (if now both the 51. Kf4
white pawns were on the a-file, the If 51. Kh4, then 51... Bf7! 52. g5
black King would go to a8 and draw Kf5! (another point about 48... Ke6!!)
easily) 57. Kf3 axb4 58. axb4 then 53. g6 BgS! (53... Bxg6 loses); and
either 58... Ke6 59. Kxe2 Kd5 60. Ke3 White can make no headway.
b5 61. Kf4, and White turns the black 51. . .
. Bf7
King and wins the pawns, or 58... 52. Ke4!
e1=Q 59. Bxe1 Ke6 60. Ke4, and If instead 52. Kg5, then 52... Kd5
White turns the black King, etc. 53. Kf5! Be6t 54. Kg6! Bg'8! 55.Kg7
or ii) 50 ... Kh7 51. h5 Be4 52. Kh4 Ke4! 56. Kxg'8 Kd3 57. Be1 c3; and
axb4 (it does not matter whether this Black draws easily.
exchange is made or not) 53. axb4
b5 54. g5 Bd3 55. g6t Kh6 56. Bd2t
Kg7 57. Kg5, followed by Bc3t and
h5-h6-h7-h8=Q, winning.
Unfortunately for White, he over­
estimated the power of the passed
pawns. From Diagram 2, the actual
game went:
46. g4 fxe3
47. Kg3?
 CJS Purdy On The Endgame

52. ..
. axb4!! worthy of vecy close study, and the
Admirably timed, as now White's student would do well to explore the
King position is such that White many alternative lines which could
threatens 53. bxa5!, in spite of the have arisen at each stage.
general plan of campaign laid down
earlier, because the black b-pawn
would then be fixed, the c-pawn Bishops Of Opposite Color,
would fall eventually, and the white With Rooks
King would penetrate the queen­ M.E. Goldstein
side---o -c mplications on both wings. [In the preceding article, A.G. Har­
We see here also the need to plan ris gave a set of guiding rules for
afresh at each step. playing with Bishops of apposite
For example, if 52... Be8 (or 52... color. In this article, Goldstein con­
Kd6), then 53. bxa5!! and either siders the complicating effect of
i) 53... Bf7 54. Bd2! Kd6 (virtually Rooks on the case of Bishops of op­
forced; e.g. 54... Be8 55. Kd4 Bb5 (say) posite color.]
56. g5!, etc., or 54... Kf6 55. g5t and A student who is aware of the es­
56. g6, etc.) 55. Kd4 Ke6 (if 55... Kc6, sentials of play for the case of Bish­
then 56. Ke5 Kc5 57. Bc3, etc.) 56. ops of opposite color often fails to
Kxc4 Ke5t 57. Kc5 e1=Q 58. Bxe1 appreciate the differences created by
Kf4 59. h6 Bg6 60. Kb6 Be4 61. h7 the presence of a Rook on each side.
Bxh7 62. Kxb7, and White queens one This was forcibly brought to my
of his pawns safely. notice in the Portsmouth Minor Open
Or ii) 53... Ba4 54. Kd4 Kf6 55. Tournament of 1923, when playing a
Kxc4 Kg5 56. Bd2t Kxg4 57. h6 Bc2 youth of nineteen named Lajos Stein­
58. Kc5, and White wins in a fashion er. Having a Rook each and Bishops
similar to that of the other variation. of opposite color, I naively assumed
53. axb4 b5 that the draw was inevitable. Steiner
54. Be1 Be8 quickly taught me that the more m<r
The only move. bile King and Bishop were sufficient
55. Kf4 Bf7 for victocy.
Again best. Endings with Bishops of opposite
If 55... Kd5, then 56. Kf5 Bf7 57. h6 color are of frequent occurrence, yet
Bg8 58. g5 Kd4 59. g6 c3 60. h7 wins. the chess literature offers relatively
If 55... Kf6, then 56. g5t! wins. few examples with a Rook added to
56. Kg5 Bg8!! each side. The presence of the Rook
This ensures the draw because if often overcomes the drawish ten­
57. Kg6!, then 57... Ke5! 58. h6 Bd5! dency that results from the Bishop's
59. Bc3t Kf4! (if 59. g5, then 59... Be4t inability to command squares of
60. Kg7 Kd4 61. g6 c3 draws) 60. g5 more than one color. The general
Be4t 61. Kf6 Ke3 62. g6 Kd3 63. Bel strategic principles of endgame play
c3 64. h7 c2 draws. are again to be observed. The player
Having seen this, White played 57. having the better centralized King,
Be3 with an immediate and foolproof pawns controlling squares of the
draw. Koshnitsky's defensive play is same color as the enemy Bishop, and

-136-
 Al l Facets Of The Endgame: Bishop Endings

the more mobile Rook usually has 31. Bc2 Nd5

winning chances. Even pre-Purdian era chess mas­
In the second international tour­ ters evidently ran short of time.
nament of his career, at Ostend 1907, [CJS Purdy was notorious for get­
the 21 year old Nimzovich played his ting into time trouble in his
favorite Hanham variation of the Phi­ games-see "CJS Purdy's Fine Art
lidor against that stolid English pro­ of Chess Annotation and Other
fessional, F.J. Lee. 71wughts, Vol. 3. "]
32. Be4
Nimzovicb "Surely you won't refuse a draw?"
32. ... Ra2!
Conceding Bishops of opposite
4 color.
33. Bxd5 Bxd5
34. Rc3

Lee
White to move

In the above position, Nimzovich

has just played 22... N (d5)e7, prepar­
ing to eliminate White's light-squared A superficial analysis of the above
Bishop by ...Bf5--with a subsequent position suggests that the draw is
quadruple attack on the isolated inevitable-White's Bishop defends
d-pawn by ...Nf5. Lee immediately the weak f-pawn and White's Rook
eliminates his d-pawn by holds the a-pawn.
23. Ne5 Bxe5! Yet the seeds of victory for Black
This liquidates the weak pawn are present. The mobility of White's
and gives White the two Bishops, all Rook is restricted. Black controls the
for the sake of Black's coming con­ second rank, and White's g-pawn is
trol of the second rank. Now: potentially weak. Therein lies the
24. dxe5 Qxe5 clue to Black's strategy-White's
25. Bxa7 Qxe2 mainstay, the f-pawn, must be re­
26. Rxe2 Rd1 t moved. That explains Black's follow­
27. Rei Rxc1 ing moves.
28. Rxc1 Rd2 34. f5!
Despite control of the second rank, 35. Kh2 Kf7
victory for Black seems far distant 36. Bc5 g5!
since White retains both his Bishops. 37. Rd3 b5
29. b4 Nd5 Black fixes the weak a-pawn,
30. Be4 Nf6 thereby releasing the black Rook for

- 137 -
 CJS Purdy On The Endgame

more important duties. Note by CJS Purdy

38. Bd4 Be4 The student will see that combin-
39. Rc3 Bd5 ing the Rook and the Bishop is the
Was the clock still ticking re- main secret. For that purpose, the
morselessly? "opposite colors" may actually be an
40. Bc5 Kg6 advantage since the attacking Bishop
41. Rd3 h5! is permanently safe from exchange.
42. Bb6 f4 The same applies to Bishops of op-
43. Bd4 Kf5 posite color in middlegame posi-
Black is now virtually a piece up, tions-the security from exchange
his King playing a decisive part in will favor an attack.
the final onslaught.
44. f3
An immediate win for Black was Two Pawns Up And No Win
threatened by 44... Ke4 45. Rd1 Bb3. The endgame student occasion­
If 44. Bb6 instead, then 44... g4 and ally sees endgames with Bishops of
...g3t. opposite color in which one side wins
44. g4 even with the number of pawns equal.
45. bxg4 hxg4 Thus the student may come to un­
46. Kg1 derrate the drawishness of the non­
White might try a last trap, viz. clashing prelates.
46. fxg4t Ke4 47. Rd1 Bb3 48. Rb1
Bc2 49. Ra1!, with a safe draw. In Nield
this, however, Black wins easily by
46... Kxg4!, with the forced gain of
the g-pawn.
46. ... Re2!
The move 46... g3 would also win.
The text move prevents the White re­
source indicated in the previous note.
47. fxg4t Ke4
48. Rd1 Bb3
49. Rf1
If 49. Rb1, then 49... Bc2.
49. ... Kxd4
And Black ultimately won.
The simple, logical style of the The diagrammed ending from a
youthful Nimzovich is noteworthy. It game in the championship of the Cor­
is doubtful whether the mature Nim­ respondence Chess League of Aus­
zo of the 1930s could have refrained tralia, J. W. Bell-A.E. Nield, will
from overelaboration in such an end­ serve as a corrective. Here White is
ing. two united pawns up, with his pieces
not badly placed; yet not only did
Black draw, but there seems to be no
way for White to win.

- 13 8 -
 All Facets Of The Endgame: Bishop Endings

58. Be1 ! does not accomplish anything in this

The idea of this is clear. We all particular position, as far as we can
know that the first rule for winning see. For example, 59. Kf3 Kg5 60. Bb5
such endings is: Kf5 61. Bd7t Kg5 62. h4t Kh5 63. Be8t
Kh6; and still White can do nothing
useful until he plays his King to h3.
RULE38
Nield gives also 59. Kh3 Kg5 60.
<¢>
Bd3 Bf2 61. Kg2 Be1 62. h4t Kg4 63.
J..l""e e p y o u r p a w n s o n
Be2t Kf5 64. Kf3 Ke5 65. Kg4 Kf6 66.
I\.. those squares opposite
Bd3 Bf2, reaching the same position
in color to that of your Bishop.
as in the game.
(reiterated for emphasis)
59. Ke4
60. Kh3 Ke5
Thus, White can win only if he can 61. Bd3 Bf2
advance his pawns in this order: h4, 62. Kg4 Kf6
then g'3-g4-g'5, and so on. If White first 63. Kh3
advanced g4, then he could never If 63. Kf3, then 63... Be1 64. Bb1
advance his other pawn. Also, if Ke5 65. Kg4 Kf6 (here Black could
White plays his h-pawn to h5 before even play 65... Bf2, whereupon, say,
getting the g-pawn to g5, then again 66. Kg'5! Bxg3 67. h5 Ke6 68. Kg6 Be5
he will find both pawns blockaded. 69. Ba2t Ke7 70. h6 Kf8-and Black
Therefore, Black plays so that af­ draws) 66. h5 Bf2 67. Kf3 Be1 68. h6
ter h4 White will be unable to play g4 (if 68. g4, then 68... Kg5) Bd2 69. h7
without playing his King to h3. Kg7; and Black draws (Nield).
Thus, we note that it is to White's 63. Ke5
detriment that one of the united 64. g4 Kf4
pawns is an h-pawn. If everything 65. g5 Bd4
were moved over one file to the left Drawn by agreement.
(except the a-pawns, of course), the Nield notes, "White can advance
position would be a win for White. his pawns to g6 and h5 in safety and
For suppose, in the diagrammed then bring his King around to f7; but
position, there were an i-file outside with ... Kh6 Black maintains the
the right-hand border, then the White blockade of the pawns, and the game
King would have the threat of reach­ is drawn."
ing hf>.-and thereafter by playing his Once the two pawns both have to
pawns to g5 and h4, White would come onto squares of the same color
have an easy win . as their Bishop, the draw is usually
59. h4 automatic in such endings, provided
White certainly gains nothing by that no useful diversion can be creat­
playing this move at once. It would ed on the other wing. The student will
be more in conformity with principle now fully understand Euwe's dictum:
for White to edge his King and Bishop
into their best possible positions first,
by playing the King to f3 and the
Bishop to d7 (say). That, however,

- 13 9 -
 CJS Purdy On The Endgame

4 . . . c5 which tends to work out in

RULE39 White's favor.
-¢- 5. Bg5 Nbd7

I
n endings with Bishops of 6. cxd5!
opposite c o l o r, material Stops the Cambridge Springs De­
means NOTHING, position fense (6. . . Qa5) and prepares a mi­
EVERYTHING. nority queenside pawn attack (b4).
-¢- 6. ... cxd5
If 6. . . exd5, then White can play
There is a limit, of course. Only in Qc2 and 0-0-0.
absurdly freakish situations will a 7. e3 Be7
plus of three pawns fail to win! On 8. Bd3 0-0
the other hand, wins with equal ma­ 9. 0-0 a6
terial are often brought off, as al­ 10. Rc1 b5
ready pointed out. In such cases, it 11. Bb1 Bb7
is necessary to be able to create com­ 12. Qd3
plications on both wings. Threatening Ne5, Nxd7, and Bxf6.
By the way, "Two Pawns Up and 12. Ne4
No Wm" is not intended as a particu­ 13. Bxe7 Qxe7
larly sensational title. Two pawns up 14. Ne5 f5
are frequently insufficient to win in 15. f3! Nxe5
Rook endings, for instance-even 16. dxe5 Nxc3
two united passed pawns. 17. Rxc3 Rfc8
The best part of chess is that mere 18. Rfc1 Qb4
material counts for so little, at all 19. Qd2 Rxc3
stages of the game-position is al­ 20. Qxc3 Qxc3
ways dominant. 21. Rxc3 Rc8
22. Rxc8t Bxc8

That "Bad" Bishop! WWinter

M.E. Goldstein
The following game is from last
year's British championship. This
kind of ending often occurs in the Ruy
Lopez, the Queen's Gambit Declined,
and occasionally the French Defense.

J. Penrose--W. Wmter
Queen's Gambit Declined
(Orthodox Defense)
1. d4 d5 J.Penrose
2. c4 e6
3. Nc3 Nf6 Now the black Bishop is wall ed in
4. Nf3 c6 by his own pawns, and the whit e
Not caring for the symmetrical line King has free access to d4. The per-

-140-
 All Facets Of The Endgame: Bishop Endings

manent threat of Kc5 then forces Not 32. . . g4 because of 33, fxg4
the black King to hover between b6 fxg4 34, Bc2, followed by Bg6. Or 32. . .
and c6. g 4 33. fxg4 hxg4 3 4 . h5, and the
The game continued: h-pawn queens.
23. b4! 33. gxh4 Bd7
Fixing the two black queenside 34. Bc2 Bc6
pawns on light squares. 35. Bd1 Be8
23. Kf7 36. e4 dxe4
24. � Bb7 37. fxe4 fxe4
25. Bc2 Ke7 Useless is 37 . . . f4, as the white
26. Bb3 Kd7 King sidles over to f3.
27. Ke2 Bc8 38. Kxe4 Kc6
28. Kd 3 Kc7 39. Bf3 Kc7
29. Kd4 Kb6 40. Kf4 Bc6
30. h3 h5 41. Bxc6 Resigns.
Puts yet another pawn on a light After 41. . . Kxc6 4 2. Kg5 Kd5 43.
square, but . . . g5 and . . . h6 would Kf6, White wins off hand.
merely postpone without preventing It is unusual to find a young player
White's ultimate breakthrough. such as Jonathan Penrose so skillful
31. g3 g5 in the ending.
32. h4 gxh4

- 141 -
 CJS Purdy On The Endgame

Chapter II
Bishops \IS. Kl\iql\ls

Bishop vs. Knight It was perhaps excusable for

T
he diagr a m m e d Bishop vs. White to hope that he might get slight
Knight ending between Miller winning chances by giving up the
and Cheney has an interesting fea­ front member of his doubled pawns
ture. in order to mobilize his Bishop . Fbr
example, 50. Bd1 Nxc4 51. Bb3 Nd.2
G.Cheney 52. Bd5. If now 52 . . . Nb1, then 53. c4
Nc3 54 Bc6; and all the pawns are
safe. But instead 52. . . c4!, and Black
will give up this pawn to win two.
Just an exercise in geometry!
Actually , White chose another way
to give up the pawn.
50. Kh4 Kg6
51. Bc2 Kg7
52. Kh3?
One of those moves based on some
A.L. Miller inexplicable "brain-wave." With Kg'3,
White to play his 50th coming a file nearer the center, White
might have escaped defeat. Of course,
The worst thing for a Bishop is to 52. Bb3 for a draw was the best.
have its pawns fixed on squares of 52. ... Nxc4
its own color. Here the trouble is very 53. Bb3 Ne3
pronounced, so we have an excep­ Threatening . . . c4 with the same
tional case in which the Knight is by two-for-one combination that we
far the superior piece. White is up a mentioned previously.
passed pawn, with the enemy King 54. Be6 Ndt
on the defensive; yet White can only And a pawn must fall.
draw. White's best play is to hold the 55. Bb3 Nxc3
position intact by 1. Ba2 Kh.7 2. Bb3 56. Bc2 b5!
(or 2. Bb1) Kg7 3. Ba2, etc. 57. axb5 Nxb5

- 142
-
 All Facets Of The Endgame: Bishops vs. Knights

58. Kg4 Nc3 Play continued:

59. Kf3 a4 61. Bc5 Nxh4
60. Ke3 a3 62. Kd2 Nf3t
6t. Bb3 a2 63. Kxd3 Kb3
62. Bxa2 Nxa2 64. Ke4 Nd2t
63. Kd3 Kg6 65. Kf4 Nc4
64. Kc4 Nb4 66. Kg5 Nxa8
65. e5 dxe5 67. Kxh5 Nc2
66. Kxc5 Nd3t 68. Kg4 Nb4
67. Kd5 Kxg5 69. Kf3 a3
68. Ke4 Kg4 70. Ke2 a2
69. Kxd3 Kf3 71. Bd4 Kc2
70. Kd2 e4 72. Bat
71. Ket Ke3 The ending Knight and pawn vs.
72. Resigns. Bishop resolves itself into the ques­
A "book" position: the black King tion whether the Knight can shield
has "the opposition" in front of his the pawn from capture by the Bishop.
pawn-an easy win. If the hostile King is far away, the a­
or h-pawn gives the most winning
chances, as the Bishop then has only
Two Minor Piece Endgames one diagonal on which to attack­
from Maroubra 1952 compared to two diagonals for a cen­
M.E. Goldstein ter pawn.
Here are a couple of misplayed Before we see where Walsh went
endings from the Invitation Tourney wrong, let's look at the prototype of
at Maroubra, popular Vince Walsh this ending. The problem was com­
being the victim in both. In his sec­ posed by the inimitable Sam Loyd
ond-round game against John Kell­ many years ago.
ner, Walsh played Philidor's Defense
with great verve; he adjourned in the Sam Loyd (1860)
following winning position.

V.A. Walsh

White to play and draw

Loyd's elegant solution, which

J.V. Kellner highlights the stalemate possibilities
After 60 K(b3)a!
... of the h-pawn, ran:

-143-
 OS Purdy On The Endgame

1. Bd7! h2 80. Bh.8 Na4 81. Bg7 Nb2 and Black

2. Bc6t Kg1 wins.] It is amusing to see how the
Not 2. . . Nf3t because of Ke2-and black King and Knight can drive the
White draws. white King right out of play. Of course
3. Bh1! ! Kxh1 White can never play his Bishop
Or 3. . . Ng2t 4. Ke2! Nf4t 5. Ket, along the long diagonal on pain of
and White draws. . .. Nc3. 76. . . Kc2 77. Kft Kbt 78. Kel
4. Kf2! Kxat, and Black wins. In this, not
We have now reached a famous 78. .. Nb2? because of 79. Kd2; and
position known to be a draw for the White draws as in the actual game.
past three centuries. [See the Har- 73. Ke3 Nb2
dy-McNabb game, Chapter 17.] 74. Ke2 .Kb1?
Kellner was familiar with this end- This finally gives away the win.
ing (which is No. 220b in Fine's Ba- Black can transpose into the winning
sic Chess Endings), but apparently variation of the previous note by 7 4 ...
Walsh had omitted to look it up. Nd1! 75. Ke1 Nc3.
75. Kd2 Nc4t
Now back to Kellner-Walsh. 76. Kd1 .Kxa1
77. Kc2
Not 77. Ket? because of 77. . . Ne3;
4 and White must abandon the c-file.
77. ... Ne3t
78. Ket Draw agreed.
The Knight can never gain a tem­
po, unlike the Bishop.
Harry Klass, whose wide knowl­
edge of the finer points of the game
is exceeded only by his love of the
After 72. Bat truly artistic, showed me the follow­
ing endgame composition by Prokop
Play went on: which may serve as a pendant to the
72. ... Nd3 Kellner-Walsh ending.
The wrong i dea! The Knight
should make for c3 so as to deprive F.J.Prokop
White of the key square e2. Hence,
72 ... Nd5! 73. Ket Nc3 7 4. Kf2 Kd2 75.
Kft Ndl! 76. Kg1. [Ed.: K.M. de­ 5
scribes 75. Kfl as weak when 75.
Kg2!? is available and forces Black
to find the "Horwitz win" as men­
tioned in Nunn 's "Secrets of Minor­
Piece Endings" on pgs. 88-89. This
win is difficult to produce over the
board. E.g., 75 ... Nc2 76. Kf2 Na4
77. Ke2 Kcl 78. Kel Nc5 79. Ke2 .Kbl White to play and win

-144-
 All Facets Of The Endgame: Bishops vs. Knights

Try your hand on this. If you can't Kxf5 51. Kc2 Kg6 52. Nxf6! And White
do it, see the end of the article. draws since the white King cannot
be driven away from h1.
¢- ¢- ¢- 49. Ne2 Kxf5
And Black wins.
My second example is from the Walsh, tired toward the end of the
1oth round game Goldstein-Walsh. four hour session, did not find the
winning line; he played instead:
V.A. Walsh 46. ... Ke5
47. Ng3 Kf4
48. Nh5t Kxg4
6 49. Nxg7 Kxh4
50. Kc4 Kg4
51. Kd3 Bg5
52. Ne8 Kxf5
53. Ke2 Kg4
54. Nxf6t
Draw agreed, as once again the
white King gets to h1.
M.E. Goldstein These two endings show how a
Mter 46. K(a3)b3 knowledge of "book draws" will save
many a half-point in tournament play.
By fine endgame play, Walsh has Although even with Mr. Purdy's as­
decoyed the white King to the queen­ sistance you can't learn the endings
side to capture a dangerous passed in ten hours, you can learn some of
pawn. Black has a clearly won game, the general principles.
but he must be careful. The easiest [The reference is to Purdy's series
way to win is: of articles THE OPENINGS IN TEN
46. . . . h5! HOURS.]
The h-pawn, whose queening
square is NOT controlled by the Bish­ Prokop Study
op, sacrifices himself to break up 1. Ke5! Bg2
White's pawns. 2. Nb6 Bb7
47. gxh5 3. Ke6!
No better is 47. g5 fxg5 48. hxg5 Regaining the opposition.
Bxg5 49. Kc2 (or 49. Ng3 h4 50. Nh5 3. ... Kd8
h3) h4 50. Kd1 Ke4 51. Ke1 Be3! 52. 4. Kd6 Ke8 (forced)
Kf1 Kf3, and Black wins. 5. Kc7 Bg2
47. . . . Ke5 6. Nc8 Ba8!
48. Ng3 7. Kb8 Kd8
If 48. h6, then not 48... gxh6? but 8. Nd6 Kd7
48... Bxh6. 9. Nb7! Kc6
48. ... Bf4! 10. Kxa8 Kc7
Not 48... Kf4 49. h6! gxh6 50. Nh5t 11. Nd6!

- 145 -
 CJS Purdy On The Endgame

[hapter IZ
Rook us. 8isl\op/Kr\iql\t

Endgame Value of on each side. The value of the Ex­

"The Exchange" change is greater in the former kind
of endgame-if we assume that there
Part I are pawns on both wings, an assump­

T
he Exchange (capitalized to tion that we shall make throughout
avoid ambiguity) is the differ­ unless we explicitly state otherwise.
ence between a Rook and a Minor The reason is that the Knight and
Piece, i.e. Bishop or Knight. the Bishop have constitutional fail­
The Rook's value relative to a ings. A Knight cannot stay on a color
Bishop or Knight is on the average when he moves, and a Bishop cannot
nearly half a pawn higher in the get off its given color. If the Knight or
endgame than in the opening. These the Bishop is alone, the failing is of­
relative values are not static, how­ ten serious. But an assistant Rook
ever; bear in mind that, as the end­ will exert his unfailing strength to
game approaches, the Rook and the tide his less reliable colleague over
pawn increase in value whereas the momentary weaknesses. Even if the
Knight and the Bishop stay at about assistant piece is only another
the same value. Knight, some strength is gained­
one Knight can stand guard when his
Opening Endgame partner has to do a chameleon act. A
Rook 5 pawns 5 pawns Bishop can also help a Knight in this
Minor Piece 31h pawns 3 pawns way. And of course a Bishop can help
a fellow Bishop so much that two
Let us now take the endgame sep­ Bishops are a law unto themselves.
arately-the Opening and the Middle­ Anything said about Minor Pieces
game will follow. assumes "the two Bishops" to be an
Immediately we must distinguish exception.
between endgames in which there Let us consider some examples.
are no additional pieces, i.e. none but Suppose we have an endgame
W + !:! vs. W + Minor Piece, and with 'it> + !:! vs. w + !J../ lfj + two ft
those with other pieces "thrown in" (pawns on both wings), and the end-

-146-
 All Facets Of The Endgame: Rook vs. Bishop/Knight

ing is a draw. If you then add to each Zukertort

side another piece of any sort,
equally well posted, the additions will
favor the side with the two pawns for 1
the Exchange-and a win may very
well result.
The only kind of endgame that one
can generalize about successfully in
this connection is the simpler kind,
i.e. � + E! vs. � + �/4)-the kind
where the Exchange has its highest
value. Steinitz
Yet even here we get disagree­ Draw with best pl ay
ment among the highest authorities.
Fine, in his Basic Chess Endings, Evidently, Cheron should have
states the following brief rules for said "passed pawn." A passed pawn
when both players have pawns. is more than just a pawn. Here the
• When there are no pawns for b-pawn, with the aid of its King (who
the Exchange, the Rook wins easily. will journey to c5), can advance be­
• When there is one pawn for the yond mid-board and thus become
Exchange, the Rook wins, but with such a menace that White will not
difficulty. have time to win any pawns on the
• When there are two pawns for other wing. If instead of the passed
the Exchange, the game is normally b-pawn, Black merely had a pawn
a draw. majority, say pawns on a7 and b7
• When there are three pawns for against a White pawn on a2, then
the Exchange, the Minor Piece wins, White should win. Certainly, a pawn­
though not always too easily. majority is a potential passed pawn.
Cheron, in his Traite Complet But don't forget the time factor. The
d 'Echecs, says: queening process (where none of the
pawns involved are more than one
''A Minor P iece plus a pawn rank beyond their home rank) in­
is cap able in spite of mate­
, volves two more moves than with an
rial inferiority, of drawing already passed pawn. In other words,
against a Rook." a pawn majority is like a passed
pawn that has to start right back
This contradicts Fine's statement, from the table, i.e. from two ranks
behind its home rank.
"When t her e is one pawn Of course, if Black did not try to
for the Exchange, the Rook advance his passed pawn (see the
·wins .. .'' diagram), adopting a policy of pas­
sive defense, then he would lose. The
When we look at the diagram pre­ actual game went thus:
sented by Cheron as an example, we 1. Kg3 Ke7
readily see the discrepancy. 2. Kf4 Ke6

- 1 47-
 OS Purdy On The Endgame

3. h4 Kd5 Capablanca
4. g4 b5
First step.
5. Rb1 Kc5 2
6. Rc1t Kd5
7. Ke3 Nc4t
8. Ke2
If 8. Kd3, then 8 . Ne5t.
. .

