Jan Timman

The Art of the Endgan1e

My Journeys in the Magical World of Endgame Studies

New In Chess 2011

For Geertje

© 2 0 1 1 New In Chess
Published by New In Chess, Alkmaar, The Netherlands
www.newinchess.com

All rights reserved. No part of this book may be reproduced, stored in a

retrieval s ystem or transmitted in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the prior
written permission from the publisher.

Cover design: Volken Beck

Translation: Peter Boel
Proofreading: Rene Olthof, Steve Giddins
Supervisor: Peter Boel
Production: Anton Schermer

ISBN: 978-90-569 1-369-4

Contents

Preface 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7

Chapter 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11
Miniature Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Chapter 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 22
Rook versus Bishop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Chapter 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 32
Preventing Pawn Promotion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Chapter 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 56
Various Promotion Combinations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Chapter 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 81
Knight Promotions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Chapter 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 99
Bishop Promotions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

Chapter 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 120
Mating Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 20

Chapter 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 147
Stalemate Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Chapter 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 168
Mutual Zugzwang . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 68

Chapter 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 183
Building a Fortress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

Chapter 11 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 199
Systematic Manoeuvres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 99

Chapter 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 224
The Disappearing Trick . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

5
The Art of the Endgame

Chapter 13 ... , ................................... 248

Three Themes . . . . . . . . . . 248
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I. Troitzky IAmelung position ........................... 248

2. Valladao Task . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 2
3.llquidation to a pawn ending. .... .. ................ 255 . . .

Chapter 14 ....................................... 260

Various Endgame Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
I. Studies with others.......... ..... .. . ..... . . . . ....... 260
2. Studies with no specific theme........... ............. 262 .

Explanation of Symbols ..... ... .... ... . ............. 266

Index of Composers ................................ 26 7
Acknowledgements ....... ..... . . . . . ............... 269

6
Preface
In my early youth, there were fascinations. I imagined how it would be
to live in a house in the woods. Castles would bring me to raptures. I
remember my excitement when I was standing with my parents on the
ferry that transported visitors to Loevestein Castle. I dreamed that on
the playground in our neighbourhood there was a castle you could
play in.
In the summer holidays, we travelled along the Rhine in a Ford
Taunus - my father behind the wheel, my mother next to him and we, the
four kids, pressed together on the back seat. Our destination was Tuscany:
Siena and Lucca, with its all-embracing city wall, which we rounded
again and again during our long walks.
Along the Rhine stood the robber baron castles with their impressive
names: Reichenstein, Rheinfels, and the Furstenberg, the ruin that tow­
ered high above the Rheindiebach village. There we used to find lodgings
in a little hotel, concluding the first stage of our trip. At home, in the liv­
ing room, I would build Lego cities and Citadel castles.

M.C. Escher's 'Metamorphosis' starts with black-and-white squares that

transform into salamanders. The salamanders become beehives. Out of
these, bees emerge, which are ingeniously transformed into birds. Then
there are fish, and in the accompanying pattern - birds again.
Here the 'Metamorphosis' arrives at things made by human hands. The
new birds transform into blocks. Out of these, houses in a South-Italian
town are formed. Now the systematic pattern of forms is broken. A bridge
connects the town with a chessboard. On a3 is a white rook, which also
exists in the other reality. On the chessboard, a position is depicted where
White perishes due to a smothered mate. And then Escher is back at the
black-and-white squares.
From building miniature cities to chess: this was a step I already made
in my youth.

The world of chess was the most fascinating. I devoured Euwe's books.
The memory of the game Reti-Alekhine, annotated in his book Practical
Chess Lessons, which never ceased to amaze me: the black rook appearing
on e3 and remaining en prise there for several moves. Another game that
was indelibly printed in my memory was Tartakower-Botvinnik from
Hans Muller's Mikhail Botvinnik, One Hundred Selected Games. It was an im-

7
The Art of the Endgame

pressive image: the white king was pinned down in the corner by a knight
on f3, supported by a mighty pawn front.

Within the romantic world of chess there was a sector that was even more
hermetically closed off from the outside world. I first became acquainted
with it when I read Herbstman's book 'De schaakstudie in onze dagen'
(i.e. 'Contemporary Chess Studies') . In this new world there was no
struggle in the sense of a man-to-man fight. The competitive element was
gone - only scientific and artistic elements remained. The studies by
Troitzky and Kubbel had a special magic. I felt that their masterpieces
were made for eternity.
It is the strict regulations that lend a special status to the endgame study.
White always has only one move to reach his goal, whether it is to force a
win or a draw. Black, on the other hand, may have alternatives, and these
extra possibilities can enhance the content and the richness of a study.
In practical chess, you think about what lies in the distance. You make
plans and calculate variations. Creating an endgame study requires a sig­
nificantly different thinking process. As a rule, the endgame study com­
poser has an end position in mind, which contains a spectacular hidden
point. Then comes the stage of retrograde thinking: he investigates the
position backwards, asking himself: 'What was White's last move?' and
then: 'What was Black's last move?' Today I still ask myself these questions
in my mind.
"
r composed my first studies in the early 1 9 70's, when I had just finished
my secondary school education and the world was my oyster. Now I look
back on this, four decades later, it strikes me how uncomplicated these
early pieces of work look. The points are nice, but the play misses the
depth of my later studies. Not until the 1 9 80s did I manage to create a few
high-level studies.
Nowadays I can hardly imagine how it was to compose studies without
a computer. In those days it was extremely difficult to work out all the tac­
tical details. You could spend days eliminating little tactical tricks that up­
set the play. And then sometimes, there would still remain questions.
Nowadays you can simply move your pieces and pawns to the right
squares on the analysis board of your computer. Tactical snags are often
useful for adding new elements to the study.

In December 20 I 0 , I decided to write The Art of the Endgame. I wanted to

show systematically, guiding the reader along themes and genres, why
endgame studies are so beautiful and so interesting. Especially the system-

8
 P refa c e

atiq were important. It had struck me that most books on endgame stud­
ies were quite random collections. My intention was to reveal the secrets
from the world of endgame studies in 1 4 chapters. I also had a second in­
tention: in 1 99 1 , my book Schaakwerk II had appeared in Dutch (later
translated as Studies and Games) . In the second part of this book I had in­
cluded a number of endgame studies. I wanted to put this material under
the microscope one more time. With the help of the computer I would
establish if the studies were correct, after polishing them up in some
cases.
In January 2 0 1 1 , I started my journey in the world of endgame studies.
I kept making new, fascinating discoveries. This stimulated me to create
new studies. Before, I had never occupied myself with the question how
you could incorporate the Novotny theme and the related Plachutta theme
in studies. Studying a number of examples pointed me in the right direc­
tion. Never before have I been so productive as an endgame study com­
poser as in the seven months that I wrote this book. It was a sensational
experience. Whether I looked at bishop promotions, systematic ma­
noeuvres, or the Valladao Task, I kept finding new paths.

The computer has refuted a number of studies by great composers. Nowa­

days there is a trend to publish such refutations as if they were great
achievements. In this book I have striven to use the computer in a con­
structive way. An impressive work of art that has been damaged, must be
repaired with great care. I have treated studies with such defects in the
same way. I corrected them in ways that did not affect the brilliant ideas.

In the first chapter I discuss miniature studies, i.e. pieces of work that are
often closely related to the theoretical side of the endgame. The second
chapter continues on this path, and is purely theoretical. Then we come
to the real romanticism. Pawn pro�otion is an important factor in chess.
First I discuss several methods to prevent pawn promotions, and after
that the various promotion combinations are reviewed. I have given a
representative picture of the themes. As with every theme, I cannot pre­
tend to be comprehensive; there are more than 7 5 , 0 0 0 studies in exis­
tence!
In the chapter on knight promotions, I have had to restrict myself even
further. This is the type of pawn promotion that is most frequent after the
regular queening, in practice as well as in studies. I have exclusively dealt
with knight promotions that repeat themselves according to a certain pat­
tern. With bishop promotions I could paint a slightly more comprehen­
sive picture again, even though I was forced to make choices here as well.

9
The Art of t h e E n d g a m e

In the chapters on mate and stalemate, I have selected those studies that
appealed to me most. That was anything but easy Each study (or game)
ends in mate or stalemate, and the motifs can be surprising and beautiful.
Mutual zugzwang and building a fortress are themes that can be of
practical use. It is remarkable how often strong players do not recognize
these themes. For this reason, I think that Chapters 9 and 1 0 are highly in­
structive.
On the other hand, the next two chapters contain themes that are en­
tirely disconnected from practice. Systematic manoeuvres do not occur in
games. Nor will a player strive to lose pieces in the course of a variation.
However, it is worthwhile to see how these offerings are justified in
studies.
In the chapter called 'Three Themes' I have dealt with three of my fa­
vourite themes. The final chapter is a collection of remaining studies.
Again, I had to make a selection here. There were studies from the past
with nice ideas, which nevertheless had to be rejected for publication, ei­
ther because the computer refuted certain variations, or because the varia­
tions were insufficiently special.

And that is how my seven months' journey came to an end. It was not
only a fascinating experience, but also an instructive one. It seemed as if I
was finally allowed to probe depths in the endgame that had been hidden
from me before. I have tried to describe and explain my findings as clearly
as possible. And I sincerely hope that the depth and beauty of endgame
studies will also move the reader.

]an Timman,
August 201 1

10
Chapter1

Miniature Studies
A miniature study is a study with white knight is also under at­
no more than seven pieces in the tack. But now White changes
starting position. With minimal tack.
material, the composer must weave 4. .tb1 xe4+ c;f;>f5-g5
the maximum amount of finesses 5. �e4-d51 c;t;>g5-h5
into the position. A classical exam­ 6. .td5-f7+ wh5-g5
ple is the following study by the 7. �f7-e8
brilliant Russian composer Mark
Liburkin.

And the black bishop is lost.

This miniature study contains
Liburkin, '64' 1933 some themes that we will touch
White to play and win upon later. It starts with a pin. The
final position shows a case of 'mu­
First of all White pins the black tual zugzwang': if it were his
knight, in order to capture it move, White would not be able to
next. win. Furthermore, the 'domina­
1. wc2-d1 'it>g6-f5 tion' theme makes an appearance:
2. <ittd 1-e2 the black bishop has no squares
Because his bishop is in the way, and is dominated by the white
Black cannot unpin his knight. pieces.
However, he has a resource in
store. I have composed a modest num­
2. �e5-f6 ber of miniature studies myself; it
3. c;f;>e2-f3 ..ltf6xh4 is not easy to build in enough
Black appears to save himself, points and finesses with so little
since after 4 . tt:lxh4+ � g S the material.

11
T h e Art of the E n d g a m e

The black king has been driven

back, but how should White pro­
ceed? After 4.l:rc3 l::th l + 5 .\t>b2
.!:[h2 + 6 .l::tc 2 Black can comfort­
ably exchange the rooks.
An elegant idea is 4.\t>b2 , as after
4 . . . l::th 2+ 5 .\t>c3 l::tx a2 6J:te5 Black
can only prevent the threatened
mate on aS by giving up the rook.
However, this idea does not work
Timman 1983
since Black can play 5 . . . cS instead
White to play and win
of S . . . l:rxa2. After 6 .bxc5 WbS
In this position it is hard to foresee White loses either his knight or his
that the black king can be drawn pawn, and in both cases the draw is
into a mating net. Yet, that is pre­ inevitable.
cisely what this study was based For this reason I initially thought
on. that there was no win for White
1. ti:Jc3-a2 here. Until I found the plausible
Now Black has two possibilities. solution to the problem of the po­
sition, which lies in 4.l:re7!. The
A) l. .. Wb3 The most obvious penetration by the black king with
move. The black king penetrates in 4 . . . \t>b3 would now be parried by
order to disturb the cooperation s.J::tx c7 l::th l + 6 . l:rc l l:rxc l +
between the white pieces. v.t•b1 7 . \t>xc l , with an easy win for
By protecting the knight with the White. 4 . . l:rhl+ s .<ot>b2 l:lh2 +
.

king, White keeps his rook mobile. 6.'it'c3

After 2 . l:re2 l:rh6 3J:lb2+ �a3
4.\t>b l l::th l + S .tt::l c l c6 he would
have no winning chances whatso­
ever. 2 ... l:lh6 3.l:le3+ �a4

6 . . Jha2 is again met by 7 .l::te 5 ,

which means that Black again has
to try his luck by pushing the
c-pawn: 6 ... c5 Hoping for 7 . l:ra7 +

12
 Min i a t u re S t u d i es

WbS 8 . l::i:a 5+ 'it>b6 , with a draw.

However White now exploits the
strong position of his rook in an­
other way:
7. 'it>c4 Threatening mate in one.
After both 7 . . . l::th4+ 8 . Wxc5 and
7 . . . l::t c 2+ 8 . tLlc 3 + White wins
without trouble.
B) I..JlbS A somewhat artificial
move which is, however, con­
Matous, 'Shakhmaty v SSSR' 1984
nected with a clear plan: Black
White to play and win
wants to carry through the saving
thrust . . . c7 -cS as soon as possible. White obviously tries to force
2.l::te3 Black to give up his rook for the
b-pawn.
1. b6-b7 J:!.d6-d8
After 1 . . . l:rb6 2..�,e3 + Black would
lose immediately.
2. �g5-f4 e7-e5
3. t2Jd2-f3+ �d4-e4
Attacking the knight, so that cap­
turing on eS brings White nothing.
However, White still has a knight
check.
4. t2Jf3-g5+ 'it>e4-f5
5. .tf4xe5
Only this rook move wins. Of Here, 5 . tb£7! also wins.
course, after the rook exchange it 5. ... J:!.d8-e8
would be a draw, and 2 . l:re7 cS The alternative 5 . . . �g8 fails to
3 . l:rc7 'it>d4 didn't bring anything 6. tb£7 We6 7 . tLlh6, followed by
either. 2... c5 8.b8'i?. After the text move White
What can White do now against cannot promote to a queen in view
the exchange of the last pawns? of 6 . . . l:rh8+ with stalemate. At the
3.tbc3 This is the answer. White same time the bishop cannot
sacrifices his b-pawn. 3 .. .l:!xb4 budge, since it has to keep control­
4. Wc2 and mate on e4 cannot be ling the squares b8 and h8. And
satisfactorily avoided. neither 6.tLlf3 We4 7 .'it>g4 l:rg8+
nor 6.tLlf7 �e6 7 .tLld6 l::tb 8 bring
This mating position can also be White anything either. So there re­
achieved without a black pawn, as mains:
we see in the following miniature 6. b7-b8l:t J:!.e8xe5
study. 7. J:!.b8-f8#

13
The Art of the Endgame

This study received second prize in

a competition organized by Shakh­
matyvSSSR.

The white king has reached the

corner, but the enemy checks are
far from being exhausted.
9. ... .!:!.g7-h7+
1 0. '>t>h8-g8 .l:l:h7-g7+
Timman 1986 11. �g8-f8 .l:l:g7-f7+
White to play and win
1 2. Wf8-e8 .l:l:f7-e7+
1. 0-0 1 3. we8-d8 .tte 7-d7+
In studies the players are always al­ 1 4. wd8-c8
lowed to castle, unless it can be And on the following move White
proved in the . starting position that captures the bishop. The white
either the king or the rook has al­ king had to march along the edges
ready moved. of the board in order to escape the
1. ... l:l.a3-g3+ rook checks.
2. �g 1 -h2 .l:!.g3-g5
3. e2-e4 In the following pawn ending, mu­
After castling kingside, White now tual zugzwang plays an important
plays the most popular starting role.
move. Black cannot save the
bishop, but he does have a long se­
ries of checks.
3. ... I;lg5-h5+
4. wh2-g3 .l:!.h5-g5+
5. Wg3-h4
The white king has to advance to
escape the checks.
5. ... .l:!.g5-g4+
s. wh4-h5 l:l.g4-g5+
1. wh5-h6 J:rg5-g6+
Timman 1988
8. Wh6-h7 l:rg6-g7+
White to play and draw
9. wh7-ha

14
 �in i a t u re S t u d i es

It is clear that the white f-pawn is

doomed. If' Black goes and collects
this pawn, the white king will have
to go for the enemy a-pawn. And in
order to advance his own a-pawn
far enough, he has to push it first.
1. a2-a4 a7-a5
But now there is a new problem:
White must lose his a-pawn.
2. 'i.t>c3-d3
Only thus. After 2 . Wb3 <it>d4 Timman 1993
Draw (dedicated to Judit Polgar)
White will be pushed back.
2. ... <i;lc5-b4 White must start with a check.
3. Wd3-d4 'it>b4xa4 1. .!:id3-d 1 +
The critical moment. If White After this move we are at a cross­
continues with the unsuspecting roads:
4.'0t>d5 he will lose: 4 . . . Wb5 S .<it>e6 A) l .....t>b2 The most natural re­
a4 6.<it>xf6 a3 7 .We7 a2 8 .f6 a l 'ii' action. 2.f5! The thematic push.
9 . f7 Wi'g 7 l O.<;t>e8 <;t>c6 l l .f8Wi' Black has to capture the pawn.
Wi'd7 mate. That is why, before he 2 ...exf5 3J:td5 This double attack
approaches the . f�pawn, White seems to solve all the problems.
must first force the enemy king to a However, Black has another finesse
worse square. up his sleeve. 3 r.:l.g2+1
•••

4. 'ot>d4-c411

This is the moment of mutual

zugzwang. If it were his move, Now a sharp exchange of pins fol­
White would lose. lows. The king has to move up­
4. ... <i;la4-a3 wards, otherwise 4 . . . it..c 6 is lethal.
5. <i;lc4-d5 4.Wh3 �fl 5 .J:� d l ! Certainly not
The rest plays itself. The white s JhfS on account of S . . . l:.f2 +
f-pawn will reach f7 and the black 6.Wg4 �h3 + , winning. 5 ... �e2
king is too far away. 6 . .l:te l i..f3 7.l:.fl �e4 8.l:.f5 draw.

15
The Art of the E n d g a m e

B) l. .. �a2 This king move has a 2. ... jLc7-b6

specific point. 2.f5 exfS 3. .lld 5 Introducing the motif of the indi­
.llg 2+ 4.Wh3 So far the play is rect pin.
identical to Variation A. 4 lb2 .. J 3. �c3-a5 tt:Je3xd5
Black's first move left this square
vacant. The bishop is protected,
while the f-pawn is taboo for the
moment. S .�h4! An instructive
move. The king will attack the
pawn from gS . S f4 6 l f5 l:tb4
•.. J

7.�g5 Winning the f-pawn.

The pins are piling up. White now

has only one king move that wins.
4. �g1-h1 11
The king moves further away from
the scene of battle. Insufficient was
4. 'iit>h 2 on account of 4 . . . �c7 + ,
while after other king moves Black
would have a saving knight check.
Timman 1994 Now there is no defence for Black.
Black to move - White wins
White remains a rook up.
This is a miniature study in the
true sense of the word. The solu­
tion is simple, but attractive. Black,
who is to move, will have to try
and eliminate the white pawn.
1. ... tt:Jg4-e3
On l . . . �d 7 , 2 . .llg 6 tt:le3 3 . .ll g 7 +
Wc8 4 . .l:!.g5 was winning.
2. �c6-c5
The only way to keep the pawn. Af­
ter 2 .d6 .ib8 there is no win; by
Timman 2008
advancing further, the d-pawn has
White to play and draw
become too weak, e.g. 3 .l:rb6 �a7
4.�f6+ 'if;>e8 S . l:tb7 tt:ldS + 6 . .l:Ixa7 Black has two extra pawns and his
tt:lxf6 7.Wg2 Wd8 ! 8 . �f3 tt:le8 and king is nicely centralized. However,
Black is just in time to eliminate White can make use of a chance
the pawn. circumstance in the position.

16
 �in i a t u re S t u d i es

1. �f2-h4!!
Pinnin'g the enemy g-pawn. If
Black captures the bishop he will
be left with the bishop of the
wrong colour.
1. ... �f6-c3+!
A pretty counter-sacrifice that can­
not be accepted either. The white
king must be within the square of
the g-pawn after the latter captures
Timman 2009
on h4.
Black to move- White draws
2. 'lt>d2-d3!
The only square for the king. It has White is far ahead in material, but
to keep attacking the enemy bishop. the black pawn is about to pro-
2. ... h7-h6 mote.
1. ... �h1 -e4
Cutting off the white rook. If
White allows the pawn promo­
tion, he will lose due to the bad
coordination of his pieces. There­
fore he has to give checks in order
to try and reach the h-file with his
rook.
2. .te7-f8+!
Not 2 .�f6+ in view of 2 . . . �g 6 !
and White does not have any effec­
tive checks left.
Protecting the g-pawn. Things look 2. ... �g7-h8
gloomy for White, but he has a re­ 2 ... �g8 would have been met by
source. 3 . .te6+ .
3. f3-f4! 3. �f8-g7+!
Forcing the g-pawn to reveal its White keeps offering the bishop.
plans. 3. ... �h8xg7
3. ... g5xh4 The same position we had two
4. 'lt>d3-e2 h4-h3 moves ago, only now without the
5. �e2-f2! white queen's bishop. The path of
The only good square for the king. the rook to the seventh rank has
After either S . �fl .td4 or S .'litf3 been opened .
.te 1 ! , it would be cut off 4. .:!.d4-d7+ 'lt>g7-h6
5. ... �c3-e1 + Now what? After S . .l:1d6+ �gS
6. �f2-g1 White would be at the end of his
Draw. wits.

17
T h e Art of t h e E n d g a m e

ending, as the course of the solu­

tion will show.
1. 'it>c1 -d2
Otherwise the g-pawn would
march on.
1. ... �e2-f3
On 1 . . .<;£;>£3 White has the impor­
tant finesse 2 .�h5 ! . He will take
on g4 next and bring his king to
the safe zone, on the lower left side
5. .l:l.d7-h7+1! of the board.
A fantastic rook sacrifice. Black has 2. a4-a51
to take with the king. Even with strongly reduced mate­
5. ... �h6xh7 rial, the strategic principles must
6. �c8-f5+1 still be adhered to. After other
The last white piece is sacrificed as moves Black would play 2 . . . a5
well. himself, after which the white
6. ... .te4xf5 rook's pawn would become a fatal
7. 'lt>f1 -g 2 weakness.
With a draw. 2. g4-g3
3. �d2-e1 �f4-e3
A follow-up study on this promo­
tion theme can be found in Chap­
ter 4.

4. kf7-c41
The prelude to a miraculous saving
clause. 4.Wfl would have been met
Timman 2009 by the winning 4 . . . \t>d4. Then the
White to play and draw
black king can march to the queen­
White is a pawn behind, but he has side unhindered and win the white
drawing chances, as he can give his a-pawn.
bishop for the enemy g-pawn In 4. ... �f3-e2
the right circumstances. There are Resourceful play by both sides. On
also chances for him in the pawn 4 . . . g2 White had prepared the re-

18
 �i n i a t u re S t u d i es

ply 5 .�fl ! . If Black then promotes However, White has a radical solu­
to a queen or rook, White is stale­ tion to his problems.
mated. This trick was discovered in 1 0. �h3xf5 g3-g2
1 8 5 6 , by the Italian Luigi 11. �f5·e4+1 <;t>f3xe4
Centurini. 1 2. �e1 -f2
After the text move White cannot And the remaining pawn ending is
take on e2 , because then the pro­ just barely drawn.
motion of the g-pawn will decide
after all.
5. �c4·d5 �e2-c4
6. �d5-g21
The only square for the bishop. Af­
ter 6.�b7 <bd4 7 .'it>d2 �ds White
would have no choice but to enter
a lost pawn ending.
6. ... �c4-d3
7. .ig2-h3 �e3-f3
Black has consistently worked on
Timman 2009
depriving the white bishop of
White to play and draw
squares. Still, there is a way out for
the bishop. Both white pieces are under attack.
8. �h3·f1 1 He will need to find a few cunning
White can now save himself in the manoeuvres to remain on his feet.
pawn ending, since he has the op­ 1. �a7·c5 d7·d6
position. But even this does not put If White now captures on d6, he
an end to his worries. will end up in a theoretically lost
8. �d3-e4 endgame.
9. .if1 ·h3 �e4·f5 2. tt:ld3·e51
The second piece is also put en
prise. Black is forced to give check.
2. �c4-d5+
3. W h 1 ·g 1 d6xc5
4. tt:le5·d7
The point of the previous moves.
White wins back the piece. How­
ever, new troubles await him.
4. ... �d5-e6
5. tt:ld7xf8 .ie6·f5
The white knight is completely
dominated by the bishop. Now
A final attempt. 1 0 .�fl �c8 ! there follows a duel between the
would lead to mutual zugzwang. kings.

19
The Art of the E n d g a m e

6. '>t>g 1 -f2 �a8-b7 1 0. '>t>f4-e3

7. <ot>f2-e3 The black king is too far away to
support the c-pawn. After I O . �b6,
both I l . tLld7 and I l .tt::l e 6 are
good enough for a draw.
A fascinating endgame. If in the fi­
nal diagram position we move all
the pieces one file to the right,
Black wins, as his king has more
squares available to it on the
queenside.
By the way, I didn't have to invent
the final part of the study by my­
self. In the tablebases, all endgames
Up to here, everything was logical. with six pieces or less have been
If Black now continues 7 .. .<it>b6, analysed to the end. It won't be
White will hold the draw by 8 .<.t>f4 long before seven-piece endgames
c4 9. r.t>xfS c3 I 0. tt::l d 7 +. Via eS and will also be included. Then we will
d3 the knight can prevent the be able to pick our miniature stud­
c-pawn's promotion. And after ies from the tablebases.
7 . . . r.t>c6 it will find its way via e6
and d4. However, Black has yet an­ In the beginning of 2 009 I visited
other, almost paradoxical king Bessel Kok in his apartment at the
move. Groenburgwal in Amsterdam. We
7. ... �b7-a7! were just drinking a glass of fine
The king steps aside in order to white wine when his phone rang.
evade all the knight checks. Now It was Garry Kasparov. For the first
White has to take care. After 8 . r.t>f4 time in many years I spoke to
c4 a situation of mutual zugzwang Garry again.
would occur, with fatal conse­ 'What are you doing these days?',
quences for White, e.g. 9.We3 he asked.
�b6 I O.Wd4 r.t>bs , and the black 'I'm busy composing endgame
c-pawn will decide the issue. studies', I replied.
Therefore, the white king too must Then he promised to send me a
step aside. new tablebase with seven-piece
8. We3-f3!! endgames; he had just received it
Now Black must either choose an from Elmar Magerramov. When I
unfavourable square for his king or came home I found the document
advance his pawn. in the mail. The tablebase contains
8. ... c5-c4 the strangest endgames; for in­
9. Wf3-f4 c4-c3 stance, it shows how four knights
Black still has no good king moves. win against a lone queen. There

20
 11in i a t ure S t u d i e s

were also a. few 'random' end­ 5. �c2-c3 �a3-a2

games, like the following. 6. l:!.b1 -b2+ 'Ot>a2-a3
7. �e5-d6+
In the corner the enemy king can­
not be mated. Therefore it is driven
up the board.
7. c;t>a3-a4
8. 'Ot>c3-c4 'itla4-a5
9. �d6-c5

White to play and win

It's amazing that White can win

here, especially as his pieces are not
very close to the black king. The
black pieces are not only far re­
moved from the battleground, but
they are also in each other's way. For
instance, without the black knight
the position would be a draw. Depriving the king of the square
1. .l:[d3-d 1 + 'itla1 -b2 b6.
2. 'it>d4-d3+ 'it>b2-b3 9. ... ..t>a5-a6
3. l:!.d1 -b1 + 1 0. .l:[b2-a2+ 'Ot>a6-b7
Forcing the king to the edge again. 11. l:a2-a7+
3. ... 'it>b3-a2 Thus White wins the rook, and
4. 'tt> d 3-c2 wa2-a3 then yet another piece.

21
Chapter 2

Rook versus Bishop

I have noticed that many top play­ lished a long article in Schaak­
ers use Dvoretsky's Endgame Manual bulletin 166 in which I used a dia­
as a work of reference, and it is true gram to indicate where the safe
that it is a good handbook, more zones for the black king were.
exhaustive than Keres' Practical Later, these conclusions were com­
Chess Endings. Recently, however, I pletely confirmed by the database. I
found a striking error in the book. also indicated that Velimirovic
Dvoretsky refers to my game made the decisive error in the dia­
against Velimirovic from Rio de Ja­ grammed position by playing
neiro 1 9 7 9 . After White's 68th 68 . . . �f8 ? (most sources errone­
move the following endgame had ously give 68 . . . �g7 ) , allowing
arisen: White to take his king to g S , so
that its black counterpart will inev­
itably be cut off on the back rank.
All this can be found in my book
Studies and Games.
As an interesting illustration I will
give you the following mini-study.

Timman-Velimirovic
Rio de Janeiro 1979

Dvoretsky makes the following

general observation about this
endgame: ' It was proven that
White should win. However the
White to play and draw
probability of its occurrence in a
practical game is rather low, while In order to draw White must start
the proof itself is quite compli­ with
cated, so we have decided to omit 1. �d3-e4
it.' It is important to ensure that the
It is not true that this endgame is bishop will be able to go to square
always winning. In 1 98 1 I pub- b7 before advancing the pawn. If

22
 Rook versus B i s h o p

the bishop covers the pawn from wasn't until after a very deep study
the fl -a6 diagonal, the position of the endgame that I discovered
will always be lost. that those key positions cannot
1. ..
. J:!.a8·b8 always be achieved.
1 . . . 0-0-0 would have been an ideal Remarkably enough, the endgame
move, but unfortunately this is im­ only occurred in my game against
possible. The position shows that Velimirovic.
Black's last move must have been a There exists an endgame of this
king or rook move. type - i.e. both sides have a rook's
2. <,t>e1 -d21 pawn, and these pawns are stand­
Certainly not 2 .a6 in view of ing nose to nose - which is won,
2 . . ..l::!:b 2, and the white king is cut namely with a white pawn on a6
off. Now 2 . . ..l:i.b2 + 3 . Wc3 .l:ta2 versus a black one on a7. It oc­
runs into 4.<;i.;b4. curred with reversed colours in the
2. ... �e8-d7 game Aronian-Kramnik, Shanghai
3. a5-a6 2 0 1 0 . Remarkably enough, during
Now this is possible. the game both players thought it
3. ... l:!.b8-b6 was a theoretical draw.
4. �e4-b7 Cheron had already concluded as
And White is safe. early as 1 948 that the endgame is
won in all cases. Only if White
Dvoretsky is not the only person were to succeed in winning the en­
who thinks that this endgame is emy a-pawn without endangering
always won. Spinhoven and his king would it be a draw.
Bondarenko are of the same opin­ In the following position, White
ion in their book De strijd tussen has to pull out all the stops to get
Ioper en toren ('The battle between the draw.
bishop and rook') . This misunder­
standing has probably been fos­
tered by the Swiss endgame theo­
retician Cheron, who had investi­
gated the endgame extensively in
the 1 960s. He gave a plethora of
variations, always with the conclu­
sion that White would win, while
emphasizing that there are certain
key positions that White must
strive for. Cheron never explicitly
states that the endgame is winning Timman 2010
White to play and draw
as such, but a brief study of his
work suggests that it is, and he If White manages to push his
probably thought so himself It a-pawn to a4 in order to cover it

23
The Art of t h e E n d g a m e

with the bishop from c6, the draw If Black had played 4... .l:!.g3 and
is in the bag. But after l .a4, both 6 ... .l:!.g5 , White would have been
l.. . .:!.cS and l . . .l':te3 will win, forced to fmd another square for his
which means that White must first bishop, since 7 .�a8 would then run
improve the position ofhis king. into 7 ... .l:!.g8! and 8 ... l:Ib8, and the
1. <;t.;>g6-f6 J::[ e5-e3 white king has been cut off. Then
2. �f3-c6! the only move would have been
Preventing the advance of the en­ 7 .�£'3. intending to meet 7 . . . .l:!.g3
emy a-pawn. with 8 .�d l !, after which Black will
2. .. . l:re3-a3 be unable to stop White executing
3. �c6-d5 his plan with the king on b4 and the
Now White will have to allow the bishop on b3 . Note that this is al­
advance, but since the black rook ways possible because of the remote
will have to vacate a3 again, White position of the black king.
has gained a crucial tempo. 7. .. . J:.h5-h6
3. ... a5-a4 8. <;t.;>c4-b4 J:.h6-b6+
4. <;t.;>f6-e5 lia3-h3 9. 'it>b4-a51
The alternative was 4 .. Jig3. I will Again the only move. Incredibly
return to this later. enough, the white king has had to
5. we5-d4 a4-a3 trek from g6 to aS in order to se­
6. 'it>d4-c4 cure the draw.
Threatening 7 . Wb4 followed by 9. . .. .!:f.b6-g6
8 . �b3 . Black has only one move to 1 0. 'it'a5·b4
prevent this. And a draw. Black will be unable to
6. ... l:rh3-h5 make any progress.
Now White will have to choose a
square for his bishop. Remarkably, most of the theoretical
investigations into endgames of
.l:I:+L\, vs �+.!.\, (without any passed
pawns) concentrate on positions
with rook pawns not separated by a
square. The position only becomes
really interesting with one or more
squares between the pawns. Some­
times the moves and conclusions
of the database are difficult to un­
derstand, but closer analysis usu­
ally showed me what exactly was
7. �d5-a8!! going on.
Improbable as it seems, this is the In the conclusions of my investiga­
only move. After, for example, 7 .� tions I will restrict myself to posi­
Black would win with 7 ...l:r.h2. tions that can be explained in an

24
 Rook vers us B i sho p

acceptal;>le and logical manner; him to build a fortress by putting

with special focus on the crucial his bishop on the g l -a7 diagonal
manoeuvres that lead to technical and advancing his pawn to a6.
positions - which are the most im­ There are two famous practical ex­
portant for endgame theory. The amples in which the pawns are
generally lengthy technical process separated.
then unfolds gradually. As this con­
tains few surprises, I will ignore it.
I will start with positions with two
rook pawns.