8. .. . b4
Second step.
9. Rb1 Kc5
10. f4 Na3 Euwe
11. Rc1t Kd4 Black to play
12. Rc7 b3
Third step. Here Black has two full pawns for
13. Rb7 Kc3 the Exchange (usually full value), and
14. Rc7t Kd4! they are united passed pawns. Yet it
Not 14 . . Kb2
. which would give appears that Black cannot save the
White time to win two pawns and game against best play. If the Minor
then give up his Rook. Piece were a Bishop, placed on any
15. Rb7 Kc3 reasonably good square, it would be
The moves were repeated, and the White who would have to fight for
game was drawn. the draw.
The Steinitz-Zukertort draw The game is from the Euwe­
seems to me to be largely fortuitous. Capablanca match of 1932. By a se­
For instance, it is very convenient ries of finesses, Black was able to
for Black that his Knight on d6 pro­ hang on to all his material for some
tects his f-pawn and thus indirectly time. Play went:
the other pawns. White should prob­ 1. e4t!
ably win if Black's f-pawn were at f6, 2. Kf4 NM!
or if all the men other than the king­ 3. Rb5! Nd3t
side pawns were moved one file to 4. Kxe4 N�t
the left-then the Knight would no 5. Kd4 f5
longer protect the f-pawn. Normally, 6. Rb2 Ng4
as we shall see, it is extremely diffi­ 7. h3 Nf6
cult for a Knight to protect pawns on 8. Rc2
both wings. Cheron's rule would be It is now clear that Black must
found to apply more often with a soon lose a pawn.
Bishop than with a Knight-in gen­ 8. Ne4
eral, with pawns on both sides, a 9. g4 Kf6
Bishop is the more reliable. 10. gxf5 Kxf5
To show how unreliable the 11. Re? Ng5
Knight is with pawns on both wings, 12. Rxa7 h5
let us take the following Euwe­ Here Euwe played 13. Ra3, which
Capablanca ending. was good enough to win had he not

- 148-
 All Facets Of The Endgame: Rook vs. Bishop/Knight

drifted a little later on (Capa ulti­ "H you are one pawn ahead,
mately drew); but the clearest win in 99 cases out of 100 the
was by 13. h4! For if 13. . Nf3t, then
. game is drawn if there are
14 Kd5 Nxh4; and White wins easily
. pawns on only one side of
by 15. Rf7t, followed by the march of the board."
the lone white pawn.
This last point illustrates an im­ There are no statistics to support
portant rule: so precise a figure-read "almost al­
ways."
RULE40 Now, nobody has ever tried to
� claim that the Exchange is normally

I
f the Rook's side can create worth more than two pawns. There­
a passed pawn, the value of fore the advantage of the Exchange
the Exchange increases, even for a pawn should equal one pawn.
though. in the process more of That advantage, then, should hardly
the e n e m y pawns become ever be sufficient to win with all
passed. pawns on one wing-or so we should
expect.
Where the defender's piece is a
Bishop, our conclusion holds, though
Part II the defender often has to play very
In Part I, we considered endings carefully. But where the defender's
with pawns on both wings. We found piece is a Knight, a loss is quite a
that even one pawn for the Exchange frequent occurrence.
is often sufficient to draw-provided Surprising! We found in Part I that
that the pawn is a passed pawn. With a Knight has difficulty in protecting
no passed pawns, two pawns are nor­ pawns on both wings-that is only
mally required for the draw. to be expected. But with all pawns
With pawns on one wing only, we on one wing, it is well known that a
should expect that a singie pawn for Knight is normally at least as strong
the Exchange will always draw. For as a Bishop.
in this case, the disadvantage is nu­ What is the special factor that
merically equivalent to one pawn; makes a Knight weaker where it has
and there is a general rule that: to fight a Rook?
It is this. Where a Rook plays a
RULE41 Minor Piece, and the Minor Piece has
� no passed pawn on its side, the

A
pawn down does not lose greatly superior mobility of the Rook
when all pawns are on one usually ensures that the Minor Piece
wing only. will always be confined to the defen­
sive.

Fine, in Basic Chess Endings,

puts it strongiy thus:

- 149
-
 OS Purdy On The Endgame

but typical. What is White's trouble?

RULE42 Obviously White has only one weak­
-¢'
T. o be confined to the defen­ ness-the square e3. Would it be as
bad if it were the g-pawn that had
sive is itself a disadvantage. moved instead of the f-pawn, i.e. if f3
-¢' and h3 were weak instead of e3? No!
The weakness of e3 is far more seri­
That rule holds throughout the ous, enabling the black King easily
game. A pawn plus is an advantage to march down to f1 and to attack
in material rather than mobility: the the white g-pawn with a second piece.
pieces, and they make up the fight­ One can say that any such position
ing strength, are equal on both sides. is a probable draw with the f-pawn
Not so when a Rook plays a Minor unmoved-and a probable loss if the
Piece. pawn has moved. The other two
Now, it is well known that a Bishop pawns matter less, though it is bet­
is, in general, a better piece for a ter if they are unmoved.
defensive role than a Knight. A Bish­ It is neither practicable nor use­
op can move from its post and still ful to give a full analysis of this end­
defend the point it was set to guard. ing. The student should try it over
The Knight cannot do that-if he has many times with a friend of equal
to move, it is goodbye. That will be skill-the result will be that Black
so even if all the pawns are on one nearly always wins. The actual game
wing. went as follows:
Even after all that, it will come as 1. h4
a shock to most good players to learn Looks weakening at first sight, but
that the diagrammed position (Vid­ Fine considers it best. The reason­
mar-Alekhine, San Remo 1930) is ing goes: if this pawn stays at home,
a loss for White. At any rate, all at­ it will be exposed when I am forced
tempts to show a draw have failed. to move the g-pawn (after the black
King gets to f1). Experiments have
Alekhlne shown passive play to be no better.
1. Ke7
2. Ne4 h6
3. Nrl
See sidebar for the drawino line.
3. Ke6
4. Nd3 Kf5
5. Nf4 Ra4
6. Nd3 Rc4
7. Nrl Rc6
8. Nh3 Ke5
Vidmar "Now if White keeps on stalling
White to move with his Knight, Black's King will get
to the first rank and win a pawn."
The position is not exceptional, (Fine). Try it out. So Vidmar played

-150-
 All Facets Of The Endgame: Rook vs. Bishop/Knight

9. h5, "fixing'' Black's g-pawn. 18. Nc5t Kd4!

9. h5 Rc2 19. Nb3t Ke5
10. Nf4 Rd2! 20. Resigns .
11. Nh3 Kd 4 An excellent exercise is to substi-
1 2. Nf4 Ke3 tute a Bishop for the Knight, first on
13. Ne6 Rd5! a light square, then on a dark one.
14. f4! Rf5! Youwillfind that it is stillquite easy
If 14... Rxh5, then 15. Nxg7 Ra5 to lose, even though theoretically the
16. Kg4 Ke4 17. Nh5; and White will game should be drawn.
draw. In any case, remember that the
15. Kg4 Rf6! f-pawn is crucial.
1 6. f5 Rf7 In the diagram, suppose both
Fine gives 16... Ke4!, as the text black pawns were on the sixth rank,
move could have been answered by or one of them on the fifth. Then
17. Nd8, with a very stiff resistance. White should draw-because the
1 7. g3? Ke4 black pawns would be near enough
to provide White with counter­
chances.
Averbach Weighs in on The secret of drawing games is
Alekhine-Vidmar not defense but counterchances.
Only if you can use pieces to hit some­
According to Batsford Chess thing will they ever show their full
Endings (pg. 3 45) and other value.
sources, endgame specialist An important general conclusion
Yuri Averbach notified the from both Part I and Part II is this:
chess world of a discovery, viz., Agains t a Rook in the endgame,
that an analyst named Leykin, a Bishop is a little better than a
writing in Chess in the USSR Knight.
1940, claimed that White can
draw by building a fortress af­ Midgame Value Of
ter 3. Kh3 (instead of 3. Nf2?). "The Exchange"
The actual game moves would I repeat the table of values given
ha ve been: 41. Kh3 Ke6 (or 41 ... in the preceding article. The figures,
Ra3 42. Kg4 Ke6 43. Kf4 g6 44. of course, represent approximates
g4) g4! Ke5 43. Kg3 g6 (not 43. .
. and do not apply in any particular
Kd4 44. h5 Ke3 45. Nd6) 44. Nf2 position.
Ra3 45. Nh3 Ra4 46. Kf2! ( 46.
Nf2 Kd4 47. Nh3 Ke3) h5 47. Opening E ndgame
gxh5 Rxh4 48. bxg6!=. The re­ Rook 5 pawns 5 pawns
sults a re conveyed to the Minor Piece 31h pawns 3 pawns
reader in a much better man­
ner in Muller & Lamprecht s ' The table indicates that the aver­
Fundamental Chess Endings, age value of the Exchange is roughly
pg. 265. 1.5 pawns in the opening and 2 pawns
in the endgame. In the middlegame,

-151-
 CJS Purdy On The Endgame

the value would normally be nearer

RULES 43-44
the opening value (11h) than the end­
�
game value (2).
I middlegame, a Rook
n the opening and in the
The difficulty, of course, is to judge
can al­
in any particular position whether it most always be given up safely
will pay you or your opponent, as the
for a Bishop and two pawns.
case may be, to exchange a Rook for Also for a Knight and two
a Minor Piece and one or two pawns,
pawns, provided that the
as the case may be. player with the extra pawns
First of all, we must distinguish has them situated safely. Very
between "quiet" positions and com­ weak isolated pawns will upset
binative positions. A "quiet" position calculations.
is one in which general positional
considerations are a fair guide, i.e.
I middlegame, the loss of a
n the opening and in the
neither King is being attacked and
there are no forks, pins, nets, or ties Rook for a Minor Piece and one
(see Guide to Good Chess) to upset pawn is usually less serious
calculations. In a combinative posi­ than the loss of a pawn.
tion, one cannot come to any deci­ �
sion without considering particular
moves (the violent ones) and seeing
where they lead. In such positions, a RULE45
Knight may be momentarily better �
than a Rook or even a Queen. A t all stages of the game, a
Therefore, we can lay down rules .ft.Bishop is better than a
for "quiet" positions only. And these Knight against a Rook. In the
are the rules that I would give as a opening and in the middle­
rough guide: game, an extraordinarily well­
placed Knight is an exception,
Rules 43-44 are often forgotten. i.e. a pawn-supported un­
Time and again, a player will give up dislodgeable advanced Knight
a pawn where he could have done in the enemy camp, on a cen­
better by giving up the Exchange for tral file or on a file on the wing
a pawn instead. of the enemy castled King.
�
Tarrasch has well said (see Rule
45) that such a Knight is nearly as there is usually some Minor Piece
strong as a Rook, and it is usually that can be exchanged for the Knight.
good play in such cases to give up a
Rook for the Knight and one pawn. The answer (to Rule 46) is usu­
That is particularly so if the Knight ally easy to come by. Rooks like open
is on the 6th (3rd) rank-he is not files, and they like the enemy to have
quite so devilish if one rank farther weak pawns. If the position is close
away. and if the enemy has no weak pawns,
Such cases are infrequent, as a Rook has little scope. The player

- 1 52 -
 All Facets Of The Endgame: Rook vs. Bishop/Knight

will probably draw-particularly if

RULE46
the piece you got for your Rook was
<¢>
a Bishop. Indeed, if you yourself have
A
vita! question to ask your­
the two Bishops, you may have an
sell when pondering the
excellent game.
merits of such a transaction
A common way of giving up the
(Rook for Minor Piece and one
Exchange, sometimes compulsorily!,
or two pawns or vice versa) is:
is for White, say, to play g3, allowing
How is the position for Rooks?
an enemy Bishop to come down to h3
<¢>
and to grab the castled Rook at f1.
Sometimes the Rook cannot move;
·with the Rook or Rooks must then go but even if he can escape, always con­
to a lot of trouble to open up the po­ sider whether it may not pay you to
sition-and that detracts from his let him be captured. That may give
chances of winning. But if the posi­ you time to take an important pawn
tion is very open and if the player or to get ahead with an attack on the
with the Rook or Rooks can use them other wing while bringing the ene­
to attack weak pawns, then the Ex­ my's attack on the kingside to a tem­
change is nearly always worth fully porary halt. If you move the Rook
two pawns. away, the Bishop at h3 lives, perhaps
Giving up the Exchange for a to take part in an overwhelming at­
pawn often halts a kingside attack tack later on.
at small cost in material. You then I wish I could get hold of a bril­
have a good fighting chance-instead liant example, played by the late
of perhaps succumbing rapidly to a A.E.N. Wallace at either board 1 or
brilliancy. board 2 of the New South Wales­
The point is that an unmoved Victoria telegraphic matches be­
Rook, or even a castled Rook on the tween 1923 and 192 7-possibly
f-file shut in by the unmoved f-pawn, against Gundersen. I have quite for­
is to some extent a liability. Before gotten the moves, but the idea must
the Rook can pull his weight, he must have influenced me subconsciously
be developed-and perhaps the en­ as I have always been particularly
emy \\ill not give you the time needed watchful for good opportunities to
for development. If you think that is give up the Exchange. I suggest to
likely to be the case, it is far better keen students the advisability of
to induce the enemy to give up an studying carefully every example
attacking Bishop or Knight (prefer­ they come across in published games,
ably with a pawn thrown in, of bearing in mind the general rules in
course) in order to win the Rook. this short article.
With one of the enemy attackers
gone, the danger to your King will
almost certainly vanish; and the en­ Error in BCE
emy will be reduced to playing for an Basic Chess Endings (BCE) is the
endgame. If you are better at the tournament player's chess bible, so
endgame than the opponent is, you it is a pity for errors in it to crop up

- 153-
 CJS Purdy On The Endgame

once again in new editions after they 2. g3 g7

have been pointed out to the author 3. d4 Bg7
in ample time for corrections to be 4. Bg2 c5
made. 5. 0-0 cxd4
The fourth edition still states that 6. Nxd4 Nc6
Rook and one center pawn against 7. Nb3
Bishop and one center pawn is al- Better 7. Nxc6 bxc6 8. c4, with ad-
ways a win, although we demon- vantage to White, as inKoshnitsky­
strated in our discussion (see chap- C. Purdy, Ukrainian invitation
ter 16, pgs. 188ff) of Jaass-Purdy, tourney 1957. A better 6th move for
Brisbane 1951 that this is not true if Black was ... Nf6. /P
the pawns are on adjacent files and 7. e6
the defender's King is well forward. 8. c3 Nge7
Fine treats only of cases with the 9. Bg5 0--0
pawns on the same file (no need to 10. Qct b5!
consider the case where each side By 10... Re8 the exchange of Bish-
has a passed pawn, for there the ops could have been avoided, but I
Rook's side must win except for freak felt that my chances of counterplay
positions) . We wrote to Dr. Fine on the queenside were more real than
many years ago, drawing attention White's chances of attack against my
to the error. kingside.
The position that we gave in the 11. Bh6 a5
article (Chapter 16) was one in which t2. Bxg7 Kxg7
the Bishop protected its own pawn. t3. Ntd2 Qb6
We did not examine the case where t4. Nf3 Ba6
the Bishop cannot protect its own With ... b4 in view.
pawn but can attack the opponent's t5. Qd2 Rac8
pawn-the pawns standing at a t6. Ra.ct b4
Knight's move from each other. Such t7. Rfdt Bc4
a position occurred in the Australian t8. Nat b3
championship tourney, Melbourne Probably the best. White threat-
1956-57. It now appears that this ened b3. Of course . . . Bxa2 is an­
type of position would be more often swered by b3.
drawn than that where the Bishop t9. axb3 Bxb3
protects its own pawn, because in 20. Nxb3 Qxb3
this case the defender's King need 2t. Nd4 Nxd4
not be well forward. 22. cxd4 a4
We give the game itself, up to the 23. e4 Rxct
position in question. The notes are 24. Rxct Rb8
by Kalinovsky, except for a few iso- 25. Rc7
lated comments by Purdy, marked/P. After 25. Rc2, Black would still
hold the initiative-so Ozols decided
K. Ozols-P. Kalinovsky to play actively.
King's Indian Attack
1. Nf3 d5

-154-
 All Facets Of The Endgame: Rook vs. Bishop/Knight

44. Bc4! Rd1 t

45. Kg2 a1=Q
4 46. Rxa1 Rxa1
47. Kf3 Kf6
48. Kxe3 Rh1
49. h5 Rxh5
50. Kd3

5
25. ... Qxb2! !
This neat sacrifice of a piece for
two pawns should have won. IP
26. Qx b2 Rxb2
27. Rxe7 a3
28. Ra7 a2
29. Bf1 dxe4
All White's pieces are tied up with­ And the game was ultimately
out chance of release. The best win­ drawn.
ning plan for Black is to establish As many other players would
pawns on f5 and g5 and then to play have, Kalinovsky thought at the time
.. .f4 and ... e3, always avoiding unnec­ that he was still in a winning posi­
essary exchanges of pawns. tion. He writes:
(To save space, we eliminate nu­
merous repetitions that were played "This or similar positions
only to gain clock time. IP) are not dealt with in avail­
3 0. h4 h5? able books of chess end­
3 1 . Ra8 Kf6 ings. Later I l earned that it
3 2. Ra5 Rd2 was treated in Chess World
33. Ra6 Kg7 some time ago. After two
3 4. Ra7 Kh6 more sessions of unsuc­
3 5. Ra8 g5 c e s sful m a n e uver ing, I
3 6. hxg5 Kxg5 agreed to a draw."
3 7. Rg8t Kf6
3 8. Ra8 Kg7
3 9. Ra6 f5 Purdy's Note
Better was 39 ... h4 40. gxh4 Kh6. Games that should have been won
40. Ra8 h4 but peter out into draws are unsatis­
41. gxh4 f4 fying as a rule, but this one has the
42. Ra6 e3? great compensation of showing an
I think . Kf6 still wins. Ozols
. . ending of great theoretical impor­
handled the defense patiently and tance not given in the books. The
cleverly, and his freeing maneuvers trouble is that all books on endgames
are well timed. are largely based on Berger's old
43. fxe3 fxe3 Theorie und Praxis der Endspiele

- 155 -
 OS Purdy On The Endgame

(long out of print); and anything that If 52... Rd8, then 53. Ke3.
Berger omitted has a good chance of 53. Bc4 Rc6
being omitted by his successors, 54. d5! Rd6
though each successor does add 55. Kd4 e5t
some original work. Thus, Berger 56. Kc5=
treated Queen endings rather briefly, [The original text above was am­
and all his successors have followed biguous in the neighborhood of
him in that. Endgame research is dif- move 53, so there has been some
ficult. speculative reconstruction here.]
In a position we gave in a preced- It is easy to see why the book posi-
ing article, the Bishop ran on dark tions, with the pawns blocking each
squares and protected its own pawn. other on the same file, are winnable.
For a draw, it was necessary that The attacker's King then has oppor­
the defender's King occupied a much tunities of taking the defender's
more forward position than it does pawn (usually after the attacker has
here, so that it could be sure of being given up the Exchange) while still
able to attack the opponent's pawn protecting his own pawn. If the
whenever the opponent threatened pawns are on the a- or h-file, how­
to give up the Exchange. A backward ever, the game is usually drawn even
post for the King would spell ruin. then. See Basic Chess Endings,
For example, suppose that in Dia- which gives eleven extremely useful
gram given here the Bishop is on e5, pages on Rook vs. Bishop with a
giving check. Then Black wins eas- pawn or pawns on both sides, despite
ily: e.g. 1... Kf5 2. Kc4 Rh7 3. Kc5 Rd7. the omission discussed here.
The trouble is that Black's Rook can
cut off the white King and attack the
white pawn simultaneously. In order Devilish Rook
to draw, White's King would need to The weakness of Bishop and
be somewhere near where the black Knight or two Knights against a Rook
King actually is. when the other pieces are off has of­
In the text diagram (i.e., No. 5), ten been stressed by us [See the two
on the other hand, the Bishop has articles on Rook vs. minor pieces
the role of persistently attacking the in "The Search for Chess Perfec-
enemy pawn while the white King tion"]. Never was it made clearer
serves as defender of his pawn. than in this fascinating ending be­
Alternatively, the white King may, tween Weber and Endzelins.
if not prevented, go to d6 and second With the Queens on the board,
the Bishop's attack on the black Endzelins had for some time been
pawn. Just by way of illustration, skating on thin ice at the precipice
from the diagram: of a volcano, so to speak; and Weber
50. ... Kf5 could draw by perpetual if he wished.
51. Ba2 Rh8 But if he wanted to try to win, Weber
Or 51... Rh3t 52. Kc4 Ra3 53. Bbl t had to stop checking and allow the
Kf4! 54. Kc5. exchange of Queens. So Weber did,
52. Bb3 Rc8 and thus the position in the diagram

-156-
 Al l Facets Of The Endgame: Rook vs. Bishop/Knight

came about. 55. Ke2 Nd5

56. Rb7 Nxb4
Weber 57. Rxa'7 Nd5?
If ... Nxa6??, Black loses both
pieces; but he ought to move his
6 Bishop-so as to play ...Nxa6 next
move and draw. Apparently Black did
not think he could lose.
58. Rd'7 Nb6
59. Rc'7 Ke5
White threatened to fork.
60. Rc6 Na8
61. Rc8 Nb6
Endzellns 62. Rb8 Bc5
63. a'7 Na8
49. Kf6 Else Rxb6 and the pawn queens.
50. Re8 Bc1 64. Rxa8 Resigns.
51. a4 Ba.3 Once again the moral is: H you
52. Rb8! win two pieces for Rook, remem­
White must keep his pawn on a ber that it is an exception to the
dark square even though it ties up a general rule that you go for swaps
Rook! with material up. In this case, the
52. ... f4 swaps were difficult for Weber to
53. a5 Nf5 avoid.
54. a6 Ne3t

- 15 7
-
 CJS Purdy On The Endgame

(l\apler H
l\inor Pieces \\Jill\ Rooks
Endgame Strategy
M.E. Goldstein

[M.E. Goldstein, whose methodical endgame play has been instrumen­

tal in securing for him the championships of Middlesex, New Zealand,
New South Wales, Victoria, and Western Australia, here discusses in­
structively one of his endgames in the Australian Correspondence cham­
pionship. -GJS Purdy.]

J\ n advanced pawn is either a Analysis reveals a positional ad­

fistrength or a weakness. When vantage for Black:
a passed pawn, by threatening to 1) White's advanced pawns, espe­
reach the queening rank, ties up sev­ cially his b-and h-pawns, are weak,
eral enemy pieces, it is evidently as Black can open up files by ... a4 or
strong. An advanced pawn that can .. g6. The player with the better posi­
.

be attacked with line-opening effect tion should endeavor to open lines

by a pawn on an adjacent file is often for the use of his Rooks.
merely weak. 2) White is weak on the dark
Consider the following position squares upon which the black
which arose in the Australian Corre­ Knights threaten to settle-for ex­
spondence championship. ample, on g5, f4, and d4 (after White's
c-pawn has been lured away from
M.E. Gold.stein c3).
3) White's backward pawn is
weak on the open d-file. That is of no
1 immediate moment since the pawn
can easily be defended as many times
as it is attacked.
4) White's unsupported d-Rook
gives the possibility of a combinative
breakthrough by ...c4.
5) White's fianchettoed Knights
are inartistic.
0. Ludlow 6) White cannot bring his King to
After White's 40th move the weakened queenside. For ex-

- 15 8 -
 Al l Facets Of The Endgame: Minor Pieces with Rooks

ample, 41. Kf2 Nf4!; and if White ex­ Desirable was 42. Rdd1, followed by
changes, then the other black Knight Re3 and Rc1.
settles on e5. 42. ... g6!
7) Black's doubled c-pawns are If now h6, the advanced pawn be­
defensively strong. comes still weaker. Black plays ... Nf7,
Despite these undeniable advan­ ... g5, and if necessary ...Ng6-h4.
tages, it is uncertain whether Black 43. hxg6t hxg6
has a forced win. A positional ad­ 44. Nd1
vantage without superior mobility A good prophylactic-to meet the
of the major pieces (here the Rooks) threatened ...Rh8-h3 with Nf2.
is not necessarily decisive. The line 44. ... a5!
opening by ...g6 is not deadly, and it Since one open file (the h-file) is
is only White's inconspicuous b-pawn not sufficient to force a win , Black
that contains the seeds of decay. forces open another file (the a-file)
The breakthrough by ...c4, a threat since the Knight at di cannot simul­
that recurs throughout the game and taneously guard h3 and a4.
that is decisive 25 moves later, is not 45. Nf2
yet adequate. For example, 40... c4!? If 45. a4, then ...c4! is deadly. For
41. bxc4! (not 41. Nxc4 because of example, 45. a4 c4! 46. bxc4 Nc5 47.
41... Rxd3, f o llowed by ...Nc5, the Nb2 (luckless wight!) Rh8, followed
double exchange of Rooks, and by ...Rh3t, ...Ke6, and ...Rdh7.
...Nd3-leadingto a Knight ending fa­ 45. ... a4
vorable to Black) Nc5 42. Nell Ng5 46. c4
(threat of ... Ngxe4!) 43. Rc2, and A desperate attempt to counter­
Black reaps no advantage from 43... attack on the b-file. Passive defense
Nge6 44. Rhh2 Nf4? 45. d4! would allow Black to occupy the
So Black decides to improve the a-file.
position of his pieces, including the 46. ... Nd4
King, before seeking a decisive break­ More exact was 46... axb3, fixing
through. the resulting weak b-pawn, followed
40. ... Nfg5 by ...Ra7 and ...Nd4.
41 . Re1 47. bxa4! Ra7
Preventinga sacrificial line-open­ 48. f4! Nge6!
ing by 41... c4! 42. bxc4 Nc5 43. Rhd1 Black maintains a pawn at e5 to
Ngxe4t! The alternative 41. f4 is evi­ restrict the g-Knight.
dently bad: Black would exchange 49. f5 Nf8!
pawns and settle a Knight on e5-- To meet 50. Nh4 by ...g5 and to
41. f4 would thus offend against 1) defend the square h7 against White's
above. Rooks. If 49... gxf5, then 50. exf5 and
41 . ... Kf7 Ne4.
42. Re3? 50. Rb2 Rb8
Probably the losing move. Black Not 50... Ra6 because of 51. a5!
now forces open a kingside file. 51. Re1 Rxa4
White must keep a Rook on the 52. Ne3 g5
first rank to meet ...g6 with Rh1 . A necessary prelude to Black's

-159-
 CJS Purdy On The Endgame

queenside attack. would wreck three years' hard work.

53. Nc2 Nxc2 65. Kg3
54. Rxc2 Nd7 65. Ne7t Kd6 66. Ng6! had possi­
55. Kf3 Rba8 bilities if Black were greedy, e.g., 66...
56. Ree2 Rxg4t 67. Kh2! Ra7! 68. Rc1 Nf8 69.
Not 56. Ra1 because of ...b6-b5- Kh3! But Black could simply play 66 ...
b4-b3. Rh7!, followed by resumption of his
56. ... b5 queenside attack, the white Knight
57. cxb5 being then a mere spectator.
Best. White secures d5 for his Black rejected the variation 65.
Knight. If 57. Ke3, then 57... b4, fol- Kg3 Kb7 66. Rb2 b4 67. Rb3 Ra8 68.
lowed by ...Ra3 and ...Nc3. Rbb2 Rah8 69. Rh2 Rxh2 70. Rxh2
57. ... cxb5 Rxh2 71. Kxh2 Kc6 72. Kg3 Kb5 73.
58. Red2 Ra3 Ne3 (else ...Ka4) c4! (but not 73... Nb6
59. Ke2 Ke7 because of 74. Nd5!)-and Black gets
60. Nd1 Kd6 a favorable Knight ending by ...Kc5.
60... b4 would permit Nc4 as well The immediate breakthrough by
as Nd5. Black therefore keeps his ...c4 is much more forcible.
queenside pawns abreast, retaining 65. ... c4!
the option of ...b4 or ...c4. 66. Rc3
61. Nc3 Kc6 Losing offhand. If 66. Kf2 or 66.
62. Nd5 RhS! Nc3, then 66... Nc5! is decisive, e.g.,
The student should note that the 66. Nc3 Nc5 67. dxc4 Rxc3! Possibly
threat of occupation of an open file the best chance was 66. Kg'2! R:xg4t
is often more effective than the ac­ 67. Kh2! Rh4t 68. Kg2! Nc5!, with
tual occupation. The file opened 19 complications favoringBlack. For ex­
moves previously is only now utilized. ample, 69. dxc4 (better Nxf6! ) Nxe4
63. Kf3 Rh3t 70. cxb5t Kb7! 71. Rc7t Kb8 72. Rdc2
A tantalizing variation runs 63... Rg3t, Black wins the Exchange or
Rh6 64. Kg3 Nb6 65. Ne7t!! Kd6 66. mates in five.
Ng8! Rh8 67. Nxf6 Nc4! 68. Rf2! 66. . .
. Rxc3
Rxd3t 69. Rf3 Rh3t, and White 67. Nxc3 Nc5!
should draw. Geometrical motif: if 68 dxc4, then
64. Kg2 Rh4 68... b4!

� � � � 68. Rb2

11:1111:11:
� ��."-'��.��· " �
2
Not 68. d4 b4! 69. dxc5 bxc3, with
an obvious win for Black. 68. Nd5
leads to a piquant finish: 68. Nd5 cxd3

� .t.,. �4),. :ft,.

� � 69. Nxf6 Rh6 70. Ng8 Rh8 71. Ne7t
��--·· -�·1r�·:f•
-� r���� �".J
Kd6 72. Ng6 Nxe4t 73. Kg'2 Rh2t!!

·£� � r� �
(desperado motif) 74. Kxh2 Nxd2 75.

� ��--··' � �r�
f6! Ke6 76. Nxe5! Nf1t! 77. Kg1 d2,

� � � � and Black wins.