Gligoric-Polugaevsky
Amsterdam 1970

Although Black eventually man­

aged to win, this position is a theo­
White to play and win
retical draw. This would be differ­
As a rule, the position is winning ent with the black pawn on f7 in­
with the black king cut off on the stead of fS , when Black would have
back rank, regardless of the posi­ the winning plan of taking his king
tion of the white king. In the dia­ to e 1 via the queenside. White's
grammed posmon, however, counterplay then would be to at­
White must watch out. After tack the black f-pawn. On f7, how­
l.Wf2? aS ! 2.<>i;>e3 �b7 ! it's a draw. ever, it is just too far away.
There is no remedy against the
threat of 3 . . . �c6. 2 . . .jt_g4 won't do
in view of 3 . 'it>d2 ! , and the black
bishop is dominated. White wins
by playing l .a5 before deploying
the white king.
Even with the black king not cut
off White can win if he succeeds in
taking his king to cS , after which
Black cannot prevent his king from
being cut off. The position is a
Reshevsky-Fischer
draw, however, if Black has the
Los Angeles 1961
king's bishop, since this will enable

25
T h e Art of the E n d g a m e

White played 5 3.�£'3? here, and In the Reshevsky-Fischer example,

now Black could have reached a the position would also have been
winning position with 5 3.. Jhe3+ winning if the black pawn had
5 4.�xe3 h5, as Evans indicated in been on its starting square. Some­
Chess Life, since White will be un­ times it's even winning if the king
able to stop the black king from of the attacking player is far away.
reaching the crucial g4-square.

How important it is to secure the

square h3 for the white king is
clear to see in the following posi­
tion:

White to play and win

Black is threatening to draw imme­

diately by playing 1 .. . Wa3 , fol­
lowed by 2 . . . �b2 .

1. l:tc2·e21 .te5-d4
2. l:te2-d2 �d4-c3
White to play and draw
3. lld2-d3
1. 'it>f3-g21 White has brought the situation
On its way to h3 . under control. Now he has all the
1. �g6-f5 time in the world to take his king
2. �g2-h3 l:lg4-c4 closer. The quickest route is via the
3. �d8-e71 fourth rank, since its eventual des­
The only square for the bishop, as tination is c8. The win will take an­
we'll see. other 6 5 moves.
3. ... llc4-c1 Remarkably enough, the diagram
4. i.e7-f81 position with the king on h7 instead
The point of the previous move. of h 1 would be a draw. In that case
The bishop is forced to hide itself; Black can save himself with 3 ...b5 !
otherwise Black would give check 4.Wg6 Wa3 ! s Jhc3+ <.t>xa2, and
on h 1 and then attack the white the white king is not in time to
bishop. reach the square e2 to restrict its
4. ..
. J:tc1 -b1 black counterpart's freedom of
5. �f8-g7(h6) movement. With the king on h5,
and Black won't make any prog­ however, the position would be
ress. winning by the narrowest margin.

26
 Rook vers u s Bishop

In the. following pos1t10n White 12.�e4! the black bishop is forced

has only one winning move. to leave the g l-a7 diagonal, after
which the win is simple.
B) 2.<.tc2 ..tcs, and now White
will have to be careful: after 3 .<>td3
Wb6 ! 4.We4 .tel! S JHI �aS !
Black would throw up an efficient
defensive line. Winning is 3.llfl !
'i.ttb6 4.'i.ttc 3!, and the black bishop
is dominated again. The white king
is threatening to penetrate via d4.
If Black has a queen's bishop, he
can build an effective defensive
White to play and win
line by keeping his b6 pawn and
1. l::t d 1 -c1 + ! his bishop on the long h l -a8 diag­
The main point is that Black can no onal, from which they cannot be
longer take his bishop to cS. driven away through domination.
1. ... Wc6-b7 With the white king on c 1 (instead
Now White has two entirely differ­ of b I ) in the diagrammed position
ent winning plans: White would have no win, since he
doesn't have the crucial rook
check. But Black would have to
play very accurately to hold his
ground.

A) 2 .l:H I �a6 3JH6 WaS

4.b4+! �a6 s.�c2 Wb7 6.�d3,
and now it becomes clear why
2.l:�hl was insufficient for the win,
since in that case, Black would be The main line goes as follows:
able to activate his bishop with 1. Wc1 -c2 �b6-c51
6 . . . .if2 . White must always make 2. c.t>c2-d3 �c5-f8!
sure to continue to dominate the 3. J::td 1 -g 1
enemy bishop. After 6 .r:J;c7 •• Keeping the bishop from g7.
7.We4 �b7 s.�ds .tgi 9.%:tg6 3. ... 'lt>c6-c5
.tf2 IO .I:t.g7+ Wb6 II.l:tg2 �e3
• 4. J:!.g1 -g6 <;f;>c5-b4

27
T h e A r t of the E n d g a m e

And White will make no progress. The solution is:

All Black's moves were the only 1. 0-0-01
moves with which to maintain the Black's problem is that his king re­
balance. mains cut off White's winning
plan involves taking the king to a4.
Sometimes White is able to win A possible continuation is: l . We6 . .

because of chance circumstances. 2 .Wb2 \t>es 3 .Wa3 �c6 4.'it>b3

�b5 S .�c3 , followed by 6.l::rb l .
The bishop cannot stay on the
fl -a6 diagonal, which means that
the white king will penetrate. An
important consideration is that the
king has the square c3 : with the
white pawn on c3 (instead of c2)
the diagrammed position would be
drawn.
White cannot win with l .l::rd l , be­
cause after l . . . �e6 2 .l::r d 2 �f3 ! his
White to play and win
own king would also be cut off
1. �c3·b4! �b6-c7
2. Wb4-a5! b7-b6+
3. 'itta 5-a6
And the white king has penetrated
decisively.

If both players have a c-pawn, there

are more winning positions. White
usually wins, for example, if the
pawns are on their starting squares
and Black has a king's bishop.
The following position is curious. 3 . .l:!.d8 is no good, as this is met by
3 . . . �d5 , and the trap is sprung be­
hind the white rook.
Here, too, it is very important that
the white pawn was still on its start­
ing square, but the endgame con­
tains an interesting paradox: gener­
ally speaking it is advisable to keep
the pawn back, but once it is on the
third rank, it is often better to ad­
vance it as far as possible. This can
White to play and win be seen in the following position.

28
 Rook vers u s B i s h o p

1. c4-c5+ Wd6-c7
2. �b1 -c2 �c3-d41
3. wc2-d3 �d4-f2
4. l:tg5-f5 �f2-g1 1
5. 'it>d3-c4
Threatening 6 . Wb4 and 7 . �b5 .
5. .ltg1 -e3
or

5. wc7-b7
And White can make no progress.
White to play and win With the pawns moved one square
1. �b3-c4 c7-c6 to the left, there are even fewer
2. 'ito>c4-d4 winning chances. The pawn struc­
Centralization. ture prevents White from penetrat­
2. ... �d7-c7 ing with the king via the queen­
2 . �a3 would have been met by
. . side.
3 . .l::!.a 2 ! . But not 3 .c4? in view of
3 . . . �d6, and Black has the situa­ The following position is also in­
tion under control. teresting.
3. c3-c41
The only way to make progress.
3. 'ito>c7-b6
4. l:te2-b2+ 'ito>b6-c7
5. c4-c5
And this position is winning, be­
cause the black bishop's move­
ments are restricted. Only if the
bishop can be taken to the g l -d4
diagonal does a position arise
which, in certain circumstances,
White to play and win
may be drawn.
1. wc3-d4
Centralization.
1. ... �d5-b7
The most tenacious defence. The
obvious move now is 2 . .l:h8 , in­
tending to create more room for
the rook for horizontal checks. But
this is met by 2 . . .�c6, with the
threat of 3 . . . d6, and Black has a
fortress. And 3 . \t>eS is met strongly
White to play - draw by 3 . . . WcS ! : now it is not the black

29
T h e Art o f t h e Endgame

king which is cut off, but the white would be blocked, giving Black
one. time to play 5 . . . dS .
2. .l:e8-e3! 3. ... 'it>d6-d5
The only winning move. White 4. <t>c2-c31
prepares a check on c3 in order to And White has the situation under
prevent Black from building a for­ control. From this point on it will
tress. take another 6 7 moves to the mate.
2. ... �d6-c6
3. J:.e3-c3+ 'it>c6-d6
And now 4.1:1g3 makes for the
quickest win. Eventually the black
king will be forced back, and then
White will have to play accurately:
if the black king goes to e 7 , White
takes his king to cS ; if Black with­
draws the king to c7, White pene­
trates via eS .

White to play - draw

With the pawn on d2 White would

have an easy win. Now he will have
to advance the pawn again.
1. d3-d41 'Ot>f8-f7
2. d4-d5!
And now White will bring up his
king.
If White starts with I . We3 , Black
can draw with very accurate play:
White to play and win
1. 'it>f2-e3 �f8-e7
Black is threatening to build an 2. we3-d4 jlc6-f3!
impregnable defensive line with The only move. Black wants to take
1 . . . rJilc7 , followed by 2 . . . d5. his king to d 7 , after which he can
1. J:.e1 -e41 make his sortie via c6 and e6.
The same principle as in the previ­ 3. �d4-c4 �e7-d7
ous position. White takes the rook 4. l:.g5-g3
to the c-file in order to upset the After 4.Wb5 We6 ! S . �b6 �e2 !
coordination in the enemy camp. Black will just be able to save him­
1. ..
. 'it>d8-c7 self
2. .l:e4-c4! 'it>c7-d6 4. ... �f3-h1 !
3. 'it>d1 -c2 Unlikely. Black has to keep his
The only way. After 3 . We2 ? rJitdS bishop on the long diagonal at all
4.'lt>d3 Wc6 ! the white d-pawn costs.

30
 Rook versus B i sh o p

5. <;t>c4-b5 �d7-e6! (or 3 .l:.a6)

Threatening to take the king to the
fifth rank.
6. .l:!.g3-g5 �h1 -f3
And the drawn position in the
comment to White's fourth move
has arisen again.

I will finish with an example of

lighter calibre.

And Black is in zugzwang. After

3. ... c;f;>f6-e7
4. Wd4-e5
Black will be forced onto the back
rank, after which the play could
continue as follows:
4. ... �e6-f7
5. .l:!.b6-h6 �f7-e6
6. .l:!.h6-h7+ .te6-f7
White to play and win 7. .l:!.h7-g7 d7-d6+
In light of the above it is not diffi­ 8. c;i;>e5-f5 �e7-e8
cult to find White's moves. 9. �f5-f6 �f7-c4
1. .l:!.d4-d6! c;i;>g8-f7 1 0. .l:!.g7-c7 .tc4-b3
2. c;i;>c3-d4 c;f;>f7-f6 11 . .l:!.c7-c3
3. .l:!.d6-b6 followed by 1 2 . .l:!:d3 and wins.

31
Chapter 3

Preventing Pawn Promotion

In a simultaneous game Lasker­ 39. ... .l:!.c4-h411
Loman, London I 9 1 0, the follow­ Thus Black succeeds in preventing
ing happened: the promotion of the h-pawn.
40. �g5xh4 g7-g5+
41 . 'Oth4xg5 �f8-g7
White resigned.
This is probably the best-known
simultaneous fragment in chess
history.

The following study was based on

Loman's 3 9th move.

34. Uf4-f8+1 �g8xf8

35. g6xh7 .l:!.c2-c1 +
36. �g1 -f2 J:tc1-c2+
37. �f2-g3 rl.c2-c3+
38. '>tg3-g4?
This seems winning, but Black has
a nasty riposte.
38. ... .Uc3-c4+
39. 'Otg4-g5 Timman 2008
White to play and win

White is a piece up, but Black has

all kinds of counterchances - es­
pecially with the pawn on e 3 ,
which will play an impo�tant
role.
1. l:1d5-d1 a4-a31
The start of a clever action. After
l . . . ex£'2 L�.xc7 .l:le8 3 .�g3 .l::t e 2
4.l:rfl White would have a techni­
cally winning position.
2. b2xa3

32
 Preven ting Pawn Promotion

Forced, as otherwise Black would

obtain another dangerous passed
pawn.
2. ... l:ta8-a4+
3. f3-f4!
Before retreating his king, White
first sacrifices a pawn. Only at the
end of the study will it become
clear why this is necessary.
3. ... .l:l:a4xf4+
4. 'oth4-g3 l:tf4-d4 8. .Jtd8-g5+! �h6xg5
This is the point of Black's play. If 9. f2-f4+
White does not take the rook, a Now the point of White's third
rook ending ensues that cannot be move becomes clear: the square f2
won. is vacated for the white king, after
5. J:!.d1 xd4 e3-e2 which the remote passed pawns de­
cide the issue. The combination
carried out by White in this study is
called a 'magnet combination'. The
black king is drawn towards the fa­
tal square gS as ifby a magnet.
This study received second prize in
the annual tourney of the magazine
The Problemist.

It is also possible to use Loman's

move to clear the way for a bishop,
which then saves the day. By way of
How can White stop the pawn an illustration I give the following
now? His task is much more diffi­ little study.
cult than Loman's.
6. J:!.d4-d6+1
The first rook sacrifice.
6. �g6-h7
6 . . . cxd6 is met by the winning
7 .�a5 .
7. J:!.d6-h6+1
The second rook sacrifice. Black is
forced to accept this one, as other­
wise his king will be caught in a
mating net. Timman 2011
7. ... 'it>h7xh6 White to play and draw

33
The Art o f t h e E n d g a m e

White is truckloads of material up,

but once the enemy h-pawn
queens, he will be mated soon.
Therefore, drastic measures are
called for.
1. tLld6-e41
With the winning threat 2. lt:Jf2 + ,
so Black has to take the knight.
1. ... d5xe4
2. .l:!.a5-h51
Loman's move.
2. 'it>g4xh5 Now comes the crucial move.
3. g2-g4+ ..t>h5xg4 4. l:tc1 -c4+!1
4. �b1 xe4 'lz- 1/z You could call this rook check a
'desperado' , i.e. the sacrifice of a
Sometimes White has to sacrifice a piece which is lost anyway, in a
rook to prevent a pawn promotion. way that yields the best result. But
then this is a very special desper­
ado. Everything revolves around
the fact that the d-pawn, which is
on the verge of promotion, comes
within reach of the white king.
4. �d4xc4
5. 'it>f3xe2 �c4-c3
6. 'it>e2-d1
Draw.

White's 4th move is known as the

Prokes, 'Schackvarlden' 1939
Prokes Manoeuvre. There are vari­
White to play and draw
ous examples with this theme
The situation looks hopeless for from tournament practice. Also,
White because his king is so far numerous other studies have been
away. But he has a miraculous sav­ made with the Prokes Manoeuvre
ing clause. as a theme.
1. �h5-g4 e3-e2 In June 1 99 3 , I was in Moscow,
An important alternative is l . . . d2, where a blitz tournament was held
in order to meet 2.Wf3 with in memory of Misha Tal. I was
2 . . . 'it>d3 . Then White has 3 . .!:.a l , working on a study that would fea­
saving the draw. The rook will keep ture the Prokes Manoeuvre via a
giving lateral checks. hidden mechanism. Eventually it
2. l:tf1-c1 + 'it>c4-d4 worked out, and I decided to dedi­
3. �g4-f3 d3-d2 cate the study to Tal.

34
 P reven t i n g Pawn Promo t i on

The climax of the study. Again

White threatens mate in one.
5. ... Wg4xh5
The same position occurs as two
moves ago, but without the white
knight. Alternatives weren't satis­
factory either, viz. :
A) 5 . . . �d4 6 .tt:lf6 + �xf6
7 . .l:!.b4+ and again the rook has
found its way towards the enemy
Timman, 'New In Chess' 1994/3
king;
White wins (in memory of Misha Tal)
B) s .. .f4 6.tt:lf6+ �fs 7 .g4+
Again we see that Black has several �e5 8 .tt:ld7+ �d6 9 . .l:Ib6+ �xb6
far-advanced passed pawns. White 1 0 .<iltxd2 �xaS+ l l .�xc2 Wxd7
must bring his king closer to the and amazingly enough, a technical
fray. bishop endgame has come on the
1. 'it>f1-e2 l:ib2-b1 board. White continues with
A thematic rook sacrifice that 1 2 .<of;ld3 , followed by 1 3 .�e4. He
White has to accept. is winning due to the fact that the
2. .!::t a 1 xb1 c3-c2 black kingside pawns are very
But now what? After 3 J:[d I c I if weak.
4..l:hd2 'iYc4+ Black has a draw 6. g3-g4+1
within reach. Therefore White has Opening the short diagonal for the
to try and find a path towards the white bishop.
black king. 6. ... f5xg4
3. l:tg8xg6 But now what?
Threatening mate in one. 7. l:tb1-h1 1 c2-c1 'if
3. ... f7xg6 8. �h4-e1 + 1
Forced, as after 3 .. .'l£ixg6 4 . .l:.b6+ Here we have the hidden mecha­
White would win right away. nism.
4. l2J h3-f4+ �h5-g4 8. ... �h5-g5
5. tt:lf4-h5!! 9. �e1 xd2+
And wins.

My interest in the Prokes Manoeu­

vre was aroused by Tim Krabbe. In
the summer of 1 99 2 we met on an
estate in the south of France, in the
vicinity of Ste Maxime. Misha Tal
had just died, and I was analysing
games for an 'In Memoriam' . After
I had showed Krabbe a couple of

35
The Art o f t h e E n d g a m e

findings, he started talking about ning study of it, but I couldn't come
the 'Double Prokes' . From that mo­ up with a good version. All the stud­
ment on it seemed as if my chess ies I published in the 1 990s turned
understanding had changed. A fas­ out to be incorrect.
cinating world opened itself before
my eyes. Was it really possible that Not until 1 8 years later did I man­
such an ingenious idea could be age to make a good, solid version.
doubled in a study? Krabbe had
thought up the following scheme.

Timman/Krabbe 1992/2010
White to play and win
Krabbe 1992
In fact, all the elements of Krabbe's
White to play and draw
scheme have been retained. Only
The intention was that White here White has two highly danger­
would draw as follows: ous passed pawns, which forces
1. tth2-h1 e4-e3 Black to act quickly. White's first
2. �g2-f3 l:Xc2-c1 move is forced.
3. tth1 xc1 e3xd2 1. tth2-h1
4. ttc1 -c4+ b5xc4 On l .�f2 Black would save him­
5. �f3xe2 c4-c3 self by l . . .e3 + ! 2 .dxe3 .l:i.c l
6. .l:tf6-f1 l:Xb8-b1 3 .�xe2 1:tc2+ .
7. .l:tf1 xb1 c3-c2 1. ... e4-e31
8. ttb1 -b4+ 'it>a4xb4 Still the best chance, even in these
9. �e2xd2 circumstances.
With a draw. 2. �g2-f3! .l:tc2-c1
The idea is fantastic, but this con­ 3. tth1 xc1 e3xd2
struction contains many flaws. 4. .l:tc1 -c4+1
Krabbe realized this all too well. He The first Prokes Manoeuvre.
asked me to make a version that 4. ... b5xc4
worked. I toiled over this for days. 5. �f3xe2 c4-c3
The Double Prokes turned out to be The white king has come closer,
a head-spinning task. My intuition but in the meantime Black has cre­
told me I could make an award-win- ated a mighty pawn pair.

36
 P reven t i n g P awn Prom o ti on

6. l:lf5-f1 l:!.b6-b1 Again, White has a powerful pawn

The second black rook offers itself pair. But first he will have to do
After other moves the white passed something about Black's threat of
pawns would quickly decide the promoting the f2 pawn.
game. 1. tt:Jf3-d2 l:!.d1 xd2
7. .l:If1 xb1 c3-c2 The best chance. The thematic 1 ... f3
would have been prospectless in
view of 2 .ttlc4+ ! �b4 3 .ttle3 , and
the black pawns are harmless. Also
insufficient was l ... .l:td6 on account
of 2 .�g2 U6xd2 3 .�xf2 .l:Ixc2
4 . .l:ig l ! .U.dd2 S . .l:ie l and White's
passed pawns decide.
2. l:tc2-c1
After 2.�g2 l:!:xc2 3 .�xf2 .l::r c l !
White would not be able to win.
2. ... f4-f3
8. l:l:b1 -b4+! 3. 'it>h2-g31 .!:.d2-d 1 1
The second Prokes Manoeuvre. Both sides make the moves we
8. ... <;i;>a4xb4 know by now. Black's last move is a
9. 'it>e2xd2 so-called 'switchback' , i.e. a piece
And wins. - in this case, the rook - moves
The solution is also identical to back to a previously used square.
what Krabbe had envisaged. White must be very careful now:
only one move wins, all other
Now that I had finally made this moves lose.
old dream come true, I decided to
find out if I could achieve more
with this theme. This led to the fol­
lowing spectacular study.

4. b3-b4+!!
The path to the black king must be
opened by force. White could not
Timman 2010 wait with this check, since after
White to play and win 4 . .l:txd l ? fxe2 5 .b4+ �a6 ! Black

37
T h e Art of t h e E n d g a m e

would gain the upper hand. After A t first glance, the situation looks
6.bS+ �aS 7 .b4+ cxb4 8 . .l:r.d8 b3 very bad for White. His bishop is
9.�xf2 l:c l 1 0 .�xe2 b2 the black threatened with capture with
c-pawn decides the issue. check, while at the same time
4. ... wa5xb4 Black is threatening a deadly dis­
Alternatives wouldn't do either: covered check with his own
A) 4 . . . �a6 S .bS+ �aS 6 .�xf2 ! bishop. How can White exploit his
and wins; far advanced passed pawns?
B) 4 . . . cxb4 S ..!:hd l fxe2 6.l::td S+ 1. �h1 -c6+1
and wins. In this line, one Prokes This bishop sacrifice forces the
Manoeuvre suffices. king onto the long diagonal.
5. llc1 xd 1 f3xe2 1. ... 'it>d7xc6
6. l:rd 1 -d4+1 2. Jlg3-g8
The first Prokes Manoeuvre. A second piece sacrifice with a spe­
6. c5xd4 cial point.
7. 'it>g3xf2 d4-d3 2. ... li'h8-h1
8. l::t g 4-g1 llc6-c1 A clever trick. Now 3 .exd8'i¥ is not
The second rook sacrifice. possible due to 3 . . . 'i¥a l + and
9. l:rg1 xc1 d3-d2 White will get mated.
1 0. l:!.c1 -c4+! 3. l:rg8-g 1 1
The second Prokes Manoeuvre. The rook has to keep following the
1 0. ... rJ.?b4xc4 queen.
11. <M2xe2 3. ... 'ii'h 1 -h8
And wins. With this I had con­ 4. l:rg1 ·g8 'ii' h8xg8
cluded a magical chapter. Otherwise Black cannot make
progress.
Related to the Prokes Manoeuvre is 5. f6-f7
a theme that was first adapted by
the renowned Georgian study
composer Gia Nadareishvili.

This is what White was aiming for.

An area of tension has occurred
Nadareishvili 1965 which Black can only remove with
White to play and draw a check. If there is no such check in

38
 P reven t in g Pawn Promo t i on

the p,osition, then the pawn pro­ Here, too, White promotes to a
motion will be unavoidable. knight. In this position it turns out
5. ... 'lig8xf7 to be winning.
6. e7xd8ttJ+ 4. ... <;i;>e6-d5
And now Black is the one who has 5. .!Dd8xb7 f5-f4
to go all out for a draw. It looks as if the win is still difficult.
For this study Nadareishvili re­ 6. .!Db7-a5!
ceived shared first/ second prize in Via the edge of the board White
the 'Vecherny Leningrad' tourney. tries to reach the d2-square in or­
Yet this isn't one of Nadareishvili's der to stop the enemy f-pawn.
best achievements. The final play is 6. ... Wd5-e4
weak. Pal Benko must have been 7. .!Da5-b3 we4-e3
thinking the same when he made
the following study.

The king controls the d2-square.

But White has another subtlety up
Benko, 'Magyar Sakkelet' 1987 his sleeve.
White to play and win
8. .!Db3-d4! we3xd4
White has two powerful passed 9. wh7-g8 f4-f3
pawns, but his king is threatened 1 0. h6-h7 f3-f2
with being caught in a mating net. 11. h7-h8'iV+
1. 'it'h5-d 1 1 And wins.
A queen sacrifice is not sufficient. This study was awarded first prize
After l .'i:Yf3 Wixf3 2 .c8'fi'+ r:J;;e ? in the tourney of the above-men­
3 .b7 ileS , White cannot win. tioned Hungarian chess magazine.
1. ... 'lic6-a8
Threatening mate in one. When, five years ago, I wanted to
2. 'ir'd1 -d81 ii.f6xd8 use this theme, I did not know ei­
3. b6-b7 ther of these two studies. I had only
Nadareishvili's area of tension had a look at an incorrect study by
again. Black has to take the b-pawn. the Bulgarian I.Ionchev. A diagram
3. ... �a8xb7 with Nadareishvili's area of tension
4. c7xd8.!D+ was given in Krabbe's Chess Curiosi-

39
T h e Art of t h e E n d g a m e

ties. For me this was the starting The second rook sacrifice forces
point for the following study. the Nadareishvili position.
5. 'i!Vc8xa8
6. b6-b7

Timman 1992
White to play and win

White enjoys a large material ad­

vantage, but his king is unsafe. In As in Nadareishvili's study, Black
order to avert the immediate dan­ now has two possibilities. The
ger, he has to sacrifice a bishop. main line continues:
1 . .ih2-d6+1 6. ... 'i!Va8xb7
The alternative l .Sit..e 6 narrowly The alternative is 6 . . . 'i!Ya6, in order
fails in view of l . . . 'ife8 2 .�f5 to wait for 7.cxd8'if WeS + 8 .'.tg5
'ti'hS , with sufficient counterplay before continuing with 8 . . . 'ifxb 7 ,
for Black. s o that the knight promotion with
1. . .. '1t>e7xd6 a fork is avoided. White must then
2. J:tc5xd5+ '1t>d6-e7 play accurately to secure the win,
Black is threatening again, but this because if the queens stay on the
time the bishop sacrifice on e6 board, a draw is inevitable. The
does work. only winning move is 9. 'ifh8 ! , in
3. �g8-e6 �e7xe6 order to force a queen swap after
After 3 .. 'i!Ve8 4.�f5 'ifhs s .�f4
. 9 ... �d6 1 0.'ifd4+ .
the white king will eventually 7. c7xd8t2H '1t>e6-d5
manage to reach the queenside, 8. tt:ld8xb7 '1t>d5-c4
where it is safe. At first sight, this looks like a draw,
4. J:td5-d8! because Black is going to capture
After two bishop sacrifices the first both queenside pawns.
rook sacrifice follows. Due to the But the white king is just in time to
last bishop sacrifice the route to g4 corner its black colleague, after
was blocked for the black queen, which the contours of the Stamma
and now g8 is made inaccessible. mate begin to take shape.
4. ... i.f6xd8 9. �g6-f5 �c4-c3
5. J:ta7-a8 1 0. �f5-e4 �c3-b2

40
 P reven ting P awn P r o m o t i on

11. 'it>e4-d3 'it>b2-a2 1. f6-f7+ Wg8-g7

1 2. 'it>d3-c2 wa2-a1 2. .itg5-h6+!
1 3. tt:lb7-c5 With this bishop sacrifice White
The knight is going to force mate lures the enemy king further for­
in the usual fashion. ward.
1 3. 'it>a1 -a2 2. ... �g7xh6
14. tt:lc5-d3 �a2-a1 3. .i;lc5-h5+!
1 5. tt:ld3-c1 a3-a2 Now it has grown into a magnet
1 6. tt:lc1 -b3 combination. The king is drawn
Mate. further out, enabling White to
With hindsight I am not so satis­ build Nadareishvili's area of ten­
fied with my version. It starts with sion on the board.
a lot of sacrifices, but all this vio­ 3. ... <it>h6xh5
lence does not really add anything 4. g6-g7
substantial to the theme. Stamma's
mate is a nice touch, but it isn't
sufficient to give the study extra
cachet. I thought it would be a
good idea to make a study where
Nadareishvili's area of tension con­
tinues to exist for a couple of
moves, and where Black just fails to
break out with his queen and
bishop due to various self-blocks
by the opponent.

Now there are two main lines:

A) 4...d2+ 5.�£'2! d l tll + 6.�el
and despite his enormous material
supremacy there is no win for
Black;
B) 4 tllc 2+ 5.Wd l ! The knight
.•.

blocks the black bishop's diagonal.

Now 5 . . . 'iVxg7 would bring Black
nothing after 6 . fxe8 'iV + �h6
7 .�d2 ! tll b 4 8 .'i1Ye6+ 'iVg6
Timman 2011 9. 'iVh3 + and there is no good way
White to play and draw
for Black to escape from the
Black has a great surplus in mate­ checks. 5 ... tlle 3+ 6.Wc l ! and now
rial. White can only try to save the knight blocks the black queen's
himself with a series of checks. diagonal.

41
The Art of t h e E n d g a m e

Also related to the Prokes Manoeu­ 34. .l:.b6-b4

vre is a famous promotion combi­
nation from a game played in Ma­
drid in the 1 9 3 Os.