68. ... cxd3
If 64... Rhxd3, then 65. Nb4t! 69. Rb1

-160-
 All Facets Of The Endgame: Minor Pieces with Rooks

Not 69. Rxb5? because of 69 ... Therefore,

Nxe4t. 1. Rgh5! Nf8
69. d2! Now White can compel a draw by
70. a3 Rh8 repetition. His only winning hope is
71. Rd1 Rd8 2. g5.
72. Kg2 ! Rd4! 2. g5 Re7
Black avoids the last trap: 72. . . 3. Nf6!
Rd3? 73. Nd5! If 3. Kf3, then 3 . . . d5 4. cxd5 (if
73. Resigns. Nf6, then Black exchanges Rooks and
plays ... dxc4) Rxd5 5. Nf6 Re3t! 6.
Kxf4 Rxh3 7. Rxh3 Ne6t, with ad­
A Finely Balanced Position vantage to Black.
[Concerning a game in the radio 3. Re2t
match France vs. Australia] 4. Kf3 Rxa2
5. Kxf4
Monsieur Biscay, quoting the gist Or 5. Kg4 a5 6. bxa5 bxa5 7. Kf5
of the report of the adjudicator, f3!, and White can, and should, force
Znosko-Borovsky, said that the ver­ a draw by 8. Rxh7t!
dict on the Kahn-Crowl game was 5. ... a5
that "White had a slight advantage, 6. bxa5 bxa5
but indecisive. " And now White should force the
That conflicts slightly with our draw by 7. Rxh7t Nxh7 8. Rxh7t Kg1)
own verdict that "Black draws at 9. h4 Rf2t 10. Kg4 Rg2t 11. Kf4, with
least." repetition of moves.
On close analysis, it appears that Or first 7. Kf5 Rt'2t 8. Kg4 Rb8,
the balance of chances is so fine that and now the draw by Rxh7t!, etc.
neither side can be credited with any If White equalizes material by 7.
superiority. Here is the diagram. Nxh7?, then he has good losing
chances after 7... Rc2.
RA. Crowl (Australia) A most instructive position.

3 Miracle!
How a player could actually lose
after being two pawns up with Bish­
ops of opposite color, without mak­
ing an absurd slip, is a mystery that
will surely puzzle most readers. It
happened in the Australian cham­
pionship tourney, Adelaide 1946-
4 7; and I can give a firsthand account
as the fortunate beneficiary.

Black is a pawn up and threatens

. . . d5. His weakness is the h-pawn.

- 161 -
 OS Purdy On The Endgame

E.W. Brose gamble; for if the plausible .. .f5 was

Black's idea, then Ke3 gave me a
drawing line that Black might over­
look.
46. ... f5
To play ...Kf6 and .. g5, but there
.

is a flaw.
47. Re7! Rxe7

Earlier in the game, White lost the

Exchange for nothing by a terrible
oversight; and Black exchanged that
advantage for two pawns. I was then
lucky enough to wrigg1e into the dia­ Black could not evade this ex­
grammed position, which contains change without losing a pawn, and
better drawing chances than might White would then probably draw. But
at first appear. if White's King were still at d3, ...e4t
40. c3 ! dx c3t would save the pawn.
The sealed move. Brose saw I now startled Brose by offering a
through my confidence trick. If 40... draw, which he naturally declined.
Bxb3 41. cxd4 exd4, then not 42. Bxd4 Where there are pawns on both sides
but 42. Bg3!, forcing the Rooks off of the board, as here, even the noto­
and obtaining a drawish ending with rious Bishops of opposite color do not
Bishops of opposite color. very often give a draw with two
41. Kxc3 Kg6 pawns down-but this position is
42. b4 Bd5 probably a draw, the reason being
43. g3 c6 that Black can never break through
44. Bc5 Re8 without a sacrifice. If only Black had
45. Kd3 h5 two adjacent pawns on dark squares,
46. Ke3 the proper color against a dark­
This move involved considerable squared Bishop, he could win eas­
mental torture. Should I play 46. h4 ily-f or example, if his h-pawn were
to stop ... h4, or should I just hope on h6 or his f-pawn on f6, Black could
Black would not do that? One thing enf orce ...g5, and all his four pawns
that deterred me from h4 was that it on the kingside could cooperate. As
would not really prevent Black from it is, White just has to play h4 and
opening a line f or his Rook if he was keep his King firm as a rock on e3.
a determined man-e.g. 46. h4 Rg8! There are, however, complications
47. Ke3 Kf5, f ollowed by ... g5! On the when Black's King comes to c4, as
other hand, 46. Ke3 was a good we shall see.

- 162 -
 All Facets Of The Endgame: Minor Pieces with Rooks

48. Bxe7 Kf7 example, 61... c5 62. Kxf5, and White

49. Bd8 Kg6 toddles back with his King. Or 61...
50. h4 B� Be6 (or ...Bg6) 62. Kg5, etc.
51. Be7 Kf7 II: 58... Be8 59. Be7 g5 60. Bxg5!!
52. Bd8 Ke6 (accepting the offer of two pawns
53. Bc7 Kd5 probably loses!) Kxb4 61. Bf6 e4 62.
54. Bd8 Kd2 c5. If 62... Kb3, then 63. Be7! Kc4
64. Bf8!; and White draws, the prin­
ciple being to prevent ...Kb3. If now
64 ... Kd4, then 65. Bg7t Kd5 66. Kc3,
etc. Notice that all of Black's kingside
pawns are fixed on light squares.
Finally, if Black's King plays down
to f1, White uses his leisure in forc­
ing all of Black's pawns onto light
squares. Then White plays his Bishop
to d4 or to c5 so as to meet ...Kg2
54. ... c5? with Kf4, drawing easily.
When trying to win, keep pawns The next few moves were:
on as many parts of the board as pos­ 55. bxc5 Kxc5
sible. After 54... Kc4!, some ingenious 56. Be7t Kd5
person may be able to find a win for 5'7. Bd8 Ke6
Black; but it seems unlikely, as the 58. Bc7 Bh3
following illustrative analysis shows: 59. Bd8 Bg4
54... Kc4! 55. Be7 Bd5 56. Bd6 g6 57. 60. Bc7?!
Be7 Bf7 (preparing the sacrifice ...g5, Meaning incorrect but "best!"
followed by ...h4!, and first placing
the Bishop where it holds up the two
white passed pawns thus created) 58.
Bd6.

On the previous move I had again

suggested a draw, but Brose de­
clined. Clearly Brose hoped that I
would allow him to sacrifice his
After 58. Bd6 (analysis) g-pawn and his h-pawn-that being
his only remaining winning chance. I
Now we give two variations. could have stopped that plan with 60.
I: 58 ... g5 59. Bxe5! gxh4 60. gxh4 Bg5, making the draw absolutely
Kxb4 61. Kf4, and White draws. For dead; but I thought I could allow the

-163-
 OS Purdy On The Endgame

combination and still draw, and 63. Ke4 Bf5t

Black would have a very, very remote 64 . Kf3 Bg6?
chance-but still a chance-of los- Black thinks h5 is a real threat,
ing.Asa matter of fact,Brose thought but it amounts to nothing with
he was taking a serious risk; for he White's King shut out. Indeed , after
knew how tired he was after six the correct move, 64... Bh3, I am not
hours of nearly continuous play, and sure if White has even a draw with
in such circumstances his enterprise 65. h5? White should, on the contrary,
showed courage deserving of a bet- play into the defensive drawing posi-
ter fate. tion with what Bishop move? Solu-
60. g5! tion at the end again.
61. hxg5 h4! After the text move, the situations
62. gxh4 are reversed. It is now White's King
Material is Nothing! who cooperates with his pawns. Nev-
Could anything better illustrate ertheless, Black has a draw-with or
the peculiarities of Bishops of oppo- without his pawns-and all the
site color? When Black was two easier because White's pawns are on
pawns up, the position was a dead the edge of the board instead of near
draw. Now the material is equal, and the center.
Black has good winning chances! 65. Kg4 Bf5t?
That is so because Black has two Here my tired opponent panicked.
passed pawns, and his King can co- Black was quite safe with 65... Kf7;
operate with them. White's passed or, if he preferred, 65... ra first, and if
pawns are on the wrong color; and 66. Bb6, then . Kf7.
. .

White's King cannot cooperate with 66. Kh5 f3

them, so they are absolutely use- 67. Bb6 Kf7!
less-provided Black plays properly. To show how difficult it is to size
White can draw with or without his up positions at sight, I mention that
two pawns, and it is just as hard, or Brose was criticized for this forced
just as easy, with them as without retreat. Black should have counter­
them. Take Black's position now attacked, it was said, with ...Kd5.
against any of your friends who are Then, however, g6 and Kg5! wins
not experts. Play ...Bh5, followed by easily.
...Bg6; then play ...f4t and ...Kf5 and 68. Kh6 e4?
...e4. The odds are ten to one on your But this was really bad-unnec-
winning. When you play ...e4, White essarily putting a pawn on the wrong
must be able to move his Bishop to color and thus giving White's Bishop
one of two drawing squares. There new squares. Black should have
are only two such squares on the moved his Bishop up and down. Then
board. Which ones are they? Solu- White would have played h5, Bf2.
tion at the end of the article. Bh4, and g6t. In his tired state, Black
62. ... f4t would probably not have found the
First ... Bh5 and then ... Bg6 would answer to that. Can you, when not
have made things more watertight, tired? Solution at the end.
but this is all right. 69. h5 Kg8
-164-
 All Facets Of The Endgame: Minor Pieces with Rooks

70. g6 Bd7 we leave the student to work out why

71. Kg5 Resigns. it does draw and why it is the only
It was a heartbreaking finish for way.
the South Australian champion, who
was beaten by fatigue plus the de­
pression caused by drifting out of a
winningposition earlier in the game.

Solutions To Queries
I: The squares are d2 and c1. Af­
ter Black has played ...Bh5, ...Bg6,
and ...f4 t, White should play Kf2; and
upon ...Kf5, then Ba5!, ...e4, Bd2!
White's Bishop then does two things:
it prevents ...e3t, and it bears on one Black to play and draw
of the black pawns so that Black's
King cannot come around to d3 with­ [Ed.: White's timing of g6 U; criti­
out first movingthe f-pawn. That puts cal to winning as U; Black's ...Ke6
both black pawns on light squares, U; to drawing. F.PH. & J.S.P. start
and the draw is easy. out with 68... Bc2 69. h5 Bf5, but
II: 65. Ba5! (to play Bd2 as ex­ according to K.M. 69 ... Bf5 loses to
plained above). 70. g6t Kf6 71. Bd8t Ke6 (Black
ill: Black will try to keep his King doesn't have f5 available for hU;
on f7 until White has played h5, Bf2, King!) 72. BM e4 73. Bf2. However,
and Bh4-then Black should play 69 ... Bd3! 70. Bf2 Ke6!! or 69 ... Ke6!
...Ke6! ! This paradoxical evacuation 70. g6 Kf6 71. Bd8t Kf5 72. g 7 Bb3
is the only possible way to draw, and 73. Bb6 and 73 ... Bg8=.]

- 165 -
 CJS Purdy On The Endgame

[hapter 14
Oueel\ El\llil\qs

The Overrated Queen very much smaller error by saying,

T
he Queen is very much overrated "Three pieces minus one pawn." By
by beginners, somewhat over­ consulting Fine's other treatise-his
rated by average players, and slightly monumental Basic Chess Endings­
overrated even by strong amateurs. we can ascertain, with a sigh of re­
One often reads notes such as this: lief, that the error in Chess the E asy
"White loses his Queen, but gets Way must have been due to mental
Rook, Bishop, and pawn in return. aberration and not to faulty judg­
By skillful play, White succeeds in ment. In the larger work, we find:
drawing ... " And so forth.
The implication always is that in "With both sides possess­
such cases a minor triumph of mind ing pawns, t he Queen is
over matter occurs. The material is roughly equivalent to three
not really equivalent, yet the player pieces. However, with no
with the Queen fails to win. other material, the ending
Indeed, it is commonly asserted Queen plus pawn versus 3
that a Queen is worth a Rook, Knight pieces is drawn, while 3
(or Bishop), and two pawns. p i e c e s plus p a w n win
Fine contradicts this in his text­ against Queen."
book Chess the Easy Way; he gives
the Queen's value as "Rook plus This indicates that the "roughly"
Knight (or Bishop) plus pawn." means that the three pieces are
Unfortunately, Fine spoils this by slightly stronger than the Queen.
following up with a very strange er­ The ending Queen versus three
ror which is certain to mislead thou­ pieces is extremely rare; but still, the
sands of readers. He says the Queen error in Chess the Easy Way could
is also equivalent to "Three pieces do harm by fostering the general ten­
plus one pawn." dency to overvalue her Majesty.
Obviously, these valuations can­ In case of any suspicion that Fine
not both be right, because that would has underrated the Queen in his
mean that a Rook is equivalent to other valuation, viz. Rook, Knight (or
two pieces. Fine would have made a Bishop), and pawn, let us see what

-166-
 All Facets Of The Endgame: Queen Endings

he says about that in Basic Chess again from Fine in Basic Chess End­
Endings. We read: ings:

"This ending (Queen ver­ "Queen and pawn are nor­

sus Rook and Piece) is mally equivalent to two
seen relatively frequently. .. Rooks. This means that the
With both sides possessing ending Queen plus pawn
pawns, the Queen is equiv­ versus two Rooks is drawn,
alent to Rook plus Bishop but Queen plus two pawns
plus pawn. H the pawns are win, while with even pawns
even, the Queen will win the two Rooks will win.
(though not without diffi­
culty), but Rook, Bishop, "With evenpawns the two Rooks
and two pawns are re­ will win"-that is the part to em­
quire d to c o n quer t h e phasize for most players.
Queen." True, that valuation applies only
in the endgame. At earlier stages, the
"Conquer the Queen" -just like best rough guide is to say that Queen
that! Most players would reckon it equals two Rooks. The Rooks, as all
quite a misfortune to have to give up our readers know, do not blossom
Queen for Rook, Minor Piece, and two fully till the endgame.
pawns-very few would know that In the opening and middlegame,
they were actually gaining material. one can also say that a Queen is al­
Unfortunately, Fine does not give most equivalent to Rook, Knight (or
any examples, except two where the Bishop), and two pawns-but if an
pawns are equal (and the Queen endgame arrives, she may lose, par­
wins) and one in which Rook, Bishop, ticularly if the piece is a Bishop
and one pawn win against the Queen. (which always works better with a
As this last one is a very favorable Rook than a Knight does).
position for the Rook and Bishop, it It must also be added that if the
is not of much value. enemy King is thoroughly exposed
Here we might mention that then the Queen's value goes up at
Queen endings are always the most least a full pawn. For example, two
scantily treated, the reason being Rooks and some pawns will not de­
that Berger treated them scantily in feat a Queen with an equal number
the first place and all subsequent of pawns if the Rooks' King is com­
end game textbooks have been based pletely exposed to checks, i.e. the
mainly on Berger. Here is a vast field King has no pawn shelter. It is very
for original research. hard to win any ending at all when
you have a badly exposed King and
the enemy has a Queen.
It must not be assumed that the
�fost players are chary of giving Queen is always a strong piece in the
up the Queen even for two Rooks. opening and the middlegame. A stag­
But listen to the truth of the matter, gering example of how weak a Queen

- 167-
 CJS Purdy On The Endgame

can be, and how strong the two Bish- 11. Ndxe4 dxe4
ops can be is the Crakanthorp-Woi- 12. 0-0 Bxc3
narski game to follow. That should 13. bxc3 Nxe5
be treated as part of this article. 14. Qxe4 f6
15. Bf4
Unwittingiy, Crakanthorp was fol­
Two Bishops vs. Queen lowing Griinfeld-Bogoljuboff,
This is one of the most curious Miihrisch-Ostrau 1923; and here he
games ever played, but it never be­ walked into the same trap. White
came widely known. It was played in should play 15. Bh4.
the first round of the Australian 15. ... Bf5!
championship tournament, Sydney However, Crakanthorp kept his
1926, between the late Spencer Crak­ head better than Griinfeld who now
anthorp, winner of the tourney, and played 16. Qd4? (16... Rad8 17. Bxe5
S.H.Z. Woinarski, the runner-up. It Rxd4 18. Bxd4 c5!).
will serve to illustrate our preceding 16. Qxf5! Nf3t
article THE OVERRATED QUEEN. 17. Bxf3 Qxf5
Crakanthorp, under duress, gave up Black has a whole Queen for two
his Queen for two Bishops in the Bishops; and, moreover, White has
opening-and drew! isolated doubled pawns. Doubtless
Notes marked /C were scribbled there is some way fo r Black to win,
by Crakanthorp in an old notebook. but the sequel shows how hard it is
to find. The two Bishops derive their
S. Crakanthorp-S. Woinarski terrific strength in this position
Queen's Gambit Declined mainly from the fact that Black him ­
(Cambridge Springs Defense) self has no other minor piece-above
1. d4 d5 all, no Bishop-that can ever threat­
2. Nf3 Nf6 en to exchange off one of the white
3. c4 e6 Bishops. In such circumstances, two
4. Nc3 Nbd7 Bishops are often at least as strong
5. Bg5 c6 as Rook and Bishop-as here.
6. e3 Qa5 We have not dissected the follow­
Black's 4th, 5th, and 6th consti­ ing play minutely; we just take our
tute the Cambridge Springs Defense, readers through it, to show that Black
which White often avoids by 6. cxd5. made no obvious error and yet failed
7. Nd2 to win .
To meet the threatened ...Ne4. 18. Rab1 g5
7. Bb4 19. Bd6 Rfd8
8. Qc2 0-0 20. c5! Rd7
9. Be2 21. h3 Qg6
Capablanca's 9. Bh4!, removing 22. Rb4 a5
the Bishop from the black Queen's 23. Rb2 h5
line of fire, is now thought strongest. 24. g4! hxg4
9. e5 25. hxg4 Rh7
10. dxe5 Ne4 Plausible but useless; better .. Qd3
.

- 168
-
 All Facets Of The Endgame: Queen Endings

first. IC Basic Chess Endings (BCE) (the

26. Rfb1 Qd3 tournament player's bible) and then
27. Rxb7 Rxb7 begin analyzing? If the game is ad­
28. Rxb7 Rc8 journed overnight, it would in my
29. Rb6 Qxc3 opinion pay better to play through
30. Bxc6 Qa1t ALL the Queen endings in BCE that
31. Kg2 Qxa2 bear even a remote resemblance to
32. e4 Qe2 your own. Although some of the po­
33. Bd5t Kh8 sitions may seem irrelevant, the tac­
34. Kg3 Qd3t tical devices they display will soak
35. f3 themselves into your mind and may
This allows Black a draw when- prove extremely helpful during the
ever necessary, but White was not long and arduous struggle-more
averse to a draw. IC helpful than a vain attempt to ex­
35. ... Qc3 haust the possibilities of your own
36. Rb7 position.
U 36. Be7, then 36... Kg7, giving However, all the positions in BCE
up Queen for Rook and Bishop-with showing Queens and pawns, al­
winning chances through the a-pawn. though very instructive and valuable,
IC do not pretend for one moment to
36. a4 exhaust the subject. Moreover, so
37. Rf7 a3 complicated are such endings in
38. Be7 a2 practice that "book wins" are often
Before forcing the draw, Black has drawn and "book draws" often won­
a last little flutter-which of course even in master play.
fails. Let's see what BCE says of the
39. Bxa2 Qe1t very common ending Queen and
40. Kh3 Drawn. pawn vs. Queen. Inter alia, this:
Perpetual check is unavoidable.
"This is a draw unless the
attacker has the c-, d-, e-,
Queen Endings or f-pawn, supported by the
Less is known about Queen end­ King, on the rank next to
ings than any other kind. The root the queening rank. As a
reason is that they run to inordinate rule, it is best for the de­
length and look very forbidding in fender to have his King as
print, besides taking up much space. far away from the pawn as
Progress is constantly being inter­ possible, unless, of course,
rupted by long series of checks. it can occupy a square in
Writers, therefore, treat Queen end­ front of the pawn-in which
ings rather sketchily. case the game is a hopeless
What is the best thing to do when draw... In general, it is im­
you get a Queen ending at an ad­ possible to advance a pawn
journment, or look like getting one? very far. T h e def ender
Is it to look up the general case in keeps on checking; and

-169-
 CJS Purdy On The Endgame

when he runs out of checks, to stop the pawn from marching from
he pins the pawn. In view the fifth rank to the seventh, and my
of the terrific number of experience is that the b- or g-pawn
possible positions of the has better chances than any other
Kings and the pawn, all of kind of pawn of getting to the rank
which require individual next to the queening rank. The b- or
attention, it is out of the g-pawn's disabilities come to light
question to s ubmit this when it arrives at the penultimate
statement to a precise rank-due to the possibilities for per­
analysis. At any rate, in ac­ petual check by the opponent; per­
tual practice it is usually petual check can often occur even if
found that the pawn can be the pawn is allowed to queen.
held back." In the diagramme d position, Kru­
ger made a false step; but the posi­
This means that the b- or g-pawn tion is interesting in that it shows
should not win even if on the rank the special drawing possibilities.
next to the queening rank-still less [Ed.: In this case it is really seen
so if farther back, for apparently its that Kruger was actually lost ac­
advance to the penultimate rank cording to K.M. See below edito­
should be preventable. rial note.]
How, then, did Pachman come to Presumably influenced by the gen­
win such an ending from Gligorich in eral principle (see above) of keeping
the Chigorin Memorial Tourney?­ his King as far away as possible,
an amusing brevity of 126 moves. Kruger played:
How, also, did Stefan Kruger lose 85. Kg1
the same ending to John Hanks in 86. Ka8 Qd8t
the Aus tralian cha mpionship, 87. b8=Q Qh8
Brisbane 1951? The second question An a m azin g f u t i l i t y , hoping
is easier to answer. against hope for Qxh8 stalemate. Of
course Hanks checked first, and
S. Kroger Kruger resigned. But had Black cho­
sen ...Ke1! on his 85th move, he could
have played 87... Qa5t, forcing the
1 routine draw with one Queen against
two. [Ed.: K.M. comments that 85 . . .

Kel!? is the best try though it still

loses. It takes accurate moves to
prove i t, so, in practical play,
Whi te may go wrong over-the­
board.] For example, 87... Qa5t 88.
Qa7 Qd8t 89. Kb7 Qd7t 90. Kb6 Qd4t.
J.N. Hanks Hanks, therefore, would have avoided
Black to play his 85th 86. Ka8 (after 85... Ke1); perhaps it
would then have been possible to ap­
Both Gligorich and Kruger failed ply the Pachman method.

- 1 70
-
 All Facets Of The Endgame: Queen Endings

S. Gllgorich several bad rrwves.-K.M.]

103. ... Qe4t
104. Kd8 Qa8t
2 105. Kd7 Qd5t
106. Kc8 Qg8t
107. Kb7 Qd5t
108. Ka6 Qa2t
109. Qa5 Qg8
110. Qe5t Kb1
111. Kb6 Qd8t
112. Kb5 Qd7t
L. Pachman 113. Kb4 Qb7t
After 99. K(b.!)at
.. 114. Kc4 Qc6t
115. Qc5 Qe4t
The ending to Pachman-Gli­ 116. Qd4 Qc6t
gorich was of real interest for theory. 117. Kb3 Qf3t
Pachman won by first getting his The next six moves are just a
pawn to the seventh rank and then merry-go-round.
pulling his King right away to the far 118. Ka4 Qa8t
corner where it could: 119. Kb4 Qb7t
a) threaten mate, and 120. Kc3 Qc6t
b) use the enemy as shelter from 121. Qc4 Qf6t
checks in the rear! 122. Qd4 Qc6t
Let's take the game from move 123. Kb3 Qf3t
100. The loser of this ending was one
of the three or four best players in
the world outside the USSR-which 3
goes to show that if there was a draw
it wasn't easy.
100. Qc7 Kb2?
[Ed.: K.M. says this loses, but
JOO . . . Qh4t, 100... Ka2, and 100...
Qf6draw.]
101. Kg8
[Ed.: This should draw. White This time White does the right
wins in a protracted struggle with thing.
101. Kh7.-K.M.] 124. Qc3 Qb7t
101. ... Qd5t ? 125. Qb4 Qd5t
[Ed.: Loses because 102. Kf8 Of course Black couldn't take the
would win.-K.M.] pawn. Now we arrive at a typical tac­
102. Kf8 Qf5t tical device: the checks must end be­
103. Ke8? cause White's King can discover
[Ed.: This draws, but 103. Ke7 check. Black is lost. Could he have
wins. Hence, in the absence of got out of it?-that's the question for
tablebases, the two players made endgame theorists. If Black couldn't,

- 1 71 -
 OS Purdy On The Endgame

then the theory of this ending will mate rank vs. Queen endings.
need revision. Keres's conclusion was that
126. Ka3t Kc2 with any pawn at all except the a­
127. Qb2t Resigns. or h-pawn there are excellent win­
Exchange of Queens is unavoid­ ning chances theoretically-and
able. Surely this ending of Pach.man's still better chances in practice be­
will become a classic! cause the defense ls so much more
What general advice can we give difficult than the attack.
about Queen endings? Only to study In view of the extraordinarily
as many of them as you can find, with painstaking research made by Keres,
a view to getting a general idea of this verdict must supersede all oth­
how to play them. A mere knowledge ers, including that inBCE. It confirms
of what is supposed to happen ''with my own suspicions expressed above.
best play" will not help you much, Diagram 14 shows the Botvinnik
especially as existing theory on the game at move 94.
subject is so nebulous.
BCE gives nearly 100 Queen end­ Ravinsky
ings-a very useful bunch.
One very common error in such
endings is to imagine a perpetual
check where it does not exist. Some­
times a King may be checked a dozen
times or more and still reach calm
water.

The Question Is Answered!

Not knowing I had written an ar­
ticle on Queen endings, the 16-year­
old expert John Purdy-who hap­
pened to be idly ffipping through the Here again the game ended at
pages of a Russian chess book-sud­ move 126, with Black's resignation.
denly called out, "Here's an article White's King laboriously journeyed
all about one Queen ending-they've to a7, just as Pach.man's did-and
crammed the analysis into 26 pages!" for the very same reason: to use the
The ending proved to be one of black King as a shelter. But the white
Queen and g-pawn on the penulti­ King's next journey was quite differ­
mate rank vs. Queen! Botvinnik­ ent because of the entirely different
Ravinsky, 13th USSR champion­ situation of Black's King. White's
ship 1945. The analyst was Keres. Queen was then on c7. After ending
The preamble states that Keres did the checks by getting his King to a7,
the research with a view toward do­ White transferred his Queen to f6 (via
ing something to rectify the very un­ Qf4t); and then his King marched all
satisfactory state of existing theory the way back: Kb7-c7-d7-e7-e6-e5-f5-
for Queen and pawn on the penulti- g5-g6. The position at move 126 was

-172-
 All Facets Of The Endgame: Queen Endings

as diagrammed. any kind of pawn.

As an appendage to the preced­
ing article, we published Keres's con­
5 clusion-to which our attention was
drawn just as we went to press:

"With any pawn at all ex­

cept an a- or h-pawn there
a r e excellent w inning
chances theoretically-and
still better chances in prac­
White to play his 1!6th tice because the defense is
so much more difficult than
White played 126. Qf4t, and Black the attack."
resigned.
In actual practice, such endings We expressed the view that such
are often drawn, largely because the endings were often drawn because,
attacker does not have enough hope thinking the general case to be a book
of winning-long successions of draw, the would-be winner did not
checks serving to discourage him. pursue his goal with sufficient opti­
�ow both the precept of Keres and mism.
the example of Pachman are there Keres had published an analysis
to give the attacker the necessary op­ in 27 pages of the ending Botvin­
timism. nik-Ravinsky 1945, and it was that
analysis that led him to the conclu­
sion given.
Queen And b- Or g-Pawn The ending happened again to
Versus Queen Botvinnik, who had meanwhile stud­
In the preceding article, apropos ied Keres's analysis, in the Hamil­
of the ending Pachman-Gligorich, ton-Russ ell Cup tourney, Amster­
Chigorin Memorial Tourney 1951, dam 1954; and he published an illu­
we expressed the view that the ex­ minating article on it in Shakmaty,
isting theory of the ending Queen and which appears in Engiish in Chess
pawn versus Queen would have to Digest for January 1955. Botvin­
be revised. Basic Chess Endings nik's opponent was Minev.
gives it as a draw in general, with Botvinnik enunciates the impor­
the exception of the c- or f-pawn. On tant principle that if the defender's
the contrary, latest evidence points King skulks in the background, as
to the ending being a win in general, he is advised to do so as not to be in
"\\'i th the exception of the a- or h­ the way when his wife wants to
pawn-we refer of course to posi­ check, then the attacker ought not
tions other than those unfavorable keep his own King near the pawn but
ones in which the defender's King rather should edge him over toward
stands in the way of the pawn, in his skulking rival. The aim is clear
which case the ending is a draw with from the accompanying diagram.

- 1 73-
 CJS Purdy On The Endgame

Mlnev cept with the c- or f-pawn, because it

was assumed that the attacker's King
would stay near his pawn.
R.J. Roycroft, the well-known Eng­
lish player, mentions inBCMthat he
published a study in the BCM of De­
cember 1950 with Queen and g-pawn
vs. Queen, White to play and win. At
the time, however, it attracted little
attention as there was no intimation
that the composer considered the
Botvtnnik general case a win-it was treated
After 91. K(d6)c5 simply as an ingenious composition
and forgotten.
Wh i t e ' s l a s t m o v e w a s 91 . All the endings so far have fea­
K(d6)c5! Black's move, in Diagram tured the b- or g-pawn. It is prema­
6, was ... Resigns. You can soon see ture as yet to assume that a center
why. Black has four possible checks; pawn also wins in general. The ·win
but in reply to each one, White's with the c- or f-pawn is given in Ba­
Queen interposes and gives check in sic Chess Endings.
doing so. Thus this diagram is worth Long ago, either in his Chess Prin­
remembering as the ideal type of po­ ciples or in his Art of Chess, James
sition to play for. Mason said that the b- or g-pawn was
Botvinnik says the endingwas for­ the best for winning this ending-he
merly considered always a draw, ex- may be right, after all .

-174-
 All Facets Of The Endgame: Heavy Artillery

Chapter I�
Hea\l\j Artiller\j

Queen vs. Rook

[Purdy's personal experience in successfully defending a K + R vs. K +

Q ending in an important game brings this article alive. In the 1990s,
positions like this have been thoroughly analyzed by means of comput­
ers; and it is possible that computers might reveal inaccuracies in
Purdy's text-but one cannot legally consult a computer in actual play,
and Purdy's discussion of plans for both sides is illuminating. -FP.H.J

S thon, we have received numer­

ince the Purdy-Hastings mara­ Shortly before the second adjourn­
ment, Purdy and Hastings arrived at
ous requests for an article on the end­ the position diagrammed, with White
ing Queen vs. Rook. A complete dis­ to move. White is in one of the snug­
cussion of the ending requires a gest positions he can reasonably
·whole book-see Queen versus Rook hope for-that is, his Rook holds the
by "Euclid." However, we shall deal third rank, shutting out the enemy
briefly with the most difficult posi­ King. If White selects the perfect
tion-Rook holding the third rank. square for his Rook at every move,
the problem of winning for Black is
Hastings one of the most difficult in chess.
The perfect play, which gives a
win in 20 moves, is as follows. We
1 have no space for all the variations,
but the student who carefully goes
through the main line will have a
good working knowledge of the end­
ing in general.
1. Re3t ! Kd4
2. Rg3! Qb2t
3. Ke1! Qc2
Purdy 4. Rh3 Ke4!
This is a position to remember.

- 17 5
-
 CJS Purdy On The Endgame

If the Rook were at g3 instead of instead of holding the third line,

at h3, White could not hold the rank White now holds the f ourth, here the
much longer; for after 5. Rh3 Kf5! d-file, which is still better. This posi­
White is in a sort of Zugzwang. tion is so hard to win that it does not
5. Rg3 Qc5!! appear in any textbook, though it is
A move impossible to find without presumably given in the brochure
previous knowledge. Queen versus Rook by "Euclid. "
6. Ke2 Qh5t But here White made a false step.
7. m Qh2t White's correct move was 106. Kb4,
8. Rg2 preventing the Queen from checking
Off at last! closer; but White chose Kb5, and the
8. Qf4t game proceeded thus.
9. Kg1 Ke3! 106. Kb5? Qb3t
10. Rb2 Qd4! 107. Kc5 Ke5
11. Rf2 Qd1t 108. Rd7 Qa3t
12. Rf1 Qg4t 109. Kb5 Qc3!
13. Kh2 Ke2 Putting White in a Zugzwang. If
14. Rg1 Qh4t now 110. Rd8, then 110 . . . Ke6!; and
15. Kg2 Qd4 White would have been forced cor­
16. Kh2 m nerwards in much the same way as
17. Rg2t Kf1 in the text. White now sealed. Hast­
Now wherever the Rook goes, it ings had 22 moves left (to beat the
can be won by a few checks. 50-move rule for a draw).
The next day, play proceeded as
Hastings follows.
110. Kb6 Qb4t
111. Kc6 Ke6
112. Rd8 Qc4t
113. Kb6 Ke7
114. Rb8 Qb4t
115. Kc7 Qc5t
116. Kb7 Kd7
White is cornered, but Black's
King is not well placed.
117. Ka6
This brings about a position
which, with White to move, would be
On the resumption next morning, a forced draw! ForWhite could check
Hastings wasted a great many moves endlessly, owing to threats of stale­
with checks. Toward the end of that mate. For example, 1. Rb7t Kc8 2.
session, it was Purdy's move in the Rb8t!! Kc7 3 . Rc8t!!
second position diagrammed. It will But here it is Black's move, and
be seen that, in some mysterious Black could have eliminated the
fashion, the scene has shifted from threats of stalemate by a series of
the South to the West. Furthermore, checks, starting with 117. . . Qa3t.

- 176 -
 All Facets Of The Endgame: Heavy Artillery

Instead, Black gave White a breath­ shown in the previous example.

er by The following would be the per­
117. ... Qd5? fect play, starting from the last dia­
118. Rb7t Kc6 gram:
119. Rb6t Kc7 1. Rc4t Kd5
120. RM Qd3t 2. Rc8 Qa3t
121. Ka5 Kc6? 3. Kb5 Qb3t!
4. Ka6!
Hastings If 4. Ka5, then 4... Kd6! produces
the Zugzwang mentioned in the first
example; and Black wins in nine
3 more moves.
4. Qb4!
5. Rc7 Kd6
6. Rb7
The third line cannot be safely
held any longer.
6. Qe4
7. Rb5 Kc6
Purdy 8. Rb6t Kc7
9. Ka5 Qa 8 t
122. Rc4t! 10. Ra6 Qb7
This surprise move made the draw 11. Rh6 Qd5t
a certainty. Black cannot take the And Black wins the Rook in five
Rook because of stalemate, and now more moves. Once the Queen can get
White has the Rook on the third line in between the two pieces, the Rook
with only nine moves to go. must fall.
The position, however, is not as The student will see that the trick­
favorable for the weaker side as the iest part is edging the Rook off the
one given in Diagram 1. The play third line. A single slip may allow the
went: opponent to play into the perfect po­
122. ... Kd5 sition shown in Diagram 1. So it only
123. Rc8 needs about four slips at the proper
White's pieces happen to be badly intervals to eat up the fifty moves!
placed. If the Rook were already at
c8 or c7, or if the white King were at Note Added
b4, then the task of driving White off C.P. Lowe, who happens to be a
the third line (the c-file) would be far connoisseur in the field of the end­
more difficult (see the previous ex­ game, inf orms us that the Rook on
ample). the third rank defense in the ending
123. ... Qa3t Queen vs. Rook does not give the
124. Kb5 longest resistance possible, but the
Black now played 124 ... Qe7?, second-longest.
which allowed White, by 125. Rc2!, Lowe is the f ortunate possessor
to play into the perfect position of the book Queen versus Rook by

- 177 -
 OS Purdy On The Endgame

"Euclid," in which the following posi- J.G. Stocks

tion is given as providing the longest
possible resistance.
5

D.Robson
Black to move

What makes Black's position so

Black requires 26 moves to win superior is White's bottled Rook.
here, whereas he can always win However, Black is not likely to win
against "Rook on the third rank" in unless he can bring up reinforce­
22 moves or less. ments. That he can obviously do by
We presume that while the oppos­ 1. .. Rxe4, giving himself a second
ing King remains at c3 the Rook passed pawn; and it only remains to
moves up and down the d-file, always check up for some forking-check type
choosing his square carefully. of combination by White. As there is
However, this position can only be no such thing, 1... Rxe4 should be
attained in exceptional cases, where­ played-whether it wins or draws
as the Rook on the third rank de­ need not be decided. Actually, it
fense is comparatively easy to get. should win. For example, 1... Rxe4 2.
Therefore, we advise the student to gxh6 (gi}t is relatively harmless here)
concentrate on that defense alone. 2. . . Rf6. Now if 3, Qd7, then 3. . . K:xh6.
Or if 3. Qc8, then 3. . . Rxh4. And if 3.
Qd5, then 3 ... Rxh4 4. hxg7 Rg4t 5.
Change Of Fortune Kh3 Rxg7 6. Qe4t Kh8 7. Qa8t Rg8;
In a Victoria-New South Wales and again Black wins. Such analysis
telegraphic match, J.G. Stocks, who need not concern the player at the
had soundly swapped Queen for time; he can reason simply that if 1.. .
Rook, Bishop, and pawn and who had Rxe4 is not good then nothing is.
good winning chances against Don­ Actually, Black did just the thing
ald Robson, suddenly spoiled his po­ that common sense rules out: 1...
sition-after which Robson, playing Re1? He was counting on gxh6 and,
with great tenacity, turned the tables. being tired, forgot that 2. gi}t might
Here is the critical position. now be effective. Then the game
should have been drawn; but after 2.
g6t Kh8 3. Qg4 (a doubtful move),
Black pursued his erroneous course
and exchanged Rooks-whereas 3 . . .

- 17 8
-
 All Facets Of The Endgame: Heavy Artillery

Rb8 {making . .Rxf1 a real threat)

. wriggling out.
would have forced 4. Rxf2, which The moral of this change-of-f or­
leaves the Queen struggling for the tune scenario is: In cross-board
draw against two Rooks. chess, utilize principles and common
After 3. . Rxf1 4. Kxf1 Rb8, Black
. sense as much as possible. Calcula­
has a terrible threat; but it is easily tion, on its own, is too much subject
parried-and then White can start to oversight.

-179-
 CJS Purdy On The Endgame

Chapter I�
Sure Dra\\ls(1J

Two Kelling Endings induce pawn advances by the other.

K
elling on the gentle art of win­ First, as in all such endings, the
ning a drawn endgame. Kings develop.
34. Nc-2 Ke7
D.I. Jones 35. Kf1 Kd7
36. Ke2 Kc6
37. Kd3 Kb5
38. Ne3 Kc6
Centralize the King is the rule.
Here it was tempting to break the
rule and play 38... Ka4 39. Nxd5 Kxa3,
as that would eliminate Black's weak
d-pawn. But 40. Kc4 would give White
a commanding position (if thereupon
40... Ka4?, then 41. Nc3t).
EK. Kelling 39. Kc3 b5
After 33... Kf8 This looks reasonable enough, but
it should have been avoided on prin­
This endgame, although slightly ciple. Rather, tempt White to make
in White's favor, should of course be the advance a4; and then play ...a6,
drawn. There are pawns on both leaving White with a weakness at b4.
wings; that is favorable to the Bishop. 40. Kb3 Be7
But no passed pawns can be made; 41. Ng4 Bd6
that is favorable to the Knight. There 42. g3
are pawns in the center; that also To prevent the Bishop from enter­
favors the Knight by limiting the ingthe White camp. Black's next move
Bishop's mobility while securing cen­ is bad-he is tempted into breaking
tral squares for the Knight. a rule because he can thereby isolate
Both sides should try to avoid ad­ three of White's pawns.
vancing pawns-since passed pawns 42. ... f5?
cannot be made, such advances 43. Ne3 f4?
would only create weaknesses. As a Pursuing his plan, Black loses the
corollary, each side should play to game. But, after 43... g6 44. Ng'2! (with

-180-
 All Facets Of The Endgame: Sure Draws!

the threat of Nh4) g5 45. Ne3 f4 46. This position occurred on the
gxf4, followed by Nf5 or Ng4, White's 49th move. To create winning chanc­
Knigbt would still enter the Black es, Kelling had to enter Black's
camp, with winning possibilities. camp-and to do that, he had to sac-
44. gxf4 Bxf4 rifice a pawn (for if 49. Rd5, then
45. Nf5 Bh2 49 ... Bf7).
Black must lose a pawn. 49. Qd1! Bxe4
46. f3! g5 50. Bxe4 Qxe4
47. Ne7t Kd6 White has exchanged off his im-
Black decides to go for a passed mobile Bishop.
pawn on the kingside at all costs, so 51. Rd6! Qb7
as to make some use of his Bishop. If ... Qe3t and ... Qxc3??, Black is
There seems to be no better chance. mated in a few moves. Passive de-
48. Nc8t Ke6 fense, however, was fatal. A better
49. Nxa7 Kf5 chance was 51... h6, making an es-
The demerit of Black's 39th move cape for the King and preserving
is painfully brought home. threats.
50. Nxb5 Kf4 52. Rb6 Qa8
51. Nd6 Kxf3 53. Qd6 Rd8
52. Nf7 Kg3 54. Qe6t Kf8
53. Nxh6 Kxh3 55. Rxa6 Qe4
54. a4 g4 Showing how futile the Queen'sre-
55. Nf5 g3 treat was.
56. Nxg3 Kxg3 56. Rd6 Re8
57. b5 Kf3 57. Qd7 Qe3t
58. a5 Bg1 58. Kh2
59. b6 Bxd4 The pawn formation g2, g3, h3 has
60. Kb4 Ke3 one virtue-it forms a fine sanctu-
61. Kb5 Be5 ary for the King (a point worth re-
62. a6 membering).
And White ultimately won. 58. Qxc3
59. Qxb5 e4
Second Endgame 60. a6 e3
E.H. Seveme 61. a7 Qa1
If 61... e2, then 62. Qxe8t.
62. Qc5 Qe5
2 Intending to meet Re6t with
... Qxc5, but finding it futile. However,
if 62 ... Kg8, then 63. Qd5t Kf8 64. Rd7;
and Black must lose his Queen.
63. Re6t Kf7
64. Rxe5
And White ultimately won.

F.K. Kelling

- 181 -
 OS Purdy On The Endgame

A Missed Draw 59. Kxh2 Kd3

60. Ra5 Re6
N. Cromarty 61. Kg"J Rf6
This position is a "book draw," but
only just, and White could easily lose.
3 If Black declines the Knight at the
outset, he does not improve his
chances. If 55... Kh3, then 56. Nf4t;
and the King must come back to g3.
If 55... Kf3, then 56. Nd4t Kf4 57.
Nxf5t Kxf5. White now continues
checking from the c-file until the
black King gets to the second rank,
R. Wade whereupon White starts checking
from the third rank and then from
Certainly White's position is un­ the first rank if the King goes there.
pleasant looking, as White is two There is no getting away!
pawns down and there are mates in
the air; but 55. "Resigns" was a seri­
ous blunder. Never resign when Winning a "Drawn Game"
your King is in a stalemate posi­ In a certain New South Wales
tion! championship tournament, nothing
In sending us the score, Kelling was more amazing than Goldstein's
jotted down on the back, "If 55. Ne2t win in the last round from Dixon. The
Rxe2 56. Rg4t!, then either draw by draw appeared to be so "dead" that
perpetual check or stalemate!" Goldstein caused amusement by con­
That is a true bill. The Rook can tinuing to play-but Goldstein had
continue checking forever. The Rook the last laugh.
has only to avoid checking at e4 or Goldstein is quite satisfied to get
checking where the black Rook can into a "drawn" endgame against most
interpose-the Rook can avoid both of his opponents. To say that his op­
these contingencies, as the student ponents "ought to have drawn" \\'ill
may verify by experiment. After not help at all. If, on the other hand,
55. Ne2t Rxe2 the results induce players to devote
[Ed.: K.M. suggests 55 .. . Kf2! more attention to endgame technique,
Then after 56. Nd4! there is 56...
, then Goldstein will have done Aus­
Bd3 57. Rc3 Be4t 58. Kxh2 Rbl 59. tralian chess a great service.
Kh3 Rb4 and Black wins,]
56. Rg4t Kf3
57. Rf4t Ke3
a player might lose his nerve; and
instead of 58. Rf3t!, he might take
the Bishop. But that would give him
a difficult task. Thus,
58. Rxf5!? Rett!

-182-
 All Facets Of The Endgame: Sure Draws!

M.E. Goldstein to the midline of the board. White had

an easy draw either by 37. Kd4 (logi­
cal because it "takes the opposition")
4 or by 37. Ke4.
Both players were short of time.
37. ... Kd5
This is in accord with the simple
rule just stated, but there is a better
move-and Goldstein would almost
certainly have found it had he not
been short of time. What is that bet­
D.L. Dixon ter move, and what should the result
White to move be with best play after that move?
Can Black force a win , or has White
With White to play his 35th move, some ingenious way of forcing a
the position was as diagrammed. draw? The position is given at the
White has an absurdly easy draw by end of the article as Diagram 5. Stu­
keeping the Rooks on; yet he inter­ dents will find Diagram 5 to be not
posed, probably because he thought only fascinating but also of immense
a clear draw could then be forced in practical value-even if they fail to
a few moves-and in that he was arrive at a final conclusion.
quite correct! Students may gain a little help
35. Re3 Rxe3t from Cheron about blocked pawns­
36. Kxe3 h4! see Guide to Good Chess, Part N.
If you can fix two units with one, 38. Kg4 Ke4
that is a definite gain. And, of course, 39. Kxh4?
the text move does that, as the black This was an easy mistake to make
h-pawn blocks the white h-pawn and under time pressure. But the edge of
restrains the white g-pawn. the board is a dangerous place for a
37. Kf3? King, just as it is for a Knight. Actu­
Another point about Black's ...h4! ally, White had a forced draw-and
was that it tempted White to go after very nearly a win-by 39. f5! Thus,
the \irtually isolated pawn. Yet in 39. f5! gx:f5t 40. Kxh4 Kf4 41. Kh5
doing so, White disregarded a lead­ Kg3 42. Kg5 f4 43. h4, and it seems
ing principle of pawn endings: that Black is lost. But no, 43... f6t!!
44. Kf5 Kxh4 45. Kxf4 Kh5 46. Kf5
RULE47 Kh4, and Black draws.
-¢- 39. ... Kxf4

A
dvance your own King as 40. g3t
far as you safely can, and If 40. g4, then Black mates in two!!
keep the enemy King back. 40. Kf5
-¢- 41. g4t Kf4
42. g5 Kf5
Time and again, such endings are 43. Kg3 Kxg5
won because one King beats the other 44. Kf3 Kh4

- 183 -
 OS Purdy On The Endgame

45. Resigns. But Black, lured by the chance of get­

The rest is easy. ting a passed pawn for the asking,
Don't fail to try the next position. gives his opponent an advantage on
one wing. And with Bishops of oppo­
site color, wing advantages are what
count. A central passed pawn is of
little use, and here the pawn has the
bad feature of occupying the most
important dark square on the board.
thus robbing its own Bishop of free­
dom. But the game is still a draw.
29. Bd3! h6
30. Bc4 Ke7
Black to play-What result? The white Bishop is hampering the
black King a little. White begins to
have hopes.
Dead Draw! 31. Ke2 g5
On move 28 of the Wade-Belton Black gratuitously creates a weak­
game in the Dominion championship ness on the side where White already
tourney, the position was as dia­ has an advantage. Things are look­
grammed. Bishops of opposite color, ing up.
level pawns, and level position! Was 32. hxg5 hxg5
it worth playing on? It was. 33. Kf3
Back again, because he now has
C.P. Belton somewhere to go.
33. ... f6
Black has three pawns fixed on
6 dark squares. That gives White fur­
ther hope.
34. Kg4 Bet
35. f4 Bd2
36. Kf5 Bet?
By exchanging pawns, Black
would have secured the draw still.
[Ed.: Purdy implies that this is the
R.G. Wade losing move, but, though perhaps
White to move not the most accurate, it still
should draw (see note to move 42),
28. e4! so does not really deserve a
In such positions, one must look query.-F.P.H.J
for the move that gives the opponent 37. e5! fxe5
some chance to go wrong. 38. fxg5 Bxg'J
28. ... d4 Black is now actually a pawn up:
Either ...Ke5 or ...dxe4 t would but with Bishops of opposite color,
have left White nothing to play for. material is of no significance. White's

- 184 -
 All Facets Of The Endgame: Sure Draws!

one wing pawn in combination with Ragtime Band

his well-posted King and Bishop can Error is the lifeblood of chess. Not
laugh at Black's two united pawns only must the greatest masters in­
which are only in the center-they evitably make errors in play, but also
might just as well be in the box. writers in books-unless they con-
39. g6 Kf8 fine themselves to exceptionally easy
40. Kf6 Bh4t positions.
41. Kxe5 Kg7 Alexander's book Chess well fills
42. Bd3 Bf6t the gap between the chess primer
If 42. .. Bg3t (a better hope), then and the expensive textbook. It is as
White would disdain the pawn and authoritative as it is deservedly popu­
play Ke6! [Ed.: K.M. slwws that with lar, but we noticed a very instructive
42. . . Be7 Black draws easily upon error in it once when someone asked
43. Kxd4 b6 44. Kd5 a5 45. Kc6 Bc5.] us to explain one of the pawn end-
43. Kd6 b6 ings treated therein.
By the rules for these endings,
Black is lost since White can obtain
a passed pawn on both wings. 7
44. b4 Kf8
45. b5 Ke8
46. Kc7 Ke7?
[Ed. : According to K.M., this
move is the turning point as, 46. . .
Bg5 47. Kb7 Bd2 48. Kxa7 Ba5 still
draws.]
47. Kb7 Kd6 White to play/Result?
48. Kxa7 Kc5
49. a4 Kb4 That's the problem set by Alex­
White threatened a5. The rest ander, who prefaces it with the re­
was: mark that endings with two pawns
50. Kxb6 Kxa4 against one are usually harder than
51. Kc6 Kb4 those with more pawns-very true.
52. b6 Kc3 First reaction: supported passed
53. Bf5 d3 pawn-why not a win? Then either
54. b7 d2 after solving it or failing to solve it,
55. Bg4 Resigns. you look up the solution and find it
White turned to good account his given as a draw.
first class knowledge of Bishops-of­ This is the book solution:
opposite-color endings; he knew what 1. Ke3 Ke5
kind of mistakes he wanted Black to 2. Kd3 Kd5
make-and played so as to give Black 3. Kc3! Ke6
a chance to make them. 4. Kd4 Kd6
5. Ke4 Ke6
6. Kf4 Kf7
7. Kf5 Kg7

- 18 5 -
 CJS Purdy On The Endgame

8. g6 ample, 7.... Kg7 8. Kf5 Kf7 9. g6t, etc.

Nothing better. Or 7... Ke7 8. g6.
8. . .
. Kh6! After 7. Kf5 Kg7, the book says
Drawn. White has "nothing better" than 8.
Now, a most valuable exercise, see g6. That, again, is wrong. With a little
if you can find the flaw. The solution shuffling, White can "change the
follows. move" and still win For example, 8.
.

Ke5 Kf7 9. Kd6 Kg7 10. Ke7 Kg611.

Solution Ke6 Kg7 12. Kf5, etc. Indeed, with
the pawns as diagrammed, White can
"How quaint the ways of win no matter where the Kings are.
paradox." But in Diagram 8, if you move all
- Pirates of Penzance the chessmen one rank or two ranks
t oward the eighth rank, then it re­
Most errors in analysis arise ally is a draw!
through taking something for grant­ Here we have two paradoxes:
ed. In chess you must always be tak­ a) that taking the opposition on
ing something for granted, or you move 7 is bad for White, and
could hardly ever come to a decision b) that if White is nearer queen­
at all-hence the inevitability of er­ ing he is worse off.
ror. On move 7, Alexander has as­ Both paradoxes arise because of
sumed as a matter of course that the the law of stalemate, which is itself
white King will take the opposition. paradoxical.
We are all taught that the opposition These draws nearly always in­
is the great desideratum in pawn volve the a- or h-pawn. There is one
endings. The trouble is that every exception. In Diagram 8, move all the
generalization in chess breaks chessmen two ranks toward the
down, sometimes. Here White can eighth rank and one file to the left­
win by deliberately giving Black the still a draw!
opposition, temporarily. U you take all these statements
for granted, they are well-nigh use­
less-they should all be verified,
8 which is quite easy, and you will then
really benefit.

Attack Plus Three Pawns, No Win!

The title proves that I am going to
talk about Bishops of opposite color.
It is quite true that one can some­
After6 Kf7
.•.
times win these endings with no
pawns up and that one may fail to
Thus, 7. Ke5! This enables White win with three pawns up.
to force the win of the black pawn The paradox is elucidated if we
and to win easily thereafter. For ex- remember that the prime desider-

-186-
 All Facets Of The Endgame: Sure Draws!

atum is the possibility of creating

chances on both wings. Two (or, as I
shall show, even three) united passed
pawns will sometimes not win; but
one passed pawn on each wing (or
the possibility of creating such a situ­
ation) will always win, even though
the enemy has a passed pawn him­
self-sometimes even if he has two.
I was recently looking at an end­
ing that could have happened in the Drawn position!
telegraphic game Shaw-Schoenfeld.
Black is three pawns up, and
White's pieces are paralyzed; yet
9 Black cannot win. Just try! You can
easily win the Bishop, but only by
allowing stalemate.
In Diagram 9, Black could play
more in accord with principles by
1. ... Bf1!
2. h4 Bh3
Black thus makes sure of keeping
Black to play at least one pawn on the K-side. E.g.,
3. g 5 hxg5
From Diagram 9, Black can, it is 4. hxg5 Bxf5
true, win by rushing up with his Suppose Black still allowed all
queenside pawns. For example, 1... his queenside pawns to play to light
b5l 2. Bd6 (necessary) a2 3. Be5 Bb3. squares (or he was forced to do so)
However, by careless play Black and he reached the following:
might get into a position like that of
Diagram 10 where White has man­
aged to exchange off the kingside
pawns and to inveigle the passed
pawns onto light squares.
[It is assumed, though rwt directly
stated, that Black can win only
with t he support of his King.
Purdy's point here is that, in th'is
case, Black's position is superior
enough to offer Black a choice be­ Black wins easily
tween starting with a pawn ad­
vance or taking care of the king­ This is a win, provided Black does­
side first. -F.PH./J.SP.] n't play ...g6? Black must leave White
a spare move. The black King plays
down to d1. To produce the stalemate

- 18 7-
 OS Purdy On The Endgame

motif, White then has to play g6. Then died a dog's death when it might have
the black King just strolls back and flickered up into a classic Rook vs.
grabs the g-pawn. Bishop ending after
(The defender's pawns are often '72.Rg8! Be5
best placed on the same colored 73. g4t hxg4t
squares as their Bishop; the attack­ 74. Rxg4
er's pawns very much the contrary.) Now if Black could play Ke5, he
...

In Diag. 9, note that after 1... Bf1 would draw easily-but his Bishop
2. h4 Bh3 3. g5 Black might win also is on e5. That's why 72. Rg8 could be
by 3 ... h5, as White's h-pawn would said to place Black in Zugzwang. If
be easily, if slowly, won-thus mak­ Black didn't play 72... Be5, he had to
ing Black's h-pawn a passed pawn. allow Rh8, . .Kg5, Ke4, with gain of
.

Actually, in Diagram 9, 1... b5! ground by White. As it is, Black can­

wins more elegantly, as White can not draw easily-but we believe he
still be prevented from exchanging can draw. Simplest is

all the kingside pawns. But care is 74. .. . Bd6

needed; and for endings with Bish­ Then of course comes
ops of opposite color in general, the 75. Re4
student is well advised to make it his cutting offtheKing-see Diagram 13.
first care to create possibilities of We digress here to give a cardi­
play on both wings rather than to put nal rule for the best u se of Rooks in
all his money on one. endgames:
A Bishop alone can never cut off a
King-a Rook can, and is most effi­
The Might-Have-Been cient when used for that purpose.
[This article, and the next, appear
to contain genuine original con­ RULE48
-¢-
use the Roo; to
tributions to EG theory. -F.PH.J
cu t off.
Purdy

Klass
After 71 ... B(e5)d6

Klass blundered (72. Re4?), allow­ After 75. Re4

ing Purdy to draw easily. So the game Is this a draw?

- 188 -
 All Facets Of The Endgame: Sure Draws!

This type of Rook vs. Bishop end­ notoriously drawish ending of lone
ing (pawn vs. pawn on adjacent files, King versus lone King.
other than a- and b- or g- and h-files) Can Black do as required? Yes, if
is not covered by Fine in the stan­ he can avoid having his King driven
dard work Basic Chess Endings farther away from the white pawn
(BCE), nor anywhere else to our than it is now. For Black merely
knowledge. Yet it is undoubtedly the needs to wait till White's King reach­
best possible formation for the de­ es c6, with the Rook threatening
fen se--we'll soon see why. Rxc7, and then to play his King to
Here's what Fine says of the end­ e5 or e4. If White's Rook has contin­
ing Rook and pawn vs. Bishop and ued to hold the e-file, Black, on the
pawn in general (BCE, p. 478): previous move, will need to have
played ... Be5-to use his Bishop as
"Even with only one pawn a shelter.
apiece, a win may be se­ In only one way can White hope
cured by skilHul manipula ­ to keep Black from playing as re­
t ion of the Rook .. . Where quired: by checking Black's King
there is n o passed pawn, AFTER getting his own King to d4
the game is won if the where it prevents ... Be5 in reply to a
pawns are in the center, but check. For that, White must at some
may be drawn with the stage play Kd4 without being checked
pawns on the a-, b -, g-, or h­ in reply AND be able to check with
files . Diagram 13 is the the Rook next move. It seems to us
model case in the center." that that is impossible against a
strong defense. If we are right, Fine
Yes, but for this "model" case, will have to revise the next edition of
Fine gives a position where the BCE slightly. Here we give some il­
pawns are on the same file. The ad­ lustrative play from the last diagram.
vantage to the attacker is that he can It is immaterial whose turn it is to
move up his pawn until the pawns move, but it would have been Black's
are on adjacent squares. You have move had the position arisen. So:
that if, in our last diagram, you put 1. Bc5
\Vhite's pawn on c6 instead of on d5. 2. Re8 Bd6
Then, certainly, White wins. All he 3. Ke3 Bg'J!
has to do is to attack Black's pawn 4. Kd3
with Rook and King and then play A mulberry bush variant is: 4. Re7
Rxc7, ... Bxc7, Kxc7. That must win Kf6 5. Rh7 Kf5 6. Rh5t Kg4 7. Rh1
because on c7 White's King protects Kf5-futile for White.
his own pawn. But that scenario can­ 4. ... Bh4!
n ot happen here. If, as soon as White Still ready to meet Kd4 with a
is ready to play Rxc7, Black's King check, either at f2 or f6.
can move to e5, e4, d4, c4, or c5, then, 5. Rf8t Ke5
as soon as White wins the black 6. Kc4 Ke4
pa"''Il, Black can win the white pawn. White is no further forward. If he
That will leave the interesting but abandons his hopes of driving the

-189-
 OS Purdy On The Endgame

black King out of play and just goes M.D. Bronn

for the pawn, we have, from here:
7. Re8t Kf5
8. Kc5 Bg3
9. Kc6 Be5
This is it.
10. Re7
And now the key move.
10. ... Ke4
Drawing.