Again White appears to have every­

thing under control, but a new sur­
prise awaits him.
Position after 30.h3 34. ... a7-a5!1
Ortueta-Sanz, Madrid 1933
The third pawn move. In spite of
Black is two pawns up, but his his enormous material advantage
structure is shattered. An easy win White cannot prevent the promo­
is not on the cards. tion of the c-pawn.
30. ... l:!.d8-d2
The introduction to a magnificent It is almost incredible that such a
combination. combination, full of study motifs,
31. tL:lc3-a4 can occur in a practical game.
This allows the combination. More However, the story becomes
tenacious was 3 1 .a4 in order to stranger still: in 1 95 3 a game from
meet 3 l . . .l:i.xb2 with 3 2 .a5. In that 1 9 3 1 between the Polish players
case it is probably best for Black to Tylkowski and Wojciechowski was
strengthen his position with published, which contained ex­
3 l . . .�g6. actly the same combination! This
31 . ... .l:!.d2xb2! game is shrouded in mystery. Orig­
A rook sacrifice which is based on inally it was claimed that it had
a series of marvellous finesses. been played in the national cham­
32. tL:la4xb2 c4-c3 pionship, but this turned out to be
33. .l:.b7xb6 untrue. There is a game between
White appears to save himself, but the same players from that champi­
now there follows a colossal sur­ onship, but here Tylkowski was the
prise. winner. No problem, according to
33 . ... c5-c4!! the sources from Poznan that had
The main point is that the white made the game public: it had been
knight is deprived of the d3- a casual game. There were sup­
square. posed to have been onlookers as

42
 P revent i n g P awn Prom o t i on

well. W�tnesses insisted that they ·

had seen the game in question, in­

cluding the combination. One of
them claimed that the game had
been published at the time in
Dziennik Poznanski, the local news­
paper, which had a weekly chess
column. This latter message turned
out to be untrue as well.
So how credible does this make
this game's authenticity? If there After 2 2. J:i.f3 , the game is said to
had been so many witnesses, this have continued as follows:
incredible combination had to 22. ... c5-c4+
have found its way to the newspa­ Understandable. Black wants to get
per column. It is also conspicuous rid of his doubled pawns.
that both players had died by the 23. d3-d4?
time the game was published. But this is very strange. The
Nothing could be checked with ei­ d-pawn could have been lost now.
ther of the two persons directly There was nothing against 2 3 . �fl .
involved. 23. ... c6-c5??
Despite all this, Tim Krabbe, who And this is simply a grotesque stra­
has published a lot about this tegic blunder. With 2 3 . . . l:td8 he
mysterious game on his website, could have gained a big advantage.
claims that Tylkowski-Wojcie­ 24. d4-d5 e6xd5
chowski has really been played. I 25. J:l.d1 xd5
quote: With a big strategic advantage.
'It is hard to think of a reason why 25. ... cJJ>g 8-h7
it would have been invented, and A strange move which, however,
it would have been very difficult fits into the scheme.
to invent it.' As for the first argu­ 26. J:td5-d7 l:ta8-d8
ment: to me it looks like an under­ 27. J:td7-b7?
standable attempt of local Poznan What is this? White puts his rook
chess players to put their city on out of play and hands the only
the global chess map. In 1 95 3 , open file to his opponent. The only
Ortueta-Sanz had already been a justification for the text move ap­
famous combination for a long pears to be that White wants to
time. What could be more beauti­ reach the position where the fa­
ful than the suggestion that all this mous combination can start as
had already occurred two years soon as possible.
earlier in Poznan? And as for the 27. ... .!::!.f6 -g6
game itself: it contains a bizarre 28. l:tf3-g3 l:lg6xg3
episode. 29. h2xg3 l:td8-d2

43
T h e A r t of t h e E n d g a m e

Now the aim has been achieved. The first move is obvious. White
The position is ripe for the combi­ must save his bishop.
nation. Tylkowski and 1. �h2-g3 d5xc4
Wojciechowski were pretty strong The technical background of this
players. It seems to me to be out of study is that both l . . .d4 and
the question that they were able to l . . . tt:lxg3 lead to lost rook endings.
play such bad moves. 2. .!:tf7xg 71
Here is the rook sacrifice, in the
Back to Ortueta-Sanz. The position other corner of the board.
after Black's 30th move is often 2. ... ti:lh5xg7
published in a stylized version, 3. f5-f6 .l:tg2xg3
without the black pawns on e6 and 4. f4-f51 !
g5 . In that case the capture on b2 If White captured on g 3 , Black
would have been the only way to would be able to scrape a draw.
win. Then you could turn the 4. .l:tg3-g5
board around in order to make an 5. f2-f41 ti:lg7xf5
endgame study of it, with the 6. f6-f7 1 -0
stipulation 'White to play and
win'. This wouldn't be a bad study, The rook sacrifice can also be made
but it should be possible to get on another square.
more out of it. There are studies by
the Italian Rossi and the Russian
Lomov, but they are incorrect. And
there is little use in trying to re­
store them, since they hardly add
anything to Sanz's combination. A
good study is the following one by
Ernest Pogosyants, the man who
composed more than 1 7 00
endgame studies.

Timman 2010
White to play and win

The situation on the queenside is

familiar. But on the other wing
there's also something going on.
First White has to stop the h-pawn.
1. �f1 -g1 .l:tb2-b1 +
2. �g1 -h2 .l:tb1 -b2
Black's counterplay is not to be un­
Pogosyants, 'Themes-64' 1977 derestimated. White has to take
White to play and win drastic measures now.

44
 P reven t i n g Pawn Promo t i on

. 3. l:tc6-g6+!! Black threatens to capture the

White sacrifices his rook in order knight on f4 with check. Therefore
to clear the way for the pawn on this knight is sacrificed first.
cS , and at the same time to open 1. tt:Jf4-g61 h7xg6
the a2-g8 diagonal. Black is forced 2. tLJc7-b51
to accept the sacrifice, as otherwise Keeping the black knight from d6
White will take on g4. and the rook from b6.
3. ... f7xg6 2. ... .!:!.b2xb5
4. c5-c6 3. c4-c5!!
The rest is well-known territory. In the circumstances, this is a quite
4. ..
. .!:!.b2xb3 special pawn move.
5. c4-c5 .!:!.b3-b5 3. ... e4xf3
6. a2-a4 tt:Jb7xc5 An ingenious defence. Black va­
7. c6-c7 cates the e4-square for the knight.
And the c-pawn queens. If White now plays 4.axb5 , Black
When I showed this study to can save himself by 4 . . . tt::lxc5 S .c7
Yochanan Afek, he shook his head tt::l e 4+ , followed by 6 . . . tt::l d 6.
and said: 'You paid a high price for
this rook sacrifice'. Afek likes max­
imum economy in a study. As little
material as possible should be used
to convey the ideas. In this study
there is an entire structure on the
kingside that doesn't play any role
in the main variation. I fear that
this is inevitable if we want to add
any new ideas to the combination.
Instead of a rook, White can also
sacrifice two knights in order to 4. <;t>g3-h311
break through. A marvellous quiet move. White
takes his time and first eliminates
the knight check, while maintain­
ing the threat to take the rook.
4. ... tt:J b7xc5
5. c6-c7
And wins.
Here again, we see this heavy
structure on the other wing, but an
extra element has been added: the
quiet move at the supreme moment.
Timman 2011 Fourteen years earlier I had already
White to play and win tried to use this idea.

45
The Art o f the Endgame

Timman 1997
With the point that 7 .axbS is im-
White wins (first publication)
possible in view of 7 . . . .!i:JxcS 8 .c7
White is a pawn up, but his queen­ .!i:Je4+, followed by 9 . .!i:Jd6.
. .

side is very weak. He will have to We have now reached the high
act decisively. point of the study.
1. f5-f6! g7xf6 7. 'it>g3-h3!!
Forced, as White's main threat was An enormous surprise. Although
2 . l:th8+ �g8 3 .f7 , while l . ..�g8 a rook and a knight down, White
would have run into 2.fxg7 + rJite7 is in no hurry to win back mate­
3 .l::.h 8 J::!:xb3 4 . .l:!.xg8 �f6 s . .!i:Je6 ! , rial, but starts by hiding his king.
and White wins. The absolute high point of the
2. l:h7xf7+ Wf8-g8 study is thus lengthened by one
Not, of course, 2 . . . '1t>e8 , in view of move.
3 .�a4+ . 7. ... g5-g4+
This is the start of the well-known 8. 'it>h3-h4
combination. And wins in two moves. The pawn
3. l:f7xb7! tt:la5xb7 on c6 decides the outcome.
After 3 .. .fxg5 4.l:ha7 .!:i:xb3 'It has become a kind of Prokes
S .l::.xaS J::!:xe3 + 6.Wg4 'lt>fl 7.Wf5 theme' , observed Krabbe when I
the rook ending should be win­ showed him this study.
ning.
4. c5-c6 l:b2xb3 After making this study I hadn't in­
5. c4-c5 dulged in the passion of compos­
The well-known theme. With this ing endgame studies for two years
pawn move the black knight is kept - until I found three messages
away from d6 and the rook from from Tim Krabbe on my answering
b6. machine in the space of thirty min­
5. ... .l:tb3-b5 utes. In the first message he said:
6. a3-a4 'Eh jan, in that study after Ortueta­
Nothing new so far. But Black has a Sanz the bishop is not covered.
nice riposte. Wouldn't it be better if it was ? ' .
6. ... f6xg5! And the second message, about

46
 P reven t i n g Pawn Promo t i o n

two minutes later: 'OK, no, I get it B) 2 . . . .l:tcl 3 .c6 �xc4 4.c7 t'rc6
,
now, the bishop cannot be cov­ (otherwise S .�cS) S .f4 e4 6.f5 ,
ered.' And finally, in the third mes­ and wins.
sage, he said: 'But still, wouldn't it 2. a2-a3
be better if the bishop was cov­ Tempting, in view of what we have
ered?' Hearing these messages, I seen before, is 2 .c6 .!:!:xb4 3 .c5 , but
felt something stir in my inner­ this fails to 3 ... tt:lxc5 4.c7 tt:le4+,
most chess soul. followed by 5 .. Jk4.
Rusty retrograde cogs started 2. ... a 7-a5
churning, picking up speed. 3. c5-c6
I decided to tackle the study on a Only now can the bishop be sacri­
larger scale. Having the bishop on ficed.
b4 instead of on b3 turned out to 3. ... a5xb4
be a good starting-point. 4. c4-c5
The result looks as follows: The Ortueta-Sanz motif. Now
4 . . . tt:lxc5 is insufficient in view of
S .c7 tt:le4+ 6.fxe4 l::rc l 7.axb4, and
wins.
4. ... b4xa3
5. c6-c7
Now the black rook has to hurry
back to the bottom rank to stop the
foremost white passed pawn.
5. ... J:!.b1 -g1 +

Timman 1999
White to play and win

As far as material is concerned, the

situation is roughly equal. White
obviously depends on his passed
c-pawns. l .c6 tt:ld6 at once yields
nothing, so he will have to start by
trying to dominate the enemy
knight.
1. �d2-b4 J:!.f1 -b1 6. �g3-h3!
There is no other defence against For two reasons, both of which
the threatened 2 .c6, as may be seen will become apparent, this is the
from: only good square for the king.
l .. J:f.g l + 2.'>itf2, and now: 6. ... l::!. g 1 -g8
A) 2 .. J:rg8 3 .c6 tt:ld8 4.�e7 ! , Interesting is the try 6 . . . a2 7 .c8'iY +
and wins; l::r g 8, which leaves White only one

47
The Art of t h e E n d g a m e

way to win, viz. 8 .tt:::l d 8 ! (threaten­ And this is the second reason for
ing mate in one) 8 . . . tt:::lx d8 9.'ifa6 , White's 6th move: the white king
and the black a-pawn is eliminated. assists in weaving a mating net. If
7. tt:le6·d8 .!:!.g8xd8 the black king had gone to fs ,
8. c5-c6! White would have taken on h7
The Prokes motif Here we see the with check and then escaped the
first reason why the white king had perpetual.
to go to h3 ; on the second rank it 1 5. ... Wh5xh6
could be checked by the rook. 1 6. 'i¥f7xf6+ �h6-h5
8. ... �d8-g8 1 7. 'i¥f6-g5
The best defence; Black turns the Mate.
game into a pawn race. A wonderful Excelsior of the black
9. c6xb7 a3-a2 a-pawn, with the newly-born
queen powerless to do anything. I
failed to find a way to create an Ex­
celsior with the hindmost c-pawn
in a similar way.

The difference between the Prokes

Manoeuvre and the combination
from Ortueta-Sanz is that with the
manoeuvre the pawn promotion is
prevented, whereas in the second
case, the promotion is realized by
the combination. I thought it
1 0. b7-b8'ti' would be a challenge to create an
This is the only way to queen 'Anti-Ortueta-Sanz study' - i.e. a
which wins, as will become clear study where White manages to
in a while. prevent the pawn promotion.
1 0. a2-a1 'il'
11. 'ti'b8xg8+ �h8xg8
12. c7-c8'ii'+ r;t;>g8-f7
1 3. 'i¥c8-d7+
Now it is clear why the b-pawn
had to queen first: the black king
must not be allowed to go to e6. If
the black king now goes to g 8 , he
loses the h-pawn with check via g7
and g8/h8
13. 'et>f7-g6
1 4. 'ii'd 7-g7+ <;t>g6-h5 Timman 2010
1 5. 'i¥g7-f7+ White to play and draw

48
 Preven t i n g P awn P r o m o t i o n

White's first move is forced. rook check on a8 or e 1 respec­

1. .!:ra2-a8 .!:rg6-b6! tively, is possible. Insufficient was
A thematic sacrifice that serves to S .'it>h l on account of S . . . bxa l �+
shield the b-file. 6 .tt::l x a l exd4 7 .tt::lx b3 d3 and
2. �e3xb6 b3-b2 White cannot save himself.
Threatening to promote the pawn, 5. b2xa 1 �+
which would even mean mate 6. tbc2xa1 �a7xd4
here. White must retreat his rook. 7. tba1 xb3
3. .l:ta8-a1 b4-b3 With a draw.
Just like in the previous study we
see a mix of the Prokes Manoeuvre Related to the Ortueta-Sanz combi­
and the Ortueta-Sanz combination. nation is an idea from a quite un­
White must find a way to get to the known study by the Russian end­
enemy king. game study composer Alexander
4. �b6-d4!! Herbstman.
A curious situation. Two moves
ago none of the white pieces was
under attack. Now all three of
them are. However, for the mo­
ment there is no reason for Black to
capture either of the pieces.
4. ... .ib8·a7!

Herbstman, 'De Schaakwereld', 1937

White to play and win

Materially White is in bad shape,

but he has a dangerous passed
pawn.
1. b5·b6 l:.d4·d1 +
The rook has to hurry in order to
stop the b-pawn.
The tension on the board is in­ 2. Wg1 -f2
creased to the maximum. White's Definitely not 2 .'iit> g 2 ? in view of
bishop is pinned, and the black 2 . . . �b l ! followed by 3 . . .�e4+
bishop cannot be taken. How can and the black rook can stop the
he save himself? dangerous pawn.
5. Wg 1 -f1 1 2. ... .l:td 1 ·a 1
The solution! Black has two ways A curious situation. Black would
to capture the bishop, but then a prefer to get rid of one of his mi-

49
T h e Art of t h e E n d g a m e

nor pieces, or both. They are hope­

lessly in each other's way.
3. h3-h4!
A tempo move.
3. ... e5-e4

Timman 2011
White wins (after Herbstman)

The black bishop is ready. The

other pieces still have to be moved
to the crucial squares.
1. e6-e7+ 'lt>f8-e8
4. 'it>f2-g2 2. ttJe4-d6+1
Now that the black pawn is on e4, Forcing Black to take with the c­
this square is accessible to the king. pawn, after which White can open
4. ... e4-e3 the path for his b-pawn.
5. 'it>g2-g3 l:ra1 -g1 + 2. ... c7xd6
6. 'it>g3-f4 l:rg1 -f1 + 3. a6-a71
7. �f4-g5 l:rf1 -g1 + A pawn sacrifice to wrongfoot the
8. �g5-h6 l:rg1 -a1 black knight.
There are no more checks, and the 3. ... ttJc8xa7
rook returns to its post. 4. b5-b6
9. h2-h3
Mutual zugzwang. White wins.

A nice study; but on move 4 there

are two duals in the position.
4.'ot>e3 wins too, and the little
move 4.h3 leads to mate in 1 4.
This is a pity, as the idea of a tiny
pawn dominating three black
pieces is spectacular, and it de­
served a better adaptation. I con­
sidered that all kinds of sacrifices The paralysing passed pawn arrives
would be possible to reach the on the scene.
spectacular position. Also, more 4. ... ttJd 1 -e31
can be done with the theme of With this knight sacrifice Black
mutual zugzwang. clears the back rank for his rook.

50
 Preventing Pawn Promotion

White has t,o accept the sacrifice, I made the following study on Oc­
since a check on dS is threatened. tober 31, the day that rumours
5. kd4xe3 l:f1-a1 were buzzing about that the great
Now Black threatens to clear the Dutch writer Harry Mulisch had
c6-square for his knight with died. While anxiously following
6... c5. the news, I was working on the
6. .\te3-c5!! ideas. When the study was finally
The only way to thwart Black's finished, the news came through
plan. Black has to accept the bishop that Mulisch had died the night be­
sacrifice, because 7.b7 :bl 8 ..1b4 fore. I decided to dedicate the work
is threatened. to his memory.
6. ... d6xc5
7. h4-h51
Not 7.h3 on account of 7. ..hS!
8.e3 f4 9.exf4 <Ji>d7! and lO.<Ji>£7
fails to 10 ...�g6+. It is very im­
portant for White to gain control
of the g6-square.
7. ... f5-f4

Timman2010
Draw (in memory of Harry Mulisch)

It is clear that White cannot pre­

vent the loss of a piece. T herefore
he goes after the enemy c-pawn.
1. .!:!.e7-d7 + 'it>d6-e6
2. l:!.d7xc7 l:!.e1-e3+
Black wants to capture the knight
8. h2-h31 with the king. To this end, he must
With this tempo move White interpolate this rook check .
comes out on top. Insufficient was 3. l2Je5-f3!
8.h6 in view of 8 .. .f3 9.exf3 <Ji>d7! Drawing the black rook to a less fa­
and Black is OK again. vourable square.
8. ... h7-h6 3. ... l:!.e3xf3+
9. h3-h4 4. �h3-g2 ::if3-c3!
Mutual zugzwang. For rook, bishop Black's best chance. After 4... .l::!.d3
and knight, White only has one 5 ..llc6+ �eS 6.l:!.xa6 l:rd2+ 7.<Ji>f3
pawn, but still Black is powerless. .!:rxb2 8.'iite3 the rook ending
9. �a2-b1 would be a draw.
10. b6-b7 1-0 5. b2xc3

51
T h e Art of t h e E n d g a m e

The only move. After 5 ..l:i.c6+ �e5 prevent promotion by hiding a

6 .l:!.xa6 l:!.c2+ 7 .�f3 �d4! the stalemate in the position. Unfortu­
rook ending is winning for Black, nately the study turned out to be
as the white king is shouldered off incorrect, for a surprising reason.
5. ... �a6-b51

Bran, 'Sovietskaya Rossiya' 1957

White to play and draw (incorrect)
The fantastic point of the previous
move. Black blocks the b-file, and The intended solution was:
now White's task to prevent the 1. tt:la7-b5
promotion of the b-pawn seems White has to try and prevent the
impossible. On 6.I:!.b7, for exam­ promotion of the d-pawn.
ple, Black has 6 . . . ii.c6+. 1. ... d3-d2
6. l:lc7-c511 Otherwise the knight goes to c3.
A calm move. White threatens to 2. ..t>d7-c8
take the bishop. Suddenly White threatens mate in
6. ... ..tb5-d7 two. There is only one parade:
Now what? 2. ... l:[f3-f5
7. l:lc5-g511 3. tt:lb5-c7+ 'it>a8-a7
Only with this rook move can 4. �e7-b4
White prevent promotion. The rook Forcing the b-pawn to promote.
aims for three different squares. The 4. ... d2-d 1 'iV
threat is 8 . .l:l:g6+ , followed by 5. �b4-c5+1
9.tf.b6. Should Black try 8 . . . 1kc6+, The point of the previous move.
then the white king steps aside, af­ 5. ... J:!.f5xc5
ter which the g 1 -square is available Stalemate.
for the rook. Finally. 8 . . . �f6 is met A charming miniature study,
by 9.l:!.g8; then the rook will have wouldn't you say? The main line is
access to the square b8. short and sweet. However, the
problem is that Black can win with
The great Russian endgame study 4 . . . d l �! . The tablebases indicate
composer Vladimir Bron has cre­ that this is mate in 7 9 moves. For a
ated a study where White tries to long time it was thought that the

52
 P reven t i n g Pawn P r o m o t i o n

ending of .1:1 + � versus � +lLJ with <:J:;;a 7 4.lt.Jxb5 + . Now Black has two
opposite-coloured bishops was a king moves:
theoretical draw. To be frank, I have A) 4 �b6 5.lt.Jc3 �xc3 6.lLJd5+,
•..

always had my doubts about this followed by 7 .lt.Jc3, with a draw;

theoretical verdict. Even with lim­ B) 4...�a6 Now White cannot
ited material, the presence of op­ stop the pawn, but he has a hidden
posite-coloured bishops provides perpetual check mechanism: 5 ...ib4!
attacking chances. This is not the dl 'if 6.lt.Jc7+ 'it>b6 7.lt.Jed5+ 'it>c6
only study that was proven incor­ 8.lt.Je7+ and the black king cannot
rect by the findings of the table­ walk out of the checks.
bases. The same fate befell studies 1. ... i.b2-a3!
by, among others, Fritz and Rinck. This bishop sacrifice enables Black
to disturb the coordination in the
In itself, the bishop promotion is white camp. White has to take the
an attractive starting-point for a bishop, as 2 .lLJe6 �xd6 would re­
new study. For example, you could sult in another theoretically win­
reverse the colours and turn it into ning ending for Black: rook +
a win study. But more can be made bishop versus two knights.
of this theme. 2. �d6xa3 d4-d3
How can White stop the pawn?
3 .lLJe6 is met by first 3 .. J;{b8 + , af­
ter which the d-pawn marches on.
3. t2Je7-f51
White sacrifices one of his knights
in order to deflect the rook to a less
favourable square.
3. ... l:tb5xf5
4. t2Jf8-e6
Bringing the other knight into po­
sition.
Timman, 'Schaaksite' 2011 d3-d2
4.
White to play and draw (after Bron)
5. �a3-b4!
Here, too, it is hard to stop the
black d-pawn. For instance,
l . .txb2? would fail to l ... d3 ! .
1. i.a3-d6
With this bishop move White
seems to solve his problems. On
l . . d3 he has planned 2.lLJe6,
.

when 2 . . . Wa7 gives nothing be­

cause of 3 .�f4 ! . The struggle be­
comes sharper after 2 d2 3.lt.Jc7+
..•

53
T h e Art o f the E n d g a m e

Bron's move, which does work in Black must take the bishop, other­
these circumstances. There are now wise White will put his rook on
two variations: the first rank.
A) S . dl 'if
.. 6.ltJc7+ 'i;; a 7 3. l:l:b8-g8
7.�c5+ %:txcS stalemate, as in Now he can play this.
Bron's study; 3. J:tc6-c7+
B) S ... d l � This is the point. 4. Wh7xh6!
6.�c5 ! Draw, since the mate threat
on c7 can only be averted with an
exchange sacrifice. The black rook
is unfortunately placed. It is im­
portant that White postpones the
knight check on c7. From e6 the
knight protects the bishop on c5.

Sometimes it looks easy to prevent

promotion. But then the position
turns out to contain an unexpected
problem.
The white king first walks to h4, in
order to force Black to take his
c-pawn.
4. .!:[c7-c6+
5. �h6·h5 l:l:c6-c5+
6. �h5-h4 l:l:c5xc4+
7. <J;>h4-h5
And now, up again .
7. ... .!:[c4·c5+
8. '1tr>h5·h6 l:l:c5-c6+
9. �h6·h7 llc6-c7+
Grinblat 'Israel Ring Ty' 1988
1 0. �h7-h8
White to play and win
And wins. Now the black rook can­
If White plays l .l:Ig8 immediately, not move to the 8 th rank, since it
there follows 1 . . .l:tc7 + and sur­ will be taken with check.
prisingly enough, the white king
cannot escape the checks. If it goes The study is based on a pretty idea,
to h8, Black plays his rook to the but its execution leaves a little to be
8th rank as well, pinning the white desired. The white sacrifices are
rook. Therefore White must give passive; neither the bishop nor the
check first. c-pawn contributes anything. My
1. J:th8-a8+ \t>a3-b2 search for a livelier version pro­
2. lla8-b8+ \t>b2xc2 duced the following result.

54
 P reven t i n g P awn P ro m o t i on

5 .�d5 ! This bishop sacrifice forces

the rook to move to the d-file.
5 . . . .l:!.xd5+ 6.'>t>xh6 .:td6+ 7 .c;i>h7
J:i.d7+ 8.Wh8 , and wins.
2. J:th8-g8 J:ta6xa4+
3. 'iit h 4-h5
First the king collects the black
h-pawn. 3 .b4 .l:!.xa2 would spoil
the win.
3. ... J:ta4-a5+
Timman 2011 4. 'iit h 5xh6 l:ta5-a6+
White wins (after Grin blat)
5. 'iit h 6-h5
The starting position already con­ The king goes back again.
tains a temptation. I .�h7 + is 5. ... J:ta6-a5+
insufficient, in view of 1 . . .<;t>a 1 6. �h5-h4 .1la5-a4+
2 . .l:!.g8 J.:txa4+ and Black has per­
petual check. The white bishop is
in the way.
1. �g8-a2+1
This bishop check forces the black
king to leave the b-file.
1. ... 'iit b 1 -a1
The side variations contain echoes
ofGrinblat's idea:
A) l . . . �xb2 2 . l.:tb8 + ! '1t>xa2
3 . l:rg8 l:i.xa4+ 4.�h5 l:.a5 +
5 . c;i>xh6 l:ra6+ 6 . '1t>h7 l:la7 +
7 . 'it>h8 and wins, entirely accord­ 7. b2-b41
ing to Grinblat's scheme. Forcing the rook to go to the b-file.
B) l . . .�c l 2 . .l::. c 8+ ! �d2 3 . .l:rg8 7. ... .1la4xb4+
.l:rxa4+ 4.�h5 I:!.a5 + 8. 'iit h 4-h5 .1lb4-b5+
9. c.t;>h5-h6 J:tb5-b6+
1 0. wh6-h7 .!:tb6-b7+
11. 'iit h 7-h8 J:tb7-b8
1 2. J:tg8xb8 g2-g1 'ii'
1 3. J:tb8-b1 +
The final point.
1 3. ... 'iVg 1 xb1
14 . ..ta2xb1
And wins.

55
Chapter 4

Various Promotion Combinations

The Ortueta-Sanz combination is 4.l:Ixh6 �b2 and Black holds. The
the most famous promotion com­ rook sacrifice opens the path to eS
bination. However, promotions can for the bishop.
be forced in all kinds of interesting 2. .. . �e8xf8
ways. Let me start with a study that Now the starting position is again
is linked to one of the miniature on the board, without the white
studies in Chapter 1 . rook. A further difference is that
the black king is on the back rank.
This is of cardinal importance for
White, as now he can threaten to
promote with check.
3. .ith8·e5
The black rook is cut off. Now
Black only has bishop checks left.
3. ... �f1 ·e2+
4. �d1·e1 1 �c1 -d2+
5. 'l.t>e1 -f21
White leaves the second bishop
Timman 1987
untouched as well. He must keep
White to play and win
preventing effective rook checks by
It looks logical to start with Black.
l .�eS . Then, on l . ..l:Ic8, 2 J�k6 !
is the winning move. But Black
has a better option . After
l . . .�gS ! 2 .�b2 l:Ic8 White can­
not win.
1. l::.d 6·f6+ 'l.t>f7-e8
After l . ..'�e7 2 ..lhfl White would
win immediately thanks to the
threat of 3.�f6+. Now the capture
on fl would yield White nothing,
because Black would put his rook
on hS . 5. ... �d2-e3+
2. l:lf6-f8+1! 6. �f2-g2!
Also insufficient was 2..�.g7 in The only square for the king. After
view of 2 . . . �h5 3 . .ith6 .l:rxh6 6 . <.t>g3 �f4 + ! 7 . Wxf4 l:Ic4+

56
 Va r i o u s P r o m o t i on C o m b i n a t i ons

8 .'iit g S l:g4+ Black would suc­ · Black sacrifices a full rook with the
ceed in giving his rook for the sole purpose of vacating the
h-pawn. e7 -square for his knight.
6. ... .ie2-f1 + 2. kh5xe8 tt:Jc6-e7
7. ..t>g2-g3! Now the promotion of the g-pawn
White consistently refuses to take is much less effective. White has to
bishops. Now 7 . . . �f4+ is not ef­ tread very carefully to achieve the
fective in view of 8 .Wxf4 l:rc4+ promotion of the h-pawn.
9 . 'it>gS or 9.�g3 and Black has no 3. c3-c4!
good rook checks left. The aim of this pawn sacrifice is to
This means that Black has run out open the long diagonal. The pur­
of resources. After 7 . . . �f2. + White pose of this will become clear
cannot take this bishop either, but eight moves later.
both 8 .'it>f3 and 8.Wh2 suffice for 3. ... .id5xc4
the win. 4. g7-g8�
Incidentally, g3 was the only This pawn promotion actually
square for the king. After 7 . 'lith 1 ? boils down to a pawn sacrifice that
�g2 + ! Black would draw in a way serves to vacate the square g7 for
we know from the miniature study the king.
in Chapter 1 . 4. ... �c4xg8

The following study is a great sac­

rificial slugfest.

5. �e8-f7+! !
After two pawn sacrifices, White also
offers his bishop into the bargain.
Timman 2005 The idea is to draw the g8 bishop
White to play and win
out to a less favourable square. Cap­
White is a lot of material behind, turing on g8 is insufficient. White
but his mighty passed pawns com­ has an extra bishop, but with the
pensate for that. wrong rook pawn he cannot win.
1. g6-g7 5. �g8xf7
Threatening to promote. 6. 'it>h8-g7 tt:Je7-g6
1. ... .l:.e7-e8+1 7. 'it>g7xf7 tt:Jg6-h8+

57
The A r t of the E n d g a m e

8. �f7·g7 a5-a4 2. <1:ld4xf3 g4xf3

9. 'it>g7xh8 a4-a3 3. c4-c51 d6xc5
1 0. 'it>h8-g8 a3-a2
11. h7-h8�
Now the purpose of White's 3rd
move becomes clear. Black cannot
promote.
This study won 6th prize in the
John Nunn 5 0 Jubilee Tourney.