Black needs to maintain the for­ John Purdy (Black to seal)

ward station of his King. A forward
station is useless to him in the stock The assembled talent convinced
position (pawns on the same file and Purdy that after 1... Kd6 (the as­
on adjacent squares) because a rear sumed sealed move-and the as­
attack on the enemy pawn doesn't sumption was correct) 2. Kd3 Kd5 3.
help, but here it is vital. A post­ Kc3 f5 4. Kd3, Black wins by ... e4t.
mortem showed that Black should That is perfectly true. White then
lose if his King retreats to e7, d7, etc. resigned!
In that postmortem, the line discussed What is the false assumption in
in this article was not properly tested. the analysis, and what is White's
Perhaps someone will test it properly drawing line? The solution is below.
now, thus contributing something Tarrasch once resigned a per­
definite to endgame theory. fectly tenable game against Noa. Ap­
parently Tarrasch was disgusted
with his own play and did not like to
Resign? be seen poring over the board, strug­
Highly instructive-as a horrible gling to find a way out against so in­
warning against ever taking quick ferior an opponent-so he noncha­
analysis seriously under any circum­ lantly resigned. Najdorf once re­
stances whatever-was the pawn­ signed a probably winning game. And
ending reached by John Purdy and many other masters have similarly
Broun, in which Purdy resigned a enrolled themselves among Folly's
drawn game. After Broun had sealed, minions.
it appears that some analysis took
place in the tournament room-un­ Solution
noticed by the Director of Play. Ap­ The absurdity is the assumption
parently the culprits did not know that White would play 2. Kd3. Far
they were behaving illegally; Broun more natural to defer that to the third
took no part. Young Purdy's object move, when the King can go to d3
was to see if his game was lost-and with the opposition. However, be­
if so, to resign without delay. sides being natural, 2. Ke3 is also
the best. The real point in this posi­
tion is not "the opposition"; it is for

-190-
 All Facets Of The Endgame: Sure Draws!

White to ensure that when Black rarely leads to correct conclusions,

plays ...e4 it will not be check, so that even by an all-master cast. Secondly,
\Vhite can push past with f4-thereby there is the published analysis of
closing up the center. It just happens some writer who goes into the ques­
that the "opposition" idea achieves tion in a more favorable atmosphere,
the desired result. that frequently finishes it.
After 1... Kd6, we have Then there is a third stage-when
1. Kd6 somebody, for his own private amuse­
2. Ke3! Kd5 ment, checks up on the analyst and
3. Kd3 f5 sometimes makes a fresh discovery.
4. Kc3 e4 Now we have to thank W. Broome
5. f4 for debunking some of our own analy­
As the spare pawn moves balance, sis on a certain pawn ending-see
the black King must be the first King page 190. In that ending, the Sydney
to move. Say junior John Purdy was so ill-advised
5. ... Ke6 as to resign.
Then not Kd4?? (see why?), but
6. Kc2 (see Diagram 14)
Other squares mig'ht be equally
good, except d4. White must never The position is actually a draw.
move to c3 if Black can answer with We "demonstrated" the draw in the
...Kd5. previous article. Broome proves that
There are several variations, the position is indeed a draw, but that
some of them quite tricky, but always our demonstration requires a trans­
\\1lite can draw. The doubled !-pawns position of moves.
are defensively an asset! If White had After
a pawn on g2 instead of on f2, then 1. Kd6
probably he would be lost, as ...e4 2. Ke3! Kd5
would create a passed pawn. we gave
[This analysis looks convincing, 3. Kd3 f5
but things are less simple than they 4. Kc3 e4
seem. The student would benefit by 5. f4
looking for a significant flaw be­ And White draws. "And White
fore referring to the next article, draws" is rig'ht, but Broome points
which is a sequel to this one.] out that White should transpose
moves: thus,
4. f4! e4t
The Critic On The Hearth 5. Kc3
\\Te are always trying to impress The trouble with 4. Kc3 is that
on students the importance of devel­ then 4... f4! wins. (How? The solu­
oping the critical attitude. tion appears below.) The immediate
In chess there are roughly three 4. f4! is the precise move. If there­
levels of analysis. First the post­ upon 4... exf4, then 5. f3 holds the
mortem at the conclusion of a game. center and draws.
If the problems are difficult, this Broome also points out that 2.

- 191 -
 OS Purdy On The Endgame

Kd3?, which White was made to play effect. Or if White saves the opposi­
in the postmortem, loses-but not the tion with h5, then his h-pawn be­
way we said it did. After comes the fatal weakness.
1. Kd6 The reason that 10. h4 is fatal is
2. Kd3? Kd5 that it leaves Black with the last
3. Kc3 f5 spare pawn move. Were Black's
4. Kd3 f4! h-pawn/h5, 10. h4 would draw easily.
as before-and not 4... e4t? The post­ It is true that so long as Black leaves
mortemists thought that 4 ... e4t his pawn on h6 White's h-pawn still
would win, and we failed to see that has another move (h5); but as soon
they were wrong. The draw after 4... as White uses that move, the h-pawn
e4t is left as an exercise; the solu­ itself becomes fatally weak.
tion appears below. Notice that White, as his best
chance, answered 5... h6 with 6. h3.
In what follows, notice that Black
Solutions similarly answers h3 with ... h6, as
After his best chance to win. That is a gen­
1. Kd6 eral rule about symmetrically op­
2. Ke3! Kd5 posed pawns in any fight for "the op­
3. Kd3 f5 position" in pawn endings. Let us put
4. Kc3? f4 it thus:
Broome's main line is
5. Kd3 h6! Don't initiate; imitate instead.
6. h3 Kd6!
7. K-moves Precisely NOT the way to play
(anywhere but e4, which would clear­ chess in general.
ly be fatal) And now for the draw after
7. ... Ke6 1. Kd6
8. K -moves 2. Kd3? Kd5
(but not to d3 because of ...Kd5) 3. Kc3 f5
8. ... Kf6 4. Kd3 e4t?
9. Kd3 Broome's solution is:
(else ... Kg5 wins by a sequence not 5. fxe4t fxe4t
hard to calculate) 6. Ke3 Ke5
9. ... Kf5 6... Kc4 only draws.
Black has now obtained diagonal 7. h3! h6
opposition-here, as often, a step to­ 8. f3 exf3
ward direct opposition on d5. If now 9. Kxf3 Kf5
10. Kc3, then not 10 .. . e4? (11. Kd4) This takes the opposition, but then
but 10 ... Kg5, winning (as already 10. Ke3; and Black cannot win. In
mentioned). So, this, if 9. Kd4, then 10. Kg4; and
..

10. h4 Ke6! White draws with h-pawn against

And Black will get the opposition Queen. Or if 9 ... h5, then
on d5. Then Black either wins the 10. Kg3 Kf5
b-pawn or plays ...e4 with winning 11. Kh4 Kg6

- 192 -
 All Facets Of The Endgame: Sure Draws!

12. Kg3 Kg5 Nimzovich undoubtedly had a reason

13. h4t for calling it an element of strategy,
And White draws. but how good was the reason?
The well-timed 7. h3! was previ­ A position with a discovered check
ously missed. in it is usually worth playing for, even
though you are unable to calculate
just what gain it is likely to bring. If
A King Escapes one could make that a general rule,
I don't suppose anything is more Nimzovich would be completely jus­
important in chess than being able tified. Unfortunately, we have all
to work out mates, but it is also of many times come across positions in
some importance to be able to judge which none of the possible discov­
when there is likely to be a mate and ered checks amounted to anything
when there isn't. at all . Therefore, "usually" is the best
one can say. So discovered check is
J.V. Kellner certainly an element of chess, but it
is hardly an element of chess strat­
egy in the sense that "the center"
and "the pawn chain" clearly are­
these are included by Nimzovich in
the same list as discovered check.
Discovered check is frequently
used for a perpetual check, and that
is the use that Stannard made of it
here:
40. Ng6 t Kh7
V.A. Stannard 41. Qc7t Bg7
After 39 B(g7)f6 (forced)
..• And White gives perpetual check
with the Knight.
Diagram 15 shows a position Ought White to expect to have a
reached in the Stannard-Kellner win here? The answer is, not with­
game. After a few moves, the players out some obviously fatal situation of
agreed to a draw; and then there was the black Queen in which she could
a prolonged postmortem, many find­ be forked by the Knight. Otherwise,
ing it hard to believe that there was seeing that we have three attack­
no win-because of the obvious pos­ ers against three defenders, we
sibilities of discovered check. ought not to expect a win It is just a
.

Students of Nimzovich (My Sys­ question of trying out all the possible
tem) will remember that he places discovered checks. If none accom­
"discovered check" in his somewhat plishes anything immediate, there is
arbitrary list of "the elements of not likely to be any win As a matter
.

chess strategy." If the terms strat­ of fact, White can win a pawn; but
egy and tactics are to have much since he is already a pawn down and
meaning, one would rather expect since winning the pawn would end
discovered check to belong to tactics. his attack, that doesn't help. Here is

-193-
 CJS Purdy On The Endgame

how the pawn is won: 40. Ng6t Kh7 obviously has the interior game be­
(not 40... Kg7??, losing the Queen in cause of his weak pawns, he ought
four moves) 4 1. Ne5t Kg7 42. Qa7t to pin the Bishop and make sure of
Be7 (moving the King leads to mate the draw.
or loss of the Queen) 43. Qxd4. Then 30. . .. Bc5!
43... Qf6, and the fun is over. 31. bxc5 Rd3t
Observe that in the play just given 32. Ke4? ?
Black had a couple of good chances An incomprehensible move, as it
of losing by permitting a little combi­ gives Black a remote passed pawn.
nation. So once again bear in mind White also acquires one; but it will
that a winning combination can be­ be relatively useless, as his Rook will
come possible at any time through be in the notoriously worst position­
an opponent's error; never attempt in front of the pawn.
to play by principle alone, dispens­ 32. Rxb3
ing with calculation. 33. c6 bxc6
34. Rxc6t Kg7
35. Ra6 Rxh3
Endgame Suicide 36. Rxa7 Ra3
With Bishops of opposite color and 37. a5
equal pawns, you would hardly credit White is helpless.
that White could lose from the dia­ 37. h5
grammed position; but John Purdy, 38. f5 g5
ably abetted by Campomanes, found 39. Ra8 h4
a way. 40. f3 h3
41. f6t Kh7!
Campomanes 42. Rb8 Rxa5
43. Rb1 Kg6
44. Ke3
If 44. Rh1, then 44 ... Ra4t, etc.
44. Ra2
45. Ke4 h2
46. Ke3 Rg2
47. Rh1 Kf5!
48. Resigns.
An instructive lesson in the weak­
ness of Rook-in-front-of-passed­
John Purdy pawn. Even so, White would have
Afte r 29K(g7)f6
... drawn had he retained his O"\\-'Il
h-pawn, giving up the useless pawn
30. Kf3? on f4 instead.
This is not very bad; but as White

-194-
 [hapler 17
fil\ll The \Vil\

Onlookers See rights and taken the position home

Least Of The Game to analyze, she would have found that

I
n the 1937 tourney for the British her sealed move gave a forced win in
women's championship, Mrs. F.F. a few moves. We leave it as a simple
Thomson had the annoying experi­ exercise for our readers.
ence of being defrauded of half a
point by interfering onlookers.
Winning A Drawn Endgame
Black: Mrs. Thomson In a New South Wales champion­
ship tourney, Harry Klass gave a
good display of endgame technique
from the diagram.med adjourned po­
sition. Klass's opponent lost because
he thought the draw was automatic.

A.E. Ramsey

White: Miss Dew

Mrs. Thomson was Black in the

diagrammed position against Miss
Rowena Drew (previous titleholder),
and she sealed 1... Qf4t. The onlook­
ers manifested displeasure, says the
B.C.M., at her dragging Miss Dew H. Klass
back to finish "an obviously drawn White to move
game;" and Mrs. Thomson agreed to
a draw. White sealed his 45th.
Had Mrs. Thomson exercised her 45. Kf1 Kf7
 OS Purdy On The Endgame

46. Ke2 f5 61. Bh1 Kf6

This weakening pawn advance is 62. d 5! cxd5
virtually forced, to free the King. On 63. Kd4 Nf7
the queenside, Black's dark squares Counterattack by 63... Ne4 is no
are also weak; but since White's better: 63... Ne4 64. Kxd5 Nf2 65. Bf3
Bishop moves on light squares, the Nd3 66. Kd4 Nxb4 67. Bxb7, and
disadvantage is slight. White wins.
4 7. Bf3 Kf6 64. Bxd5 Nd8
48. Kd3 Ke6 65. Be4?
49. Nc4 Nf7 Missing a forced win by 65. b5! If
50. Ne3 Nxe3? thereupon 65... axb5, then 66. Bxb7!;
Impatient to simplify and hasten and the a-pawn must queen.
the anticipated draw, Black makes 65. Ke6
the very move that gives him losing 66. Kc5 Kd7
chances. This wish to "simplify" has 67. Kb6 Kc8
lost countless games. The exchange 68. Bg2 Kb8
unisolates White's d-pawn, with se- There were some repetitions,
rious consequences. Black was quite which we omit. How is White to win?
safe either by leaving the d-Knight Klass discovered the way during the
on his blockade square or by with- second adjournment-it is quite pi­
drawing it and then maneuvering his quant.
other Knight, always watching for d5. 69. Bd5 Kc8
51. fxe3 Nd6 70. Kc5! Kc7
52. b4! 71. Kd4! Nc6t?
When is a pawn advance merely Black's only chance was 7 1... b6.
weakening, and when is it justified The rest is clear.
for the purpose of fixing or weaken- 72. Bxc6 Kxc6
ing enemy pawns? The point is, can 73. Ke5 Kb5
the opponent take advantage of the 74. Kf5 Kxb4
"weakness" created? Here not, Black 75. Kx g5 Kxa5
is immobilized by White's threat to 76. Kf4! b5
enter with the King if the Knight goes 77. g5 b4
away. 78. K e3
52. Nb5 And White wins because after
53. a4 Nd6 78... Ka4 79. Kd2 Ka3 80. g1} b3 81. g7
54. h4 g5 b2 82. Kc2 Ka2, White queens with
55. hxg5 hxg5 check.
56. g4 Kf6
57. a5! Ke6
58. e4 fxe4t City Of Sydney Tourney
To permit e5 would let White's A playoff match of four games for
King enter. the City of Sydney championship be­
59. Bxe4 Kf6 tween H. Klass and B.Y. Mills was
The pawn ending would be lost. narrowly won by Klass, who thus re­
60. Bf3 Ke6 tained his title won the previous year.

-196-
 All Facets Of The Endgame: Find the Win

After winning the first game, Klass Premature Resignation

drew the next three. Hardy resigned to McNabb in a
In the second game, Klass offered position which would at least have
a draw in the diagrammed position, given McNabb some very interesting
it was promptly snapped up. Klass adjournment analysis to win. The po­
afterwards pointed out that the offer sition is chosen for our next study, it
was highly ill-advised. will be found very entertaining. In-
deed, as pointed out by Kelling, it is
Klass to all intents and purposes the same
as a composed study by the Rev. Gil­
bert Dobbs (USA); and the solution
3 runs exactly the same after a couple
of preliminary moves. The whole idea
is based on the well-known draw with
King, Knight, and a- or h-pawn
against lone King. (The Horwitz win.)

4
Mills
White to move Black to move/Draw

White must obviously play either In the small diagram above, Black
1. Kd2 or 1. Rb7t and 2. Rc7 to stop draws by ... Kf8, because the Knight
the pawn from queening. The former is on a dark square-if the Knight
line gives Black a quick win by 1... were on a light square, the first
e4, threatening ...e3t. In the latter move would be . Kf7. If we imagine
. .

case, we have the following variation: that the white King has just cap­
1. Rb7t Kf6 tured a Bishop in the corner, we see
2. Rc7 b4 that Diagram 5 is based directly on
3. Kd2 e4 this position.
4. Rxc2 Nxc2
5. Kxc2 e3
And Black has a book win. Klass
had forgotten that in such a position
the two isolated pawns (two ranks
apart and across the middleline) are
not only safe from capture but actu­
ally win, unaided, against the oppos­
ing King.
By missing this win Klass nearly
,

let himself in for a drawn match, as White to play and win

Mills had a good chance to win in
the fourth game. Evidently, the white King must
take the Bishop when the black King

- 19 7 -
 CJS Purdy On The Endgame

is not on e8 or e7. And don't forget the Knight on f5. To do that will be a
that in certain positions the Bishop test of whether the idea has been
can leave the corner-the black King properly grasped.
need not always move. Don't fail to
try Diagram 5.
7

Did You Try Diagram 5? Black to move/Loss

The previous position occurred in
a New Zealand championship tour­ In the preceding article, I pointed
ney-see page 197. out that Diagram 4 is a draw with
If you did not try Diagram 5, do it White to move. Or rather, I pointed
now before reading this article. out that after 1. Kxh8 it is a draw
Very few positions in chess can with Black to move-Black plays 1...
be solved by absolute demonstration; Kf8, because the Knight is on a dark
but when such positions do occur, it square. It all comes to the same thing,
is very hard, and sometimes impos­ for if White moves his Knight about
sible, to solve them in any other way. instead of taking the Bishop, then
The positions in question are almost Black simply oscillates with his King
invariably blocked pawn positions, or moves his Bishop out when that is
with no pieces other than the Kings. safe. You canno t gain or lose a
In this game, however, each side has move with a Knight.
a Minor Piece as well. Here is the But Diagram 7 is a win for White
diagram again. with Black to move. For the black
King must at once lose control of f8,
J.L. Hardy and then Kxh8 wins for White: if the
black King has to go to f7 (a light
square), he can be forced away by
6 the Knight.
In Diagram 7, imagine the Knight
on h5 or on some other square com­
manding g7. Then White, to move,
wins by Ng7 (obviously).
Observe also that in Diagram 7 the
black King and the Bishop are on the
prongs of a phantom Knight fork. The
ff. McNabb Knight would win the Bishop and the
White to play and win game if he could check from g6. Now
we have some KEY SQUARES in
NOTE: In Diagram 5, the Knight is mind, and that is the secret of all ret­
actually on f5; but for this article, I rograde analysis-the method I
accidentally had it on e6, as here. spoke of. The key squares so far are
That is a blessing in disguise. After e7 (a danger square), g7, and g6. The
studying this article, the student can main reason g7 is a key square is
easily apply the method to win with that whenever the Bishop moves out

-198-
 All Facets Of The Endgame: Find the Win

of the corner White can never win

without moving the Knight to g1 or
to f6 to obstruct the Bishop.
Now we want a link between g1
and g6. The Knight can command g1
from h5, f5, e6, or e8, but of these the
only square that will be safe for cer­ White to move
tain from the black King is h5. That
at once suggests f4 as a new key The result would be the same if
square, for from there the Knight Black had played 1... Kd8 or 1... Kd6,
commands both g6 and h5. So the for 2. Nh5 would then have left the
link we want is f4. black King no choice but to return to
d7: if 2... Kd5 or 2 . Ke5, then 3. Kg8
..

Ke6 4. Nf4t forces Diagram 8.

Now 3. Kg8? gives only a draw (3.. .

Ke7!), and further Knight maneuvers

merely go round the mulberry bush.
But in all positions where progres­
sive moves are impossible, one
White to move wins should ask, "How would this go with
(The black King Is anywhere) the other side to move?" Thus sup­
posing White to miss his turn on
Diagram 8 shows the Knight on move 3, Black would clearly lose. For
f4. With the move, White wins no mat­ if 3. . B-moves, then 4. Ng? wins. And
.

ter where the black King is. If the if 3... Kd8, then 4. Nf4! Kd7 (the Bish-
King is on e7, then Ng6t forks. If op cannot budge because of Ne6t­
K/e8, then 1. Nh5 Ke7 (forced) 2. Ng? g1) 5. Kg8; and the black King must
produces Diagram 4. If K/f6, then move into Diagram 8-for if the Bish­
Kxh8 wins because the black King is op moves, then 6. Nh5 B returns 7.
on the Knight's color. Kxh8, followed by King out.
(NOTE: With Black to move, Dia­ So in Diagram 9, White needs a
gram 8 would be a draw if the black move that does nothing.
King stood on e8-Black would sim­ 3. Kf8! !
ply move out his Bishop.) Zugzwang! If 3.. Kd8, then White
.

Now return to the original dia­ wins by 4. Nf4 as above. If 3 ... Ke6,
gram. Naturally, we try then 4. Kg8 Ke7 5. Ng? gives Diagram
1. Nf4 Bat 7. If 3. . Kd6, then 4. Kg8 Ke6 5. Nf4t
.

2. Nh5 Kf6 (forced) 6. Kxh8; and White wins

Once the Bishop is out, White ob­ because the black King is on the
viously can do no good until his Knight's color.
Knight can again command the dark It is remarkable that in an actual
diagonal. game there should occur an endgame
2. . .
. Bh8 study more attractive and artistic
Forced. Now we have Diagram 9. than 90% of those that are actually
composed and published. To com-

- 199 -
 OS Purdy On The Endgame

pose a difficult study with so few

pieces is every composer's ideal.
The joke is that, in the original
position, Hardy resigned!

-200-
 All Facets Of The Endgame: Interesting Endgames

[l\apler Ii
ll\lereslil\q El\llqal\\es

The Gol dstein-Crowl Ending Now we have a book draw, Rook

T
he intensely interesting position and a-pawn against Bishop not of the
diagrammed arose in a Mel­ color of the queening square. Here
bourne Annual Tourney. The colors Black can play ...Ka6; but even if it
were reversed. were White's move and the black
King were at a8, White could do noth­
M.E. Goldstein ing better than to play into a posi­
tion like that diagrammed below.

1 A Book Draw

F.A. Crowl
White to play

Crowl, having tried everything he Either to play/Draw

could think of, tried
77. Rc5!? Ka7 In the diagram, if 1. a6?, then 1...
The Rook cannot be taken. Ba7 (see below) 2. Rf8t Bb8; and now
78. Kc7 Be3 if 3. Kb6, then it is stalemate. Alter­
79. b5 axb5 natively, if 2. Kc6, then 2 ... Be3, etc.
Had Black taken the Rook instead It is better to keep a6 for the King.
of the pawn, he would have been lost For example, 1. Ka6 Kb8! (now that
in a few moves! the white King blocks its own pawn,
80. Rc6 b4 the black King can and must vacate
81. Rb6 b3 the corner-we now have a stock
82. Rxb3 ending of Berger's) 2. Rb7t! Kc8! (2...

- 20 1 -
 CJS Purdy On The Endgame

Ka8 loses) 3. Rb6 Kc7! 4. Ka7 Bf2, the solution-which readslike a com­
and Black draws. posed study. Careful search failed to
reveal any alternative; and the edi­
tor then sent his line down to Watson,
Note Added asking for his okay before publish­
Look at the "A book draw" Dia- ing. Watson wired back, "Solution
gram 2. After 1. a6, the draw is not correct," so we have every reason to
obtained by 1... Ba7?, but by 1... Bg1 believe that there is one, and only
or by 1... Bd4 since the square a7 one, solution. The longest variation
must be kept for the black King if is just over a dozen moves.
checked. For example, 1. a6 Bg1! 2. First note that with the Rook and
Kc6 Be3, etc. The draw is easy. the Bishop off, White easily wins the
The move 1... Ba7 would be cor- black pawn. So Black can never al­
rect if there were no white pawn, but low the exchange of Rook for Bishop
it loses in the actual position-as in the corner. Secondly, White's
pointed out by C.G. Watson: 1. a6 Ba7 whole idea is to get his King to b6 (to
2. Rf8t Bb8 3. Ka5! Ka7 4. Rf7t Ka8 win the pawn); therefore White must
5. Kb6 B-moves 6. a7, etc. try to drive the Bishop off the key
diagonal.
1. Rf3! Bg1
Watson Study Not 1... Bd2 because of 2. Rf7t.
etc. If 1... Bd4, then 2. Kd5 Bg1 3.
No. 120: C.G. Watson Rf7t forces the text line.
Adapted from Crowl-Goldstein 2. Kd5! Kb7
Melbourne 1933 If 2 ... Kb8, then 3. Rf6 forces the
text line. If 2... Bh2, then 3. Rf1!.
3. Rf7t Kb8
3 If 3... Ka8 or 3... Kc8, the reply is
the same; but (after 3... Kc8 4. Ke4)
if 4... Bh2 (threatening ... Bd6), then
White can win only by Ra7-but
easily.
4. Ke4!!
The beauty of the solution is that
the white King, whose only desire is
White to play and win to get to b6, goes right back here to
get there!
C.G. Watson discovered that 4. ... Kc8
Crowl had plenty of chances (even Trickier is 4.. . Bh2, but White
as late as move 77) of playing into a reaches the text by 5. Rf1! Kb7 6.
position like that of Endgame 120, in Kd5 Kc7 7. Rf6, etc.
which Watson saw that there was a 5. Rf1! Bh2
beautiful win He immediately wrote
. Not ... Ba7? because White can
to the editor, asking for his ''verdict;" then force the exchange of Rook for
and the editor managed to discover Bishop in the corner.

-202-
 All Facets Of The Endgame: Interesting Endgames

6. Kd5 Kc7! [This move should have lost the

7. Rf6 game. White would have done well
Not Kc5 because of ... Bd6t. to follow Purdy's advice, already
7. ... Kb7 published in a 1931 issue o f 'Lius­
8. Rb6t Ka7 tralasian Chess Review," o f con­
And now the second pretty phase sciously looking for com binations
of the solution, the "love-chase" at every move. Here Black's threat
theme. is not too hard to see if one is look­
9. Rg6! Bf4 ing for threats. Having seen it,
10. Re6! Bh2 White might well play 46. Nel, en­
If 10... Bg3, then 11. Rf6 Bh2 12. abling his King to stop the h-pawn
Rf1, etc. If 10... Bd2 or if 10... Kb7, if necessary-with an early draw
then 11. Kc5. likely. -F.PH./J.S.P.]
11. Re1! Bg3 46. ... Bxg4!!
12. Rf1 47. Bxg4
Back again! 47. hxg4 loses [
12. ... Bh4 47. Nxg4
13. Kc6 48. Kf3 Ne5t
And White wins. 49. Ke4 g4!
50. Nf4
Not 50. N:xh4 because of 50... Kg5!
New Zealand Championship 50. ... g3
1933/34 The logical plan was to force an
entry for the black King; therefore
First Endgame 50... Nf7! 51. hxg4 Kg5! , and Black
This remarkable endgame oc­ wins.
curred in the fifth round of the New [The text move, creating a sup­
Zealand championship tournament. ported passed pawn, probably also
wins if followed up correctly, see
R.I. Jones the note to move 53. -F.PH./J.S.P.J
51. Ng-2 Ng6
52. Ne3 Kg5
53. Ng-2 b5
When your opponent cannot do
anything, get your own pieces as well
posted as possible before breaking
up. Here ...Ne7-f5-d4 first, as Dids­
bury points out.
54. cxb5 axb5
55. a4 bxa4
R.G. Didsbury /Ed.: 55. . . c4! is brutal and wins
White to play his 46th move quickly.-K.M.]
56. bxa4 Ne7
46. Ke4? 57. a5 Nc8
A natural oversight! 58. a6 Na7

- 203 -
 OS Purdy On The Endgame

59. Ne3 Nb5

60. Ng2
A second pawn goes; but White, 5
rightly, has confidence in his a-pawn.
[Safer appears 6 0. Kf3; and if
60..Kf6, then 61. Nc4 Ke7 62. Ne3
Kd7 63. Nf5-and White will rwt
lose. -F.PH./J.S.P.J
60. Nc3t
61. Kd3 Nxd5
62. a7 Nb6 Again Black shies. We think the
63. Ke4 d5t position is a win. The reader should
64. Ke5 d4 try to knock out our analysis, which
65. Ke4 Na8 goes as follows:
Black is two pawns up, but para­ 71. ... Kf5!
lyzed! 72. Kd3!
66. Kd3 Nc7 If 72. Nxh4t? then 72 ... Ke4 leads
67. Kc4? into the variation shown in the next
Too risky, as it gives an entry to note.
Black's King. White wished to avoid 72. Ke5
losing his passed pawn by 67. Ke4! 73. Nxh4 Kf4
c4 68. Kxd4 Nb5t. But White thereby 74. Ng2t Kf3
g-ets two pawns, and the resultant po­ 75. Nh4t Kt?
sition is drawn since the black King 76. Ke4
is kept out. If 76. Kc4?, then 76... Ke2 77. Ng2
67. ... Na8 d3! 78. Nf4t Ke3; and Black wins.
Why not try ...Kf5, the only chance 76. . .
. Ke2
to win? Black could easily work out 77. Ng2
that he could always draw if need be. If 77. Ng6, then 77.. Kd2 78. h4 g2
.

68. Kd3 Nb6 79. Ne5 d3 80. h5 Ke2; and Black wins.
69. Ke4 Na8 77. ... Kd2!
70. Kd3 Now if both sides queen, Black
Legal point: after this move, Black wins the white Knight by . ..Qc6t.
could claim a draw-but White could Black is then a piece up, with his King
not, it being Black's move. and Queen well placed; and he wins
[The law has since been changed in all the variations that we can see.
so that rww White could claim a And if 78. Nf4 or if 78. Kf3, then Black
draw before playing his 70th by wins by 78... c4.
announcing his intention of mov­ [By "both sides queen, " the follow­
ing 70. Kd3. -F.PH./J.S.P.] ing sequence appears to be in­
70. ... Nc7 tended: 78. h4 c4 79. h5 c3 80. h6 c2
71. Kc4 81. h7 cl=Q 82. h8=Q; but 1WW if
Drawn by agreement. Black wins the Knight as suggested
(82...Qc6f 83.Kf4 Qxg2), then White,
with 84. Qxd4f, will draw either

-204-
 All Facets Of The Endgame: Interesting Endgames

by perpetual check or by winning [It is not at all clear th.at Black

Black's last pawn. There may be can win as suggested above: e.g.
rrwre subtle attempts available to 79. Kel Ke5 80. h6 Kf6 81. Nf4 d2f
Black, but it is not clear how a win 82. Ke2 K/7 83. h7Kg7 84. Nh5f, with
can be demonstrated. -F.P.H./ a draw. Purdy obviously felt, at
J.S.P.J the time, th.at Black should win;
but the busy magazine editor left
NOTE ADDED: (from the next is­ the issue apen for comment, and
sue of the magazine) A.W Gyles there seems to have been no fur­
points out that in our "win" line above ther correspondence.
(see Diagram 5) White can draw sim­ In any case, this is an instruc­
ply by counterattack: 71... Kf5! 72. tive endgame, typical of practical
Kxc5!, followed by Kc6, etc.-evi­ play. It illustrates how expert
dently easy to overlook, as no one players, commentators, and read­
else spotted it. ers can be deceived by apparently
But our claim to a win for Black, simple positions. -F.PH./J.S.P.J
although not properly demonstrated,
appears to be correct. Black's proper
course is Second Endgame
71. ... Na8 In the last round of the New Zea­
72. Kd3 land championship, Dunlop and Se­
(this produces a position which had verne were still battling on amidst
appeared three times before, but only breathless silence after all the rest
a player whose turn it is to move can had finished. If Dunlop won, he would
claim a draw by repetition) be champion. When the following po­
[Again, White would now be able sition arrived, Dunlop seemed to have
to claim the draw before actually the title in his pocket, for he was a
making his 72nd move.-F.P.H./ pawn up and was bound to win an­
J.S.P.J other.
72. Kf5!
73. Nxh4t Kf4 J.B. Dunlop
74. Ke2!
If 74. Ng2t, then 74. .. Kf3 75. Ne1t
Kf2 76. Kd2 c4 77. h4 c3t 78. Kd1 6
c2t!; and Black wins.
74. c4
75. Ng2t Ke4
'76. h4 d3t
'77. Kd2 Kd4!
78. h5 c3t
79. K-retires Ke5!
and Black wins. E.B. Severne
There are other variations. This After Black's 44th move
win ( . . . Kf5) was also available to
Black on moves 66, 68, and 70. We give the moves without pro-

- 205 -
 OS Purdy On The Endgame

gressive comment, asking the stu­ I once went over all my mistakes
dent to find where Dunlop went in tournaments and found that I had
wrong. The answer appears below. lost many vital points through not
45. Nf2 b4 invariably asking myself both these
46. Nd1 Nd3 questions. And it should be so easy!
47. cxb4 axb4 Dunlop saw the obvious threat,
48. Ne3 Nxe5 and that put him off looking for any
49. Nc2 Nd3 other (Nd4). After 49... Nc6 50. Kg5
50. Nd4 Nf4t d4, White must resign. See the sav­
51. Kg5 Nxg6 ing of labor, besides the win of the
52. Nxe6t Kf'l game!
53. Kxf5 Ke/ In the text play, 50... Nc1 also fails.
54. Nd4 Kd6 The sacrifice on move 54 was practi­
55. Kxg6 Kc5 cally obligatory.
56. Nb3t Kc4
57. Kf5 d4
58. Ke4 d3 Endgame
59. Ke3 Kc3
60. Nd2 Drawn. E.M.Cowen
Move 49, was where Severne set
his opponent one of those little puz­
zles that save lost games. With only
two moves (...Nd3 and ...Nc6), it is
extraordinary that a strong player
could choose the one that allows the
opponent a forced draw instead of
the one that wins with ridiculous
ease. Yet that is what often happens.
There are three lessons to be
drawn from this blunder: A. Francis
1) The triangle with its apex in White to play
the rear is the very weakest forma­
tion for pawns; In the position diagramm ed, White
2) (more general) a material ad­ had to lose the Exchange; and he
vantage is often deceptive if the op­ played 41. Rxe5?, which left him no
ponent has an advanced pawn; game whatsoever.
3) (more general still, and abso­ White should have played 41.
lutelyvital) never ask yourself, "What Rg6t!! If thereupon 41... Kxe7?, then
is his threat?;" but always ask, 42. Re6t, etc.; and the black King can
"What are all his threats?" never escape the checks except by
And never, in any circumstances capturing the Rook and leaving White
whatever, make a move without first stalemated. Or if 41... Nxg6, then 42.
asking yourself, Rd7t, etc.
Alternatively, if 41... Kd5, then 42.
"Can he then do anything (nasty)?" Rd6t! Ke4 43. Rxe5t Kxe5 (forced)

-206-
 All Facets Of The Endgame: Interesting Endgames

44. Re6t! Kxet}-stalemate. temporary barrier behind which the

But M.E. Goldstein has pointed out Black forces may develop-the pawn
that Black has one way of escape, being ultimately bound for c5. Gold­
and that is by finding a way to f2. stein's move is to meet ... c6 with Ba6!
41. Rg6t Kd5 So Black must hit the Bishop by the
42. Rd6t Kxd6! "book" move ... a6 instead. But that,
43. Rd7t if followed up as Hastings does here,
If 43. Re6t, the black King again is quite satisfactory.
finds sanctuary at f2. Note that 11. Ne5 Nxe5 12. dxe5
43.... Ke6 Ne8! gives an equal game.
44.Rd6t Kf5 11. ... a6!
45. Rf6t Ke4 12. Ba4! Rc8!
46. Rf4t Ke3 "Book" is 12 ... c5?, which gives
47. Re4t Kf2 Black a terrible game. The text is a
48. Rf4t Nf3t! preliminary that saves everything­
And Black wins. as pointed out by the writer many
years ago.
13. 0-0 Ne4!
An Amazing Game 14. Bg3?
Giving Black the two Bishops-a
M.E. Goldsteln-G.H. Hastings serious error. Had White exchanged
Match, 2nd Game Bishops, the variation would have
Queen's Gambit Declined been properly tested.
(Orthodox Defense) 14. ... c5
An amazing game that supports 15. Bxd7?
Goldstein's own dictum, to which we Intending to give Black "hanging
heartily subscribe, that book knowl­ pawns." But Black's two Bishops
edge of Rook endgames is far more against two Knights will be more than
valuable than book knowledge of adequate compensation. This is not
openings. In the notes, we incorpo­ Goldstein!
rate a complete endgame article­ 15. Nxg3
one of the most useful we have ever 16. hxg3 Qxd7
published. 17. dxc5 Rxc5
1. d4 d5 Black would have obtained a
2. c4 e6 greater advantage by 17 .. bxc5! , de­
.

3. Nc3 Nf6 priving White's Knights of central

4. Bg5 Be7 squares.
5. e3 0-0 18. Rfd1 Bf6
6. Nf3 h6 19. Nd4 Rfc8
7. Bh4 Nbd7 20. Qd3 g6
8. Rel b6! 21. Rc2 Qg4
9. cxd5 exd5 22. Rcd2 Rc4
10. Bb5 Bb7 23. Nce2 Qe4
11. Qe2 24. Qb3 b5
If 11. 0-0, then 11... c6, erecting a 25. Nc3 Qe7

- 20 7
-
 CJS Purdy On The Endgame

26. Nde2 Qb4! previous note). After the text, Black

27. Nxd5 still has good winning chances, but
Goldstein knew that this would nothing clear.
lead to an endgame very much in 34. Rxd4 Rc3
Black's favor; but he could not resist 35. b4 Rctt
the temptation of putting an end to 36. Kh2 Rc4!
the passive woodshifting to which he 37. Rd6 Rxb4
had been condemned for so long. 38. Rxa6 Rb2
Also, he now eliminates the specter 39. f3 b4
of the "two Bishops." 40. g4 b3
27. Bxd5 41. Rb6
28. Rxd5 Qxb3
29. axb3 Rc2
30. R1d2 Bxb2 9
31. Rd8t
This simplification looks like play­
ing Black's game. What about 31. g4
to reduce the horrible immobility of
White's kingside? But Black would
preserve a big advantage, being able
to make a remote passed pawn and
having a strong Bishop against a 41. ..
. Rb1?
Knight. The is a very common type of Rook
31. Rxd8 ending. When the pawn has reached
32. Rxd8t Kg7 the rank next-but-one to the queen­
33. Nd4 ing rank, the problem is whether to
push the pawn one more rank, with
the Rook in front of it on the queen­
8 ing rank, or to leave it where it is.
The first course has the advan­
tage that the enemy Rook must re­
main always on the file of the pawn
and that the enemy King can never
expose itself to a safe check from the
Rook. It is the proper course when­
ever:
33. ... Bxd4 1) a safe check can be forced or
Ceding a valuable option. By 33... 2) the enemy has a weak pawn on
Rc1 t 34. Kh2 Rb1 35. b4 Bc3 36. Rd6 the other wing-a pawn that our King
Rxb4 37. Rxa6 Bxd4 38. exd4 Rxd4 can ultimately capture.
39. Rb6 b4, Black seems to have a Here, as usual, neither of these
winning position, with his pawn plus, conditions is present; and the only
owing to the ghastly position of way Black can hope to win is by
White's King-hemmed in by his im­ bringing his King over to the queen­
mobile set of doubled pawns (see the side-and that is useless unless

-208-
 All Facets Of The Endgame: Interesting Endgames

there is a square in front of the of the article; nor does it invali­

passed pawn where the King can date Black's plan stated below,
shelter from checks in the rear. For which is the only possible winning
that strategy to work, the pawn must attempt-though not guaranteed
stay where it is. Here it was particu­ to win in all similar positions, as
larly bad to take the Rook off the sec­ Purdy points out. -F.PH./J.S.P.]
ond rank where it was exerting some It may be possible for White to
pressure. draw the game after 41... Kf8!, but
The only thing to try was a sortie we have shown that it is certainly
with the King. For example, 41... Kf8 not easy. Note that it is far better to
42. Rb7 Ke8 43. e4 Kd8! 44. Rxf7 have the passed pawn on the b-file
(White must accept the sacrifice to than on the a-file, as the black King
get a passed pawn of his own to pit can come to the pawn's support more
against Black's) Kc8!, and the white easily.
Rook cannot stop the b-pawn! Every one of our readers who
If 45. Ra7, then simply 45... Ra2. If plays much chess will get this type
45. Rf4, then 45... Rxg2t!-though of ending many times in his life, and
White could perhaps draw the end­ we cannot give better advice than to
ing with Rook and pawns against study the analysis given from the dia­
Queen. grammed position with the greatest
All this upset was caused by 43. care-it is based on a demonstration
e4. Imagine, say, 43. Kg3 played in­ of the position by G. Koshnitsky.
stead; and we might find that White's
passed pawn was not far enough ad­ The plan, in the diagrammed
vanced when the time came to do his position, is: Leave the pawn
stuff. Suppose we try 43. Kg3 Kd8 44. on the rank next-but-one to
Rxf7 Kc8 45. Rf4 (best) Re2 46. Rb4 the queening rank and
b2!! 47. e4 Kc7! 48. e5 Kc6 49. e6 Kc5 threaten an invasion o f the
50. Rb7(or Rb8) Kc4 51. e7 Kc3, and queenside with your King,
Black wins easily by ... Rxe-pawn. giving up a kingside pawn.
[Clearly Black does not win in the After the enemy Rook grabs
position reached because White the kingside pawn, you have
has the move and can draw by time to move your Rook from
checking from the rear. Black has in front of your pawn and to
no time to hide Ms King on bl with push the pawn onto the rank
White threatening to queen the next to the queening rank.
e-pawn. It is perhaps surprising You will then be defending
that 1w reader commented (Purdy the pawn with the Rook still,
alw a ys acknowledged correc­ but the Rook will be at the
tions), but this seems to be another side of the pawn instead of
example of the power of the printed shut up in front of the pawn.
word. Chess books should be read A win is then quite likely.
critically, according to Purdy.
Purdy's error here does not de­ 42. Kg3 b2?
tract from the instructional value It seems incredible that Hastings

- 209 -
 CJS Purdy On The Endgame

should have lost with Black from this Rook And Pawn vs. Rook.
position. Of course the game should A pawn, other than the a- or
be drawn, but Black must play for h-pawn, past the middle of the board
the draw. with a clear file between the pawn
43. Kf4 Kf8 and the enemy King regularly wins.
44. Ke5 Ke7? There is an exception, however, with
Easy draw by 44... Rei. Just as a the c-, d-, e-, or f-pawn. If the enemy
joke, note the terrible trap 44... Rei! King is on the "narrow" side of the
45. Rxb2 Rxe3t 46. Kf6???? pawn, as here, and at least as close
45. Rb7t Kf8 to the queening rank as the pawn and
46. Kf6 Kg8 if the enemy Rook can put at least
47. e4 h5 three clear files between itself and
48. gxh5 gxh5 the pawn (as Black has done with
49. Rb8t Kh7 his last move), the defender can
50. Kxf7 h4 draw. The method is by lateral
51. Kf6 h3! checks, and the three files are nec­
52. gxh3 Rf1 essary so that the checked King can­
Arranging to fight two isolated not approach the checking Rook.
pawns instead of two united pawns.
Black still has hopes of a draw. See Chapter 1.
53. Rxb2 Rxf3t
54. Ke6 Kg7! If the black King were at g7 or g8,
By 54... Rxh3?, Black would give White could now check on the file,
White a book win with 55. Rg2, shut­ driving the black King a file farther
ting out the black King. away and giving an easy win . But
55. e5? here the reply would be .Kf5.
..

White evidently thought he could To cope with the forthcoming lat­

still give up the h-pawn for a book eral checks, White decides to bring
win The right move was 55. Rh2, but
. his own Rook to the queenside so as
it would not be easy to win. to be able to interpose appropriately.
55. ... Rxh3 But then the black King is free to ap­
56. Ke7 Kg6! proach the pawn whenever the white
57. M Ra3! King leaves e7, and the draw is very
easy.
58. Rf6t Kg7
10 59. Rd6 Ra5
Anywhere along the a-file will do
now-checks have no point.
60. e6 Ra7t
[Ed.: 60. . . Ra7t is fatal. 60 ...
Ra8! is calledfor.-K.M.]
61. Rd7
[Ed.: 61. Ke7! +-from K.M.}
Drawn position 61. ... Ra6
The position is now a book draw. See the NOTE ADDED below.

-210-
 All Facets Of The Endgame: Interesting Endgames

62. Ke8t Kf6 The position is a draw as stated

63. e7 Kg7?? previously. Hastings now played 61...
Black makes a terrible blunder, Ra6, on which we did not comment.
doubtless due to fatigue, in voluntar­ There followed 62. Ke8t! Kf6 63. e7,
ily withdrawing his King from the whereupon Black played 63... Kg7,
pawn. Black had the easiest of draws giving White an easy win.
by 63... Ra8t 64. Rd8 Ra7 65. Rd6t We gave 63 ... Kg7 a double query,
Kg7; for if now 66. Rd8, then simply which was uncalled for because af­
...Kf6 again. And if 66. Rc6, then ter 63 ... Ra8t 64. Rd8 Ra7 65. Rd6t
Black checks and plays ... Kf7. The Kg7, White also has a win-which
same pattern holds if 66. Kd8. we leave for the student to find for
64. Rd8 Ra7 himself as an exercise.
65. Rc8! Kg8 It was really 61... Ra6? that lost
Black has lost a tempo. If 65... Kf6, the game. Clearly 61... Ra8! was more
then White checks and plays Kd8; logical, for it was more important to
and then the Rook can interpose to a keep the white King from advancing
check. than from retreating. On the 8th rank,
66. Kd8 Resigns. the white King can help his pawn to
A tragic finish! queen.
Here is how the draw is obtained.
61. ... Ra8!
Note Added 62. Rb7 Kg6!
We have noticed an error in our Or 62... Ra1. But 62... Kg8? loses
notes to the game above. Hastings' (the analysis of the win is long and
decisive error was not his 63rd move, complicated).
but his 61st. As the ending is highly 63. Kd7
instructive and as it will repay the Or 63. Rb1 Ra7t 64. Kd8! Kf6! 65.
closest study, we give the position Rei (if 65. e7, then 65 ... Ra8t 66. Kd7
with Black to play his 61st move. Ra7t 67. Kd6 Ra6t! , etc.) Ra8t 66.
Kd7 Ra7t 67. Kc6 Ra6t! , etc., and
Hastings Black draws.
Or 63. Rc7 Kg7 64. Rd7 Kg-6 (not
64... Rb8 because of 65. Ra7! ) 65. Rd1
11 (if 65. Rd8, then 65 ... Ra7t 66. Ke8
Kf6! ) Ra7t 66. Kd6 (if 66. Ke8, then
66... Kf6! ) Ra6t, etc.
Or 63. Kd6 Rd8t! 64. Ke7 Ra8
(back again).
63. ... Kf6
64. Rc7
Or 64. e7 Kf7 65. Rc7 Re8 66. Kd6
Goldstein Ra8! , etc.
After 61. Rd7 64. ... Kg7!
Not 64... Ra1? because of 65. e7
Rdit 66. Ke8! , etc.

- 211 -
 CJS Purdy On The Endgame

65. Kd6t Kf8 game should be drawn, as explained

And Black draws. further. Black preferred, not without
In the classic words of Lajos Stein­ justification, to keep the Queens on,
er, "Not so easy." But the student who so that White's King should be a li­
takes the trouble to study this analy­ ability for White rather than an as­
sis will have a real grasp of the intri­ set. Black has to cope with White's
cacies of a type of Rook ending of threat of d6t. Play proceeded:
very common occurrence. 39. ... Rc5!
It will be seen that Hastings could 40. Qb8t Rc8
well be excused for losing, after all. The gain of an important tempo,
as Black's Rook is now safe.
41. Qe5 Qxh3t?
A Fighting Ending Black made this blunder without
The diagrammed position arose hesitation, realizing immediately that
shortly after the resumption of a he should have begun a checking
game between Nield and Purdy in the campaign by ... Qf1 t-a line not de­
first round of the 1939 Australian void of winning chances. Now Black
championship. At an early stage, has not even a draw! The win of the
Purdy had been in a winning posi­ pawn causes the initiative to change
tion; and now, although a pawn down, hands, Black can now make nothing
he was still optimistically struggling at all out of the position of White's
to squeeze out a win-on the score King.
of White's very exposed King. You 42. Kd4 a5
must picture both players in shirt 43. d6 RdS
sleeves, in sweltering heat; they have 44. Qd5t KhS
just snatched a hurried meal and 45. Qf7?
some adjournment analysis after four But White falters in turn. Black
hours' battling in the afternoon. cannot play 45... Rxd6t?, but he is
not without a defense. Correct was
C.J.S. Purdy 45. Qe6! That forces the exchange of
Queens, and a very instructive end­
game develops. We advise students
12 to follow the analysis carefully, as
similar positions occur with some fre­
quency. The black Rook must be
given up, and Black's passed pawns
then try to race the white Rook. Thus,
45. Qe6! Qxe6
46. Rxe6 h5
47. Kd5 h4
A.E. Nield 48. Kc6 Kh7
Blaek to play his 39th move 49. d7 g5
50. ReS Rxd7
If 39... Qxh3t 40. Kd4, Black would 51. Kxd7 h3
have to exchange Queens; and the 52. Re2 g4

-212-
 All Facets Of The Endgame: Interesting Endgames

43. Kd5 Rdtt

44. Kc6 Rc1t
45. Kb7! Rd1
46. Kc7 Rc1t
and now the point-where the dou­
bled pawns come in handy-
13 47. Rc4!
(the coup known as bridge-building
[Lucena], one of the most iinportant
devices in Rook endings)
47. ... Rd1
48. d7
Again Black's Rook must be given
up, and it will be found that White
After 56- g3 (analysis) draws at least.
As it would hardly be possible to
At first sight, this looks like an work this all out over the board, the
easy win for Black, as he cannot pos­ student can see the iinportance of
sibly be prevented from queening. having a good knowledge of end­
And so it would be if all the chess­ games, especially of Rook endgames,
men on the kingside were moved one not only because Rook endgames are
or more files to the left. But with the the commonest kind and very diffi­
black King on the edge of the board­ cult, but also because they are full of
there's a hint!-it is an easy win for little rules and special stratagems
White. How? We leave that as a siin­ that can be learned from books.
ple exercise. The remainder of the game is in­
Now, if White did not have a pawn­ teresting as a fight. Both (!) players
majority on the queenside, the play were now playing for a win.
would be exactly the same; but the 45. Qc8!
result would be a draw. And if the 46. Qe7 h6
reader will now go back to Diagram 47. Re5 Qb8
12 and play 39... Qxh3t 40. Kd4 Qxb3 48. d7
41. axb3, he will find that just such a More courageous than prudent,
position can come about. White no but not fatal.
longer has a queenside majority be­ 48. ... Qxb2t
cause his pawns there are doubled 49. Ke4?
and cannot win. Worn out by toil and heat, White
Black himself, in the variation un­ makes a fatal miscalculation. White
der discussion, could win by getting could have drawn by 49. Kc5, but he
his Rook behind the d-pawn, but for still thought he could secure a full
a little catch: point. Black's next move wins the
39. ... Qxh3t passed pawn in all variations.
40. Kd4 Qxb3 49. ... Qg2t
41. axb3 Rc1 (say) 50. Kf5 Qh3t
42. d6 Kf7 White had thought to go now to g6

- 213 -
 CJS Purdy On The Endgame

and f7, but that spills the Queen. After 37. Nxd5 exd5, White can
51. Kf4 Rxd7 draw easily by 38. f5!, as the break­
And Black won: through by e4 is then always avail­
52. Qe8t Kh7 able when the need arises.
53. Re3 Rd4t But 37. Nxd5 is actually answered
54. Re4 Qh4t by 37... Kxd5!, and the brain-racking
55. Resigns. begins. It looks as if Black is then
threatening ...Ke4, but actually that
is no threat at all-f3 is not a point
Note Added of entry, and White can easily guard
In the analysis after White's 45th d3. If White were to consider ...Ke4 a
move, Black's 47th and 48th moves threat and were to play 38. Kf3, then
should be transposed. Otherwise, she would lose rapidly; after 38 . . f5!
.

Nield points out, White could win the Black would be threatening ...b4!, and
h-pawn by 48. Re5 followed by Rh5. after 39. Ke2 to forestall that, the
After 47... Kh 7!, Re5 would be met by black King would just walk ar ound
...Kh6. This exemplifies an important to a5 until ...b4 would be safe again.
elementary maxim: Don't get your However, White must not move
forces separated. her g-pawn, as she would then turn
...Ke4 into a threat.
White's best move, then, for a
Two Curious Endgames start, is to do nothing, say 38. Kd2:
37. Nxd5 Kxd5
First Endgame 38. Kd2
The position diagrammed oc­ Now if Black's King starts off for
curred in a New South Wales Wo­ a5, White simply moves up her
men's championship game, Mrs. e-pawn and is then always ready for
Redies-Mrs. Gheysens. The prob­ counterattack by d5. That suggests
lem is whether White can exchange 38... f5; but then when Black's King
Knights and draw. gets to a5, White's King can be ready
to go to a3-and White can always
Mrs. Gheysens retain the opposition.
The trouble with ...f5 is that it
blocks a possible avenue for Black's
King.
Can't we first immobilize White's
pawns and thus deprive her of pawn­
moves? Then White can be depriv ed
of the opposition when the Kings
stand snarling at each other on a3
and a5. Yes, try 38... h5! ! That threat­
ens ...h4, immobilizing the pawns­
Mrs. Red.Jes after which ...f5 and the King tour
White to play her 37th would win easily for Black.
38. ... h5!!

-2 14-
 All Facets Of The Endgame: Interesting Endgames

Now ...h4 has to be stopped. If 39. Black felt hopeful of winning this
h4, then 39... Ke4! (threat ...Kf5) 40. ending, in spite of Bishops of oppo­
Ke2 g6 (a valuable spare move) 41. site color, because his Queen and
g3 f5; and White is in Zugzwang. Bishop could combine for attack,
If instead 39. g4, then 39... hxg4 whereas White's couldn't. The crux
40. hxg4 Ke4 41. Ke2 g6; and White of the situation is the pawn on e5, a
is in Zugzwang again. dark square-it prevents White's
Finally, if 39. g3, then 39... Ke4 40. Bishop from ever achieving the same
Ke2 f5, etc.; and Black wins. mobility as Black's. The sealed move
If 38. Kd1, instead of Kd2, the play is not difficult.
is exactly the same; and 38. Ke1 42. ... a5!
would lose by the immediate ...b4. Black's extra pawn was useless
[The [U,St a.ssertion is not correct: except for one purpose: to die. Thus
Black's strategy for the rrwve 38. Black gets a passed pawn. Black now
Kel must be the same a.s for the has what it is absolutely necessary
moves 38. Kd2 and 38. Kdl. It is to have for any hope of winning end­
only when the white King gets (],S ings with Bishops of opposite color:
far away a.s 38. Kfl that the tactic chances on both wings.
38... b4 works.] 43. bxa5 b4
The value of flank-pawn advances 44. Qf2!
for fixation purposes is the lesson to Through careless analysis in the
be drawn from this intricate little adjournment, Black had missed this
ending. Also: for winning purposes, strong move which holds everything.
keep open as many avenues as pos­ As White now has the second rank
sible. protected against checks to the King,
he can answer ...b3 with Bc3, immobi­
Second Endgame lizing the passed pawn. There is no
The second endgame is vastly dif­ advantage, therefore, in pushing the
ferent; it occurred in a New South b-pawn-that would only put the
Wales championship tourney. Black pawn on a light square and thereby
to seal his 42nd move. increase the mobility of White's
Bishop.
C.J.S. Purdy At the same time, White's move
Qf2 prevents the breakup by ...g5 be­
cause Black's f-pawn is under fire.
Black began to feel much less hope­
ful, but he found the only aggressive
move.
44. . .. Kg8!
As ...g5 is again possible, White
decides to plug.
45. g5 h5
Black has gained something. Pos­
H. Klass sibilities of a breakthrough are ruled
Black to move out forever, but White's h-pawn is

- 215 -
 CJS Purdy On The Endgame

fixed on h4, depriving White's King Great perils attend the royal tour
of one possible flight. Also, the square here embarked upon, but they must
g4 is now held permanently by Black; be faced. The idea is for Black to get
and there are the makings of a mat­ so far into the enemy camp with his
ing net. King that he can afford to exchange
46. Bd2 Kg7 Queens. Meanwhile, Black must try
To make White move. White's to stop White's King from also com­
Bishop is the only unit that can af­ ing into the center.
ford to move. 66. Kg'J Qd3t
47. Be3 Qd5 67. Kh2 Ke8
The first step is accomplished: 68. Ba3 Qa6
capture of the a-pawn, the pawn 69. Bb2 Qc4
could not have been held forever. 70. Kg1 Qe4
48. Bd4 Qxa5 71. Kh2 Qc4
49. Qd2 Qd5 72. Kg1 Qd3
We now pass over 14 moves to the 73. Kh2 Bb7
next episode. See Diagram 16. 74. Ba3 Kd7
75. Kg1 Qd5
Black's Queen has the jitters
again, if that white King got out, it
would be terrible. Some of the ma­
neuvers are just to gain on the clock.
76. Kh2 Qe4
77. Kg'J Qd3t
78. Kh2 Qd5
79. Kg'J Qd1
80. Kh2 Qd3
Black to play his 63nl 81. Bb4 Bc6
Another sealed move. Black's 81st
White has for some time annoy­ is to make a safe way up for the King
ingly maintained his Bishop on the via c8, b7, and a6-and from there,
diagonal where it stops Black's King God willing, on to b5 and c4 (Queen
from entering the game. Black has moves thrown in at timely intervals).
gone to a lot of trouble to get his 82. Ba3 Kc8
Queen over to a4 merely to play ...b3 83. Kg1??
without allowing Ba3. From a3 the A gruesome error which brings
Bishop would restrain both the the game to an end before it has re­
b-pawn and Black's King. Now, at ally got into its swing. It was neces­
long last, Black nerves himself to ad­ sary for the Bishop to oscillate. Sub­
vance the b-pawn. sequent wrangling never actually es­
63. ... b3 tablished whether Black could have
64. Bd4 Qc4 forced the win, but we think there
Black's Queen hurriedly regains should be some way to do it.
the center. The King can wait. 83. Qh3
65. Bb2 Kf8 84. Qh2 Qf3

-216-
 All Facets Of The Endgame: Interesting Endgames

85. Qd2 Qg3t 29. f4!

86. Resigns. Not necessary yet, but it will be to
White loses his Queen if 86. Kf1. support the outpost Knight after
This ending shows what tortuous ... Ke7-d6. If not played now, Black
maneuvers are sometimes necessary might forestall it by . . .f5, driving
in endgame play. One proceeds from White's Bishop to f3 in front of the
one small gain to another, in easy pawn.
stages, until winning chances begin 29. Ke7
to loom up mistily. 30. Kr.! Kd6
31. Ke2 g5!
Threatening . . .f5 32. Bf3? g4.
A Lively Endgame White wants to maintain the long di­
Most players would find it hard to agonal, so
believe that in the hands of two rec- 32. Bf3!
ognized masters of endgame play the Goldstein points out a humorous
diagrammed position could lead to event that could easily happen: 32.
anything but a draw. Unfortunately, Nc4t Kc5 33. Kd3 b5? 34. b4#! !
the wealth of adventure that is ob­ 32. ... f6
tainable in endgames is buried trea­ Goldstein now decides on fire-
sure for the majority of players. works.
True, the play was not altogether 33. fxg5!? fxe5
accurate, but that only shows what Not 33... Kxe5? because of 34. g1)!
unexpected difficulties lurk in such 34. gxh6
positions. The game was won by M. White, though getting only two
Goldstein from L. Steiner in a Sydney pawns for the piece, has three passed
"A" grade interclub competition. The pawns (unfortunately not all on dif­
time limit of 28 moves per hour and 7 ferent files).
moves per quarter-hour thereafter 34. Nf'7
naturally increases the possibility of 35. h7 Ke7
error, but fluctuating fortunes always 36. Bh5!? Nh8
make for entertainment. 37. g4 Kf6
38. h4 Kg7?
L. Steiner Intent on grabbing the advanced
pawn and short of time, Steiner miss­
es the obvious ... e4 followed by ...e5,
shutting out the white King and mo­
bilizing the black Bishop. Thus
Black could have won, and we see
that White's impetuous rush (moves
36/7/8) was unsound. Whether some
safer line would have drawn is hard
to say.
39. Kd3 Bd7
M.E. Goldstein 40. Ke4 Bc6t
After 28 Kf8
.•.
41. Kxe5 Bd5

- 217 -
 OS Purdy On The Endgame

42. g5 Kxh7 Which piece should Black take?