4. jLf7-c4+11
A magnificent bishop sacrifice,
with which White keeps the a7-g l
diagonal closed. As a result, the
white a-pawn can march on freely.
4. 'it>d3xc4
5. a5-a6 <Ji>c4·d3
Timman 2008 6. 'it>g1 -f1 1
White to play and win
Otherwise the black c-pawn be­
It is clear that the passed a-pawn is comes too dangerous.
White's chief trump card. How­ 6. ... c5-c4
ever, before he can march with this 7. a6·a7 c4·c3
pawn, White first has to take a few 8. a7-a8'if c3-c2
precautionary measures. The black c-pawn is still danger­
1. h5·h61 ous, and White has to keep playing
The point of this pawn sacrifice accurately.
will become clear after a few 9. 'iYa8xf3+
moves. First the f-pawn is eliminated with
1. ... g7xh6 check.
The alternatives lose: 9. ... �d3-d4
A) l . . . g5 2..�d 5 ! �f6 3 .lt:Jc6 and 1 0. 'iff3-f2+ 1
the a-pawn decides. By playing the Forcing the king back to the third
g-pawn to g 5 , Black has deprived rank.
his bishop of the square g 5 , and as 1 0. ... <Ji>d4-d3
a result the bishop is dominated; 11. li'f2·e2+ <Ji>d3-c3
B) l . . .g6 2 .li:Jxf3 gxf3 3 .'\t>fl 1 2. 'ife2-e3+ �c3-b2
�d8 4.a6 �b6 5 .�g8 and White 1 3. 'ife3·b3+
obtains a decisive passed h-pawn. And wins.

58
 Va r i o u s P r o m o t i on C o m b i na t i ons

In tht; following study White must Black would emerge as the win­
first sacrifice a bishop and then ner.
create a 'nerve square' in order to
force promotion.

4. b2-b3!1
An astonishing quiet move. White
Timman 2009
postpones the choice of pawn
White to play and win
capture. Meanwhile, Black's mate
White is an exchange down, so threat is prevented.
speedy action is required. How­ 4. ... ::tc5-h5
ever, l .h6 achieves nothing as The alternative was 4 ... tt:ld3 in or­
Black has the simple 1 . . .hxg6 der to play for a perpetual check.
2 .hxg7 Wf7. So first of all, the After 5 .hxg7 �c l + 6.Wa2 .l:i.c2+
black king must be diverted. 7. �a3 l:c 1 White decides the
1. �b7-c8+ �e6-e7 game by creating space for his king
On l ... �f6 , 2 .�d8+ is immedi­ with 8 .b4.
ately decisive. 5. h6xg7 ::th5-h1 +
2. �c7-d6+! 6. �b1 -b2 .l:th1 -h2+
This bishop sacrifice deflects the 7. Wb2-c1
black king from the kingside. The king can only run away along
2. ... �e7xd6 the first rank. Otherwise it would
3. h5-h6! be caught in a mating net.
The time is ripe to set up the 'nerve 7. tt:Jb4·a2+
square' . Black can take neither of 8. Wc1 -d1 tt:Ja2-c3+
the two pawns. But he has a hidden 9. �d1 -e1
resource up his sleeve. It looks as if the white king escapes
3. ... tt:Jd3-b4! to the kingside without trouble.
A subtle knight move. Black However, Black manages to stop
weaves a mate threat into the po­ the g-pawn by sacrificing his rook.
sition. Consequently, White can­ 9. .l:th2-h1 +
not resolve the 'nerve square' : 1 0. �e1 -f2 tt:Jc3-e4+
4 . gxh7 would fail to 4 . . . .Uh5 and 11. �f2·g2 tt:Je4-f6
4.hxg7 to 4 . . . l:!.g5 . In both cases 1 2. �g2xh1

59
T h e Art of t h e E n d g a m e

The king has travelled a long way 1. b3-b4

and ends up in the opposite corner. Threatening 2 .b5 , after which
1 2. ... h7xg6 White would have a winning ma­
Threatening 1 3 .. .<i.te 7 , after which jority on the queenside. Black's re­
Black could look to the future with ply is therefore forced.
confidence. But White has a 1. ... c7-c6
method at his disposal to cut off 2. f2-f4!
the king. The point of this pawn sacrifice is
that the b8-h2 diagonal is ob­
structed. The intention behind this
will become clear later on.
2. ... g5xf4
The alternative was 2 ... .id8 in or­
der to pre-empt the coming break­
through combination. White then
wins with a thematic variation:
3.fxg5 + WxgS 4.h7 .if6 S .b S !
cxbS 6 . c 6 ! dxc6 7 .�e7 ! and the
h-pawn decides.
1 3. �c8-a6! 'Ot>d6-e7 3. b4-b51 f4-f31
1 4. �a6-c4 1 -0

In the following study, White also

tries to take profit from two re­
mote passed pawns.

4. 'it>h2-g 1 1
Not 4.gxf3 on account of 4 . . . cxb5
5 .c6 �d8 and Black has a bishop
check in reserve.
Timman 2010 4. ... c6xb5
White to play and win
5. c5-c6! �e7-d8
White has an extra pawn, but the 6. �a3-e7+!
far advanced h-pawn is threatened And wins. Black must allow the
with extinction. Swift action is promotion of either the c- or the
called for. h-pawn.

60
 Va r i o u s Promo t i on C om b in a t i on s

The Austrian judge Alois Wotawa is time for White to lay his cards on
known for his original ideas in the table.
endgame studies. In the following 3. f2-f3!1
study he has built in a hidden The point of the fianchetto. The
mechanism to achieve the pawn white bishop lies in ambush in or­
promotion. der to carry out the combination.
3. ... a4-a3
4. �g2-h31
Just in time.
4. ... �c8xh3
5. g3-g4 1 -0
The problem is that White can also
win with I .f4 or l .�e4+ 'it>xh6
2 .f4. The white pawns are too
strong. We could start the study
one move later, but then it would
be a rather dull affair.
Wotawa, 'Schach Magazin' 1949
White to play and win
Wotawa published another study
The intended solution is: on this theme in the same issue of
1. g2-g3 Schach Magazin.
A quite remarkable little move.
White vacates the square g2 for the
bishop. With this late fianchetto
White prepares a combination
with which the black bishop will
be cut off from play.
1. a7-a5
2. �a8-g2 a5-a4

Wotawa, 'Schach Magazin' 1949

White to play and win

First of all, White will have to put

things right on the queenside.
1. Wd1 -c1 a4-a3
2. '1t>c1 xb1 a3-a2+
3. 'it>b1 -a1
The white king is again boxed in
Black's counterplay is assuming on the queenside, and the threat is
threatening proportions. It's high 4.�a7.

61
The Art of t h e E n d g a m e

3. ... .th2·b8 The only way to make progress.

4. f2·f3 White threatens 5 .c7 �e6 6.�b 7 ,
Not a fianchetto, but it still works. winning the bishop.
4. ... Wh5·g5 4. ... �d5-e6
5. ..tb6·g 1 !I Wg5-f5 5. c6-c7 �e6-c8
6. e2·e3! 'lt>f5-e6
7. �g1 ·h2 �b8xh2
8. f3·f4 1 ·0
This version is correct, but it makes
a rather crude impression. With his
king White collects a knight that
has no further function.

It seemed to me that the theme

with the white bishop lying in am­
bush deserved a better execution.
Black has just managed to hold his
ground. Now White sets an am­
bush with his bishop.
6. .\:i.a8-g2!!
Wotawa's fianchetto.
6. ... h6·h5
The only way to try and stop White
executing his plan. On 6 . . . f3
7 .�xf3 'it>g7 , 8 .h3 ! would have
been winning.
7. f2-f31 h5-h4
Timman 2009 8. �g2-h31!
White to play and win
The point of the previous. Black
A quite natural starting position. cannot take the bishop.
Since l .cxd6 �d7 yields nothing, 8. ... h4xg3
White has to make a breaking move. 9. h2xg3
1. d4-d5! d6xc5
2. d5xc6 .te8·f7+
Heading for d5 . Otherwise Black
cannot stop the passed pawn.
3. wa2-a 1 1
Only much later will it become
clear why this is the right square
for the king.
3. ... �f7-d5
4. �b7-a8

62
 Vari o u s P r o m o t i o n C o m b i n a t i ons

Now Black has a choice between 2. ... �e5-f6

two options: 3. d6-d7 id6-d8
A) 9...fxg3 1 0.gxc8 g2 l l .�f5 ! Again Black has apparently man­
g l 'iV+ 1 2.�bl Now the point of aged to avoid the loss of his bishop.
White's 2nd move becomes clear. However, White has a nice trick up
The queen cannot stop the b-pawn. his sleeve.
B) 9 �b7 If White now contin­
... 4. �b8-g31 a5-a4
ues with 1 O .gxf4, he cannot win 5. �g3-h4
due to the fact that Black has a Basically this is the same principle,
passed pawn on the kingside. but now both kings are involved in
Therefore White must play very the combination. The black bishop
accurately: 1o.ge6 + �g7 l l .g4! is drawn to a critical square on the
�f6 1 2.c8'iV �xeS 1 3.�xc8 'it>g5 short d8-h4 diagonal.
1 4.�d7 and White wins by the 5. ... �d8xh4
smallest of margins. 6. g2-g4+
And wins.
Wotawa later elaborated on the
theme in a slightly simpler study, The top Latvian player Herman
with less material, as follows: Mattison was a gifted endgame
study composer. With little mate­
rial he could weave subtle ideas
into a position.

Wotawa, 'Schach Magazin' 1951

White to play and win

A perfectly natural starting positio­

Mattison, 'Magyar Sakkvilag' 1925
n. The first move is obvious.
White to play and win
1. d5-d6 �d4-e5
2. �c7-b8 The intended solution is:
This principle, to move the bishop 1. h6-h7 g7-g6
one square further away in order to Black must grant White a mighty
achieve promotion, stems from the pawn pair to avoid the promotion.
Czech composer Jindrich Fritz. 2. h5xg6 ttla6-c7
Later in this chapter we will meet 3. �a3-f8 ttlc7-e6
him again. 4. g6-g7 ttle6xg7

63
T h e Art of t h e E n d g a m e

5. �f8-b4+ �e1 -d1 With this pawn sacrifice White

6. �b4-c3!! �f6xc3 clears the d2-square for his king.
7. �e3-d3 2. ... �c3xf6
And wins.
The problem, however, is that the
white passed pawns are so strong
that he can win in many ways, all the
more so since the black king is so far
removed from the scene. For exam­
ple, 3 ...tb4+ and 3 .�e4 are also
more than sufficient for the win .

It seemed to me that the elegant

'petite combinaison' with which
White forces the promotion de­ White still cannot promote his
served a better fate. I managed to pawn in view of the knight check.
make an addition. Now Mattison's move follows, in a
subtler version.
3. �a3-b2!1
This bishop sacrifice has an aston­
ishing point. 3 . . . �xb2 will be met
by the quiet move 4.�d2 ! . White
rules out the knight check without
attacking the enemy bishop. The
subtle point is that any black knight
move is met by 5 .c3 , closing the
long diagonal for the black bishop.
3. ... t2Jg7-f5+
Timman 2008
The best defence. Black is going to
White to play and win
block the long diagonal in order to
It's clear that White has to try and prevent the pawn promotion.
promote his h-pawn. Black has 4. 'it>e3-d2 tLlf5-d4
nothing to fear after l .hxg7 'lt>f7.
1. h6-h7 �e1 -c3
The only way to parry the promo­
tion threat. Black must keep the
knight check in reserve. After
l . . . tZ:lxfS + 2.'�f4 kc3 3 . �xf5
White would reach a winning
bishop ending. The h-pawn is too
strong.
2. f5-f6+!

64
 Va r i o u s P r o m o t i o n Com b i n a t i ons

5. �d2-d 1 11 1 3. 'itJb4xb5 'itJf5-e4

White must retreat his king further 1 4. �b5-b6 �e4-f3
in order to avoid being disturbed 1 5. 'itJb6xb7 �f3-g2
by more knight checks. 1 6. a5-a6 'itJg2xh2
5. ... �e7-f7 1 7. a6-a7 �h2-g2
The king is on its way to eliminate 1 8. a7-a8'ii
the dangerous h-pawn. In the end the other rook pawn is
6. e2-e3 the one that promotes. The finale
Winning the knight. Now a pawn still requires precision.
ending ensues by force. 1 8. ... h3-h2
6. 'ft;f7-g7 1 9. 1i'a8-g8+
7. il!.. b2xd4 ilL.f6xd4 And wins. On the next move White
8. e3xd4 d6-d5 will capture on dS .
A necessary measure. On 8 . . �xh7 .

White would have broken through In the following study White

decisively with 9 .dS ! , when one of achieves the promotion with two
the white queenside pawns will successive bishop sacrifices.
promote.
9. Wd 1 -d 2 wg7xh7
1 0. b4-b5
There is a dual here. First 1 0. �c3
and only then l l .bS is also good.
1 0. ... c6xb5
11. �d2-c3
White's plan is clear. He is going to
collect the black b-pawn.
11. �h7-g6
1 2. wc3-b4 Wg6-f5
Timman 2009
White to play and win

There are only minor pieces on the

board. Apart from the black king's
bishop, they are all more or less
scattered over the board. White is a
piece up, but the black a-pawn can
become very dangerous. Evidently,
White must give a discovered
check with his knight.
1. tt:lf7-d8+ 'itJe6-e5
2. tt:ld8-c6+1
Now a race starts. The black king is White sacrifices his knight in order
going after the white h-pawn. to create a passed d-pawn.

65
The Art of t h e E n d g a m e

2. ... d7xc6 7. �e2-d2! �f5xe5

The alternative was 2 . . . 'iit> f5 . After 8. h6-h7 a3-a2
the follow-up 3 .�xh7 + 'iit> e 6 9. d7·d8'iY a2·a1 'iV
4.�g8+ 'iit> f5 White must pull out If Black took on d8, the white
all the stops to stay on top. The queen would prevail over the
only good move is 5 .�b6 ! in order bishop pair after the promotion.
to force Black to take on c6. There 1 0. il'd8xf6+ �e5xf6
follows 5 . . . dxc6 6.d7 a2 7 .�d4! 11. h7-h8'if+
�xd4 8 .d8"iV ..tfl + 9.<.t>d2 ! �c3 + And wins.
1 0 .'iit> c 2 ! a l "if' l l ."iVd7+ and Black
gets mated. In the next study White strives
3. d6·d7+ �e5-f5 for domination of the enemy
4. �g8xh7+ �f5·e6 bishop.
And now for the bishop sacrifices.
5. �h7-f5+!!
Insufficient for the win was
5 .�b 1 , since after 5 . . . 'iit> x d7
6 .�xf4 �h3 7 .�c l a2 ! 8 .�xa2
�f5 the white king's bishop would
be cut off from play. The text move
forces the black king to go to an
unfavourable square.

Timman 2010
White to play and win

White is a piece up, but he is run­

ning the risk that both his pawns
will disappear from the board. So
he has to try and prevent one pawn
being traded off. But first of all he
has to choose the right square for
his bishop.
1. �d4-c51
5. ... �e6xf5 The introduction to a magnificent
6. �c7-e511 combination. Insufficient for the
Absolutely stunning. win was l .�c3 on account of
6. ... �g2-f1 +! l .. .'iit> g 7 2 .�a5 �e8 3 .ti:lc3 d4
Bishop sacrifices from both sides. 4.gxf6+ 'iit> f7 ! and White cannot
White cannot take, as then the protect both his pawns from
black a-pawn promotes with elimination.
check. 1. ... �h8-g7

66
 Vari o u s P ro m o t i on C o m b i n a t i ons

Now �he threat is again 2 . . . �e8 , 6. ... �f7-h5

with an immediate draw. 7. 'it>f1 -e1
2. �c5-e71 The black bishop is still domi­
An important move. White forces nated. The promotion of the
his opponent to capture on gS. b-pawn can no longer be pre­
2. ... f6xg5 vented.
3. �e7-d6!!
The brilliant point of the previous
move. Now on 3 . . . �e8 White
plays 4.�xe 5 + , followed by
S .tLld4, with an easy win. There­
fore Black has no choice. He must
take the bishop.
3. _ c7xd6

Timman 1990/2011
White to play and win

This is an old study that I cor­

rected several decades later. It
looks as if the promotion of the
d-pawn can be achieved without
trouble.
1. d6-d7 tt:Jc4-e5+1
Insufficient was l . . lad6, since the
.

4. tt:Je2-d4!! endgame that arises after 2 . Wxg5

The second piece sacrifice. White �g7 3 .d8'i!\V l:f.xd8 4.lhd8 is won
prevents Black advancing his according to the tablebases, e.g.
d-pawn and also deprives the 4 . . . tt:Jes S .�e6! �c6 6J:t:g8+ �h7
bishop of the squares c6 and e6. 7. 'it>f6, and the king penetrates
The b-pawn cannot be stopped any with devastating effect.
more, even though it is only on bS. 2. l:te8xe5 �b7-c8
4. ... e5xd4 The point of the previous move.
5. b5-b6 d4-d3 Because the white rook has been
Now White has to watch his step. deflected from the 8th rank, Black
He has only one winning move, can pin the pawn that is about to
and all other moves lose. promote. However, White has an
6. '>t>g2-f1 1 unpinning mechanism at his dis­
The right square for the king. After posal.
6 .�f2 g4 7 .b7 g 3 + Black would 3. J::t e 5-e7+ �h7-g6
win. 4. l:te7-e6+!1

67
T h e Art o f t h e E n d g a m e

1 0. 'fgc3-c7+
From here on several 'waste-of­
time duals' will crop up. This is a
term that signifies that other moves
are also sufficient, but they only
delay the winning process, e.g. by
move repetition. At this point there
were the duals 1 0. 'if d3 + and
1 0 . 'if c2 + . I will not mention them
in the following.
A lot of traffic on the c8-h3 diago­ 1 0. ... Wh7-g8
nal! Black cannot take the rook. 11. Wh5xg5
4. ... �g6-f7 Finally White has found time to
5. d7-d8"i¥ take the g-pawn. Black now has to
The point of the previous move. move his bishop to the long diago­
The rook check served to unpin the nal in order to avoid immediate
white passed pawn. defeat.
5. ... �c8xe6+ 11. ... �h3-g2
6. '.t>g4-h51 1 2. 'ifc7-d7
The only correct king move. White Methodically White tries to gain
has to leave the g-pawn in peace territory.
for the time being, since after 1 2. l:te6-c6
6 . Wxg5 l::l. c S + 7 .Wh6 .ixh3 1 3. "ft'd7-e7
8 ."iid 6 l::l.c 8! Black would build an
impregnable fortress.
6. ... �e6xh3
From this moment on everything
goes according to the tablebases,
though of course I did not know
that when I composed the study.
7. �d8-d5+
Driving the rook to a less favour­
able square.
7. ... l:rc6-e6
8. 't\Yd5-f3+ 'i.t>f7-g7 So far I had found all the moves by
The only square for the king. White myself in 1 99 0 . It seemed to me
cannot capture the bishop, because that White should be winning, al­
then there is a rook check on h6. though I couldn't see clearly how
9. 'fgf3-c3+ �g7-h7 to go about it. What I didn't realize
Black keeps preventing the capture was that here a situation of mutual
of the bishop. But now White can zugzwang has arisen. Both sides'
drive his king to the back rank. pieces are on their best squares.

68
 Various Promotion Combinations

13. ... �g2-h1 The Novotny theme signifies that a

The most tenacious defence. piece is placed on the intersection
14. Wg5-f5 between a diagonal of the oppo­
Now that Black has no check on nent's bishop and a file (or rank)
h3, this is strong. The black rook on which his rook operates. This
now has to move to a less favour­ theme occurs in problems and
able square. studies as well as practical games.
14. ... .klc6-h6 In the Czech problem composer
15. �e7-d8+ Anton Novotny's original problem,
Here there is a real dual. 15.'ele3 the interference was performed by
also leads to mate in 41. a bishop, in his second problem it
15. ... Wg8-h7 was performed by a knight. Both
16. 'fUd8-g5 problems were published in the
Forcing Black to play his bishop. Leipziger Illustrirte Zeitung in 1854.
16. �h1-f3 One year later he came up with the
17. 'fUg5-e7+ �h7-g8 following problem, where the
18. <o1tf5-g5 l.:!.h6-c6 theme was repeated.
19. <o1tg5-f4 �f3-h1
This is almost the same position as
a few moves earlier, only now the
white king is on f4 instead of f5.
20. 'fi'e7-e8+ Wg8-g7
21. 'fi'e8-e5+ <o1tg7-g8
22. 'ii'e5-h51

Novotny, Leipziger lllustrirte Zeitung 1855

Mate in 3

This looks a little like a position

from a game. First White will have
to set things right on the
queenside, since the interference
move Ltg4 fails to I..Jlxc3.
This attack on the bishop is lethal. 1. .!:[c3-c5!
22. ... l:Ic6-c1 Threatening mate in 1 .
A last desperate line of defence. 1. ... �a7xc5
23. <o1tf4-e3 .!:[c1-f1 Black is forced to close the c-file
24. �h5-g6+ <o1tg8-f8 for the rook, as 1 ..Jlxc5 was not
25. we3-e2 possible in view of 2.�f2+ and
And White wins the bishop. mate.

69
The Art of t h e E n d g a m e

2. .ille 2-g4! 1. ... J:!:d3-e3

After the interference on the 2. g6-g7 .illd 1 -b3
queenside, White moves the battle Everything appears to be ready for
scene to the kingside. This interfer­ the Novotny, but closer inspection
ence move with the bishop is the shows that White has to be patient.
only move that wins. Black is 3. .l::tc 6-c7+!
mated on g3 or h2. After 3 . .l:!.e6 .l:!.xe6 4.g8'ii' i1i,c4+
White cannot win. Therefore he
In studies, the Novotny mostly introduces an intermezzo. With the
serves to enforce a pawn promo­ rook check he protects the e-pawn,
tion. As an example I give a very which gives him time for a
good study made by the Hungaro­ zwischenzug with the king.
American grandmaster and study 3. ... wb7-a6
composer Pal Benko. The only square for the king. After
3 . . . �b6 4Jlc3 White would win
immediately.
4. 'it>f1 -f2
On this square the king is better
placed, as there will be no bishop
checks later on.
4. ... l:te3-e4
5. l:!.c7-c6+
It's time for the Novotny theme.
5. ... 'it>a6-a5

Benko, 'Magyar Sakkelet' 1977

White to play and win

In this miniature, as so often, a

pawn promotion is the main point
of the Novotny theme. White has
to choose whether he wants to
promote the e- or the g-pawn. In
the first case the intersection of the
Novotny is on e6, in the second it
is on g6.
1. e6-e7 6. l:tc6-e6!!
Insufficient is l .g7 on account of With his king on a safe square,
I ..Jig3 2 .e7 ii,hs 3 . .t:i.g6 .C:xg6 (not White's interference is now lethal.
3 ...i1i,xg6, as then White has the 6. ... l:te4xe6
winning intermediate move 4.c;t>f2) 7. g7-g8� l:te6-f6+
4.e8'i¥ Itf6+ S .�e l i1i,xe8 6.g8'i¥ 8. wf2-g1 !
i1i,d7 ! and Black has built a fortress. Another safe square for the king.

70
 Vari o u s P r o m o t i o n C o m b i n a t i on s

�- ... �b3xg8 and 2 . �£'2 �h5 ! 3 .�e 1 l:i.b8 is also

9. e7-e8� prospectless. The only road to the
And wins. Black loses a piece. draw runs as follows: 2 .�f5 ! l:rf8
Benko received first prize for this 3 .�d3 �e4! 4.�xe2 �c2 5 .�c4
study in the competition organized b2 6.�a2 and White holds by a
by the Hungarian magazine. thread.
The only thing missing in this 2. J::t e3xe8 b3·b2
study is the thematic sacrifice with 3. �h3-f5 �a8-e41
which one of the enemy pieces is
drawn beyond the intersection
point.

So far we have seen a knight and a

rook occupy the intersection point.
A bishop can also do it.

Here is the Novotny. But now

White has an incredible saving
clause.
4. .l:!.e8-a8+!!
This anti-Novotny sacrifice en­
ables White to stop both passed
Timman 2011 pawns.
White to play and win
4. �e4xa8
White is a pawn behind. He must 5. �g1 -f2 �a8xg2
try and eliminate the far-advanced 6. �f2xe2
e-pawn. Draw.
1. .l:!.g3-e3
Insufficient was 1 . 'it>f2 in view of I have thought for a long time that
l . . . l:i.e8 2 .�e 1 �d5 , and Black a Novotny with a queen as interfer­
keeps both pawns. ing piece would be very unlikely.
1. ... .l:!.h8-e81 Until 2 004, that is, when Forster's
The thematic sacrifice, drawing the voluminous biography of Amos
white rook beyond the intersection Burn came out. In this book,
point. A dangerous alternative was Forster included a casual game that
1 . . . �f3 . Then White has to play was played in 1 9 1 0. The combina­
very accurately to hold his ground. tion it contains had been lying hid­
He has to avoid the rook ending, den for 94 years.

71
T h e Art of t h e E n d g a m e

MacDonald-Burn Timman 2011 (after MacDonald-Burn)

Liverpool 1910 White to play and win

It looks as if the open g-file gives White is a rook up, but the open
White a decisive attack. However, g-file is giving him headaches.
Burn comes up with a fantastic re- Black's attack can become danger­
source. ous, especially with his far ad­
33. ... 'i/Vc8-g4!! vanced passed pawns in the cen­
A pure Novotny with the queen. tre.
Not only is the g-file obstructed 1. .!lh1 -h411
for the moment, the short d 1 -hS The thematic sacrifice, drawing the
diagonal is also temporarily closed black bishop beyond the intersec­
off. tion point g 5 . Black has to take the
White has to liquidate to an end­ rook, as otherwise he will be left
game. with too little material for the
34. .t:lg1 xg4 tt:lh4-f3+ attack.
Now White could have obtained an 1. ... ..td8xh4
advantage with 3 5 .�g3 tbxd2 2. 'iVc1 -g511
3 6.l:txg 5 + <;;t>h 6 3 7 J:tf5 . From Burn's Novotny with reversed
here on he can activate his king via colours.
h4 in order to gain the initiative. 2. ... llg6xg5
However, MacDonald played Under these circumstances, too,
3 5 .�g2 and after 3 5 . . . tbxd2 this capture represents Black's best
3 6.l:txg 5 + �h6 3 7 .h4 tbxb3 Black chance.
eventually managed to win the 3. tt:lh5-f6+ wg8-g7
game. 4. tt:lf6xd7
White has regained the queen.
Burn's 3 3rd move belongs to the Now, however, a pawn race starts,
most beautiful moves in the en­ with strong passed pawns on both
tire history of chess. I have incor­ sides.
porated it into the following 4. ... .t:lg5xg4+
study, and added a few extra 5. 'it;>g2-h3 e4-e3
points. 6. wh3xg4

72
 Va r i o u s Promo t i o n C o m b i n a t i ons

It's a close call. The immediate

6.b7 was insufficient for the win in
view of the following attractive
line: 6 . . . e2 7 .tle l �g3 ! 8 .<J;;lxg4 d2
9 . Ihe2 dl � 1 o . �xg3 'i!i'b3 +
l l .�g2 'iWxb7 1 2 J:f.e7 + �h8 ! and
White cannot win.
6. ... e3-e2

Herbstman 1929
White to play and win

A typical Novotny position. White

has to try to promote one of his
queenside pawns.
1. c6-c7 jLf1 -h3
Black has to try and stop the
a-pawn with his rook.
2. f4·f5!!
Here is the pawn Novotny. Black
7. .l:f.d 1 -g 1 1 has to capture with the bishop,
Again, the g-file is the scene of ac­ closing off the 5th rank for his
tion. The rook is lying in ambush. rook.
On 7 . . . d2 White now has 8 .'ot>f3 + , 2. �h3xf5
winning. 3. a6-a7 �f5·e4+
7. ... i.h4·f2 4. wh1 ·g1 J::l: h 5-g5+
8. b6·b7 i.f2xg1 5. �g1 -f21
9. b7-b8'fi e2·e1 � Not S .'ot>fl on account of S .. J.::tfS+ ,
The race has ended in a tie, but followed by 6 . . J::tf8 , and Black wins.
with the new queens on the board 5. ... l::%.g5-c5
White has a mating attack. Now 5 . . JHS + would fail to
1 0. �b8-f8+ 6. �e3 , attacking the bishop.
With mate next move. 6. tt:Jb4-c61
This knight Novotny immediately
In principle, you cannot achieve an decides the battle.
interference of a diagonal and a
line with a pawn - after all, the It seemed interesting to create a
pawn is already on the line of the study where the Novotny theme is
rook. But it can interrupt a rank on linked with an en-passant capture. I
which the rook is standing, and did realize that this would merely
this motif has served as the basis be a pseudo-Novotny. The result
for a study. looks like this:

73
The Art of the E n d g a m e

Timman 2010 Timman 2010

White to play and draw White to play and draw

The white pawns are still more First two introductory moves.
than one square away from promo­ 1. tt:lg3-f5+ 'lt>h4-g5
tion, and so White has to take 2. tt:lf5xd4
some unusual measures to main­
tain the balance.
1. �c8-f5!1
A stunning bishop sacrifice. The
g-pawn is protected and the black
king is decoyed to fS , so that
White can promote with check
later on. Insufficient was l .g7
il.a2 2.. �.f5 <itf4 ! 3 .c7 il.ds , when
the white king is trapped in a mat­
ing net.
1. ... �g5xf5 Black can capture the knight in two
After I . . J:te l 2.g7 .!:i.e8 3 .ild7 ways:
White would reach a tenable A) 2...lhd4 3 .g 7 jLa2 4.c4 ! !
bishop ending. l:Ld2+ s .�gl �xc4 6.c7 with a
2. g6-g7 �b1 -a2 draw;
Things look gloomy for White, but B) 2...exd4 3 .il.f5 ! ! The same
he has a miraculous escape. sacrifice as in the previous study.
3. c2-c4!1 3 ... <>itxf5 4.g7 l:Ld2+ S .Wgl ! ..ta2
No matter how Black takes this 6.c4 ! ! dxc3 7.c7 and again Black
pawn, the c-file will be closed. cannot win.
3. d4xc3
4. c6-c7 The move c2-c4 never fails to be
Draw. spectacular, but it never results in a
real Novotny. The en-passant cap­
The following study is a slight ture closes off the c-file for the
elaboration on this theme. black rook.

74
 Various Promotion Combinations

It is, however, possible to make surprises, and Rinck uses many su­
studies where a pawn is moved to perfluous pawns for his adaptation
the intersection point of two diag­ of the theme. Hardly a typical situ­
onals. T his is called the 'Plachutta ation. Usually Rinck was a past
theme', after the Austrian problem master when it came to economic
composer who invented it in 1858. use of material.
A Plachutta is also possible when a Probably the study received first
file and a rank are crossed by an prize because in the 1930s there
enemy piece. The famous French were not yet many studies with the
endgame study composer Henri Plachutta theme. However, I
Rinck has made an adaptation of thought it ought to be possible to
this theme. make a better adaptation of the
theme. In general, in studies with a
queen on the board you have to
use a lot of material to get every­
thing right. But in this case it
turned out to be possible to build
in many interesting points from a
natural position.

Rinck, 'Gazette de Lausanne' 1932

White to play and win

White has two passed pawns that

are about to promote. With two
thematic piece sacrifices he creates
a Plachutta.
1. �f1-c41
Timman 2011
Drawing the queen beyond the
White to play and win
critical point.
1. ... 'iYe6xc4 Black has a large material superior­
2. l:ta1-h11 ity, but the white g-pawn is very
And now the bishop. dangerous, and the a-pawn can be­
2. ... �c6xh1 come dangerous as well.
3. d4-d51 1. a5-a6 c3-c2
And mate is unavoidable. Black has to look for counterplay.
This study received first prize in On l...'i:Yds the most convincing
the tourney of the Swiss newspa­ win for White is V;t>h 1 �h3
per. To be frank, this surprised me. 3.�fl.
The study does not contain any 2. 'it>g1-h2!!

75
The Art of the Endgame

A marvellous quiet move. It was 5. 'iia 2xc4

tempting to play 2 . a 7 . Then 6. a6-a7 �c2-e4
2 . . . �e6 is met by the winning
3 .l:he6 c l 'iV+ 4.�fl . However,
Black has 2 . . . jLe4 ! , winning after
3 J:he4 c l 'iY+ 4.'it>h2 'ltca3 ! . With
the text move White prepares fur­
ther actions.
2. ... c2-c1 'ii
Black sacrifices his c-pawn to
create counterthreats. Useless was
2 . . . �e4 in view of 3 . �fl and the
bishop has to go back again.
3. l:te1 xc1 it.f5-e4 7. d4-d5!
The Plachutta. Black must capture
with the bishop in order to avoid
immediate mate. It is astonishing
how safe the white king is, though
protected by one pawn only. Because
the black bishop had to go to e4, the
4th rank is closed off for the queen.
7. �e4xd5
8. g7-g8'ii + �d5xg8
9. a7-a8'i!t+ fi'c4-c8
1 0. 'iia 8-a5+ �d8-e8
Now what? On 4.�e2 , 4 . . . 'i¥g8 is 11. l'ia5-e5+ �g8-e6
sufficient. l l . . .�f7 is met by 1 2 .'ile7#. The
4. .Uc1 -c211 capture on g8 has blocked an im­
This rook sacrifice not only serves portant escape square for the king.
to prevent mate, but it also forces 1 2. �e5-h8+
the black bishop to leave the long And wins. In three moves the
diagonal. queen has moved from one corner
4. ... j_e4xc2 square to another.
5. .tb5-c411
Just like in Rinck's study, White Analogously to the Double Prokes,
sacrifices a rook and a bishop to re­ there are also Double Novotnys.
alize the Plachutta. The queen is With this I mean Novotnys with
drawn beyond the critical square the same piece, so Herbstman's
b3 and now it can only try to stop study does not meet this demand.
the g-pawn. Insufficient was S .g7 Fritz composed a magnificent
in view of S . . . �b3 , after which study that also served as a source of
White runs out of options. inspiration for Benko.

76
 Various Promotion Combinations

Black has built up a new line of de­

fence. But now comes the second
Novotny.
7. .!:1e6-g6!!
The intersection point has been
shifted. White wins.

There is no doubt that this con­

struction, with a rook as the inter­
fering piece, is the most manage­
Fritz, 'Lidova Demokracie' 1950
able.
White to play and win
Eight years after Fritz, the Hungar­
Black has dangerous counterplay. ian study composer L.Zoltan suc­
White has to play sharply in order ceeded in activating a third
to gain the upper hand . Novotny in one study.
1. e6-e7 l:tg5-e5
2. g6-g7 c4-c3
Opening the a2-g8 diagonal and at
the same time bringing one of his
own passed pawns closer to the
promotion square.
3. l:!.c6-e611
The first Novotny.
3. ..
. �a2-d5+
4. f2-f3!
Otherwise Black can take on f2
Zoltan 1958
with check .
White to play and win
4. �d5xf3+
5. '>t>h1-g1 l::[e5-g5+ It is clear what is going to happen
6. '1t>g1-f1 �f3-h5 here.
1. c6-c7 l:!.b5-c5
2. e6-e7 �e2-b5
3. l:tf6-c6!
The first Novotny.
3. ... .!:!:c5-e5+
4. '1t>e1-d1 �b5-e2+
5. '>t>d1xc1 �e2-g4
6. l:tc6-e6!
The second Novotny.
6. ... J:!.e5-c5+
7. ..t>c1-b1 �g4-f5+
8. d2-d3!

77
The A r t of t h e E n d g a m e

White has to sacrifice the pawn, The first Novotny.

and now Black gets the opportu­ 2. ... �g2xe4
nity to build up a new line of de­ Forced, since after 2 . . Jhe4 3 .ti:Jb8
fence with his rook and bishop. Black would get mated quickly. He
8. ... �f5xd3+ cannot cover both the a6- and the
9. �b1 -a2 �d3-b5 c6-square at the same time.
1 0. J:re6-c61 3. e7-e8tt:J!
The rook is back. This third The Phoenix Principle. The pro­
Novotny is absolutely lethal. moted knight aims for the square
c7. Insufficient was 3 .e8Wf? on ac­
The Israeli endgame study com­ count of 3 . . . �b7 and Black draws
poser Noam Manella has made a the long end of the stick.
study in which a knight functions 3. ... �e4-b7
as the interfering piece twice. The 4. tt:le8-c7
version he published in The
Problemist in 1 99 1 contained some
snags. However, it wasn't hard to
repair the study.

4. ... J:te2-e6
With all his might Black will try to
protect the squares a6 and c6.
Manella, The Problemist' 1991 5. tt:Jd7-b8 d6-d5
(correction Timman 2011)
6. tt:lb8-c6+!
White to play and win
The second Novotny. There are two
To the original version, a white mate patterns now:
pawn on hS and a black knight on A) 6 ...ihc6 7.ti:Jxa6#
h6 have been added. The white B) 6 ..Jhc6 7.ti:JxdS#
knight has been moved from gS In both cases, the promoted knight
to g 3, and also a first move by gives mate.
White with the e-pawn has been
added. I invented a new challenge: a Dou­
1. e6-e7 .l:!.f2-e2 ble Novotny with a queen as the
Black has to try and stop the pawn. interfering piece. Here is the re­
2. tt:Jg3-e411 sult:

78
 Various P romoti on C om b i n a t i ons

The point of the preceding checks.

With this third check Black averts
the mate threat. If White moves
away his king, Black will take on eS
with the bishop and he can look to
the future with confidence.
4. i¥e5-g3!!
The second Novotny. The intersec­
tion point is shifted from eS to g 3 .
4. ... il.h2xg3
Timman 2011
The best chance. After 4 . . . l:l:xg3 +
White to play and win
5 .� f2 Black would lose quickly.
Black not only has a huge material 5. e7-e8ir+ �h8-h7
advantage, but he also has danger­ 6. ire8-e4+
ous threats. These are presumably Now the point of the second queen
the necessary conditions for a sacrifice becomes clear. The e-file
queen Novotny to work. has remained open for the new
1. �f5-e51 1 queen. Without this check White
The first Novotny. White threatens would be lost.
mate in one move in two ways. 6. ..
. 'it>h7-g8
Black has to give a number of The king cannot stay on the h-file.
checks. 6 . . . g6 would be met by the devas­
1. ... tt:lb5-a3+ tating 7 .'i¥g2, when White threat­
2. �c2-d2 tt:la3-b1 + ens to capture the rook with check.
The move 2 . . . b I lLl+ has no power. 7. 'it'e4xb1
The most convincing win for It is clear that White is in control.
White then is 3 .'it>d l .l:!d3 + 4.�e2 However the battle isn't over yet,
�d2+ S .�fl .l:.dl + 6 .�f2 �d2+ since Black can maintain his dan­
7 .'ii e 2, and Black is finished. The gerous b-pawn.
white passed pawns decide. 7. �g3-b8+
3. Wd2xe3 .l:!.h7-h3+ 8. 'it>e3-d2 .l:!.h3xb3

79
T h e Art of t h e E n d g a m e

Black has blocked the passed pawns This i s a quite complex position.
on the queenside for the moment Black has a considerable material
and is also supporting his own advantage and he also threatens to
b-pawn. Nevertheless, there is a promote his a-pawn. White will
forced win for White. have to rely on his a- and g-pawns.
9. 'lt>d2-c2 llb3-b5 1. tt:la8-c7+ �e8-d8
1 0. 'iWb1 -a2+ '.t>g8-f8 2. g5-g6 a2-a1 'iW
11. '.t>c2-b1 3. g6-g7 �a1 -a2
The king has taken the place of the 4. c3-c4!
queen. As a result the queen has Preparation for the Plachutta. The
free rein. White wins. queen is deflected from the a-file.
4. ... 'ifa2xc4
In Fritz's and Zoltan's studies, the 5. tt:lc7-d5!!
two intersection points of the in­
terfering piece were on a rank, in
mine it was on a diagonal.

There is also such a thing as a Dou­

ble Plachutta, with a knight as the
interfering piece. With the follow­
ing study, the brilliant Polish study
composer Jan Rusinek won first
prize in the tourney organized by
the magazine Problem.
The first Plachutta.
5. ... 'ifc4-c1 +
6. tt:ld5-f41
The point of this move is that after
6 . . . 'ilfxf4+ 7 . �g6 Black has no
checks on the g-file.
6. ... �c1 -g1
7. tt:lf4-g21!
The second Plachutta. The intersec­
tion point has been shifted from
dS to g2. Now Black has nothing
Rusinek, 'Problem' 1976 better than a check on c I , after
White to play and draw
which White can repeat the moves.

80
Chapter 5

Knight Promotions
Late summer 2 0 1 0. In a hotel 2. ... f4xg3
room in Antwerp I was thinking 2 . . .f3 would be met by 3 .b8'iV,
about the possibilities of studies with mate soon.
with multiple knight promotions. 3. b7-b8tt:J!
Quite soon I had finished the fol­ The 'Phoenix Principle' : the knight
lowing study. that has been captured on g3 rises
from its ashes on b8.
3 . ... c4-c3
4. tt:Jb8-c6
This was the intention of the
knight promotion. By offering his
knight to be captured, White
avoids the stalemate.
4. ... d 7xc6
5. d6-d7 c6-c5
6. d7-d8tt:JI
The procedure is repeated. The
Timman 2010 previous knight has been sacri­
White wins- Mate in 13
ficed, and a new knight arises.
This study has a main line in which 6. ... c5-c4
White gives a forced mate in 1 3 . 7. tt:Jd8-e6
This means that in a certain way, it Again White is in time to avoid
can also be seen as a problem. First stalemate.
of all White has to prevent the pawn � _ f7xe6
on c3 from becoming dangerous. 8. f6-f7 e6-e5
1. �d1 -c1
Insufficient was l .b7 f3 ! 2 .gxf3
c2+ 3 .Wc l c3 and whatever White
does, Black will be stalemated.
1. ... c3-c2
Now l .. .f3 fails to 2.exf3 ! c2 3 .tLle2
c3 4.tLld4 with mate next move.
2. b6-b7
It is time to think of promotion.
After 2 . ttJfs f3 White would not
have time to give mate.

81
T h e Art of t h e E n d g a m e

9. f7-f8ll'l! vistas were opened up before me.

The third knight promotion. In the end I had seen so many
9. .. . e5-e4 knight promotions that my head
1 0. ll'lf8-g6 started swimming.
This knight too is sacrificed imme­ I shall give a short overview, in
diately. chronological order where possible.
1 0. ... h7xg6
11. h6-h7 g6-g5 Berland's study may indeed be the
1 2. h7-h8'ii' first with the type of knight pro­
Finally the purpose of all these motion that I had discovered in
knight promotions becomes clear; Antwerp. His study - and in fact
this promotion enables White to many that will follow in this chap­
work his way to the enemy king. ter - can also be seen as a problem.
Incidentally, White could also have I will not mention this aspect again
promoted to a bishop. in the following.
1 2. g5-g4
1 3. 'ifh8xc3#

In Antwerp I was the guest grand­

master during the 4th Inventi tour­
nament. It seemed a nice opportu­
nity to show this study to the
audience. It was received with en­
thusiasm. Two weeks later I met
Yochanan Afek in Wageningen
during a competition game for the
Herland, 'Deutsches Wochenschach '
club of the same name for which
1913. White to play and win
we both played. Afek was clearly
less impressed. This looks a lot like my construction.
'I hope you didn't spend too much First White promotes his a-pawn.
time on this ' , he remarked. At my 1. a5-a6 �h2-g1
negative answer he continued: Black is going to lock himself in.
'This theme was first used in the 2. a6-a7 h3-h2
beginning of the previous century 3. a7-a8ll'll
by the Jewish-Romanian study Black has one move left with his
composer Sigmund Herland. Many hindmost h-pawn.
versions have followed after that, 3. ... h4-h3
sometimes with a total of five 4. ll'la8-b6 c7xb6
knight promotions.' 5. c6-c7 b6-b5
This stimulated me to investigate 6. c7-c8ll'l! b5-b4
the theme further. As I delved 7. ll'lc8-d6
deeper into the subject, new chess The familiar pattern.

82
 Kni g h t Promo t i ons

,7. ... e7xd6

8. e6-e7 d6-d5
9. e7-e8�
And wins. An elegant study, but
with a modest set-up.

It is not certain that Herland was

the first to compose such a study.
In Harold van der Heijden's data­
base the following study can be
found. Starting up the mechanism again.
6. f7xg6
7. f6-f7 g6-g5
8. f7-f8tt:ll g5-g4
9. tt:Jf8-e6 d7xe6
1 0. d6-d7 e6-e5
1 0. d7-d8tt::l !
And now White can choose the
route along which he wants to give
mate.
This study differs from Berland's
in that Black has active counter­
F. Fritz play; he has to put his king into a
White to play and win
stalemate position, which is a
A pure pawn ending. White has to well-known theme in the world of
promote his h-pawn. endgame studies.
1. h3-h4 'it>b6-a5 It is unclear when this study was
He cannot stop the h-pawn. There­ made. In the database it has been
fore the king heads for a4 to wedge adorned with the label 'source un­
itself in. known' .
2. h4-h5 Wa5-a4 Who was F. Fritz? He should cer­
3. h5-h6 b5-b4 tainly not be mixed up with
4. h6-h7 b7-b5 Jindrich Fritz, who featured so
This is given as the main line in the prominently in the previous chap­
database. The alternative 4 . . . a5 de­ ter. When I asked Van der Heijden
mands more of White. I will return himself, he couldn't give me a sat­
to this later on, in another context. isfactory answer. In any case we
5. h7-h8tt:JI know for certain that the great Rus­
This is necessary to prevent stale­ sian Alexey Troitzky did not know
mate. him. This is proved by the follow­
5. ... a7-a5 ing study, which has been errone­
6. tt:lh8-g6 ously labelled incorrect.

83
The Art of the Endgame

7. ... g7-g6
This little move was considered to
be the refutation of the study. Black
postpones the choice of either box­
ing in his king or letting it out via
the escape square g S . 8 .4::lxc7 is met
by 8 . . . g5 , with inevitable stalemate,
while 8 .4Jb6 'it>gS doesn't lead to a
win either. However, White is not
obliged to move his knight. Now
Troitzky, '64' 1934/Timman 2011
we come to my relocation. With the
White to play and win
king on g I White would have two
My contribution to this study is winning moves - in other words,
very modest: I have moved the the study would contain a dual.
white king from g i to h i . Never­ 8. rJih1 -g 1 !
theless, we will see that this is an The only move that wins.
improvement. 8. g6-g5
1. �f3-d1 9. tt:la8-b6 c7xb6
Threatening to capture on a4, 1 0. c6-c7 b6-b5
which results in a pawn ending by
force.
1. tt:la1 ·b3
2. �d1 xb3 a4xb3
3. a3-a4 rJih6xh5
4. a4-a5 wh5-h4
From this point onwards the study
is, apart from my relocation of the
white king, a mirrored version of
the previous one.
5. a5-a6 g5-g4
6. a6-a7 h7-h5 11. c7-c8tt:l! b5-b4
7. a7-a8tt:l! 1 2. tt:lc8-d6 e7xd6
1 3. e6-e7 d6-d5
1 4. e7-e8tt:l!
And again White can choose how
he will give mate here.
Remarkably enough, there are
more imitators of Berland's study
than of Fritz's or Troitzky's. The
reason is that more knight promo­
tions are possible if the enemy king
is anchored in the corner.

84
 Kni g h t P ro m o t i ons

Four years before Troitzky's study, a 6. f7-f8tb! g5-g4

follow-up on Herland's study was 7. tbf8-e6 d 7xe6
published. 8. d6-d7 e6-e5
9. d7-d8tb! e5-e4
1 0. tbd8-c6 b7xc6
11. b6-b7 c6-c5
1 2. b7-b8tb!
The fourth knight! Now White
again has two routes to give mate.

We see that the number of knight

promotions has risen from two to
four. Now, study composers faced a
challenge. Would it be possible to
Novikov, '64' 1930
build in a total of five knight promo­
White to play and win
tions? The English chess author and
Anyone who knows anything study composer Harold Lommer
about retrograde analysis will no­ worked intensively on this task.
tice that the bishop on b I has to be Lommer had a habit of making re­
a promoted piece, since the bishop cord-breaking studies. For example,
on c8 can only have been captured in I 9 3 3 he had won first prize in a
on its starting square. The first tournament organized by the Journal
move of the solution is obvious. de Geneve with a study that contained
1. Wd2-c1 all possible promotions (in study
The enemy forces on the queenside circles this task is given the term
must be kept under control. Allumwandlung, in short: AUW) . In
1. e6-e5 itself, that study is not so special.
2. h6-h7 e5-e4 Much more attractive was the one
3. h7-h8tb! e4-e3 Lommer published a little later.

tbh8-g6 f7xg6 Lam mer, 'Feuille d'Avis de

4.
Lausanne' 1933. White wins
5. f6-f7 g6-g5

85
The Art of t h e E n d g a m e

Black threatens mate, and so the The first move is not very elegant.
knight promotion is forced. In 1 96 8 , in an article entitled 'The
1. d7-d8tt:J �c6-d5 Phoenix principle in the
Now White has to pay attention. End-Game' in the endgame study
After 2.g8'ii �e3 + ! 3 .'iixd5 Black magazine EG, Lommer returned to
would be stalemated. this study. On the task of realizing
2. g7-g8�!! �d5-e4 five knight promotions in one
study he wrote: ' . . . It confirmed
my belief that the first promotion
could only be achieved by the cap­
ture of a Black Rook or a Black
Queen' . Now, more than four de­
cades later, we can conclude that
this still holds true.
1. f7xg8tt:J c3-c2
2. tt:Jg8-h6 g7xh6
3. g6-g7 h6-h5
4. g7-g8tt:JI h5-h4
3. e7-e8l::t ! ! 5. tt:Jg8-f6 e7xf6
For the same reason White must 6. e6-e7 f6-f5
promote his third pawn to a rook. 7. e7-e8tt:J! f5-f4
No AUW, since the regular queen 8. tt:Je8-d6 c7xd6
promotion fails, but still quite 9. c6-c7 d6-d5
beautiful because the starting posi­ 1 0. c7-c8tt:J d5-d4
tion is so natural.

Lommer took up any new chal­

lenge eagerly. In 1 93 5 he pub­
lished the following study.

The first four knights have pro­

moted without a problem. But
now the troubles start. White has a
prosaic win here with 1 1 . �g4:
l l . . .h3 1 2 .�xf4 h2 1 3 .t2lb6 axb6
Lommer, 'Journal de Geneve' 1934
1 4.a7 bS 1 S .a8'ii' b4 1 6.�g3 b3
White to play and win
1 7 .f4! b2 1 8. 'iVf3 and mate next

86
 Kni g h t P r o m o t ions

move. This problem is typical for This passage raises several ques­
studies with a record number of tions. Firstly, Lommer's memory
knight promotions: the construc­ must have played tricks on him,
tion around the black king has to since he had published his study
be very solid. Otherwise White one year earlier. It is interesting that
will be able to find his way to the in Stalinist times it should have
black king with a regular queen been possible for a Russian to get
promotion. hold of the Journal de Geneve and
11. tt:Jc8-b6 also to react to it in a letter. In chess
The thematic path. study circles, Russian emigrants
11. ... a7xb6 must have done some courier work.
1 2. a6-a7 b6-b5
1 3. a 7-a8tt:J Somewhat puzzling is the fact that
The fifth knight. three studies by Liburkin have
1 3. ... b5-b4 found their way into the original
1 4. tt:Ja8-b6 b4-b3 Dover edition of 1234 Modern End­
1 5. tt:Jb6-c4 b3-b2 game Studies. Besides the incorrect
1 6. tt:Jc4-d2# study that Lommer also gives in his
article, the other two look as
Not only Lommer was working on follows:
knight promotions. In Russia,
Liburkin was interested in manu­
facturing a correct and ultimate
version. Lommer writes:
'As far back as in 1 93 5 I published a
5 -fold effort in the Journal de Geneve,
but some time later I received a let­
ter from the great Russian com­
poser M.S. Liburkin, telling me that
he had found a "cook" and had cor­
rected it for me. At that time I was
busily engaged in the publication of Liburkin 1938
White to play and win
1234 Modern Endgame Studies and
just had the time to replace it by Again White has to start with the
Liburkin's under his own name. I ugly capture.
thought I was doing this great com­ 1. f7xg8tt:J c5-c4
poser justice, all the more as his 2. tt:Jg8-h6 g7xh6
composition was far superior to 3. g6-g7 h6-h5
mine. Unfortunately I did this per­ 4. g7-g8tt:J! h5-h4
son a great disservice, for later, 5. tt:Jg8-f6 e7xf6
when I had more time, I checked it 6. e6-e7 f6-f5
carefully, finding two bad cooks. . .
' 7. e7-e8tt:J! f5-f4

87
The Art of the E n d g a m e

8. l2Je8-d6 c7xd6 noeuvred to c3 , and there is only

9. c6-c7 d6-d5 one route available.
1 0. c7-c8ttJ d5-d4+ 1 6. b4-b3
11. �c3xc4 d4-d3 1 7. lbb6-a4 b3-b2
1 8. l2Ja4-c3 b2-b1 '*'
1 9. .!:td1 xb1 d2-d1 'if+
20. l2Jc3xd1
The stalemate is prevented and
Black has to move the bishop.
20. ... it.g1 -e3
21 . l2J d 1 -f2#

1 2. wc4-d5
There is a dual here. First 1 2 .tt::lb 6
also leads to the desired result. In
this phase of a study such a dual is
not really a big problem.
1 2. d3-d2
1 3. l2Jc8-b6 a7xb6 Liburkin 1938
White to play and win
1 4. a6-a7 b6-b5
Liburkin started this study with an
admirable pawn move, namely
l .d8tt::l . However, there is some­
thing wrong with the starting po­
sition. Just as in Novikov's study,
the bishop on g I has to be a pro­
moted piece. And that is not all.
When I showed this position to my
wife Geertje, she immediately re­
marked that the position could not
have been reached with regular
play. Geertje is an expert on retro­
1 5. a7-a8ltJI grade problems. She pointed out
The fifth knight! that Black still has all his pawns and
1 5. ... b5-b4 can therefore not have a promoted
1 6. l2Ja8-b6 piece. This makes the starting posi­
The finale requires a delicate tion illegal and the solution,
touch. The knight must be rna- however refined, irrelevant.

88
 Kn i g h t P romotions

'Ge�rtje is probably the only per­

son to notice this after all these
years' , Afek remarked. Yet it is pos­
sible that Lommer had discovered
it as well. With uncommon fierce­
ness, he devoted himself to the task
of composing a study with five
knight promotions. He describes
this as follows:
'I tried to clinch these five elusive
horses in 1 945 and again in 1 946 Lommer, 'Szachy' 1965
White to play and win
in the British Chess Magazine, but
both versions were found incor­ The pawns are again poised to pro­
rect. I cannot remember how many mote.
times I have tried to achieve this 1. f7xe8tt:J d4·d3
quest in the last 3 0 years(!) , but 2. tt:Je8·f6 g7xf6
each time I failed dismally. As I 3. g4-g5
dealt with the knight promotions, I Unfortunately, there is already a
thought of the stubborn mule and second solution here. With 3 . Wf4
continued working whenever I was White achieves the same result.
in the mood.' 3. ..
. f6xg5
These are all too familiar signs of 4. g6-g7 g5-g4
addiction. If Liburkin's last study 5. g7-g8tt:JI g4-g3
version had been correct, Lommer 6. tt:Jg8-f6 e7xf6
would probably not have gone to 7. '>t>f5-g6 f6-f5
all this trouble. 8. e6-e7 f5-f4
9. e7-e8tt:J! f4-f3
Incidentally, Lommer wasn't the 1 0. tt:Je8·d6 c7xd6
only one working on knight pro­ 11. c6-c7 d6-d5
motions. In 1 9 3 S , the Dutch study 1 2. c7-c8tt:JI d5-d4
composer Th.Kok published a 13. tt:Jc8·b6 a7xb6
study with four knight promo­ 1 4. a6-a7 b6-b5
tions, and eight years after that the
Finn V.Karhia followed suit with
the same number. I won't give
these studies, as they do not add
anything significant to the one by
Novikov.

Lommer's last attempt looked as

follows:

89
The Art of t h e E n d g a m e

1 5. a7-a8ttJI 8. e7-e8tLll f4-f3

The fifth knight. 9. t2Je8-d6 c7xd6
1 5. ... b5-b4 1 0. c6-c7 d6-d5
1 6. t2Ja8-c7 11. c7-c8t2J! d5-d4
And the knight will make its way 1 2. t2Jc8-b6 a7xb6
to f3 . 1 3. a6-a7 b6-b5

A dual on move 3 - this signifi­

cantly devaluates the study. There
is, however, a fully correct study by
Lommer with four knight promo­
tions, where the final promotion is
to an ordinary queen. The fifth
knight is hidden in a side variation
- the 'try', as it is called in study
circles.

1 4. a7-a8�
Unfortunately, 1 4.a8tb is insuffi­
cient. After the continuation
1 4 . . . b4 1 5 .tbc7 b3 1 6.tbe6 bxa2
White would not be in time to give
mate.
14. ... b5-b4
1 5. �a8xf3 b4·b3
1 6. 'ijyf3-g4!
Lommer, 'Szachy' 1965 A pretty quiet move.
White to play and win
1 6. ... b3xa2
The solution develops quite sche­ 1 7. J:ra1 xf1 + �g1 xf1
matically. 1 8. 'ijyg4-d1
1. f7xg8t2J d4-d3 Mate.
2. t2Jg8-h6 g7xh6 Thus, in 1 965 we could take stock:
3. g6-g7 h6-h5 only one Liburkin study with five
4. g7-g8ttJI h5-h4 knight promotions had stood its
5. t2Jg8-f6 e7xf6 ground.
6. �f5-g6
The king again steps aside to yield Six years after that, the Norwegian
passage to a black pawn. study composer Jarl Ulrichsen cre­
6. ... f6-f5 ated a study with five knight pro­
7. e6-e7 f5-f4 motions, without duals.

90
 Knight P r o m o t i ons

Zinar became involved. However,

his first two versions in 1 9 8 3 and
1 984 contained irritating duals.
This must have discouraged Zinar,
who was a great expert on pawn
endings. For many years he
stopped working on the theme.
Only in 2 0 0 8 did he come up with
two fully correct studies with five
knight promotions. The first was
Ulrichsen, 'Stella Polaris' 1971 published in the Ukrainian
White to play and win
Problemist, the second in the Uralski
The pawns are standing in line, like Problemist.
narrow gates, which the Phoenix
knights can dance through. The
main line is clear and contains no
surprises:
1. h7xg8tt:J �e1 -a5
2. .l:.b3-b1 + ..ia5-e1
3. ttlg8-h6 g7xh6
4. g6-g7 h6-h5
5. g7-g8ti:JI h5-h4
6. ti:Jg8-f6 e7xf6
7. e6-e7 f6-f5
f5-f4 Zinar, 'Ukranian Problemist' 2008
8. e7-e8ti:JI
White to play and win
9. tt:Je8-d6 c7xd6
1 0. c6-c7 d6-d5 The construction in the lower left­
11. c7-c8ti:JI d5-d4 hand corner is the same as
1 2. tt:Jc8-b6 a7xb6 Novikov's. On the construction in
13. a6-a7 b6-b5 the upper right-hand corner, Zinar
1 4. a7-a8ti:J! b5-b4 remarked in an article in the January
15. ttla8-b6 b4-b3 20 1 1 issue of EG: 'Top right is the
16. ti:Jb6-c4 b3-b2 shape H - Cyrillic "N" for Novikov!
1 7. tt:Jc4-d2# - just the right dedication.'
1. g7xf8ttl
It was a very heavy construction that The correct knight promotion.
Ulrichsen used; all black pieces are Now Black must move his remain­
on the board without joining the ing knight.
play. I am not sure if this was what 1. ... ti:Jh8-g6
Lommer had been searching for. 2. ti:Jf8xg6 h7xg6
More than a decade later, the 3. h6-h7 g6-g5
Ukrainian study composer Mikhail 4. h7-h8tt:J g5-g4

91
T h e A r t of t h e E n d g a m e

5. tt::l h 8-g6 f7xg6 In this study, White's handling of

6. f6-f7 g4-g3 his bishop is quite subtle.
1. c7xb8tt::l e4-e3
2. tt::l b 8-a6 b7xa6
3. �a5-b4!
The right square for the bishop.
3. ... a6-a5
4. b6-b7 a5xb4
5. b7-b8tt::l ! b4-b3
6. tt::l b8-c6 d 7xc6

7. f2xg3
White has to spend a tempo to take
the pawn.
7. ... g6-g5
8. f7-f8tt:J g5-g4
9. tt::lf8-e6 d7xe6
1 0. d6-d7 e6-e5
11. d7-d8tt::l e5-e4
1 2. tt::l d8-c6 b7xc6 7. 'Wt>c5-b6
13. b6-b7 e4-e3 The king has to step out of the way
1 4. d2xe3 again.
Again a time-out, to take the pawn. 7. . .. c6-c5
1 4. ... c6-c5 8. d6-d7 c5-c4
1 5. b7-b8tt::l 9. d7-d8tt::l ! c4-c3
And the knight will give mate on b4. 1 0. tt::l d 8-e6 f7xe6
11. f6-f7 e6-e5
Here is number two: 1 2. f7-f8tt:J ! e5-e4
1 3. tt::l f8-g6 h7xg6
1 4. h6-h7 g6-g5
1 5. h7-h8tt::l l
The fifth knight. Now White can
again choose his route towards the
mate on a3 or c3 .
This was what Lommer had been
searching for all those 3 0 years!
But Zinar had also needed a quarter
of a century to get the job done.
Zinar, 'Uralski Problemist' 2008 Now, three studies with five knight
White to play and win promotions had been made with-

92
 Kn i g h t P r om o t i on s

out duC).ls. But still Zinar wasn't sat­ 7. d6-d7 ..t>g3·h3

isfied. He considered that a great 8. tt:Jd8-b7 �a8xb7
number of knight promotions on 9. d7-d8tt:J
the same promotion square could And the third.
also be realized, drawing inspira­ 9. ... �b7-a8
tion from an old study. 1 0. tt:Jd8-b7 �a8xb7
Stalemate.