Black again has a piece for two In other words, will his Knight pre­
pawns, but he cannot get his Knight fer to play against a Knight or against
out free of charge. White's King is a Bishop? A Bishop's normal slight
now a host in himself. superiority disappears if the play is
43. Kf6 Be4 confined to one half of the board be-
44.Bg4! Ng6 cause the squares the Bishop com-
45. h5 Nf8 mands on the other half of the board
46. Ke7! Kg7 are well-nigh useless to it. Can it be
47. g6! e5 said that the Knight actually becomes
48. h6t Kxh6 superior in such cases? Probably one
49. Kxf8 Kxg6 would find that much depended on
50. Ke7 Kg5 the pawn position. It would be a valu-
Equal material, but the better able piece of research to try out this
King wins! ending both ways.
51. Bc8 Bc6 We should say that Najdorf s chief
52. Kd6 Kf6 reason for playing ...Bxe2 rather than
53. b4! e4 ... Bxb5 was that it gained him a
54. Be6 Kg5 tempo for his King to come in. After
Black is helpless. ...Bxb5, Bxb5, Black would have to
35. Bd5 Kg4 move his Knight; and then it would
36. Bxc6 bxc6 again be attacked. Moreover, it was
37. Ke5 Resigns. extremely difficult to estimate the
Bishop's defensive value to White,
whereas Black could at least be sure
Piquant Endgame that a Knight would equal a Knight.
In an 8-game match with Fine, Another question is, Should the
Najdorf was down two points after ending be a win? If the white pawns
two games. In the third game, the dia­ were united, it would certainly be a
gram shows the position with Black draw-on the general rule that, other
to play his 43rd move. things being equal, a pawn plus does
not suffice where the pawns are all
Najdorf on one wing and none passed. Does
the isolation of the white pawns turn
White's game into a loss? That, too,
18 would need prolonged research to
answer. At least one can say that it
is very hard for White to draw. Here
is how the game went.
43. Bxe2
44. Kxe2 Kf6
45. Nd6 Ke6
46. Nc4 Kd5
Fine 47. Nb2 Kd4
Mter 43. K(e3)f2 48. Nd3

-218 -
 All Facets Of The Endgame: Interesting Endgames

If 48. Kf2, then 48... Nf4 49. Kg3 63. ... Nf5
Nd3 forces the white Knight out of 64. h3
play. If 64. Nf2, then 64... Nd4, Zug-
48. ... f5 zwang.
49. Nb4 Nf4t [After 64. Nf2 Nd4 there appears to
50. Kd2 be no clear Zugzwang, thaughBlack
Not 50. Kf2?? because of 50... obviously maintains winning
Nd3t. chances. It seems more likely that
50. ... Ne6 White avoided 64. N/2 because then
51. Nc6t Kd5 64... Ne3f wauld drive the white
52. Ne7t Ke5 King to the back rank, cramping
53. Ke3 Kf6 White severely-probably deci-
54. Nd5t Kg5 sively. The text continuation ap-
55. Kf2 Kh4 pears to offer better chances of sur-
56. Kg2 g6 vival; but in fact it gives Black a
57. Ne7 Kg5 more or less forced winning line-
58. Kg3 f4t White may have overlooked this,
59. Kf2 Nd4 but in any case the defense is very
60. Nd5 Nc6 difficult.-FP.H./J.S.P.]
64. Ne3t
65. Kh2 Nc2
19 66. Kg2
If 66. Nd2, then 66... Ne1 wins.
66. Ne1t
67. Kf2 Kxh3!
68. Kxe1 Kg2
69. Ke2 h5
70. Ng5 h4
71. Ne6 g5
61. Nc3 72. Resigns.
Could White do better by retain­ After 72. Nxg5, there follows 72...
ing his pressure on the f-pawn until h3 73. Nxh3 Kxh3. Thereafter Black
Black was forced to regroup to free wins the f-pawn. Since the black King
his King? Thus, 61. Kg2, waiting. A is then on the rank next-but-one to
good position for more research. the queening rank and is in front of
61. Kh4 his pawn, Black wins. There's just a
62. Kg2 Ne7 little trap for beginners. If 74. Kd3,
63. Ne4 then not 74... Kg3?? but 74 .. . Kg2.
It has been said that White could
still draw by 63. Ne2 Nd5 64. Ng1,
keeping out the black King. Black, Solution
however, has a win. See if you can After 63. Ne2 Nd5 64. Ng1, Black
find it-<mr solution appears at the to play and win.
end of the article. A very interesting
study.

-219 -
 OS Purdy On The Endgame

the 3rd game of the match requires

amendment. Our 68... Nd4 will not do
(because of h3). Instead Black must
play 68... g4 at once. If 69. Ne2, then
69.. . Kg5.

The Confident Kibitzer

Endgames are skipped by many
readers. An endgame that has some
Black to play and win human interest is likely to be skipped
by fewer. In a tricky ending in a City
The first phase consists in play­ of Sydney championship, C.J.S. Pur­
ing ...h5 and ...g4 at a moment when dy-E. Daroczy, human interest was
White must exchange pawns, thus provided by a very confident kibitzer.
giving Black a passed pawn.
64. . .. g5 E.Daroezy
65. Kf2 Ne3
66. Ke2
Not 66.h3? because of 66... h5,
with Zugzwang.
66. .. . h5
67. Kf2 Nf5
68. Kg2 Nd4!
[Ed.: See the Note below.]
69. Kf2 g4
70. fxg4 h:xg4
Second phase: playing to advance C.J.S. Purdy
the pawns safely. That can be done After U .. N(f3)xe5
.

if Black's Knight can play to e4-that

is easy. For example, In the diagram, Black has just of­
71. Kf1 Nc2 fered his Knight with 44... N(f3)xe5!
72. Ne2 Ne3t in a valiant attempt at counterplay.
73. Kf2 Nd5 Had Black played 44... Nxh2, White
74. Ng1 (forced) would have won fairly easily by Nd4t.
74. ... Nc3 White, after long calculation, re­
75. Ke1 luctantly declined the Knight and
Or 75. Kf1. played 45. Nd4t Kf6 46. Kb6. Daroczy
75. . . . f3 then sealed, and the kibitzer (a good
76. Kf2 Ne4t player)-completely disgusted by
And Black wins. Purdy's caution-burst out with, "Is
this the new chess, Mr. Purdy? No
Note Added other master in the world would have
Our solution to the puzzle set in refused that Knight."
the course of the Knight ending in Purdy said he thought his play

-220 -
 All Facets Of The Endgame: Interesting Endgames

was clearer. "Taking the Knight was In this case, however, the capture
quite clear," said the kibitzer; he then of the Knight would have been an out­
showed 45. fxe5 Kxe5 46. Kb6 Ke4 right error, strange as it may seem.
47. Kxa6 Kd3 48. Kxb5! This does Start again from Diagram 21:
win. 45. fxe5? Kxe5
46. Kb6 g5!
"Is this the new chess, Mr. 47. Kxa6 f4
Purdy? No other master in 48. gxf4t gxf4
the world would have re­ Now the obvious play is
fused that Knight." 49. Kxb5 Ke4
50. Kxc4
Purdy pointed out that Black (no better is 50. Net because of 50...
would not lay his King down, but ... g5, c3! 51. Kc4 c2 52. Nxc2 f3, and the
etc. The kibitzer, however, said he f-pawn queens)
had analyzed that to a clear win also. 50. f3
Then everyone had to leave. 51. Nd4 f2
Now, we do not publish this end­ 52. Ne2
game just to prove the kibitzer wrong. Perhaps the kibitzer saw this far
The endgame is published as highly and reasoned that Black cannot now
interesting in itself, and the kibitzer's queen because of the forking check,
excitement just provides an extra overlooking that 52... ft=Q pins the
touch-we rather envy anybody who Knight and wins for Black!
can get so worked up about other Far from having a clear win,
players' games. White, it seems, must scratch for a
Even had the kibitzer been theo­ draw if he takes the Knight. What is
retically right, it is a mistake to play the drawing line for White after 45.
a sharp line with a risk of oversight fxe5 Kxe5, and how should the play
when there is a quiet way to win­ continue for both sides? The solution
perhaps slow but not requiring fine appears at the end of the article.
calculation. Mostly the sharp play will Some readers may wonder how
win for you quickly, but occasionally the game did proceed, so here are
you will find that indeed you have the moves. Black's sealed move was
miscalculated and thrown away the obvious enough.
game. The object in tournaments is 45. Nd4t Kf6
to win points. Try the sharper wins 46. Kb6 c3 (sealed)
rather in practice games. 47. Kxa6 Nc6
48. Ne2!
" ... we rather envy Not 48. Nc2 Ke6 49. Kxb5 Kd5 50.
Ka6, as Black gains time with ... Nd4.
anybody who can c2
48.
get so worked up 49. Nc1 Ke6
about other play­ 50. Kxb5 Kd5
51. Ka6 Nd4
ers' gam.es." 52. b5 Ne6
53. Ka7 Resigns.

-221-
 OS Purdy On The Endgame

After 53... Nc5 54. b6 Kc6, 55. Na2! Black can effectively play ... Ke4-d3---
threatens a fork! whereas if 47. Kxc4, then ...Ke4 with
the old threat of ...g5, etc.
Solution However, at move 46 there is a
As before, draw by the surprising h4 (stopping
45. fxe5? Kxe5 ...g5). For example, 46. h4 Ke4 47.
46. Kb6 g5 Kb6 (or 47 h5) 47. . Kf3 48. Nd4t Ke4,
.

47. Kxa6 f4 etc.

48. gxf4t gxf4
and now Note To Solution
49. a4! f3 There is one tricky variation that
50. axb5 fl we might have added to the solution.
51. Ne3 Ke4 Starting from the diagram, we gave
52. Nf1 45. fxe5 Kxe5
(if 52. Nxc4, then 52... Kd3 53. Ne5t 46. h4 Ke4
Kd4! 54. Nf3t Ke3 55. b6 f1 =Qt, Black but we did not say what happens if
queening with check) then
52. c3 47. a4 bxa4
53. b6 c2 48. Kxc4
54. b7 c1=Q The answer is 48... g6!, threaten­
55. Ng3t Kf3 ing .. Kf3, etc. If thereupon 49. Kc5,
.

56. bS=Q Qh6t then

57. Qb6 Qxh2 49. Kf3
58. Nf1 50. Kb6 Kxg3
and White draws. For example, 51. Kxa6 f4
58. Qf4 52. b5 f3
59. Qc5 Kg'l 53. b6 rl
60. Ne3t Kf3 54. Ne3 Kf3
61. Nf1 And Black draws at least; for if
with repetition of moves virtually 54. Nf1, then 54... a3, etc.
forced for both sides. In this, if 48. Kb6 instead of 48.
The immediate 46. a4 is not clear: Kxc4, then 48... Kd3 draws at least.
46... bxa4; and now if 47. Kb6, then

- 222 -
 (l\apler I?
El\dqal\\e \\/izards

Rubinstein's Black Magic nating one; but a central post, in it­

A t the great tourney of Carlsbad self, means nothing to a Rook. Un­
1"'\.. 1929, Rubinstein won from Mat­ like the other pieces, a Rook can com­
tison a Rook ending that will never mand the same number of squares
be forgotten while chess lives. So from a side or corner square as from
"hopelessly drawn" did the ending the center; and the only test for the
appear, that the editors of the tour­ Rook's well-being is his mobility.
nament book united in asserting that Here the white Rook commands no
if the event had occurred 300 years useful squares, either on the rank or
earlier Rubinstein would certainly on the file-at least he won't as soon
have been burned for dealing with as Black has played up his King and
evil spirits. freed his own Rook. The backward
black d-pawn will then not be a weak­
Rubinstein ness, but rather a useful obstruction
to the white Rook. In short, we can
conclude that the game is even.
Black's initial plan is clear: step up
with the King.
20. ... Kd7
21. c4
With the object of keeping Black's
d-pawn permanently backward. The
move, however, gives White a fixed
pawn (the c-pawn itself) which may
Mattison become a target for ...b5 at some later
Black to move time. It was safer to avoid the cre­
ation of any weakness and at once to
Superficially, the position might develop the King (Kd2-e3). White,
appear advantageous to White, as his however, is apparently under the il­
Rook has a "dominating'' look. For a lusion that he has a slight pull.
Queen, Bishop, Knight, or King, the 21. .
. . g6!
square d5 would certainly be a do mi- Not yet ...Ke6. One must always
 CJS Purdy On The Endgame

be ready to alter or to interrupt a point of endgame strategy. It might

plan if any tactical objection to it appear that the plan is not feasible
crops up. Here the immediate ...Ke6 since White can play g3, but Rubin­
would encourage White to post his stein has everything cut and dried.
Rook aggressively on the queenside, 27. g3 f4!
via d3. So Black, before completing 28. Re2t
his first plan, starts his second­ If 28. g4, then 28... hxg4 29. fxg4
which is: to open a file for the black Rh8 30. Kd3 Rh4 31. Rg2 Ke5 32. Re2t
Rook. Kf6; and Black wins a pawn (if 33.
22. Kc2 Rg2, then 33... Kg5) while retaining
This move and the next one are the initiative.
not the right development for the 28. Kf5
white King-he ought to centralize. 29. Re4 fxg3
But White, after playing c4, already 30. hxg3 Rg8
has a sneaking fear that his queen­ At last the Rook moves-the first
side will be opened up later on; and, time that he could do so effectively.
therefore, he thinks it wise to use his Now we see that the g-pawn is fixed
King for queenside protection. See after all, on the third rank instead of
how one inexactitude breeds an­ the second. Not exactly fixed since it
other. can push to g4, but thereupon Black
22. ... Ke6 would have the possibility of ...h4 (un­
23. Kc3 less the black King is on f5, then
Better Rd3 (threatening Rb3 to tie ... hxg5 is necessary), making a dan­
the black Rook-a plan that the text gerous remote passed pawn.
move puts out of court). 31. Rf4t
23. f5! [Ed.: Interestingly, 31. g4f can
24. exf5t gxf5 be played at tkis point if White is
25. Rd2 only interested in a draw.]
In opening a file for his own Rook, 31. . . . Ke6
Black has of course also opened one 32. Re4t
for White-which White now pre­ Before playing g4, White drives
pares to use. Note, however, that the the black King out of play-quite in
e-file is not very effective for White accord with principle. In the tourna­
because the black King holds the ment book, Brinckmann suggests as
squares at the business end of it-a the "last chance" 32. g4 at once; and
useful point to remember. thereupon if 32... h4, then 33. g5. How-
25. ... b5! ever, after 33 . .. Rxg5 34. Rxh4 b4t ,
What White foresaw has hap­ White must give up his a-pawn (after
pened. This move threatens nothing, ...Rg2t) as the only alternative would
but it creates the possibility of a be to give Black the "second rank
break on the queenside at an oppor­ absolute" (White's King being con­
tune moment. fined to his back rank forever).
26. b3 h5 32. ... Kd7
Threatening to fix the g-pawn. 33. g4
Fixation of weak pawns is a cardinal Now 33.. . h4 can be answered by

-224-
 All Facets Of The Endgame: Endgame Wizards

Re2, etc., stopping everything. pawn and conserve his own.

33. ... Rf8! 35. a4
34. Re3 A waiting move, based on the idea
Tying up a Rook to the defense of that ...h3 is not a real threat; also,
a pawn is always risky, and a mas­ the move eliminates the weakness at
ter of endgame play would have seen a2 and forces Black's hand on the
that a better chance was to give up queenside. We shall return later and
the pawn: 34. gxh5 Rxf3t 35. Kc2 look at other possibilities.
Rh3. Now the h-pawn must fall; but 35. bxa4
with three pawns against four, all on 36. bxa4 Re8!
one wing and none passed, a draw 37. Kd2
can usually be obtained with care. Looks like a saver, as 37... h3 per­
Here the task would be tricky, as mits the white Rook to get back via
White's King can get tied up by ...b4 b3 (say) and 37... Rxe3 loses a tempo
unless White is prepared to concede and brings the white King into the
Black a passed pawn. square of the h-pawn. If 37. Rd3, then
34. ... h4! 37... h3 38. f4 (if 38. Kd2?, then 38...
h2) h2 39. Rh3 Re3t; and the h-pawn
queens.
If 37. Rd3 h3 38. Kc2, then 38...
Re2t 39. Kc3 h2 40. Rd1 Re3t, to be
followed by ... Rxf3, winning.
37. Rxe3!
38. Kxe3 d5!
39. Resigns.
For after 39. cxd5 there follows:
39. h3
The position diagrammed might 40. Kf2 h2
seem to offer White plenty of hope­ 41. Kg2 c4
has he not two united passed pawns? 42. g5 c3
Why should Black's single pawn, 43. g6 c2
even though well up, be devastating? 44. g7 c1=Q
If it were Black's move now and if 45. g8=Q Qg1t
Black played ...h3, White could give 46. Kh3 h1=Q#
up his f-and g-pawns for the h-pawn Now go back to the second dia­
by 1... h3 2. f4 Rxf4 (not ... Rh8 be­ gram. We find that ... Re8! was actu­
cause of Re1) 3. Rxh3 Rxg4 4. Rh7t, ally a threat, and White's 35th did
with a fairly sure draw. As it is nothing about it. The obvious play is
White's move, surely White must be 35. Kd3, to hurry the King over. Still,
all right? But Black's threat of ...h3 however, 35... Re8! (threat ...h3); and
can wait a little; Black will first ex­ White's Rook or King must move,
tract the last ounce of benefit from thereby returning the tempo appar­
White's Rook being tied. ently gained. For example, 35. Kd3
Note that 35. g5 is useless (35. g5 Re8! 36. Re2 Rxe2 37. Kxe2; and now
Rf5 36. g6 R g5)-Black would win a 37... bxc4 38. bxc4 d5! as in the game.

- 225 -
 OS Purdy On The Endgame

In this, White cannot improve mat­ 1906) was the winner of the interna­
ters by first exchanging pawns him­ tional tournament at Hastings 1895,
self, as Black is then threatening ...c4 ahead of Lasker, Tarrasch, Chigorin,
with the same effect as ...d5 in the and Steinitz. Pillsbury's last-round
game. game against Gunsberg illu strates
Therefore, try 35. Kd2, to leave some important endgame themes.
room for Rc3 if ...Re8. Then the threat Before we discuss the endgame fea­
of ...h3 becomes real-thus 35... bxc4 tures of the game in detail, we ask
36. bxc4 h3 37. f4 Rxf4 38. Rxh3 Rxc4; the student to play through the game
and the weakness of the pawn on a2 to get a feel for its structure.
forces White to play Ra3, giving up
the g-pawn-after which the two Pillsbury-Gunsberg
united black pawns win . Hastings 1895
To eliminate the weak white Queen's Gambit Declined
c-pawn, suppose 35. cxb5 axb5; and (Slav Defense)
now 36. Kd2. Then, however, a new 1. d4 d5
Gorgon enters in the shape of 36... 2. c4 c6
d5 (if 37. Ke2, then 37... Re8!). 3. e3 g6
Finally, if 35. cxb5 axb5 36. a4 Schlechter's Variation. In the tour­
bxa4 37. bxa4 h3, then Black soon nament book, Lasker objects to it
obtains a win with two united passed because it creates a target on the
pawns against a single passed pawn, castling side. See the next note.
as in the 35. Kd2 variation. 4. Nc.3 Bg7
5. Nf3 Nf6
6. Bd3 0-0
7. Ne5 dxc4
Which move lost the game for 8. Bxc4 Nd5
Mattison? Was it his 34th, or some 9. f4 Be6
move before that? Certainly by about 10. Qb3
move 30 or earlier, White was already Lasker notes, "So far White has
doomed to lose a pawn-and that af­ treated the opening to perfection; but
ter making no obvious error. As for here 10. Bd3, soon to be followed by
the black magic, nothing can fully h4, seems preferable."
explain that. This article will not en­ Thus White would utilize the
able anyone to go to his club and play moved pawn in front of Black's cas­
an ending as well as Rubinstein. But tled King and the fact that White him­
it will have achieved much if it gives self has not castled. This recalls
those of our readers who don't yet Pillsbury's own advice: "Castle be­
care much for endgames an inkling cause you must, or because you will,
of the uncanny subtlety that is pos­ never because you can."
sible in them. The text move permits Black to
force a drawish ending.
10. b5!
Pillsbury Crashes Home 11. Bxd5 Bxd5
Harry Nelson Pillsbury (1872- 12. Nxd5 Qxd5

- 226 -
 All Facets Of The Endgame: Endgame Wizards

13. Qxd5 39. h6 f4

Anxious for a full point, Pillsbury 40. h7 Resigns.
nevertheless walks into a drawish po­ Now set up the position in the dia­
sition rather than submit to disad­ gram below; and play the game
vantage. The art of playing to the througt again from move 18, com­
score is to ignore it. paring our remarks given below with
13. ... cxd5 the opinions you formed when play­
14. Nd3 ing througt the unannotated moves.
"White must try to reserve his
Knigtt for the ending," says Lasker, Gunsbe rg
"as the abundance of obstructions
leaves little scope to the Bishops and
the Rooks."
14. Nd7
15. Bd2 Rfc8
16. Ke2 e6
17. Rhc1 Bf8
We now ask the student to play
througt the rest of the game, which
is unannotated, and to draw his own
conclusions as to what errors Black Pillsbury
made and how he came to lose. After After 17 . Bf8
..

that, let the student turn to our de­

tailed discussion of the ending. White has a few small advantages,
18. Rxc8 Rxc8 the main one being that his King is
19. Rc1 Rxc1 already centralized. The second one
20. Bxc1 Bd6 is that White's queenside pawns are
21. Bd2 Kf8 both unmoved, whereas Black's are
22. Bb4 Ke7 disjointed (in particular, Black is
23. Bc5 a6 weak on the s quare c5). White's
24. b4 f6 Knigtt is situated better than Black's
25. g4 Bxc5 Knight; against that, Black's Bishop
26. bxc5 Nb8 is much more mobile than White's
27. f5 g5 Bishop-so that White is almost
28. Nb4 a5 bound to play Bb4 at some stage if he
29. c6 Kd6 is to achieve anything.
30. fxe6 Nxc6 The obvious play is 18. Bb4 at
31. Nxc6 Kxc6 once, but Pillsbury elects to swap off
32. e4 dxe4 the four Rooks. That costs him two
33. d5t Kd6 whole tempi-his Bishop is dragged
34. Ke3 b4 back to c1 and has to return to d2
35. Kxe4 a4 before it can resume its part in the
36. Kd4 h5 game. So Pillsbury's decision is quite
37. gxh5 a3 interesting. Certainly 18. Bb4 is the
38. Kc4 f5 move one would suggest to the ordi-

-227-
 OS Purdy On The Endgame

nary player, for it is against principle kingside pawns on the fourth rank,
to be the first to exchange Rooks where they would have aggressive
where there is only one open file. possibilities. The move in accord with
Apart from the two tempi, of elementary principles, King develop­
course, there would be no doubt ment by 25. Kf3, would have little
whatever as to the wisdom of swap­ meaning unless it threatened 26. e4;
ping the Rooks. Where slight end­ but that would not be a real threat
game advantages exist, their effect because of 26... dxe4t 27. Kxe4 f5t
normally increases with exchanges 28. Kf3 Nf6-White is then vulner­
of pieces. able in the center.
18. Rxc8 Rxc8 25. ... Bxc5
19. Rc1 Rxc1 "And here," wrote Lasker, "25.. .