For this study, Shmulenson re­

ceived first prize in the 7th USSR
Team Championship. The play is
varied and interesting - only the
number of knight promotions can
be improved. The Russian end­
game composer Sochniev achieved
four promotions, but Zinar man­
aged to surpass that.
Shmulenson, 1970
White to play and draw

The three white centre pawns will

promote to knights.
1. e5-e6 tLlb7·d8!
The best try. An attractive side vari­
ation was l . . . �b5 2 .lue5 h2+
3 . Wh l �a4 4.e7 ..tc2 5 .tt:ld3 ! (in
order to control the e4-square after
promotion) S . . . �a4 6.tt:le5 with
move repetition.
Zinar/Didukh, 'EG' 2009
2. e6·e7!
White to play and draw
Not 2.tt:lxd8 ? in view of 2 . . . .-tbs ,
and Black controls the board - White has an army of seven pawns.
White will be mated within a cou­ Of these, only the b-pawn will not
ple of moves. promote to a knight.
2. h3-h2+ 1. d7-d8tt:J l:.e8xd8
3. Wg1-h1 �a6-b7 2. e7xd8tt:J J:lh8xd8
4. e7xd8tt::l l 3. c7xd8tLl tt::l b8-c6
The first knight promotion. The first three knight promotions
4. ... �b7·a8 have taken place. Now White ad­
5. tt:Jd8-b7 �a8xb7 vances his pawn front in order to
6. d7·d8N achieve the next three as well.
The second. . . 4. e6-e7 tt:Jc6xd8
6. ... �b7-a8 5. e7xd8tt:J 'it>f2·g3

93
The Art of the E n d g a m e

6. e5-e6
Another pawn on its way to a
knight promotion.
6. �g 1 xb6
7. e6-e7 �b6xd8
8. e7xd8tt:J 'it>g3-h3
9. d6-d7 �h3-g3

Zinar 2010
White to play and draw

Again the battle will take place on

the long diagonal.
1. e7xd8tt:J j_b7-a8
2. tt:Jd8-b7
The first Phoenix knight is about to
rise.
1 0. tt:Jd8-c6 2. ... �a8xb7
Here the Phoenix Principle makes 3. d7-d8tt:J! �b7-a8
its appearance. White sacrifices a Maintaining the tension on the
knight in order to clear the promo­ long diagonal.
tion square for the next knight. 4. d6-d7
1 0. ... �a8xb7
11. d7-d8tt:JI
The sixth knight.
11. ... �b7-a8
1 2. tt:Jd8-b7 �a8xb7
Stalemate.
Zinar made this study together
with Sergey Didukh, a brilliant
young study composer from the
Ukraine. The number of knight
promotions in this study is unpar­
alleled. Yet it makes a somewhat ar­
tificial impression, especially since 4. ... tt:Jb5-d6
five of these six knight promotions Black does all he can to eliminate
are captures. Doubtlessly for this the stalemate threat. Now the the­
reason, Zinar started working on a matic 5 .tLlb7 is met by the win­
study that would make a more nat­ ning 5 . . . t2Jxb7. Therefore White
ural impression. This is what he must first take the knight.
came up with: 5. c5xd6 �h3-h2

94
 Kni g h t P r o m o t i ons

Black can only wait. It is clear that White must support

6. 8d8·b7 �a8xb7 his foremost a-pawn. He has a
7. d7·d881 �b7·a8 choice between two king moves.
8. d6·d7 �h2·h3 1. wb5·a6
9. ti:Jd8·b7 �a8xb7 Three moves further on we will see
1 0. d7·d881 �b7·a8 why this is the only winning move.
11. 8d8·b7 1. ... e2·e1 8
After the fourth knight promotion The knights make an appearance
comes the fourth thematic knight again, but this time they won't save
sacrifice. Black.
11. ... �a8xb7 2. �h4xe1 d2xe1 8
Stalemate. 3. �a6·b6 e3·e2
4. a5·a6
There is a difference between mi­
nor promotions on one and the
same square and minor promo­
tions on different squares. The for­
mer serve as protecting moves, the
latter as preparation for avoiding
stalemate.

So far, a total of 6 3 knight promo­

tions have been seen here. In all of
them, except in Shmulenson's
study, one element has been lack­
ing: the lively play that is (usually) A situation of mutual zugzwang
the hallmark of a good study. The has arisen. Black is forced to make
following study by Zinar will a thematic knight sacrifice.
please true study lovers. 4. ... 8e1 ·f3
5. �h1 xg2 e2·e1 8
6. �g2·h3!
Now that Black does not have any
dangerous passed pawns left, the
white bishop can leave the long di­
agonal, and head for b 7 .
6. ... 8e1 ·d3
7. �h3·c8 ti:Jd3·c5
8. �c8·b7+
And mate next move.
For this study Zinar received the
Zinar 2010 special prize in the traditional
White to play and win
Moscow Tourney.

95
T h e A r t of t h e E n d g a m e

Even prettier is the study that Zinar The white king returns in time to
made one year later. He returned to lock up his black colleague in the
the theme of the scattered knight corner.
promotions. 4. Wg2xh1
5. \tle1 -f1 g7-g5
6. h5xg6 h6-h5
7. g6-g7 h5-h4
8. g7-g8tiJ!
The first knight promotion. There
are no surprises in the remainder
of the solution.
8. ... h4-h3
9. tLlg8-f6 e7xf6
1 0. e6-e7 f6-f5
11. e7-e8tiJ f5-f4
Zinar 2011
1 2. tLle8-d6 c7xd6
White to play and win
1 3. c6-c7 d6xc5
There's only one way to stop the 14. c7-c8tiJI c5-c4
black h-pawn. 1 5. tLlc8-b6
1. 0-0-0 <it;>g5-h4 And wins.
The best . chance. After I . . . Wxg4 With his. last promotion White will
2 . .l:ih i 'it>h3 3 .<;t>di 'it>g2 4.'it>e i get a queen, as he did in Berland's
Wxh i S . Wfl gS 6.hxg6 hS 7.g7 study.
h4 8.g8tLl! h3 9 .ti:Jf6 exf6 I O.e7 fS
I l .e8� f4 1 2 .We i Wg2 I 3 .�g6+ In this study Zinar completed the
White wins easily. This winning so-called 'Valladao Task' . The re­
method is not possible when the quirements for this task are that
white g-pawn remains on the the main line of the study should
board. contain all 'special' chess moves :
2. J:rd1 -h1 wh4-h3 castling, an en passant capture
3. Wc1 -d1 Wh3-g2 and underpromotion. See also
4. \tld1 -e1 Chapter I 4. An incredible
achievement.

What could I add to all this knight

violence? I thought that it might
be interesting to try and make a
study with four knight promo­
tions, rounded off with a bishop
promotion. The result looks like
this :
(see diagram next page)

96
 Knight Prom o t i ons

6. g7-g8ttJ! h5-h4
7. t2Jg8-f6 e7xf6
8. e6-e7 f6-f5
9. e7-e8ttJ! f5-f4
1 0. t2Je8-d6 c7xd6
11. c6-c7 d6·d5
1 2. c7-c8ltJ !

Timman 2011
White to play and win

I didn't manage to avoid the short­

coming of the ugly capture on the
first move either.
1. h 7xg8ttJ d6-d5
2. �e4-f3
White cannot take the pawn. More­
over, the bishop has to stay on the So far things have developed sche­
long diagonal in order to prevent matically. Now Black has a choice.
Black breaking loose. 1 2. ... d5-d4
2. ... d5-d4+ The alternative was 1 2 . . . dxc4. In
In reply to the alternative 2 . . . dxc4 that case White wins by 1 3 .ttJb6 c3
White has a clear-cut winning 1 4.�c4! cxd2 I S .�xd3 axb6
method: 3 .'it>c2 c3 4J:rd l cxd2 1 6.a7 bS 1 7.a8'iV b4 1 8 .�e2! and
S .ti:Jh6 gxh6 6.g7 hS 7 .g8ti:J! h4 White will again give mate by
8 .ti:Jf6 exf6 9 .e7 fS 1 O .e8ti:J! f4 bringing his bishop to fl .
I I .ti:Jd6 cxd6 1 2.c7 dS 1 3 .c8ti:J d4 13. t2Jc8-b6 a7xb6
1 4.'it>b l d3 1 S .ti:Jb6 axb6 1 6.a7 b5 14. a6-a7 b6·b5
1 7 .a8'iV (or il) 1 7 . . . b4 1 8 .ile4 b3
1 9 . .1xd3 b2 2 0 .�fl ! ..te3
2 I .ii.xg2#. This scheme, where a
bishop is used to temporarily block
the back rank, we will also encoun­
ter in the main variation.
3. 'it>c3-b41
Only much later will it become
clear why this is the only square for
the king.
3. ... d4-d3
4. t2Jg8-h6 g7xh6
5. g6-g7 h6-h5 1 5. a7-a8j_1!

97
The Art of t h e E n d g a m e

The only winning move. After 1 6. ... c4-c3

1 5 .aS'iY? bxc4 1 6.�e4 f3 ! White 1 7. i£.e4xd3 c3-c2
wouldn't be able to give mate in With 1 7 . . . cxd2 Black could have
time. The problem is that the pawn postponed the mate for one move,
on d3 is protected, so that the but the text demands more of
white bishop cannot reach the White.
fl -square in time. Also, here it be­ 1 8. �d3·f1 d4-d3
comes clear why the white king 1 9. �b4-b3 (a4,b5,c4)
had to be played to b4. If it were on The king has to move to a light
b2 or b3 , then Black would now square, otherwise Black would give
react with 1 5 . . . b4, after which the a deadly bishop check.
maw would be unavoidable. 1 9. c2-c1 �
1 5. ... b5xc4 20. l:ra1 xc1 �g1 -d4
16. �a8-e4 21 . �f1 xg2
The point of the bishop promotion. Mate. On this study I worked in­
The newly-promoted bishop is on tensively for approximately twelve
its way to fl in order to give mate. hours.

98
Chapter S

Bishop Promotions
In the preceding chapter, I showed Of course, this can be achieved in a
how a bishop promotion can serve much more elegant way.
to avoid stalemate. This type of
promotion can also have the oppo­
site effect: supporting self-stale­
mate. Let me show you a modest
example.

Rusinek, 'New Statesman' 1971

White to play and draw

Here, too, the white king is in

grave danger. Especially the c-pawn
is in the way.
Timman 2011
1. a6-a7 i.c4-a6+
White to play and draw
2. b6-b7 tt:Jc3-e4
The white king is in the way of the Threatening mate in one.
g-pawn's promotion, and it also 3. g7-g8tLl+
threatens to be caught in a mating The first minor promotion.
net. For example, White would be 3. ... 'it>e7-e8
mated by force after l .h8'ilf? tlxdS . 4. tt:Jg8-f6+
1. tt:Jd5-f4+1 J::lf5xf4 This was what it was all about. The
Otherwise White will queen his knight that threatened to give mate
h-pawn. is deflected.
2. h7-h8i.! 4. ... tt:Je4xf6
And despite his huge material su­ After 4 . . .<�"Jdxf6 S . �b8 White
periority (or maybe because of it!) would manage to bring his king
Black cannot avoid stalemate. into safety.
Incidentally, White could not play 5. a7-a8�11
l .lLJe7 +?, as then Black wins with A very elegant solution to the
l . . . !il.xe7 2 .h8� !il.f6 ! and the problems in the position. Black
stalemate threat is eliminated. must now try to eliminate the

99
The Art of t h e E n d g a m e

stalemate threat and weave new Again, White has to compete

mate threats into the position. against a large material superiority.
5. ... tt:ld7-e5 At this point it is hard to imagine
On 5 . . . �fl White could draw with that this position will culminate in
6.b8tLl ! . stalemate.
6. 'it>c8-b8 tt:le5-c6+ 1. b5-b6+ <;t>c7-b7
7. �b8-c8 .ia6-f1 2. �f1 -c41
Threatening to give mate from the A strong bishop sacrifice which
other side. Now 8 .b8tb? wouldn't Black has to accept, as otherwise a
do on account of 8 . . . tbe7 + 9.<.t>b7 check on dS follows.
�g2 + 1 0 .<.t>a7 tbc8 + and the pro­ 2. ... �g8xc4
moted bishop is lost. 3. .l:i.d1 xd7+!
8. b7-b8J::m This rook sacrifice Black has to ac­
A real treat of minor promotions. cept as well, since after 3 . . . Wa8
This rook too will serve to con­ 4.axb8'if+ Wxb8 S .l:i.b7 + White
struct a self-stalemate. easily holds the draw.
8. ... �f1 -a6+ 3. ... ttlf8xd7
9. .l:i.b8-b7! 4. a7-a8�+11
Draw. Black has no time to threaten A tremendous surprise. Insufficient
mate since neither of the two pro­ was 4.a8'ii + in view of 4 . . . <.t>c8
moted pieces can move, and nor and Black wins.
can the white king. 4. ... �b7-c8
Not surprisingly, with this study 5. b6-b7+
Rusinek won first prize in the New The point. The bishop is walled in
Statesman tourney. after all. Wherever the king goes,
White is stalemated.
It is even more elegant if the
bishop promotion takes place at a This same principle can be dou­
moment when the self-stalemate is bled.
not yet in sight.

Gurvich, 'Trud ' 1928 Hurtig, 'Schackvarlden' 1943

White to play and draw White to play and draw

1 00
 B i s h op P r o m o ti on s

White starts on the queenside. clear that White has to promote his
1. a 7-a8i.+l 'Ot>b7-c8 pawn.
2. b6-b7+ 1. h7-h8it+!
The first bishop is walled up. Why not 1 .h8'it + ? The answer to
2. ... 'l.t>c8-d7 this question will be revealed at the
3. �a5-d5+ end of the study.
The battle scene is moved to the 1. ... tt:ld3-b2+
kingside. 2. Wd1 -c1
3. ... Wd7-e7 Threatening mate. Now Black has
4. �d5-e5+ we7-f7 to drive the white king away from
5. 'ii'e 5-g7+! its current position with checks.
This queen sacrifice prepares the 2. f2-f1 'iV+
second bricking-up. 3. ttld2xf1 i.d8-g5+
5. ... 'l.t>f7-g7 4. ttlf1 -e3!
6. h 7-h8.Jt+! Keeping the b 1 -h7 diagonal closed
The second bishop appears on the for the time being.
board. 4. ... .Jtg5xe3+
6. ... wg7-f8 5. wc1 -c2 .Jte3-d41
7. g6-g7+ Black sacrifices his bishop in order
With stalemate on the next move. to open the diagonal.
For this study A.Hurtig received 6. �h8xd4 e4-e3+
second prize in the tourney of the 7. 'l.t>c2-c1 c4-c3
Swedish chess magazine. Now it becomes clear why White
couldn't promote to a queen. In
Sometimes the self-stalemate is no­ that case he would have been
where in sight at the moment of hopelessly lost in this position.
the bishop promotion. 8. �d4xc3 l:la4-b4
8 . . . l:!.c4 would have been met by
9 Jha2 + �xa2 with stalemate. Af­
ter the text move, Black threatens
to win with 9 . . . .l:!.b3 .
9. J:ra5-a3!
Preventing this.
9. ... l:r b4-c4
There is nothing better.
1 0. J:ra3xa2+ Wa1 xa2
Stalemate.

Rokirovkin 2009
In the great majority of studies,
White to play and draw
bishop promotions serve to pre­
Black has a huge material superior­ vent stalemate. This theme does,
ity, but his king is in danger. It is however, allow for a lot of choice.

101
The Art o f t h e Endgame

The following study by Benko is called the Excelsior Principle. Al­

very good. though White now has a big mate­
rial advantage, the win is far from
easy. His king's bishop is still
walled in.
5. ... lLld8-c6
6. i.. b 8-g31
The only square for the bishop. On
6 . . . ltJeS + White can now play
7 . 'ii>fs with a win.
6. ... lLlc6-e7+
7. Wg6-f7 lLle7-g6!
Another fantastic resource. White
Be nko, 'Ma gyar Sakkelet' 1988 cannot take the knight.
White to play and win
8. i..h 7-g81 lLlg6-e5+
In this miniature study White has 9. 'it>f7-e81
trouble freeing his bishop. He has And wins.
to start pushing his passed pawn. For this study Benko received sec­
1. b2-b4 lLlg3-e2 ond prize in the tourney of the
The only way for the knight to try Hungarian chess magazine.
and stop the b-pawn. After l . . ltJe4.

2.'1itfS ! tt:Jdl 3 .�g6 tt:Jxf3 4.'ii>e4 If I think of bishop promotions, I

the knight would be dominated, am immediately reminded of
while the black king is too remote Vasily Smyslov.
to stop the b-pawn.
2. b4-b5 lLle2-d4 Smyslov as an
3. b5-b6 lLld4-c6 endgame composer
4. b6-b7 It is a striking fact that Smyslov
Black has managed to stop the pro­ started composing studies at a very
motion of the b-pawn, but this young age. This craft lay dormant
does not mean that his troubles are during the time of his great suc­
over. After 4 . . . tt:Jb8 S .f4 ltJd7 6.f5 cesses. It was only at an advanced
tt:Jb8 7 .f6 gxf6 8.Wh6 fS 9 . .ixf5 age, when blindness stopped him
the white bishop would be freed, playing practical games, that he
and White wins by bringing his picked it up again. For 60 years he
king to d6. had produced very little in the area
4. ... lLlc6-d8! of endgame studies.
Black is playing for stalemate. In 2 0 0 1 a little pocket-sized book­
5. b7-b8i.. l let called Moi Etiudy was published.
The white pawn has travelled the It contains 5 7 studies, 48 of which
entire road from starting square to are from after 1 99 8 . I got the
promotion square. This motif is booklet, autographed by Smyslov

1 02
 B i shop P r o m o t i ons

himself, via Genna Sosonko. 'To quire great precision, and the
dear Jan Timman' , said the inscrip­ pawns on the h-file make the study
tion. I was struck by the steadiness uneconomical. Smyslov and
of his handwriting; it had not been Pervakov, who edited the booklet,
affected by his blindness. In 2 0 0 5 a may have decided to leave it out.
second edition appeared, contain­ The book does contain a variety of
ing a total of 1 1 4 studies. This follow-up studies on the same
means that in four years he had theme: a black queened pawn sur­
made exactly another S 7 studies. rounded by white minor pieces,
Curiously, Moi Etiudy does not con­ followed by a minor promotion for
tain the first study that Smyslov had White.
made back in 1 9 3 6, a year after his All those studies are more eco­
well-known game against Gera­ nomical than the original one, but
simov. That study looks as follows: I find them less impressive. An ex­
ample:

Smyslov, '64' 1936

White to play a nd wi n Smyslov 1999
White to play and win
It goes as follows:
1. ..ltc2-b1 ! a2-a11!V+ The solution goes as follows:
2. c;t>a5-b5 �f2-g3 1. .tf5-b1 a2-a1 'ii
3. g6-g7 ..\ig3-b8( 1) 2. wc5-b4 c;t>f3-f4
4. g 7-g8�!1 �b8-f4 3. g5-g6 \t>f4-g5
5. .tg8-a2 ..if4xd2 4. g6-g7 c;t>g5-h6
6. f5-f6 �d2-f4 5. g7-g8�1
7. f6-f7 �f4-d6 Here, too, meant to prevent a stale­
8. \t>b5-c6 �d6-f8 mate.
9. wc6-c7 5. ... c;t>h6-g5
And wins. 6. �g8-a2
It's a great study, but it does have a And now the picture is complete.
few snags - which is understand­ 6. c;t>g5xg4
able if you consider that Smyslov 7. \t>b4-c3 \t>g4-f3
was only 1 S . The end does not re- 8. c;t>c3-d2 c;t>f3-g2

103
The Art of t h e E n d g a m e

9. wd2-e2 'it> g2xh2 The situation is critical for White.

1 0. �e2-f2 'it> h2-h1 He must be careful not to be out­
11. ti:l a3-b5 (c4) tempoed, e.g. 8 .�f5 ? hS 9 .�c l h4
And wins. But not 1 1 .tt::lc 2 in view 1 0 .'it>d 1 d4, and this is a situation
of 1l...�xbl! 1 2...�.xb 1 h 2 , and of mutual zugzwang.
stalemate is inevitable. 8. �b1 -c1 h7-h5
9. �d7-b5!
In 2 0 0 9 I started working on a The only bishop move to take con­
variation on Smyslov's original trol of the situation.
study with the intention of making 9. h5-h4
it economical, adding some ele­ 1 0. �b5·f1 d5·d4
ments and creating an interesting 11. wc1 ·d 1
ending. This is how it looks: And another situation of mutual
zugzwang has arisen. White to
move would lose, Black to move
cannot make progress, since after
11. . .d3 12...�.h 3 d2 13 .�fl he
would even lose.
It had been my intention to show
this study to Smyslov, but I never
had time to do it.

In the years when Smyslov was in­

active in the area of studies, he cre­
Timman 2010 (a fte r Smyslov ) ated a fantastic twin study.
White to play and draw

The first few moves are obvious.

1. a6-a7 �d6-b8
2. a7-a8'ii'+ wa4·b4
White seems to have run out of op­
tions, since 3 .'iVa6 is met by 3 . . . g2.
3. �f7-g6!
An improbable move. Now 3 ... hxg6
can be met effectively by 4.'i!Ya6 af­
ter all, since the white queen can
now go to either g6 or fl .
Smyslov, 'Pravda' 1976
3. ... g3·g2
White to play and wi n
4. �g6-b1 g2-g 1 � !
The well-known motif. A simple bishop ending is on the
5. �b1·f5 il.g1 xd4+ board. It's hard to imagine that
6. wa1 -b1 �d4-a7 White will soon promote to a
7. �f5xd7 'it> b4-c3 bishop.

1 04
 B i sh op Promo t i ons

1. f6·f7 �c1 -a3 The exchange of bishops would

2. .lic3-g7 not help Black.
After L�.b2 ..tf8 3Si7e2 '>i;?dS 8. �g7-h6+ 1 -0
Black will save himself.
2. ... f4-f3! In the second study Smyslov shows
3. g2xf3 that the defensive plan can also be
White has to take the pawn, since successful.
after 3 .f8'iV �xf8 4 .�xf8 e2 +
5. �f2 fxg 2 he would even lose.
3. ... Wc4-d3l
The point of the preceding moves.
On 4.f8'iV? Black has planned the
reply 4 . . . e2+ S .'>i;?f2 �cS . If White
takes the bishop, then Black pro­
motes his e-pawn to force stale­
mate. If he plays 5 . We 1 , then
S . . �xf8 follows, with a draw after
.

6.�xf8 We3 .
4. f7-f8�1 Smyslov, 'Pravda' 1976
White to play a nd draw
Removing the stalemate from the
position. The question remains if We have switched colours. The
White can win with his unusual main reason why White can draw
bishop pair. is that his bishop eventually has
4. ... e3·e2+ enough squares.
5. Wf1 ·f2 e2-e1 'if+ 1. d5·d6 c3-c2
6. Wf2xe1 Wd3-e3 2. �e1 -d2 j_d4-b2
3. c5-c61
Same procedure.
3. ... b7xc6
4. Wf5·e6 c2-c1 �
5. d6-d7+ �c8·d8
6. '>ite6-d6 c6-c5
7. 'iit d6xc5 Wd8-c7
8. �d2·e1 !
And Black cannot win.

Shortly after Smyslov had made

Black defends as stubbornly as pos­ this twin study, he showed it to the
sible. After 7 ...txa3 '>i;?xf3 he would American grandmaster Lubosh
also win the h-pawn. Kavalek.
7. f3-f4! we3xf4 'What do you think of it?' , the for­
8. We1 -f2 �a3-c1 mer World Champion asked.

105
T h e Art o f the End g a m e

' Brilliant' , Kavalek replied. 1. ... lLl h6xg8

'Do you really think it's good?' 2. h 7xg8�
Smyslov asked once again. And The same principle.
again Kavalek professed his enthusi­ 2. �d5xg8
asm. Later he told me that he had 3. itd8xg5 �g8-d5
been quite surprised by Smyslov's 4. g7-g8� �d5xg8
apparent lack of self-confidence. I 5. g6-g7
think that things are slightly differ­ A fourth pawn is getting ready to
ent. If you have completed such a promote to a bishop.
beautiful work, you are fully aware 5. ... .ig8-d5
of its worth. But you still want to 6. g7-g8.i �d5xg8
sound out the opinion of an expert; 7. .ih5-g6
his enthusiasm will strengthen your This square has been cleared for
deep feelings of satisfaction. the boxed-in bishop. Mate has be­
come unavoidable.
Just as with knight promotions, 7. ... .ig8-d5
there once was a contest to include 8. 1i.g6-d3+
the greatest number of bishop pro­ And mate follows.
motions in one study. And again, Not a very impressive study. The
Lommer and Zinar were the great­ play contains no highlights and
est rivals, albeit in different peri­ half of the bishop promotions are
ods. Lommer achieved a total of achieved with captures.
four promotions. Before Zinar composed his study
with a series of bishop promotions
he had set himself another goal: a
pawn ending with a melange of
bishop and knight promotions.

Lemmer, 'British Chess Magazine' 1945

White to play and win

The white pawns are poised to pro­

mote.
Zinar 1980. Black to move ­
1. f7xg81i.
White wins (incorrect )
Not l .fxg8"ii' in view of l ...ltJ£'7
2 .'ifxf7 .l:hb6+ 3 .'ifxd5 .l::rb 8+ The intended solution is:
4.'>t)xb8, stalemate. 1. ... f2-f1 ttJ

1 06
 B i shop Prom o tions

Threatening mate in one. In most to prevent this with 1 . 'iY a3 , Black

cases the pawn structure around brings along another knight to re­
the white king serves to play for inforce the threat.
stalemate, and in that event the op­ 1. �b4-c3!!
posing side has to try and give This queen sacrifice has the same
mate in time (see for instance function as the knight sacrifice
Troitzky's study in Chapter 5 and from Zinar's (incorrect) study.
one of my own studies in Chapter White attacks the enemy g-pawn,
8) . In the present case, stalemate and therefore Black has to accept
motifs do not play a role. White the queen sacrifice.
has to try to prevent an enemy 1. ... tt:Je4xc3
knight from mating him. 2. d7-d8�! 1
2. b7-b8� Not 2 .d8'i!V in view of 2 . . . lt:lde4
As after 2 .b8'i¥? lt:lg3+ 3 .'ilfxg3 3 .'i\Yxc7 fl lt:l 4.�f4 lt:le2 , and
Black is stalemated. White cannot prevent Black giving
2. ... 'iith 1 -g2 three knight checks on g3 and then
3. e7-e8tt:J being stalemated.
After 3 .e8'i!V h l lt:l Black would win 2. tt:Jc3-e2
immediately. 3. i.d8xc7 f2-f1 t2J
3. ... h2-h1 t2J
4. tt:Je8xg7
And wins.

For this creation Zinar was inspired

by a fantastic study by Troitzky and
Korolkov.

4. b7-b8�!!
The climax of the study. The criti­
cal square g3 can only be protected
by bishops.
4. ... tt:Jd2-e4
5. i.c1 -f4
And the three white bishops keep
TroitzkyjKorolkov 1938/1939 the three black knights under con­
White to play and win
trol.
Actually, White must take great For this study Troitzky and
care not to lose. Black is threaten­ Korolkov received first prize in the
ing mate on g3 and if White tries Chigorin Memorial Tourney.

1 07
The Art of t h e E n d g a m e

In Zinar's version, White's second After l . . .fl WV 2.b8Wi Black will

bishop promotion fails, but there is have no chance.
a second black knight promotion 2. b7-b8�1
plus a white knight promotion. Not 2 .b8'i!Y on account of
However, the study is incorrect be­ 2 . . . tbg3 + 3 .'YWxg3 l:txgS + 4.\t>xgS
cause Black can win with the simple with stalemate.
1 . . .fl Wi. Both sides can get two 2. ... �h1 -g2
queens and then the white king is Preparing the second knight pro­
too exposed, despite the defensive motion.
pawn construction. More generally, 3. e7-e8ttJI
something else is wrong with this Certainly not 3 .e8�? h 1 tb ! and
concept. The purpose of the white White gets mated.
knight promotion is to attack the 3. ... h2-h1 '1!1
enemy g-pawn and thus to clear the The best chance in the circum­
escape square h6 for the king. How­ stances. Now there follows an in­
ever, White can also attack the pawn teresting finish.
with his newly promoted bishop. 4. tt:le8xg7 tt:lf1 -g3+
I decided to try and construct a Otherwise White would play his
new version of the study with the promoted knight to fS .
same idea. Soon I came to the con­ 5. �b8xg3 �g2xg3
clusion that this cannot be realized 6. �h5-h6
in a pawn ending. A heavier con­
struction was needed.

After two promoted minor pieces

have been swapped, the white king
Timman 2011 must flee upwards.
White to play and win (after Zinar)
6. ... 'ifh1 xe4
White has enough material to win The best chance. After 6 . . . 'ir'xh4+
if Black decides to get two queens. 7.tbh5 + �f3 8.e5 ! the white pawn
The black rook will deprive the mass would march on undisturbed.
newly-promoted white pieces of 7. tt:ld7-f61
the crucial square eS. White sacrifices his bishop in or­
1. b6-b7 f2-f1ttJI der to keep the g-pawn. It is vital

1 08
 B i shop Prom o t i ons

that the solid construction on the A highly unusual knight ending.

kingside remains intact. White has no fewer than six extra
7. ... 'i!Ye4xa4 pawns, but his king is again
8. a6·a7! trapped inside the well-known
Not 8 .tt:lf5 + in view of 8 .. Jhf5 . pawn construction, in this case on
Therefore White also sacrifices his the queenside. Both sides will have
a-pawn in order to make his to push their passed pawns.
kingside front as strong as possible. 1. g5-g6 c3-c2
8. :a5xa7 2. g6-g7
9. tt:lg7·f5+ 'it>g3-f4 At first sight this looks like an easy
1 0. g6-g7 :!.a7·a8 win for White. However, Black has
11. tt:lf5·e 7 a resource which enables him to
play for stalemate.
2. ... tt:ld7-f6!
3. tt:le4xf6
And now Black can promote in two
ways. In both cases White reacts
with a minor promotion.
A) 3 cl 'if 4.g8.1:1.! Not 4.g8'iW?
•.•

in view of 4 . . . �a3 5 .tt:ld7 'ilfc2

6.tbc5 'iWxa4+ 7 .tt:lxa4, stalemate.

The promoted knight has the final

word. White wins because his
kingside pawns can march on in
closed formation.
The Israeli study composer Gady
Costeffhas also adapted this theme.

4.. /ot>a3 s.tt:ld7 'iYc3 6J:tg4 'ifcs

7.tbc5 'i:YaS+ 8.tt:la6 and wins.
B) 3 c l tt:l 4.g8..i! and wins.
.•.

Only not 4.g8'iW? on account of

4 . . . <;,i;ta l !, followed by S . . . tt:lb3 +
with stalemate.
Costeff has added a rook promo­
tion to the theme. For this study he
Costeff, 'Shahmat' 1986 received shared first/ second prize
White to play and win
in the Czerniak Memorial.

1 09
The Art o f the E n d g a m e

Let's go back to the attempts to in­ of the square h8. However, Black
corporate a record number of pro­ has still another resource.
motions into a study. 8. l:!:e8-f8
Almost half a century after 9. e6-e7 l:if8-f5
Lommer's study, Zinar came up 1 0. e7-e8.i
with the following piece of work. The fifth and last bishop promo­
tion. White wins.

These studies with whole series of

bishop promotions are not very at­
tractive. Lively play is only possible
in studies with two bishop promo­
tions.
It is a remarkable coincidence that
in the two studies by Lommer and
Zinar, the bishop promotions took
place on the same square. This was
Zinar 1983 also the case in a pawn ending
White to play and win
with two bishop promotions by
Five of the seven white pawns are Zinar.
going to promote.
1. f6-f7 l:!:g8-h8
The well-known recipe. If White
queens, Black will first play
2 .. Jla I ! and after 3 .jLxa I sacrifice
on hS.
2. f7xe8� l:!:a8xe8
3. d 7xe8� l:!:h8xe8
4. �b2-d4
The best square for the bishop. If
necessary, White is prepared to sac­
Zinar, 'Shakhmaty v SSSR' 1988
rifice the bishop on f2 .
White to play and win
4. ... J:re8-h8
Black keeps forcing White to pro­ The black king is hedged in, which
mote to bishops. allows for a stalemate motif.
5. e7-e8� l:!:h8xe8 1. f7-f8�1
6. e6-e7 J:re8-h8 Also here, promotion to a bishop is
7. e7-e8� J:rh8xe8 necessary to avoid stalemate.
8. e5-e6 1. ... g3-g2
Now the point of the promotions 2. �f8-c5 wa2-a1
becomes clear. The long diagonal is White has avoided the first stale­
opened, and the rook is deprived mate trick, but now the threat is re-

I IO
 Bishop Promotions

newed. Black wants to play 3 ...a2,

followed by 4... gl 'if.
3. �c5-e3
2...a2 is met by 4.�cl with mate.
3. ... g2-g1 't!Y
4. �e3xg1 a3-a2
5. �g1-d4+
White must sacrifice his bishop to
prevent the stalemate.
5. ... e5xd4
Timman2010
6. e4-e5 f6xe5
White to play and win
7. f5-f6 e5-e4
8. f6-f7 d4-d3+ First of all, White puts his rook in
9. e2xd3 e4xd3+ ambush .
10. �c2xd3 �a1-b2 1. l:rd8-d1 + t2:la3-b1
2. ttd1-h11 a4-a3
3. �a7-g1!
The point of the previous move.
Black is forced to move his knight,
and after that, White will put his
bishop on a better square.
3. ... t2:lb1·c3
4. �g1·d4+ t2:lc3·b1

Black has weaved a new stalemate

trap into the position. Now
II.f8'if a I'if 12.'ifg7 + is met by
the nasty 12 ...�xb3!.
11. f7-f8�1 a2-a1�
12. i.f8-g7+
And wins.

I thought it would be interesting to 5. 'it>h2-g11

compose a study where the bishop White repeats the motif with his
promotion is not intended to pre­ king.
vent an immediate stalemate; the 5. ... t2:lb1·c3
promoted bishop would have to 6. cJtg1·f2+ t2:lc3·b1
serve to block a line and then pre­ 7. Wf2-e1
pare mate. As an introductory The king approaches, in order to
study, I constructed the following. give mate!

111
The Art of the Endgame

7. ... t2lb1-c3 This way White is going to bring

8. .itd4xc3 'it>a1-b1 his king closer.
9. 'Ot>e1-d2# 4. ... 'Ot>g2-g1
The idea of the study is to put the 5. 'it>b6-b5+
rook in ambush and then use two The alternative was S.<.t>c6, with the
other pieces to temporarily block same consequences. Later on, the
the back rank, preventing stale­ king also has an alternative route.
mate. Eventually mate follows by 5. ... rJtg1-g2
the three pieces in cooperation. 6. 'Ot>b5-c5 'Ot>g2-g1
7. 'Ot>c5-c4+ \t>g1-g2
You can also do something similar 8. 'it>c4-d4 'it>g2-g1
on a diagonal. Then the bishop is 9. 'Ot>d4-d3+ 'it>g1-g2
the piece lying in ambush. 10. 'it>d3-e3 'Ot>g2-g1
11. �e3-e2+ �g1-g2
12. �a7-b8
Now that the white king has come
closer, the bishop temporarily
leaves the cr ucial a7 -g 1 diagonal in
order to put the rook into position.
12. 'it>g2-g1
13. .l:rf8-f1 + 'it>g1-g2
14. .l:rf1-f61 'it>g2-g1
15. .ib8-a7+ �g1-g2
16 . .l:rf6·b6
Lommer1975
White to play and win

First of all White puts his bishop in

ambush.
1. .ite5-b8 g7-g6
2. 'Ot>d8-c7 'it>g2-g1
3. �b8-a7+ 'it>g1-g2
4. 'it>c7-b6

This was the point of the 14th

move. Now the rook will edge its
way through and give mate.
16. ... 'Ot>g2-g1
17. .l:!.b6-b3+ 'it>g1-g2
18. .l:rb3-e3! 'it>g2-g1
19. .!:te3xg3#

112
 B i shop Prom o tions

Thi� was the last study that 6. �d4-e51

Lommer made. White sacrifices his bishop in or­
der to create a passed d-pawn.
Let's get back to bishop promo­ Again, this leads to the Phoenix
tions. In the following study White Principle, since White will pro­
again has to put his rook in am­ mote to a new bishop before long.
bush first. 6. ... d6xe5
7. d5-d6 e5-e4
8. d6-d7 e4-e3
9. d7-d8�1
Other promotions would be of no
use to White.
9. ... e3-e2
1 0. �d8-a5
Now it becomes clear why the king
had to go to a6. The passage for the
promoted bishop had to be kept
free.
Timman 2010 1 0. ... e2-e1 'if
White to play and win
11. �a5xe1
1. J:t.c1 -h 1 1 g4-g3 Again the back rank has been
2. h2xg3 a4-a3 closed by a bishop, this time with
3. �b6-a61 devastating effect.
The king makes way for the 11. tDb1 -d2
bishop, which is heading for g 1 12. �e1 xd2+ b2-b1 'if
again. Only seven moves later will 1 3. �d2-c3#
it become clear why the king could
not go to aS or c 7 . Later I found a study by Costeff
3. ... d7-d6 which contains a nice addition to
4. .ia7-g 1 1 tt:Jb1 -c3 this theme.
5. �g1 -d4+ tt:Jc3-b1

Cos teff 1979

White to play and w in

1 13
The Art of t h e E n d g a m e

White's first move is anything but

obvious.
1. l:!.b1 -a1 1
Only seven moves later will the
purpose of this little rook move be­
come clear. I.l:!.d 1 looked obvious,
but after l . . .dS 2 .e6 d4 3 .e7 d3
4.e8'if d2 5 Jhd2 ti:Jf3 6JH'2 g 1 'iY
7.l:hf3 Black has 7 . . .'if d I ! and ,

White cannot win.

d7-d5 Timman 2011
1. ..
.

White to play and win

2. e5-e6 d5-d4
3. e6-e7 d4-d3
4. e7-e8�1 If White queens, he cannot win.
The familiar bishop promotion. An illustrative line is 1 .d8'iif c3
4. ... d3-d2 2 .l:t d 1 a3 3 .l:td4? ltJd2 4.'iVc7
5. �e8-a4 d2-d1 'if b1 'if S.'ifxc3 + 'ifb2 6.'ifxd2 �b l
6. �a4xd1 tt:lg1 -h3 7 .'iVe1 + 'ifc 1 8 . .l:.d 1 a 1 'if+ and
Black wins.
1. d7-d8�1
On its way to g I .
1. ... c4-c3
2. �d8xb6 a4-a3
3. �b6-g1 ! tt:lb1 -d2
4. �g1 -d4+ tt:ld2-b1
5. �d4-e51
Again, the thematic bishop sacri­
fice.
5. ... d6xe5

7. �d1 -c2+1
On its way to b 1 . That is why the
white rook had to go to the corner
square.
7. ... tt:lh3-g1
8. �c2-b1 tt:lg1-h3
9. �b1 -e4+ tt:lh3-g1
1 0. l:!.a1 -a2
And wins.

The same can be done with two

bishop promotions. 6. b5-b6

114
 B i s h o p Promotions

White has to push this passed 1. �b3-d5+1 e6xd5

pawn. After 6.d6 e4 7 .d7 e3 stale­ 2. g6-g7 J:!.a3-b3+
mate would be unavoidable. 3. 'it>b6-a6
6. ... e5-e4 If the king moves to the c-file,
7. b6-b7 e4-e3 Black will save himself with a
8. b7-b8�1 check on c3 .
Like a Phoenix, the bishop has 3. ... d5-d41
risen from the grave. Re-introducing the stalemate motif
8. ... e3-e2 into the position.
9. �b8-g3 4. e5-e6
This is what it was all about. White An important alternative was 4.g8..t.
is in time to temporarily close off As White threatens mate on dS , he
the back rank again. now wins the rook on b3 . But Black
9. ... e2-e1 � can just save himself with 4 .....\b.es
10. �g3xe1 Ql b 1 -d2 S .�xb3 .ikf6 6.�d5+ �b8 7.�xf3
11 . �e1xd2+ b2-b1� �h4 8.<it>b6 �d8+! 9.Wc5 has.
1 2. �d2xc3# 4. ... .!:!.b3-b2
5. e6-e7
Is it possible to have two bishop
promotions alongside each other? I
have spent a lot of time trying to
find an answer to this question. The
following study is a first attempt, al­
though the bishop promotions oc­
cur in the side variations.

Both pawns are about to promote

and Black now has to choose how
to organize his defence from here.
5. ... l:!.b2-b7
The best defence. On S . . . l:rb l , both
6.e8� and 6.g8� would be win­
ning. Now both bishop promo­
Timman 2009 tions fail to 6 . . . l:ra7 + 7 .�b6
White to play and win
l:rb7 + , and White cannot win.
White cannot play l .g7 on account 6. g7-g8Qll
of l . . . .ti.xb3 + 2 .�a6 .l:ib7 3 .g8'iW The only correct promotion. After
.ti.b6+! with stalemate. This theme 6.e8lLJ Uxg7 7 .lLlxg7 .tg3 Black
will crop up again and again. would not need to worry.

115
T h e Art of t h e E n d g a m e

6. . .. .!:.b7-b1 The white passed pawns are dan­

Black keeps playing for stalemate. gerous, but after l .�d6 �a2 Black
White cannot promote his e-pawn. can just save himself.
7. ti:lg8-f6 l:. b1 -e1 1. �f8-g7+!
8. e7-e8� With this bishop sacrifice, White
But now he can. renders the 6th rank inaccessible to
8. ... .!:. e1 xe8 the black rook .
9. ti:l f6xe8 1. ... .!:. f7xg7
2. b5-b6
An unbelievable situation: Black
has a rook and two minor pieces
for a handful of pawns, and yet he
is lost. He still has one chance left:
2. ... �b1 -a2
3. b6-b7 i .a2-g8
Now 4.b8'tW runs into 4 . . . tt:Jxe3!
S .dxe3 .l:rg6 + ! and stalemate is un­
avoidable. The logical 4.a7 is also
insufficient for the win after
The complications are over and a 4 . . . tt:Jxe3 ! S . dxe3 .l:rxg2 and Wliite
knight vs bishop ending has arisen. doesn't have an effective minor
The finish still requires precision. promotion.
9. .
.. �b8-g3
1 0. �a6-b6 �g3xf2
11. ti:l e8-c7+ 'lt>a8-b8
1 2. a5-a6
And wins.

The same stalemate motif also oc­

curs in the following study.

4. g2-g31!
A very subtle little move. The point
is that after 4 . . . ti:Jxe3 S .dxe3 .l:!.xg3
White has at his disposal 6.b81t! ,
winning. This means that the
rook's range of action is restricted.
4. ..
. ti:lc2xe3
Timman 2011 Black cannot postpone this knight
White to play and win sacrifice any longer.

1 16
 Bishop Promotions

5. d2xe3 J:lg7-g4
6. a6-a7 .!:f.g4-g7
Now 7 .a8�? is insufficient on ac­
count of 7 .. Jhb7! 8.�xb7 �d5
9 .�c8 i.e6 1 0.i.a6 �c4 and
again, White cannot avoid stale­
mate.
7. g3-g4!
A pawn sacrifice, enforcing a situa­
tion of mutual zugzwang. Black has
to help himself to the pawn.
7. ... J::t g7xg4 Screening off the b-file.
11. ... .!:f.b1 -g1
1 2. tt:Jb4-c6
And wins. Black has to give his
rook for the b-pawn, after which
the knight vs bishop ending is an
elementary win.

It's possible to make an even more

beautiful version of this theme.

8. a7-a8tt:J!
Just as in the previous study, a
knight promotion provides the so­
lution to the problem.
8. ... l:lg4-g7
On 8 . . . .ll g l , 9 .tLlb6! would have
been winning.
9. tt:Ja8-c7!
An indirect defence of the b-pawn. Tm i man 2011
White to play and win
Again, a situation of mutual
zugzwang has arisen. Black has an enormous material
9. ... l:lg7-g1 advantage, but he cannot stop the
1 o. tt:Jc7-d5! white queenside pawns.
Centralization. Now on 1 O . . l::!.g 7 ,
. 1. b5-b6 Wg8-h8
both l l .tLlf4 and l l .tLle7 are win­ Making room for the bishop.
ning. 2. b6-b7 ke6-g8
1 0. l:lg1 -b1 The stalemate construction is
11. tt:Jd5-b4 ready.

117
T h e Art o f the E n d g a m e

3. a6-a7
White could not queen immedi­
ately, since after 3 .b8� l:rg6+!
stalemate would be unavoidable.
3. l:tg2-g1
4. h2-h3!

6... �d5 After 6 . . . �f7 7 .�e5 +

Wg8 8 . h4 l::lh 3 9 .'tt> g 5 White
would have a clearly winning posi­
tion. 7.i..e5 + and wins.
I consider this study to be the best I
have composed so far.

Studies with bishop promotions

Mutual zugzwang. The rook is on both sides are rare. Inevitably,
forced to a less favourable square the bishop promotions will have
on the g-file, and then minor pro­ different motifs: the attacking side
motions will become possible. wants to avoid stalemate, whereas
There are now two main lines: the defending side strives to
A) 4...l:.g3 5.a8tLl! Not 5 .a8� in achieve just that.
view of 5 .. J:tg7! with mutual
zugzwang. 6 .h4 is met by
6 . . . 1:t.h7 + 7.Wg5 l::r g 7 + and White
does not have the square h4 for the
king. And if White continues
6 .b8�, Black has the little saving
move 6 . . . �d5 ! . White does not
have a check on eS;
B) 4...l:t:g 7 5.a8�! Again a situa­
tion of mutual zugzwang has
arisen. The black rook has to go
down again, as after 5 .. Jhb7 Timman 2011
6 . ..txb7 �dS White would pre­ Wh ti e to play and win
vent the stalemate threat by
7 .�a6 . 5 .. J�g3 6.b8�! The sec­ Again, the kings are facing each
ond bishop promotion. The other in the upper right-hand cor­
newly-created bishop pair reigns ner. Both sides have dangerous
supreme. passed pawns. In order to realize

1 18
 B i shop P r o m o t ions

promotion, pieces will have to be Here too, the correct move. This
sacrificed. blockade motif - in a slightly differ­
1. t2Jd6·f7+1 ent form - was invented by Starnma.
With this knight sacrifice White 5. ... �f7-g8
closes off the f-file for the enemy Preparing a stalemate construction.
rook. At the same time, he wants to 6. b6·b7 a2·a1 i.!
promote his b-pawn with check.
1. .. . ii.g8xf7
2. b5-b6
It looks as if Black has run out of
options now. What can he do
against the threat of 3 .b7 ? 2 . . . b2
runs into the devastating 3 .�b I ! .
2. ... �a2·e2!
A stunning rook sacrifice. Black
threatens to play his rook to e8 ,
and therefore White has to accept
the sacrifice.
3. �d3xe2 The point of the previous move.
A somewhat uncommon bishop White cannot promote to a queen or
ending has emerged on the board. rook on account of stalemate. More­
Soon it will be a four-bishop ending. over, he does not have a waiting
3. b3·b2 move with his bishop, since then the
4. i.e2·d3 a3·a2 compromised position of his king
would seal his fate: Black promotes
his pawn and gives mate on g 7 .
7. b7-b8.tl
A real treat for lovers of bishop
promotions. The stalemate has
been removed.
7. .tg8-c4
8. i.b8xa7 i.c4·d3
A final attempt. After 8 . . . �xa6
9 . .1d4+ �g 8 I 0 .�g 7 Black would
have no chance to survive.
9. �a7-d4+
5. i.d3·b1 11 And wins.

1 19
Chapter 7

Mating Patterns
In the first chapter I gave two min­
iatures in which the black king was
mated in the middle of the board.
This is the most attractive mating
pattern.

Wotawa has made various studies

where the mating pattern seems to
appear out of the blue. A good ex­
ample is the following study.
3. l:.b7-f71!
Magnificent. White deprives the
black king of escape squares.
3. ... a2-a 1 'if
3 . . �xf7 4.�d5 + and White wins.
.

4. �b7-d5+ �e6-e5
5. f2-f4#
It's impressive if in the main varia­
tion the composer manages to draw
the king from the corner to the cen­
tre, where it is then mated.
Wotawa, 'Deutsche Schachzeitung ' 1961
Ghenrikh Kasparya'n has tried to re­
White to p lay and win
alize this theme in a study, but un­
The white pieces seem to be badly fortunately this study is incorrect.
positioned for stopping Black's far
advanced passed pawns. White
even runs the risk of coming off
worse.
1. J::h 7-b7 a4-a3
2. '>t>c6-c51
The first step in the direction of
the mate, even though such a mate
is nowhere to be seen at this
point.
2. ... a3-a2 Kasparyan 1962
Now what? Wh tie to play and win (inco rrect )

120
 M a t i ng P a t terns

First of all White has to eliminate �xd2 1 O .ltJxd2 �b6 1 l . We5

the g-pawn. WaS , followed by 1 2 ... 'it>b4, Black
1. lt:Jh7-f6+ 'it>g8-f7 effectively blocks the white passed
2. tt:Jf6xg4 tt:Jg7-e6+ pawns.
3. 'it>g5-f5 6. c3-c4 �e1 -c3
He must centralize his king as 7. tt:Jc6-a71
much as possible. Surprising, but strong. The knight
3. ... �f2-e1 aims for the critical square bS.
Black besieges the white passed 7. ... tt:Je6-d4+
pawns. White cannot push the 8. 'it>f5-e4 tt:Jd4xb3
c-pawn, as after 4.c4 ltJd4+ the 9. tt:Ja7-b5+ 'it>d6-c5
b-pawn will be lost. It looks as if Black is going to elim­
4. tt:Jg4-e5+ Wf7-e7 inate the last pawn, but:
5. tt:Je5-c6+ 1 0. �h6-f8+1 'it>c5xc4
The aim of the knight checks is 11. tt:Jb5-a3#
clear. White wants to control the
square d4. Can this study be saved? I fear that
5. ... 'it>e7-d6 the long march by the black king
After this king move, White even­ cannot be maintained. But still it is
tually manages to catch the enemy possible to make a good study out
king in a mating net. However, of it.
with 5 . . . Wd7 ! Black could have
held the draw. After 6.c4 tLlcS
7. lDeS + Black again has to find the
right square for the king.

Timman 2011
(correction of a study by Kasparyan )
White to play and win

First White must centralize his king.

1. Wg6-f5 �h4-e1
In Endgame Challenge, Nunn gives Now everything is clear.
7 . . . �c8 , after which White reaches 2. c3-c4 �e1 -c3
a technically winning endgame 3. tt:Jc6-a71 tt:Je6-d4+
with 8.tLlf3 �c3 9.kd2 . Better is 4. 'it>f5-e4 tt:Jd4xb3
7 . . �c7 , as after 8 .lDf3 £i.c3 9.£i.d2
. 5. tt:Ja7-b5+ 'it>d6-c5

12 1
The A r t o f the E n d g a m e

6. �h6-f8+! �c5xc4 How to continue?

7. ttlb5-a3# 4. �d6-b41
Preparing a mate on the long diag­
onal. For this purpose 4.�a3 ? was
not suitable, because then Black
would close the long diagonal with
4 . . . d4! , after which White cannot
avoid losing the bishop.

All the pieces are playing - the

study is still lively.

In the following study the mate is

given by the bishop.

4. ... b6-b5!
A subtle defence. 4 ... d4 would
have been met by the devastating
5 . l:h b 7 . But now Black does
threaten to close the long diagonal.
5. �b4-c51
Threatening to unpin the rook by
6 .l:rc7.
5. ... b7-b6
6. �c5-a3
Timman 1984 The last finesse. White first pro­
White to play and win
vokes . . . b7-b6 , and then eventually
The first move is obvious. retreats his bishop, to a3 , since the
1. 'itg3-g4 passage to a7 has now been cleared
Protecting his own knight and at­ for the white rook.
tacking the enemy knight. 6. ... b5-b4
1. .. . �e8-f71 7. �a3xb4 �f8xe7
The best defence. Black gives up 8. �b4-c3#
the knight, hoping to set up a pin It is very rare for a rook to remain
which is unpleasant for White. en prise for so many moves.
2. �h2xd6 �h6-f8
The point of the previous move. Later, I encountered the same mat­
3. ttlh5-f4+ �g6-f6 ing pattern in a study by Matous.

122
 M a t i n g P a t t erns

With the sacrifice of a knight and

a pawn, the same mating position
appears on another side of the
board. It does not make any dif­
ference that one of the pieces
which blocks the king's escape
route is a knight instead of a
bishop.

The following study is a very subtle

Matous, Chavchavadze MT 1987 one.
White to play and win

White can only hope to win if he

can keep his pawn - either that, or
he must succeed in giving mate. At
this point, mate is nowhere in sight.
1. tt:lc4-e5+ 'it>d3-e4
2. g5-g6 �c1 -h6
3. tt:le5-f7 ..ih6-g7
Now what? The g-pawn looks
doomed.
4. ..ig4-d1 1
Sochniev , 'Vecherny Leningrad' 1987
For the moment the pawn is indi­
White to play and win
rectly protected.
4. ... 'it>e4-f5 The main question is whether
5. ..td1 -c2+ 'it>f5-f6 Black will manage to exchange
6. tt:la8-b61 pawns.
The other knight comes to the res- 1. 'it>g3-f4
cue. A natural and strong move.
6. tbh4xg6 1. ... e7-e6
7. tbb6-d7+ c;fo>f6xf7 Black cannot wait too long with
8. ..tc2-b3# this, otherwise White will consoli­
date his position.
2. d5-d6 e6-e5+
3. 'it>f4-g41
Again White chooses the best
square for the king.
3. �f6-e6
4. tbh4-f5
For the time being, White has
managed to maintain the pawn,
but Black keeps harassing it.

123
The Art of t h e E n d g a m e

tLlc5-e4 Timman 1984

4. ...
White to play and win
5. �f1-c4+
The bishop comes into action. White can capture the rook, with
5. �e6-d7 check even, but that wouldn't
6. �c4-b5+ �d7-e6 bring him anything : L�xe8 +
7. d6-d71 <fio>gS 2..�.d2 + �f6 and Black is
Driving the black knight to a ready to give his knight for the
square where it blocks the black white pawns, keeping his bishop,
king. which is a reliable weapon in the
7. ... tLle4-f6+ battle against the bishop pair.
8. '>tg4-g5 tLlf6xd7 1. �e1 -d2
9. �b5-c4 Threatening mate from both sides.
Mate. 1. ... �c8xf5+
The only defence.
2. e4xf5 tLle7-c6
The sensational knight move
2 .. .tbg6 would save the draw after
3.fxg6 .ld.g8 4.g7 <fio>g6 , but it loses
after 3 .�c6, which renews the
mate threat. The stalemate attempt
3 ... tt:lf4+ 4.�xf4 .ld.e3 + turns out
to be a useless demonstration after
S .�g2.
3. �b5-a41 -
For this study, Sochniev received A very subtle move. White now
first prize in the 'Vecherny Lenin­ threatens mate on d I . Insufficient
grad' tourney. was 3 .�xc6 in view of 3 . . . .l:!.e3 +
4.<fio>g2 l:re2 + with a draw.
If the harassed king is on the edge 3. ... .!:te8-e21
of the board, in principle no Again, the only defence. Black vol­
blocking pieces a.re needed for the untarily places his rook in an un­
mate. pleasant pin. Another method to

1 24
 M a t i n g P a t terns

prevent immediate mate is not 9.�b2. Now Black should definitely

available. not take on fS , but both 9 .. J:tb4 and
4. .ia4-d1 tt.:lc6-d4 9 ... .l:Ie3 + lead to a draw.
5. ... tt.:ld4-f3
6. f5-f6!

Black has built up a shaky defence.

Nevertheless, it is not so easy to This push is only effective when
breach. Black cannot be brought the queen's bishop is on e3. There
into zugzwang as he can always it blocks the black rook's path to
play his knight to f3 . stopping the f-pawn, and also, after
5. jLd2-e31 6 . . . <J.ig6 7 . .ixe2 Black has no
The mate motif is disturbed by a check on g l .
stalemate motif. In combination 6. ... .l:!.e2-h2+
with the next move, this is the only The only reasonable try. After both
winning method. After S .�c l tt:Jf3 6 . . . .r:.xe3 7 .f7 and 6 . . . tt:Jg5+ 7.�g3
6.'J.ig3 tt:Jd4, White cannot make tt:Je4+ 8 .�f4 Black runs out of
progress, for example: moves.
A) 7.'�f4 �h6 8 .'it>g4+ �g7 7. \fr>h3-g3 Wh5-g6
9 . .ixe2 tt:Jxe2 1 0 .ilb 2 + <it>f7 8. �d1 xf3
1 1 . �f3 tt:Jg 1 + and White cannot The rook has no good squares. On
catch the knight without losing his 8 . . . .l:i.h8 , both 9 .�e4+ and 9 . ..tcs
f-pawn. Incidentally, 7 ... �h4 is are winning.
also insufficient for the draw in 8. ... .!:[h2-h7
view of 8 . ..txe2 tt:Jxe2 + 9 .<it>eS 9. �f3-e4+
tt:Jxc 1 1 0 .f6 tt:Jd3 + and the knight And wins. Black can still try the
stops the f-pawn via cS or f4; cheap trap 9 . . . �h5 , but then
B) 7.�b2 �gS (not 7 ... tt:JxfS+ White has a large range of winning
8.Wh3 and Black loses a full rook, continuations.
ending up in a 2 bishops vs knight
ending) 8 .Ld4 l::le4 (the only The Russian endgame study com­
move to prevent White reaching a poser A.S.Gurvich created a fa­
theoretically winning position) mous miniature in which a mating

125
The Art of t h e E n d g a m e

net is woven around the black king Interestingly, this mating net can
in a surprising way. also be constructed if Black has
much more material.

Gurvich, 'Shakhmatnaya Moskva' 1959

White to play and win Timman 2011
White to play and win (a fter Gurvich )
It seems like an impossible task.
White's a-pawn is very hard to pro­ Material is equal. White has the
tect, whereas Black has a strong bishop pair and a strong passed
passed pawn. However, the positio­ pawn, but the black pawns aren't
n contains a subtle combination. exactly harmless either. White has
1. �h7-g81 l:!.a2-b2 to act energetically, otherwise he
The a-pawn was taboo on account may end up in trouble.
of mate in one. 1. �h6-f811
2. �g8-b311 A bishop sacrifice Black has to ac­
Blocking the black b-pawn. cept.
2. ... .l::!: b 2xb3 1. ... �e7xf8
3. l:th8-a81 On 1 . .J:rc2, 2.�b7 ! is winning.
2. e6-e7! �f8xe7

The point of the bishop sacrifice. The same theme that was discussed
3. �a5xa4 earlier in this chapter. The starting
4. '1t>a7-b6# position is again reached, but this

1 26
 Mating Patterns

time without the queen's bishop 4 . .ili.b3!! Black would get mated in
and the e-pawn. So the a2-g8 diag­ the way that we have already seen.
onal has been opened for the other However, he has a better defence.
bishop. By the way, after 2.l:lh7 jg_f8 3.e7
3. i.g8-b3!1 .ili.xe7 4..l:he7 .:tc3 Black would
Gurvich's bishop sacrifice turns have sufficient counterplay.
out to be equally effective here. 2. ... J:!.a3-c31
3. ... �b2xb3 Now the battle will revolve around
4. .l:Ih8-a81 wa5xa4 the square e7.
5. �a7-b6# 3. .l:Ih8xh5+
The study has gained in content Now the intention of the pawn sac­
with the extra sacrifices. But still rifice becomes clear.
Gurvich's study is better, because it 3. ... f6-f5!
shows the brilliant idea in all its Black also sacrifices his f-pawn. Af­
simplicity. In my study, the fact that ter 3.. .<>t>xa4 4.jg_xe7 J::[c7+ S.'�b6
the black knight doesn't move is .l:txe7 6.l:!.h7 he would have been
especially annoy ing. Therefore I left without a chance.
decided to make a study in which 4. .!:.h5xf5+ �a5xa4
Gurvich's mating position features 5. i.d6xe7 .l:!.c3-c7+
in a side variation. 6. �a7-b6 J:!.c7xe7
7. .l:If5-a5+
Necessary in order to break the
blockade on e7. Insufficient for the
win was 7.�cS. as becomes clear
from the following variation:
7...d3 8 ..:tds .:te8 9 .l:!.fs l:!.c8 +
10.Wb6 b3 ll.e7 b2 12Jh5+
<.t>b4 13..:tbS+ <.t>c3 14 . .:tb3+
Wd2 1sJhb2 <.t>cl, followed by
16...d2, and White cannot win.
7. ... 'lt>a4-b3
Timman 2011
8. l:ta5-a7
White wins

Here Black's passed pawns are even

more dangerous.
1. f5-f61
The intention of this pawn sacrifice
will soon become clear.
1. .
..
g7xf6
2. �g3-d611
The well-known bishop sacrifice.
After 2 ... �xd6 3.e7! �xe7

127
The Art of the Endgame

It looks as if White wins comfort­ With this rook sacrifice White

ably, but Black has a fantastic re­ forces the passage of his e-pawn
source: with check.
8. ... tt:Jc2-e31 1 3. ... '>t;b2xa2
Indirectly protecting the rook. 1 4. e6-e7+
White has to take the knight. And wins.
9. f2xe3 d4xe3
An interesting situation. If White The Russian endgame study com­
takes the rook, the black pawn will poser Viktor Evreinov is from the
queen undisturbed, after which he same generation as Vladimir Bran
draws easily. and Vladimir Korolkov, but he has
1 0. l:ra7-c71! left behind a much smaller body
A very strong move. The black king of work. Whereas Bran and
is cut off and can therefore not Korolkov each have more than
support the e-pawn. 400 studies to their names,
1 0. ... 'it>b3-a3 Evreinov hasn't even made 1 00 .
Black wants to march on with the This is mainly due to the fact that
b-pawn. Now 1 1 Jk 1 would be in­ he only started to compose seri­
sufficient for the win due to ously when he was 5 0 . To my
1 l . . .b3 1 2 . .tre l b2 1 3 ..l he3+ '>ita2 . mind, he composed one of the
After White has been forced to give most beautiful mate studies imag­
his rook for the b-pawn, the black inable.
king marches to the kingside, just
in time to conquer the white
h-pawn.
However, White has a hidden trick
up his sleeve.
11. %:!.c7xe71 e3-e2
1 2. ne7-a7+ 'itta 3-b2

Evreinov, 'Shakhmaty Riga' 1959

White to play and win

The white bishop is lying in am­

bush. It will play the starring role
in the mating combination.
1. l:l:d4-e41!
An absolutely stunning first move.
White opens up one diagonal for
13. %:!.a7-a2+1 his bishop and closes another for

128
 M a ting P a t t erns

till' black queen. The sacrifice of

his rook - with check! - is neces­
sary in order to get everything
right.
1. d5xe4+
2. �f3-g2!
Walking away from all the queen
checks.
2. e4-e3+
3. Wg2-g1
Evreinov, 'Ceskoslovensky Sach' 1960
White to play and win

The black passed pawns cannot be

stopped. Therefore, White has to
search for mates.
1. �d7-b5 c3-c2
2. J:!.e1 -e71
Forcing the h-pawn to move for­
ward, after which the black king
loses one escape square.
2. ... h7-h6
3. J:!.e7-e31
Again, the only square for the king. Suddenly there is a concrete mate
The position is intriguing. Black is threat: 4 . .l:.h3+ gxh3 5 .g 3 + , fol­
a full queen up, but he cannot avert lowed by mate on e 2 .
the catastrophe along the a l -h8 di­ 3. ... c2-c1 ltJ!
agonal. The only defence.
3. ... �g8-f7 4. �b5-f1 1
A desperate attempt to keep the
long diagonal closed.
4. e6xf7 'iVc6xc5
5. f7-f8'i¥+ 'ii'c5xf8
6. f6-f7+ 'iVf8-g7+
7. Wg1 -h2!
For the third time, the only square
for the king.
7. 'it'g7xa1
8. f7-f8�
Mate.
Depriving the black king of the
Evreinov also worked on another h3-square. Now Black has to pro-
ingenious mating position. mote his other pawn, too.