20. Bxc1 Bd6 Nxc5 26. bxc5 Bc7 would have left
21. Bd2 Kf8 the game perfectly even, a Bishop
22. Bb4 Ke7 being more suitable to stop advanc­
Black has used his two tempi to ing pawns than a Knight." In those
centralize his King-White's advan­ days the Knight was generally
tage is now very small indeed. thought to be practically always su­
23. Bc5 a6 perior-hence Gunsberg's choice.
Very likely, Black could safely 26. bxc5 Nb8
have exchanged. In any case, it would Intended as a merely temporary
have been better to move the a-pawn retirement, preparatory to ...Nc6, this
two squares. The best formation for move momentarily disturbs the bal­
two pawns is the phalanx (abreast) ance in the center and gives Pillsbury
because then they are mobile. The the opportunity for some beautiful
safest phalanx is that of pawns that forcing play. Lasker gives 26... a5,
have never moved at all; but if one and certainly no win for White is then
pawn comes forward two ranks, it is demonstrable.
(as a rough general rule) usually 27. f5! g5
good for the rear pawn to join the Capturing on f5 would lead to the
advanced pawn-hence the common loss of the d-pawn, giving White two
maneuver of "straightening up the united passed pawns. The game is
pawns." Lasker states that 23... a5 now purely combinative--everything
would have "reduced White's chance had to be calculated to a nicety.
for a win to zero." 28. Nb4! a5
24. b4 29. c6! Kd6
Having gone to some trouble to 30. fxe6! Nxc6
play his Bishop to c5, White natu­ H 30 ... axb4, then 31. e7, etc.
rally consolidates it. This would be 31. Nxc6 Kxc6
the obvious play if only to free the 32. e4!
Knight. At last, the point-White must ob­
24. . .. f6 tain two united passed pawns. Even
25. g4 so, it is not yet clear that Black's wing
Threatening nothing more, at the majorities are worthless.
moment, than to "straighten up" the 32. ..
. dxe4

-228-
 All Facets Of The Endgame: Endgame Wizards

33. d5t Kd6 33. Rfxf6t Kg7

34. Ke3 b4 34. Rf5 Rxb4
In vain is 34... f5, as White must One of the first maxims of Rook
then queen one of his three united play is: Take. Taking something
pawns. frees the Rook for other work.
35. Kxe4 a4 35. Rxg5t Kf8
36. Kd4 36. Rxd6 Rf7
Here Black played 36... h5? which Advanced passed pawn versus
allowed White to win without trouble two retarded passed pawns-very
by 37. gxh5. If 36... Ke7!, then White frequently a draw, even though, as
wins only by the skin of his teeth. here, the two pawns are remote from
Thus, 37. Kc4 b3 38. axb3 a3 39. Kc3 the white King and united. It's like
f5 40. gxf5 h5 41. b4 a2 42. Kb2 a1=Qt an ace in bridge being given the same
43. Kxa1 g4 44. b5 h4 45. b6 � 46. value as two kings. The time factor
hxg3 hxg3 47. d6t, and White must enters.
queen a pawn with check. 37. Rd8t Ke7
This beautiful endgame win was 38. RggS Rd4
a worthy finale to Pillsbury's won­ Adjourned here.
derful effort. 39. RgeSt Kf6
40. Re6t
As nearness t o que ening i s
We're Off To White's asset, get nearer! With 40.
Beat The Wizard d6! White would threaten Re7; he pre­
To beat Watson in an end-game fers to leave the Rooks on, for com­
savors of Lese majeste-like baffling plications. Of course, it is good to
the Wizard of Oz with a card trick. complicate in a losing game; but if
Stefan Kruger thus did it in the Aus­ your position is merely unfavorable,
tralian championship. it pays in the long run to play for
restoring the balance-and in a
Kruger simple way if possible.
40. ... Kf5
41. Re3 b5
4 Already reducing the time lag.
Now Black should win-not easily of
course.
42. Kh2 a5
43. Rd6 b4
44. Rde6 Rxd5
Now it is clear that White is theo­
retically lost, but it is surprising
Watson how near he comes to justifying his
After Black's 30th tactics.
45. g4t Kg5
31. Bxd6 exd6 46. Kg3
32. Re6 Rxa4 As I am annotating by candlelight

-229-
 CJS Purdy On The Endgame

in a blackout and have the "flu," I now tied to the back rank until the
had some excuse for thinking 46. King can emerge again. But now
Re5t? might work; but a little analy­ Watson's fatigue tells. Had he sim­
sis shows that White's pawns would ply played 56. Rb5 for repetition, it is
ultimately fall. very doubtful if Black could force a
46. Rf6 win. If then 56... Rf8, trying to get
47. h4t Kg6 back again, 57. Kh4; and now Black
48. Re7 Rf7 cannot restore his King to g8, as the
49. R7e6t Kg7 King will simply be checked back to
50. h5 Rdd7! h8 again.
Kruger continues in good style. 56. Kh4? a4
The moment has come to interrupt The pawns are able to move a
the advance with a defensive mea- rank further up. Thus they get too
sure. near queening for White's coun­
51. h6t Kg8 terplay to be effective.
With mating possibilities created 57. Rge5 Rg8
at last, Watson is in his element; but 58. Rb5 b3
he was tiring. 59. Re1 Raa8
52. Re8t Rf8 60. Rb7 Rgb8
The importance of Black's 50th 61. Rc7 b2
(defense of seventh rank) is now 62. Resigns.
clear. Because Ree7 is too late to force
53. R8e5 Ra7 mate, Black being able to protect h 7
54. Rb5 simply by queening.

Two Watson Endgames

First Endgame
Some say that C.G. Watson is
lucky-that's superficial. Ingenious
"saves" are part and parcel of his
style. If our theory is right, you would
expect to find quite a number of near­
54. ... Rc8? misses in his games-brave attempts
Kruger falters. Better was 54... that ended only in glorious failure.
Rf1, threatening ... a4, etc. That would Well, you do. Here is a really amus­
force White's King back to g2, delay­ ing example from the Melbourne 1922
ing the counterplay. Then 55... Rf6 tourney. This was one of the two
56. g5 Rc6, followed by King out (or games Watson lost, and he has re­
if 57. Kf3, then 57... Rc3!, exchanging membered it with enjoyment long af­
one pair of Rooks). That would cer­ ter forgetting the details of his ten
tainly win. wins-for it was a Puckish game, and
55. Rg5t Kh8 Watson is, above all, the Puck of
A sad necessity. A black Rook is chess.

- 230 -
 All Facets Of The Endgame: Endgame Wizards

'�d those things do best 44. f8=Qt! Rxf8

please me 45. Rd5t!! Kc6
That befall preposter­ The Rook fights alone against two
ously." Rooks and a Queen. Of course the
point is clear now, but few players
Crakanthorp would have had the imagination to
build up the position from the first
diagram.
46. Rc5t Kb6
47. Rc6t Ka5
Can the Rook keep checking for­
ever like this? If it can, the game is a
draw by perpetual check or by stale­
mate. Actually, Watson gave three
more checks and then resigned-we
leave our readers to see why.
Watson And now it's time to mention that
White to play Crakanthorp, on move 43 at any rate,
saw the whole thing coming. The re­
Most players would resign here, lentless Viner, in the same circum­
but even a resignable position has stances, would have crushed it with
possibilities for Watson-he played: 43... Rf8 instead of queening. But
41. Be4 Kd6 Crakanthorp had a Puckish streak
The reply Watson had counted on, too, and he went to the trouble of
since it wins the Bishop outright and working out that he could let Watson
the black pawn can then queen bring off his ingenious coup and then
safely. wriggle out. The sight of the specta­
42. Rf5 dxe4 tors excitedly whispering and ges­
43. f7 b1=Q ticulating to each other when they
imagined that Watson was forcing a
draw and the thought of the letdown
they were shortly to get would fill
Crakanthorp with impish giee.

Second Endgame

He Beats Reti
[from the London 1922 tourney]
The onlookers still didn't know
what Watson was up to and thought Richard Reti was one of the
he was giving a shocking exhibition world's greatest players then. Wat­
of bad taste in not resigning. Now son's game with Reti went to 92
Watson brought out his secret wea­ moves, and we diagram the position
pon. after Black's 74th ( 74... Nc2).

- 23 1 -
 OS Purdy On The Endgame

Watson Nd4-perhaps Reti turned down this

line because he still cherished faint
hopes of a win.
8 75. ... Ne3
Still better seems 75... Nd4; for if
thereupon 76. Nc8, then 76... b5 77.
Nxa7 b4-and White's Knight finds
himself shut out. After the text move,
76. Nc8 should draw still-but Reti
goes for tricks.
76. Ne8!? Ke7!
77. Ng7 Ng2!
Perhaps Reti had missed this ear­
An interesting Knight ending. lier; he cannot now play his threat­
White can equalize the material at ened e6, and f5 will be refuted by
once by 75. K.xh6; but Black's passed ... Nh4t, giving up the Knight-after
pawns, being so remote from White's which the remote pawns should win .
King, outweigh White's central Things are getting desperate.
passed pawns which find Black's 78. Kf5 a5
King in front of them-even though 79. Ke4 a4
they are more advanced and White's 80. Nf5t Kf8
King can assist their further advance. Nicely safe from checks.
75. Kg6 81. Nd4
Maroczy's note, his final one on White has some hopes again.
this game, is "Kxh6 offered many What now?
more chances." Of what? Black could 81. ... Nxf4!
utilize the tempo to play 75 ... a5. If Quite crushing. If the Knight is
thereupon 76. Nc4, then 75. Kxb.6 a5 taken, the win is clear cut after ... a3.
76. Nc4 a4 77. Nxb6 (if 77. Kg6, then The e-pawn must fall, and Black will
77... Nd4!) a3 78. Na4 a2 79. Nc5t win on one wing or the other.
Kd5 80. Nb3 Nd4 81. Na1 Ke4 82. Kg5 82. Nb5 Ng6
Ne6t. Black wins one pawn and has 83. Kf5 Kg7
time to win the Knight on a1 and to 84. e6 Ne7t
get out again-meanwhile giving up 85. Ke5 h5
his own Knight for White's remain­ And after a few more moves,
ing pawn (but not too soon). Or 76. White resigned.
Kg6 Nd4! If then 77. f5t, not 77...
Kxe5? but 77... Nxf5!, giving up the
Knight. Endzelins Again
The logical move was 75. Nc8!­ M.E. Goldstein
to be sure of annihilating one of the
remote pawns, and doing so before Endzelins displayed his endgame
the other one can get too far along. wizardry against Dutchman Baay. A
Black could then, if he so desired, peaceful Caro-Kann reached the fol­
force a draw with 75... Kf5 76. Nxa7 lowing position after 33 moves.

-232-
 All Facets Of The Endgame: Endgame Wizards

L. Endzellns 41. Qg3 g5!

42. f5
Leads to quick disaster. But 42.
fxg5 Ng6 leaves White's pawns too
weak.
42. Qd4
43. fxe6 fxe6
44. a4
White suddenly finds himself with­
out a good move, the a-pawn being
threatened by ... Qa1t.
J.G.Baay 44. Ng6
Blaek to move 45. Qh2 Qxe5
46. Qxe5 Nxe5
At first sight you will say, "Queens And Black won after another 10
still on the board, pawns balanced­ moves, penetrating to b4 with his
no interest, must be a draw." But King.
Black has some small assets: White's A very fine ending which the stu­
queenside pawns on the same color dent should find most instructive.
as White's Bishop, the white a-pawn
under pressure, control of the dark
squares, and the fact that Queen and Game Of A Lifetime
Knight combine better than Queen
and Bishop. CJS Purdy-D. Hamllton
Let's see how the magician pulls Australian Championship
the rabbits out of the hat. (Brisbane 1967)
33. ... a5! [We bring this c oll ec tion o f
Preventing b4 and fixing the endgame articles by CJS Purdy to
pawns on light squares. a close with an example of what
34. Kh3 Qc5! Nimzovich c alled "Heroic Defense."
Centralizing the Queen. For some background on endings
35. Qe1 Nb6 with Rook v s. two minor pieces,
Aiming for h4. see the articles ROOK AGAINST
36. Qd1 Nc8 TWO MINOR PIECES and MINOR
37. Kg2 Ne7 PIECES FOR A ROOK in "The
38. Qd3 Nf5 Search for Chess Perfection." The
39. g4 Nb4t! game displayed here was also re­
40. Kf1 print ed a s Ga m e 49 in "The
If 40. Kg3 or 40. Kh3, then 40. .. Search.... "]
g5!, undermining the support of the
e-pawn. The Queen and Knight may This was the most astonishing fin­
then threaten mating attacks. ish in Purdy Senior's long career, and
40. ... Kc8 probably that of Hamilton's then
To free the Queen, which need no short one, even though the astonish­
longer watch d6. ment was for Hamilton mixed with

- 233 -
 CJS Purdy On The Endgame

grievous disappointment. Still, he 72. ... a4

saw the funny side, and congratu- 73. h4 Nc2
lated his opponent with genuine en- There was a simpler procedure:
thusiasm. 73 . a3 74. h5 a2 75. Rf1 Bd4 76. h6
. .

The first diagram shows the posi- a1=Q 77. Rxa1 Bxa1 78. h7, and now,
tion in C.J.S. Purdy-Hamilton after provided Black makes the right tour
the second adjournment. with his Knight, he can retain both
pieces and stop White queening.

10 "Very slow resign­

er, this Purdy."

Three considerations, luckily for

Purdy, deterred Hamilton:
a) it took some calculation;
b) the ending Bishop and Knight
After 68. Rgd4 versus King is at best very tedious;
and
At first sight, it seems that Black c) (most important) Black saw an
should win on sheer material. This obvious way he could get a whole
is why I spent the whole of my first Queen, and White not. The game
Thursday morning (free day) analyz­ looked a pushover. It might have
ing 68... Kc7 69. Rd7t Kc8 70. Rd4. I needed an Averbach to see the fal­
came to the conclusion that because lacy in c).
of Black's bad King, White had a 74. h5 a3
draw in all variations, and covered 75. Rf1 a2
several sheets of paper with them. 76. h6 Bd4
Hamilton came to the same con­ 77. h7 Na3
clusion, and correctly gave up his Too late now for ...al =Q, but the
pawn plus. He had sealed his 68th. text move seems to force White to
68. ... Bg5 resign, so what the hell.
69. Rg4 Bd8 78. Rd1
70. Rd4 Bb6!
This worried me, as I simply had
not had time to analyze it. Rook will
draw against Bishop and Knight
with level pawns all on one wing,
but this is different. Easy draw if
Black's passed pawn were a rank
further back; but given where it is,
it's a runner.
71. Rxh4 Kb5
72. Rf4
The only hope.

- 234 -
 All Facets Of The Endgame: Endgame Wizards

ter, but it wouldn't matter! We re­ Black is a Queen and Knight up,
turn to this point later. but his King is one square too far
79. Kf7 Nb1 away, and so is his Knight. Black
Looks a killer. "Very slow re­ must allow either stalemate or pro­
signer, this Purdy," some onlookers motion.
thought. Now return to diagram 11. Instead
80. Kg8 of 78... Bh8, try 78... Bb2. Then 79.
First shock. If 80... a1=Q, then 81. Kf7 Nb1 80. Rd8; and the same old
Rxb1 t forces a book draw. joke is on, as Ra8 is one threat. Or if
80. ... Bb2 78... Bc3, then 79. Rc1 Nb1 80. Kf7
81. Rd8! a1=Q 81. Rxc3 Qxc3 82. Kg8; and
Just the same if White's King were White still draws.
only at f7. So the tempo lost by 78... Bh8 did
81. ... a1=Q not matter.
Nothing else to be done now.
82. Rb8t Kc6 Exercises
83. Rxb2! QaSt
Taking the Rook would be useless
even with the white King still on f7,
as Kg8 would follow, with book draw.
84. Kg7 Qa7t
85. Kg8 Qg1t

�� � � ���
�
� � � �
� � � � r 12
�
� � � � � � �
�
� � � : � � � After 87. Kh8
�
� � � � � � �
�
� � � � �
� � [From the final game position af­
� � � � � �
�.:� �
� .... � � LY- � ter 87. Kh8, solve the following two
puzzles:
i) Move the black King to d7;
86. Rg2!! then, Black to move and win;
White cannot afford to keep his ii) Move the Knight to c2; then
Rook-nor can Black usefully avoid Black to rrwve and win.
taking it. These two exercises demonstrate
86. ... Qxg2t the force of Purdy's comment after
87. Kh8 Drawn. move 87 in the game.]

- 235 -
 End
of the
Game
 CJS Purdy on the Endgame

RULES FOR PLAYING ENDGAMES

Chapter 1.
§ + ft vs. §
1) The defense draws against any
pawn that is not more than one rank
beyond mid-board (a white .ft on
rank 5, e.g.) when the defender's \t1
stands on the queening square and
the defender's .§ is on the rank next­ Either to move/Draw
but-one to the queening rank (a de­
fending black K on rank 6, e.g.­
Cheron). If the .ft is on the rank next
to or next-but-one to the queening Chapter 2.
rank, the defender draws, provided
that he can safely give check-to do ft Endings
that the defender should try to get
his .§ behind the .ft . 4) The Square.
Look at the .ft, and run your eye
2) The defending .§ is strongest along the diagonal that leads from
when it has a clear interval of three the .ft to the queening rank. If the
or more files or ranks between itself enemy Wis "in" the square of which
and the opposing .ft. In future con­ that line is the diagonal, or if the en­
siderations, Rules 1 and 2 will be emy W can get into that s quare be­
used many times-they should be fore the .ft advances, then the en­
absorbed thoroughly. emy \t1 can stop the pawn from
queening-otherwise not.
3) [For an a- or h- .ft on the 2nd,
3rd, 4th, or 5th rank in positions such 5) The Opposition.
as that of the diagram] The defender "When the \ts are on the same rank
draws by threatening to exchange or file, with one square between
.§ s, provided that his W is no far­ them, as in the diagram, we notice
ther from the queening s quare than that each one prevents the other one
the attacker's \t'. The defender then from advancing, either straight
draws with or without the move. ahead or diagonally. Therefore, the

-238-
 End of the Game

first W that has to move is at a dis­ 8) When the 'it's are on opposite sides
advantage; for wherever he goes, he of the ft s' file, the W who can first
gives the opposing W the option of attack the enemy .ft from the rear
either advancing or of again "taking wins the enemy .ft and saves his own.
the opposition," provided there is no
obstruction.

Whoever moves wins

Either to play/White wins

9) When the 'it's are on the same side
of the .fts' file, as in the below dia­
6) [For any .ft. other than an a- or h­ gram, if one W can reach the rank
.ft.] occupied by the enemy .ft then he
6A) W in front of his .ft. and on the wins the enemy .ft and saves his own.
rank next-but-one to the queening The W in question can always do
rank: Always a win. what is required if he can reach, with
6B) W in front of his .ft. and not on the opposition, the rank occupied by
the rank next-but-one to the queen­ his own .ft..
ing rank: A win only if the W has
the opposition or if he can advance
or if his ft is more than one rank
behind him.
6C) <if? not in front of his .ft.: A win
only if the W is on the rank next­
but-one to the queening rank and his
.ft. can go to the rank next to the
queening rank without checking.
6D) The defending W can move onto
the square immediately in front of Black to move/White wins
the .ft.: Always a draw.

7) [For an a- or h- .ft.] Against an a­ 10) Two passed .fts always protect

or h- .ft., the defense always draws if each other from capture unless they
the def ender can either reach the are not beyond mid-board, away from
queening s quare or if he can keep their home rank, and have two files
the opposing W confined to the file between them.
of the .ft..

-239-
 CJS Purdy on the Endgame

11) Advance the ft that cannot be 16) In ft endings, when you have a
blockaded by an enemy ft . chance of taking the opposition or of
preventing the opponent from taking
the opposition, do not neglect to do
so without good reason.

17) When in doubt, retreat your W

and gain distant opposition rather
than advance and lose the opposi­
tion.

12) In symmetrical ft positions, the Chapter 6.

player who has to move first will run
out of moves first-unless he can Endgame Prin­
blockade both the enemy ft s with his
first move. ciples
13) For purposes of gaining moves 18) Seek out the enemy's weaknesses
with .ft s, always move a ft that has and if possible induce him to create
already moved rather than an un­ more, but do not try to force him by
moved .ft. playing weakening moves of your
own-that is, until you have obtained
14) Once thews can no longer inter­ a crushing advantage and can afford
fere with each other's business, i.e. to launch out. All the time, your first
the game has become a race, never care must be to avoid giving coun­
calculate from move to move in the terchances.
ordinary way. Simply count up the
moves the first queener will queen 19) "U! and ft versus lone "U! is nor­
in, and then see how far the oppo­ mally a win only with the c-, d-, e, or
nent's queening ft will get in the f- ft that is on the rank next to the
same number of moves. queening rank.

15) In a position where it is your move 20) The rule for the defending player
and where the opponent must arrive in 'fk endings is to have his \f;' either
at his present position every alter­ very close to the scene of action (so
nate move, imagine that you have al­ that he can afford to exchange tks),
ready played two, four, six, or eight or, failing that, as far away as pos­
additional moves (neglecting ft ex­ sible (so as not to limit his choice of
changes) and that you then have to checks). The middle way is often fa­
make your move. Then ask, "Do I tal.
win?" If the answer is no, ask, "Can I
lose or gain a move somewhere?" 21) In a middlegame, one feeble piece
need not be a serious handicap; for

-240-
 End of the Game

it is only one unit among many. But 28) For the case of ws on the same
with very few pieces on the board, side of the file of the two blocked 1 s
the immobility of one means a pro­ (and each 'ill on its own .ft 's side),
portionately bigger burden to carry the 'lf( who occupies a limit of the
for the others. enemy ft wins that .ft and holds his
own .ft.
22) To get the best out of a .Q., avoid
clogging his diagonals with ts.

23) Before ever beginning to think of Chapter 8.

making a passed .ft, get all your
pieces into as good positions as pos­ � Endings
sible.
29) Place your t:! behind a passed
24) To win, exchange off pieces; to ft .
draw, exchange off .ft s.
30) Put your t:! BEHIND a passed ft
if the ft has crossed or can cross the
middleline. Put your t:! IN FRONT of
Chapter 7. a passed .ft if the .ft is held in its
own half of the board .
.ft Endings
31) Avoid putting a § in a position
25) Avoid ft moves while you are fol­ of IMMOBILE DEFENSE. In the case
lowing Rule 23 because .ft. moves cre­ of an attacked, unmoved a- or h- ft ,

ate lasting weaknesses and thus usually prefer to give up the ft rather
make your task harder. than to tie up a a for its defense­
instead, use your a for attack in
26) Try to free your position from some way.
weaknesses; and, if possible, make
it hard for the opponent to do like­
wise.
Chapter 9.
27) If you are a .ft. up: Exchange pie­
ces {not ts) wherever you can do so 4J Endings
without losing in position. Exception:
do not rush an exchange that will 32) Queen a .ft. if you can.
produce .Q.s of opposite color {though
if such an exchange is pressed on 33) In the endgame � and .ft vs. �.
you, evasion may be worse than ac­ if the .ft. is on the rank next to the
ceptance). Corollary: refrain from queening rank and if it can be sup­
forcing an exchange that will give ported by its W on that same rank/
your opponent two .Q.s against .Q. and Always a win, unless the defender's
�- W can occupy the queening square
and cannot be ousted therefrom.

- 24 1 -
 CJS Purdy on the Endgame

43) In the opening and in the middle­

Chapter 10. game, a .§ can almost always be
given up safely for a .Q. and two is.
A Endings Also for a 4) and two is, provided
that the player with the extra .ts has
34) As far as possible, place your .1t s them situated safely. Very weak iso­
on the squares of opposite color to lated .1t s will upset calculations.
that of your own .Q..
44) In the opening and in the middle­
35) Use your W to command squares game, the loss of a .§ for a Minor
of the color not controlled by your Piece and one .1t is usually less seri­
own .Q.. ous than the loss of a ft .

36) Attempt to conduct maneuvers 45) At all stages of the game, a .Q. is
on the wing not occupied by the en­ better than a 4) against a .§ . In the
emy <;ti. opening and in the middlegame , an
extraordinarily well-placed 4) is an
37) Endeavor to secure a passed .1t , exception, i.e. a ft -supported undis­
or at least complications, on both lodgeable advanced 4) in the enemy
wings simultaneously. camp, on a central file or on a file on
the wing of the enemy castled w.
38) Keep your .1t s on squares of the
opposite color to that of your .Q.. 46) A vital question to ask yourself
when pondering the merits of such a
39) In endings with .Q.s of opposite transaction ( .§ for Minor Piece and
color, material means NOTIDNG, po­ one or two ft s or vice versa) is, "How
sition EVERYTHING. is the position for .§ s?"

Chapter 12. Chapter 16.

�vs. A/4J Sure Draws(?)
40) If the .§ 's side can create a 47) Advance your own w as far as
passed .1t, the value of the Exchange you safely can, and keep the enemy
increases, even though in the process W back.
more of the enemy .1t s become
passed. 48) Use the .§ to cut off.

41) A ft down does not lose when all

ft s are on one wing only.

42) To be confined to the defensive

is itself a disadvantage.

- 242 -
 End of the Game

Explanations of Files, Ranks, Symbols,

and the Queening Square.

Black Side

White Side

Explanation of Symbols
x = captures

t = check

+- = White has a winning advantage

-+ =Black has a winning advantage

± =White has a definite advantage
+ = Black has a definite advantage

;!; = White has the better game

+ = Black has the better game

oo = Unclear

11 = with the idea

N/f3 = Knight on f3 (for example)

•

Only some of these sysmbols may have been used

in the current book.

- 243 -
 CJS Purdy on the Endgame

BIBLIOGRAPHY

Euwe, Dr. Max & Hooper, David A Guide to Chess Endings (Dover 1976)

Fine, Reuben Basic Chess Endings (David McKay 1941)

Matanovic, Aleksandar eneycl(Jj}edia of chess endings, vol. 2 (Rook and

. Pawn) (Sahovski Informator 1985)

Miiller, Karsten & Lamprecht, FrankFundamental Chess Endings (Gam­

bit 2001)

Miiller, Karsten & Lamprecht, Frank SECRETS ofPawn Endings (Every­

man Chess 2000)

Nunn, John Secrets of Minor-Piece Endings (Henry Holt 1995)

Purdy, C.J .S. Guide to Good Chess, 12th printing (Thinkers' Press 2002)

Speelman, Jon; Tisdall, Jon & Wade, Bob Batsford Chess Endings (Batsford
1996, corrected edition)

Many acknowledgments were made in this book to Andre Cheron's

Traite Complet d'Echecs. Half of Cheron's work is devoted to the
endgame and forms one of the very best treatises on that department
of the game. [Andre Chiron (1895-1980) went on to write a large
book, Nouveau Traite Complet d'Echecs, devoted entirely to the
endgame; he later revised and erpanded the work into a 4-volume
German version, Lehr- und Handbuch der Endspiel An English trans­
lation of the German edition would be rrwst welcome.]

-244-
 End of the Game

THE THINKERS' PRESS, INC. PURDY LIBRARY OF CHESS

Each volume of the Library is a stand-alone selection of material from the

published writings of C.J.S. Purdy.

Published Titles
1. Guide to Good Chess (12th ed., 2002)
2. The Search for Chess Perfection: The Life, Games, and Writings of
CJS Purdy (1997)
3. How Purdy Won: The Correspondence Chess Career of a World
Ch<lmpion (2nd ed., 1998)
4. CJS Purdy's Fine Art of Chess Annotation and Other Thoughts Vol.
1: 100 Annotated Games, Purdyisms, The Purdy Player (1992)
5.Extreme Chess: The World Chess Championships of1935, 1937, 1972
(An all-in-one-volume algebraic-notation reissue of Purdy's books: How
Euwe Won, The Return of Alekhine, How Fischer Won) (1999)
6. Action Chess: CJS Purdy's 24 Hours Opening Repertoire (An edited
version of the series of articles THE OPENINGS IN 10 HOURS plus material
on the Colle System) (2001)
7. CJS Purdy's Fine Art ... Vol. 2, 120 Annotated Games plus 13 Articles
& Studies
8. CJS Purdy's Fine Art ... Vol. 3, Purdy on Purdy, Annotations to 70
over-the-board games of CJS Purdy and to 30 games of John Purdy,
Purdyisms (revised), The Purdy Player, The Road to 2000 (How to Become
an Expert), and several other articles.
9. CJS Purdy on the Endgame by CJS Purdy, with additions and correc­
tions by Frank Hutchings, John Purdy, and GM Karsten Millier. Includes a
short course on endings and a much longer course on All Facets of the
Endgame.

Titles Under Consideration (all in algebraic notation)

10. CJS Purdy's Fine Art ... Vol. 4: Annotations to 100 games from 1936-

1945 and to 35 games from 1976 - 1979.

11. Chess Bits and Obits Articles on the History of Chess and Obituary
Notices for 14 Chess Masters (appreciation of each master's career plus a
few of the master's best games-68 annotated games in all)

And a Supplemental Title (see next page)

- 245 -
 CJS Purdy on the Endgame

Supplemental Book

12. Chess World Editors: Lajos Steiner & Maurice Goldstein. Approxi­
mately 320 pages of sensational games (about 300) and articles by top edi­
tors and masters from the Land Down Under.

- 246 -
 End of the Game

COLOPHON

This book was set in Century 725 BT condensed, with

DIAGRAM being the chess diagram font C.R. Horowitz®
from Thinkers' Press, Inc.
•

Cover: Blue Sky Communications

Layout and design: Nate Long
Keyboarding: Nate Long, Ralph Tykodi
Compiling: Ralph Tykodi, Frank Hutchings
Design proofing and Printer Preparation: Bob Long
Proofing: Nate Long, Karsten Muller, Ralph Tykodi, Bob
Long, Frank Hutchings, John Purdy

Special Thanks to . . .
For contributions concerning the Vidmar-Alekhine, San
Rerrw 1930 game:

Phil Millett
Jim Byrd
Maurice Carter
Karsten Mtiller

-247-
 ARES

C magazine, whose primary goal is

hess for 2003 in BBUAREB

to bring the enjoyment back to chess in

printed form. A 64-page {minimum) pub­
lication with games, stories, articles... and
a minimum of news.
Friendship around the country, tips and
columns by new writers and experts on
their subject, from all over the world.
A subscription is $30.00/4 issues in
the USA; $37.50 international; $45.00
international airmail.
For those adults who do not have time,
a quarterly publication eases that lost en­
joyment of a game which brought so
much to us.
Here is a sampling of the first issue:

• Behind the scene in Bahrain

• Louis Paulsen-Hypermodernist!
• Developments in Rook Endings
• A recent overview of the King's Indian
• Dealing with The Old Guy
• Rex Stout on GAMBIT
• Best Recent Games... and a ton more

BBUAREB
Thinkers' Press, Inc.
1101 W. 4th St.
Davenport, IA 52802 USA

1-800-397-7117
This logo is not only a symbol for Quality
books, but a recognized mark throughout
the world.

Thinkers' Press, Inc., since 1973, has pro­

duced fine chess books, books on magic, fic­
tion, poetry, religion, and science ... though
our primary emphasis is on chess. Some of
our hits, at many fine stores near you, are:

• Practical Chess Analysis by Buckley

• Leonid Stein, Master of Risk Strategy by
Gufeld & Lazarev
• Super Nezh by Pishkin
• How Purdy Won by Purdy
• Guide to Good Chess by Purdy
• The Richter-Veresov by Gufeld & Lazarev
• The Great Evans Gambit Debate by Rhode

We are on the internet at:

www.thinkerspress.com

Thinkers' Press, Inc.

1101 W. 4th St.
Davenport, Iowa
52802USA
1-800-397-7117

Should you be interested in

retail chess merchandise,
request our free color catalog.