1 29
The A r t o f the Endgame

4. b2-b1 if
5. g2-g3+ 'oth4-h5
6. l:te3-e5+ g6-g5
7. �f1-c41
The mating net is shifted. The
deadly check on f7 cannot be pre­
vented. Black now starts to play for
stalemate.
7. 'iVb1 -g6+
8. '>t>f6-e7 tt:Jc1 -d3
Timman 2011
9 . .l:!:e5-e3!
White to play and win (after Evreinov)
The only square for the rook.
9. ... tt:Jd3-e5 Black still has all his pawns. The
1 0. l:te3xe5 a7-a6 passed pawns threaten to become
very dangerous, so White has to
search for a mate.
1. llb1 -e1
White brings his rook into play.
The alternative 1 .�d1 was no good
in view of 1 . . . g 3 , giving the black
king some air.
1. ... �c8-a61
With this surprising bishop sacri­
fice, Black manages to create coun­
terplay. After 1 ... d5 White would
11. h2-h3! weave a mating net around the en­
With this final little move White emy king, as follows: 2 .g3 + Wh5
manages to avoid stalemate and 3 .�d3 (threatening 4.�xg6 +
wins. hxg6 5 .l:re8 with mate) , 3 . . . 'it>h6
4.h3 or 4 . .l:.h 1 , followed by 5 .h3 .
A pretty study, but you can sense that Black is helpless.
more can be got from this theme. 2. i.f1 xa6
Apparently Evreinov felt the same. White has to accept the bishop sac­
He made three other studies with rifice. After 2 .Ite7 h6 3.g3+ Wh5
this theme; one in the same year, the 4 . .:es + g5 5 .�xa6 b l 'iY he would
other in 1 967 and the third in 1 980, have no more than a draw. It is im­
not long before he died. Although portant to keep the check by the
Evreinov won prizes with these last g-pawn in reserve.
two studies, none of them turned 2. ... c3-c2
out to be correct - a good challenge 3. lle1 -e7!
for me. I set to work and improved Now this Evreinovian thematic
Evreinov's idea as follows. move is good.

130
 M a ting P a t t erns

3. h7-h6 in another way. 6.<>t>xg6 ! Here it be­

4. J::i. e7-e3 comes clear why White had to wait
with the g-pawn check. 6 ...b 1 'iV+
7.�xh6 White has allowed the
queen promotion with check. He
now threatens mate in one. 7 ...g3
The only way to prevent it.

There are two main lines here:

A) 4 ... c4 Black sacrifices a pawn
now in order to play for stalemate
later on. 5 ..ixc4 c l tt:l ! Here, too,
the only way to prevent the mating
combination. 6 ..ifl ! b 1 if 7.g3 + 8 .id3 ! ! A surprising bishop sacri­
•

'li;lhS 8 . .l:!:e5 + g5 9 . .ic4 'ii'g6 + fice. Again there is a threat of mate

1 0. �e7 tt:ld3 Black is also going to in one, so Black's reply is forced.
sacrifice his knight. 8 ...'ihd3 9.hxg3 + ! An important
zwischenzug. After 9 Jhd3 tt:lxd3
I O.hxg3+ �xg3 l l .a6 c4 1 2 .a7 c3
1 3 .a8'if c2 1 4.'iff3 + <>t>h2 White
would not be able to win. 9 ... Wg4
1 OJhd3 tt:lxd3 1 1 .a6 and wins.

Variation B looked like an interest­

ing starting-point for a new study.

1 1 ..!::!: e 3! The only square for the

rook. On l l . .l:!.e2 tt:lf4! saves Black.
1 1 ... d5 1 2 ..ixd5 tt:le5 1 3.llxe5 a3
1 4.h3 ! Evreinov's final move.
White avoids stalemate and wins;
B) 4... cl ttJ After this immediate
knight promotion the play takes on
a different character. s .�fl ! d6!
Preventing the check on eS . Now Timman 2011
White has to corner the black king White to play and win

131
T h e A r t of the E n d g a m e

White is a large amount of material fice becomes clear: he wants to de­

behind, but he can create a power­ flect the white d-pawn to c3.
ful mating threat. 5. d2xc3
1. .l:!.f8-f3 Alternatives were no good:
Mate on h3 can only be prevented A) 5 Jhc3 fl 'iV+ 6 . .txfl c l tb
with heavy material sacrifices. and White cannot win. His rook is
1. ... �a2-b3! less favourably placed, so he cannot
This queen sacrifice draws the work with the little d2-d3 move;
white rook beyond the critical B) S .�xg6 .teS ! and Black is
e3 -square. winning, as the mate threat is
2. .l:!.f3xb3 tt:lg4-e3 gone.
Black sacrifices his knight in order 5. ... f2-f1 �+
to be able to push his foremost 6. ike2xf1 c2-c1 tt:ll
g-pawn. Here too, Evreinov's knight pro­
3. l:tb3xe3 g5-g4 motion is the only way to prevent
The first mate threats are warded immediate mate.
off But White has more trumps. 7. 'it>f6xg6
4. 'it>e7-f6! Not 7.g3+ �hS 8 . .tg2 in view of
Depriving the black king of the 8 . . . b 1 if 9 . .te4 a2 ! and Black
gS-square. As in the previous stud­ comes out on top.
ies, White is threatening 5 . .l:rh3 + 7. b2-b1 �+
gxh3 6.g3 mate. Again Black has to 8. 'it>g6xh6 g4-g3
make huge sacrifices to avert the 9. �f1 -d3!
new mate threat. All still as in the previous study.
After 9 . . . 'iVxd3 1 0 .hxg3 + �g4
l l . .l:rxd3 tbxd3 1 2 .c6 White pro­
motes with check and then reaches
a winning queen ending. However,
Black has a more tenacious de-
fence.
9. ... ii'b1 -b7!

4. ... itb4-c3+!
Only with this bishop sacrifice can
Black stay afloat. After 4 . . .fl 'i¥+
S .�xfl c l tD 6.�xg6 b l 'i¥+ 7 .d3 !
he would soon be mated. Here the
intention of the black bishop sacri-

1 32
 M a t i n g P a tt erns

Black. continues the struggle for the In the following study, White sacri­
square e4. I O.l:le4+ now runs into fices a knight and three pawns in
I O . . . 'ifxe4 l l .hxg 3 + <;tJxg3 order to open the diagonal for his
1 2 .he4 dxcS 1 3 .�bl t2Jb3 and bishop.
the draw is inevitable.
1 0. .id3-f5!!
White keeps playing for mate.
With the text White deprives the
black king of the g4-square.
1 0. ... tt:Jc1 -e21
Both promoted pieces rush to the
aid of the black king. Now White
cannot increase the pressure. How­
ever, he still has another trump.
11. c5xd6
Timman 2009
This passed pawn is going to de­
White to play and win
cide the issue.
11. ... a3-a2 1. h6-h7 �c1 -b2
1 2. d6-d7 'it'b7-b6+ On l . tt:J£7 , 2 .eS is winning.
. .

1 3. l:te3-e6 g3xh2 2. c2-c31 i.b2xc3

3. e4-e5!
The second pawn is sacrificed.
3. ... ..tc3xe5
4. h7-h8"i¥1
And the third.
4. ... .ie5xh8

The situation remains tense.

14. l:te6xb6 h2-h 1 1¥
1 5. d7-d8�+ �h4-g3+
16. 'it>h6-g71
Not 1 6.�g6 in view of 1 6 ... l2Jf4+
1 7 . <;tJg7 'ifhS ! and White cannot
win. After the text move Black can 5. tt:Jf5-g7!
promote a fifth pawn, but after This was White's intention. The
1 6 . . . a l 'if 1 7 .'ifd3+ <;tJf2 1 8 .'iff3+ square fS is vacated for the bishop.
'>tie l 1 9. l:lb l + White emerges With the immediate S .�d3 g3
with an easily winning endgame. White would achieve nothing.

1 33
T h e Art of t h e E n d g a m e

5. �h8xg7 The first mating attack has been re­

6. �b1 -d3 ..tg7-d4+ pelled, but the white king is still in
7. �g 1 -h 1 g4-g3 an awkward predicament. The bat­
8. �d3-f5 tle will now be between the white
Mate. king's bishop and the knight.
4. ... ttlg2-h4
A nice piece of work, but it's an­ 5. jf_f3-e4
noying that the black knight does The black king too turns out to be
not move. The following study is unsafe. However, he has a nice fi­
more interesting. nesse in reserve.
5. ... f7-f51
6. �e4-b1 1
Only move. Not until 1 1 moves(!)
later will we see why the alterna­
tive 6.�c2 is not sufficient.
6. a6-a5
7. �b1 -c2

Timman 2011
White to play and draw

Black is threatening mate, so White

is forced to sacrifice his rook.
1. l:rb1 -b8+1 j!_d6xb8
2. d5-d6
The pawns are slowly set in mo­
tion, in order to open up the long A situation of mutual zugzwang
diagonal for the queen's bishop. has arisen. This will also be ex­
2. ... �b8xd6 plained at the end of the study.
The starting position, only without 7. ... a7-a6
the white rook and the pawn on This is the little move that White
dS , is on the board again. This is a wanted to provoke.
separate theme: in the main varia­ 8. �c2-d1 1
tion of a study the same position The obvious move was 8.�a4, in
appears on the board with the order to besiege the black king
same player to move, but without a again via a detour. However, Black
certain piece. I will return to this gets the upper hand after 8 ... tbf3
theme in Chapter 1 2 . 9 .�d7 tLld4 1 0.�e8 f4 1 l..�J7 +
3. d4-d5+ wh8-g8 'lt>xf7 1 2 .�xd4 ( I 2.';ilxh7 is met
4. ..ta1 -g7 by the winning 1 2 . . . tLlfS )

1 34
 M a t i n g P a t t erns

1 2 ... �g8 1 3 .g6 �f8 + . The black 1 4. ... �b6-c7

passed pawns decide. With the text But White still cannot play his
move White takes the crucial g-pawn. He will also have to give
f3 -square from the black knight. his d-pawn in order to free his
8. ... c4-c3! king.
A pawn sacrifice to make more 1 5. d7-d8�+ �c7xd8
squares available for the knight. 1 6. g5-g6 h7xg6
1 7. �h6xg6

9. d2xc3
The only move. After 9 .�a4 t2Jf3 Now it becomes clear why White
1 O.�d7 , 1 O ... tiJeS is winning. An had to play 6.�b 1 ! . If the black
illustrative variation: 1 1 .�e6+ pawn were on a7 here, Black could
tbf7+ 1 2.�xf7+ �xf7 1 3 .g6+ have played his bishop to b6, win­
hxg6 1 4.hxg6+ '>1;;>g 8 1 S .�xc3 f4, ning. Without this possibility, Black
and again the black passed pawns has only one method to maintain
decide. the balance, despite his great mate­
9. ... lLlh4-g2 rial surplus.
1 0. c3-c4 lLlg2-e3 As for the alternative 1 7 .�c 2 : in
11. c4-c51 study composers' circles such a
The play is becoming very sharp. move, which looks perfectly natu­
Black has to take the pawn. ral but is still insufficient on ac­
11. ... �d6xc5 count of a later subtlety, is called a
1 2. d5-d61 'logical try'.
The consequence of the previous 1 7. ... ..id8-g5
move. White sacrifices a piece in Otherwise the white h-pawn can­
order to eventually free his king. not be stopped.
1 2. ... lLle3xd 1 1 8. �g6xg5 'oti>g8-h7
1 3. d6-d7 �c5-b6 Draw.
14. �g7-d4
Finally the queen's bishop can I have made a version of this study,
move again. which features the Excelsior theme.

1 35
T h e A r t of t h e E n d g a m e

1 5. d6-d7 .2.c5-b6
1 6. �g7-d4 �b6-c7
1 7. d7-d8'iY+
Excelsior!
1 7. ..
. �c7xd8
1 8. g5-g6 h7xg6
1 9. 'it>h6xg6 �d8-g5
20. 'it>g6xg5 'it>g8-h7
Draw.

Timman 2011 In the tourney of Het Tijdschrift van

White to play and draw
de KNSB in I 948 , the Dutch end­
The white king again finds itself in game study composer ).Fernhout
a mating net. White will have to try received first prize for the follow­
to open the long diagonal for the ing study.
queen's bishop.
1. c2-c3!
On its way to d8 !
1. ... tbg2-h4
The alternative I . . . �d6 eventually
transposes.
2. kf3-e4 f7-f5
3. ke4-b1 !1
In this position, this bishop move
is even more mysterious. It takes
no fewer than I S moves before its
Fernhout 1948
intention will become apparent.
White to play and win
3. ... .tc7-d6
4. b7-b8'ii' + �d6xb8 The intended solution was:
5. c3xd4 c5-c4 1. e6-e7 �d8xe7
Black must keep the long diagonal 2. 'ii'b 3-a3 �b6-d8
closed for a little longer. The rest of 3. �d4-c5 'ii'e 7-e5
the main line is known. 4. �c5-b6 'ii'e 5-e7
6. d5-d6 �b8xd6 5. �b6xd8 'ii'e 7-d6
7. d4-d5+ �h8-g8 6. �d8-e7 'ii'd 6-d8
8. �a1 -g7 a6-a5 7. .te7-f8
9. .ltb1 -c2 a7-a6 with mate.
1 0. �c2-d 1 1 c4-c31 Unfortunately this vanatwn con­
11. d2xc3 tiJh4-g2 tains a number of duals, which is
1 2. c3-c4 tiJg2-e3 why that first prize was withdrawn
1 3. c4-c51 �d6xc5 later. On the first move I . 'ii'b4 is
14. d5-d6 tiJe3xd1 winning, on the 4th move the al-

136
 Mating Patterns

ternat�ve 4.'f!Va4 is also sufficient. to try to give mate, but he must

On move 6 there are all kinds of watch out for stalemate, and even
methods to force mate. On the last his own king will face mate threats.
move, 7. 'iV a8 forces mate in a 1. ka5-b4
more elegant way. The first queen sacrifice. After
The Dutch endgame study composer I . ..ir'xa4 2 ..if8 Black would be
Jan van Reek tried to repair the study, mated.
but his attempts weren't too success­ 1. . . . 'it'a8-b8
ful. He put the white queen on d3 , Now 2 .�f8 doesn't work on ac­
removing the first dual. Next, his count of 2 . . .�g 3 and Black wins.
main line continued with 3 ... �b6 2. �a4-a81
(instead of 3 .. .'i¥e5 ) , but here, too, The second queen sacrifice.
the fun is over after 4.'ii'a 8+. On 2. ... l2ld5xe3
4 ...'ii'd 8, S ..id4 is a nice coup de 3. d2xe3 b7-b6
grace, but the prosaic exchange on The first counter-sacrifice of a
d8, followed by manoeuvring the queen.
bishop to f8 , is also quite sufficient. 4. .lkb4-c5!
I thought that not much had been Only in this way can White force
left of the attractive motifs in the the win. Alternatives do not suf­
original study. After delving deeper fice, see:
into the matter I managed to con­ A) 4 . .if8 ? 'i¥g 3 and Black wins;
struct a good version with the B) 4 . .ie7 'ifd6 ! and the white
same motifs. bishop is caught in a bizarre way -
stalemate is inevitable;
C) 4 ..ia3 . This quiet little bishop
move is hard to refute. Black saves
himself in the following long varia­
tion: 4 . . . 'ii'd 6 S .'ii'dS .ie7 ! 6.'ii'xf7
�f8 + 7 . 'iit g 6 �g3 + 8 . Wxf6
.ie7 + ! (forcing the white king to
go to a less favourable square)
9.We6 (if White takes, Black will
have a 'kamikaze queen') 9 ... .ixa3
1 O.'ife8+ 'iitxh7 I I .'iVg6+ 'ifxg6
Timman 1990
1 2.hxg6+ 'iitg 7 (here it becomes
White to play and win
clear why Black's 8th move was
The kings are condemned to each useful: the square g7 has been va­
other's company in the same way cated for the king) I 3 . WeS �cS
as before. Anchored in the corner, 1 4.Wxe4 '>t>f6 I S .'>t>d3 ii.f8 ! 1 6.e4
they cannot move. No other sub­ WeS and Black can prevent the en­
stantial changes have been made to emy king from penetrating.
the position either. White is going 4. ... �b8-d61

137
T h e Art of t h e E n d g a m e

The second counter-sacrifice. Now Here we see the difference with

5 .'i!Vc6 or 5 .'ii'xe4 would not yield variation C after White's 4th move.
anything in view of 5 . . . �e7 , fol­ The long diagonal is inaccessible to
lowed by 6 ... �f8 + , and after the the black bishop.
bishop and the queen have been 11. ... �h7-g8
captured, Black is stalemated.
5. �a8-d511

1 2. h5-h6 kc5xe3
13. h6-h7+ �g8-h8
The third queen sacrifice, and at 14. g6-g7+ <of;>h8xh7
the same time the climax of the 1 5. �f6-f7 1 -0
study. All pieces are hanging, but
neither of the two sides shows any Mating patterns with the king in the
intention of capturing anything. corner are essentially the most natu­
5 . . . 'ifxd5 would obviously be met ral. There is a well-known motif
by 6 . .tf8 and mate. where bishop and knight keep the
5. ... �d8-e7 enemy king captive in the corner.
The third counter-sacrifice. Black is The French endgame study com­
still hoping for stalemate. poser E.Ratner was the first to work
6. �d5xf7 out this theme.
Vacating the g6-square for the
white king. The rest is rather pro-
saic.
6. .ie7-f8+
7. �h6-g6 �d6-g3+
8. �g6xf6 .if8xc5
9. 'iff7-e8+
White is aiming for a queen ex­
change, otherwise he won't
achieve anything.
9. �h8xh7
1 0. tie8-g6+ 'ti'g3xg6 Ratner, 'La Strategie' 1920
White to play and win
11. f5xg6+

1 38
 M a t ing P a t te rns

White must lose a knight. How­ and advances his pawn, and White
l'Ver, he manages to drive the black has no way to make progress.
king into the corner. Therefore he has to opt for another
1. lbd4-b31 '.t>b6xb7 waiting move.
2. lbb3-a5+ Wb7-a8 3. ith6-f8!
There's nothing to be done about
i l. After any other king move, Black
loses the bishop.
3. lba5-c6
rollowed by 4.�a6 , and the white
king marches to c8 undisturbed,
after which the bishop will give
mate.

That was quite easy. The adaptation

hy the Dutch endgame study com­
poser C.].de Feijter has more con­
tent. Black is in zugzwang.
3. ... e3-e2
4. 'iiio> g3-f2
Winning the e-pawn. After that,
the king will have free rein.

The Finnish endgame study com­

poser Julius Gunst elaborated on
this theme later on, with the fol­
lowing result.

De Feijter 1932
White to play and win

White begins by walling in the en­

emy king.
1. lbd5-f6 i.h5-d1
The best defence. Black wants to
rush his passed pawn as far as pos­
sible.
2. i.c1 -h6 e4-e3
Gunst, 'Tij dschrift van den KNSB' 1949
Now White has to watch out. If he
White to play and win
plays the careless 3 .�g2 , then
3 ... �e2 ! follows. Black transfers The starting position is clearly
the bishop to the fl -a6 diagonal based on Ratner's study. Thanks to

139
The Art o f t h e E n d g a m e

the unfortunate pawn on c7 Black 6. ... .ih2-g3+

cannot prevent his king from being 7. �f2-e2!
driven into the corner. Certainly not 7 .<:J;; g 2? in view of
1. .ic4-a6+ �c8-b8 7 . . . �e l ! , followed by 8 . . . �d2 , and
Otherwise the bishop will be lost. the bishop makes it to the safe di­
2. tt:Je5-c6+ 'itb8-a8 agonal.
3. wh1 -g2 7. ... �g3-h2
White cannot allow the black pawn 8. 'ite2-f3 .ih2-g3
to reach the £'2-square. 9. �f3-g4
3. ... .ig5-h4 Good alternatives are 9 .�a6 and
If the black pawn had been on fS , 9.�e4.
the draw would have been inevita­ 9. ... �g3-h2
ble. Black could have played his 1 0. �c8-a6! �h2-g3
bishop to a square on the c l -h6 di­ 11 . �g4-f5
agonal, after which he has enough Now it becomes clear what White
tempo moves. With the pawn on has achieved with his king triangu­
f4, his movements are restricted. lation. As the black bishop is on g3 ,
4. �g2-f3 �h4-g3 1 1 . . .f3 fails to 1 2 . <;t>g4 and the
f-pawn is lost.
11. ... .ig3-h2
1 2. �f5-e6 f4-f3
13. 'iti>e6-d7

Now White faces a problem. After

S .�g4 .th2 he cannot follow up
with 6.'l£;>fs because then Black
holds the draw with 6 . . f3 7 .<:J;;e 6
.

f2 8.�d7 fi 'iY 9.�xfl <:J;;b 7 and Just in time.

the black king escapes just in time. 1 3. ... f3-f2
5. �a6-c81 14 . ..t>d7-c8 f2-f1 �
The same tempo move as in De 15. i.a6-b7
Feijter's study. Mate.
5. ... �g3-h2
6. wf3-f2 This mating net with jL+ttJ is es­
The start of a king triangulation. pecially effective, since the attack-

1 40
 M a t ing P a t t e rns

ing side al:ways has a tempo move

with the bishop. There are situa­
tions where the defending side can
save himself by avoiding the mat­
ing net with all his might.

A mutual zugzwang situation.

Black cannot make any progress;
B) 2 ...h3 Another try to tempt
White into capturing on a l . White
doesn't fall for it. 3.f6+! Not 3.Wa2?
on account of 3 ... �b2 ! 4.f6+ Wf8.
Timman 2009
Black maintains his knight on a 1 and
White to play and draw
will eventually liberate it. 3 ...'it>xf6
If Black succeeds in liberating the Otherwise, with his pawn on f6
knight on a 1 , he will win easily. White would capture on a 1 after all.
Therefore the first move is obvi­
ous.
1. <;t>d2-c1
On its way towards collecting the
knight.
1. ... tt:lc5-a4
2. �c1 -b1
Now there are two main lines:
A) 2 ... tbxc3+ If White now took
the knight, the play would trans­
pose to Gunst's study with reversed
colours. However, White has 4/,t>a2 ! �c l S.�b l ! The white
something better. 3.c>t>b2 ! ! White king keeps oscillating around the
leaves the knight in peace. The rea­ knight. 5 ... �b2 Otherwise he will
son is that now Black cannot gain a not make progress. Now, however,
tempo with his bishop, whereas a piece is lost. 6.�e8 The point of
the knight can always be captured the check on the third move: draw.
at a later stage. 3 ... �b4 4.�f3 !
Again, the only move. After 4.�g4 Bishop and knight can also force
Wf6 5 .�h3 <it>g5 White would end the enemy king into a mating net,
up in zugzwang. 4 ... �f6 s .�g4 if the defending side has pieces and
'wt>gs 6.gh3 pawns around his king.

141
The Art of the E n d g a m e

Now Black has the clumsiest

bishop pair imaginable. It is im­
portant that the white king is on
the corner square. Had he played
3 . Wa7 , then Black would now have
had the pinning move 4 . . . iLe3 ! .
4. ... c6-c5
5. �d4-e5!
The only good square for the
bishop. Now it becomes clear why
Timman 2011
White couldn't take on b7 on
White to play and win
move 3 : Black would have a check
White's task is clear. on dS in that case.
1. h5-h6 5. ... b7-b5
Threatening mate in one. There is The king's bishop cannot be saved.
only one defence. Therefore Black sets his queenside
1. ... J::. d 2-d41 pawn majority in motion.
Interrupting the long diagonal of the 6. tLlf5xh6 b5-b4
white bishop and opening the c l -h6
diagonal for his own king's bishop.
l ...l:Id8+ fails to 2 .Wa7 , when Black
will be irrevocably mated.
2. i.c3xd4 i.c1 xf4+
Now where should the white king
go?
3. 'i.tb8-a8!!
Unbelievable, but true. In the near
future the white king has to avoid
all contact with the black bishops.
The reason will soon become clear.
3. ... �f4xh6 Now White has two plans that
4. tLle7-f5 both look promising: he can either
bring his king to f8 or advance his
f-pawn. If he chooses the first plan,
things can continue as follows:
7 .�b7 c4 8 .Wc6 and now the lines
fork:
A) 8 . . . c3 9 .Wd6 c2 l O .i!.f4! (not
1 0 .�b2 ? in view of 1 0 . . . a4
1 1 . We7 a3 and Black wins) 1 0 ... b3
1 1 .�c 1 i!.ds 1 2 . \t>e 7 ! gxh6
1 3 . Wf8 and mate follows. It's nice

1 42
 M a t i n g P a t t e rns

to see how the white bishop tem­ This is the square White had to go
porarily leaves the long diagonal in to.
order to slow down the advance of 11. ... c3-c2
the black pawns; 1 2. j.e5-b2 h7-h5
B) 8 . . . a4! 9 .�d6 a3 1 0 .�e7 a2 1 3. tt:Jc6xa5 h5-h4
l l .�f8 al 'iY 1 2 .�xa l c3 and the 1 4. tl:Ja5-b3
white bishop is cut off from play. Just in time. The knight stops the
After 1 3 .<�:Jxg8 h5 ! 1 4.4:le7 �h7 c-pawn, enabling White to use his
White cannot win. bishop to stop the h-pawn.
7. f2-f41 1 4. ... �g8-h7
The second plan is the correct one. 1 5. �b2-c1 !
7. ... c5-c4 Holding off the king.
Black can also try 7 . . . .te6, in order 1 5. ... g7-g6
to immediately sacrifice the 1 6. f5-f6! h4-h3
hishop. After 8.f5 �xf5 9.4:lxf5 c4 1 7. j.c1 -f4
I O.�b7 a4 l l .Wc6 c3 1 2 .4:le3 a3 And wins. The white king, which
1 3 .4:lc2 a2 1 4.4:la l ! White wins in has been standing on the corner
the nick of time. square for 1 4 moves, will arrive
8. f4-f5 back in time to keep its colleague at
With the devastating threat of 9 .f6. bay.
8. ... j.g8-f71
Only by sacrificing his bishop can It's a pretty sight if the enemy king
Black prevent mate. is drawn to the corner to be mated.
9. tl:Jh6xf7+ �g8

Kuznetsov/Sakharov 1954
White to play and win

1 0. tl:Jf7-d811 It is clear that White will have to

The knight also has to go to the try and win one of the black bish­
rim. After other knight moves, ops. However, l . .l:rb8 is insufficient
Black could achieve a draw. in view of l . . . �xa2 2.l:rxh8 hxg 6 ,
1 0. ... c4-c3 and Black saves himself Therefore
11. tl:Jd8-c6 White must first give a check, in

1 43
T h e A r t of the E n d g a m e

order to drive the enemy king to a

less favourable square.
1. Ub1 -b5+! 'it>g5-h6
2. .!:!:b5-b8 �g8xa2
An important alternative was
2 . . . �g 7 . In that case White con­
tinues 3 .h6+ �xh6 4 . .l::rx g8 and
now:
A) 4 . . . itb2 5 .gxh7 'it>xh7 6.l:rg3 ,
followed by 7 .I:i:a3 and wins;
B) 4 . . . �g 7 . Now White has to Timman 2010
White to play and win
be careful, as 5 . gxh7 Wxh7
6 . .l::r e 8 itb2 leads to a theoretical The problem for White is that his
draw. However White wins after a-pawn is hanging. For example,
5 . lhg 7 ! 'it>xg 7 6 . gxh7 Wxh 7 l .gxh7 would not suffice for the
7 . 'it>xa3 and the black king will win on account of l ... gxa2 .
be held off by its white counter­ 1. h4-h5
part. Preventing the capture on a 2 .
3. l:.b8xh8 �a2-g81 1. ... 'it>g7-h6
A surprise. The black bishop re­ The best defence. Black renews the
turns in order to lock in the rook. threat of taking on a2 . Bad was
4. �a4xa3 <it>h6-g7 l ... hxg6 in view of 2 Jhg6 + Wh7
It may seem as if White has not 3 . .C:g2 and the white a-pawn is pro­
achieved anything, but . . . tected, after which he wins easily.
5. h5-h6+1 �g7xh8 2. l:.d6-d81
6. g6-g7 Only so. After the tempting 2.g7+
Wxg7 3 .h6+ Black would deliber­
ately allow White to lock him up
after 3 . . . Wh8 . The stalemate motifs
would then render the win impos­
sible.
2. ... �g8xa2

Mate.

I thought that a better study could

be built around this mating posi­
tion.

1 44
 M a t i n g P a t t erns

3. J:id8-h8!
The only way to worry Black. On
.L.<;;i;>xhS , 4.g7 wins. Black has to
give his bishop for the g-pawn,
and then the white rook and king
will prevail over the enemy pawns.
However, Black again has the
self-incarceration trick.
3. ... �a2-g8!
4. �d4-c3
The king has to rush to the Timman 1983/2011
White to play and win
queenside in order to stop the
f(xemost a-pawn. White is a pawn down, but in spite
4. a3-a2 of the reduced material he has dan­
5. �c3-b2 a4-a3+ gerous attacking chances.
6. 'Ot>b2-a1 1. :i.c7-c4
The white king is bricked in. Now With this attack on the bishop
the same will happen to its col­ White wins an important tempo;
league, with the difference that the he will manage to give a bishop
latter will not be stalemated, but check on f4. His rook will land in a
mated. pin, but that is less important.
6. �h6-g7 1. ... �e4-d5
7. h5-h6+1 �g7xh8 2. �c1 -f4+ �d8-c7!
8. g6-g7 Black sacrifices his bishop, with
Mate. the sole purpose of blocking the
c-file. After 2 . . .<lt>a8 3 .�f3 he
would be quickly mated.
3. �f4xc7+ �b8-a8
Not 3 . . . <;;i;> c 8 in view of 4.�g4+
and mate.
4. �d1 -c2!
A quiet move. White is going to
bring his bishop to e4. The move
4.�f3 was insufficient for the
win. After 4 . . . ..txf3 S . .l:!.f4 �dS + ,
followed by 6 . . J:I:b8 , Black
A magnificent spectacle. As if by a achieves an endgame that is just
magnet, both kings have been tenable.
drawn to the corner squares. 4. ... l:tb5-c5
The best defence. Black is playing
The mating position in the corner for stalemate.
can also be completely different. 5. �c2-e4 .!:!.c5-c61

1 45
T h e Art of t h e E n d g a m e

bishop because of stalemate. Now

the study reaches its climax.
6. �b3·b21
A quiet unpinning move. Black is
forced to exchange on e4.
6. ... �d5xe4
7. �c4xe4 �c6xa6
Otherwise a back-rank mate fol­
lows. But now mate comes from
the other side.
8. �e4xb4
A curious situation. The rook in­ And mate on b8 cannot be pre­
terrupts the pin by the bishop vented. The black rook is on the
which, in turn, keeps the white most unfavourable square imagin­
rook pinned. White cannot take the able.

1 46
Chapter S

Stalemate Patterns
Stalemate patterns have in com­ This is hardly one of Kubbel's best
mon with mating patterns that studies, but it is nice to see how
they are very attractive when they Black, due to the mate threats
occur with the king in the middle against his own king, is forced to
of the board. The great Russian deprive the white king of its last
endgame study composer Leonid squares.
Kubbel wove many stalemate mo­
tifs into his studies. In Kubbel's study, White's king did
not move. It is more attractive if it
moves to the middle of the board
in order to be stalemated. Kubbel
also composed studies with this
motif.

Kubbel, 'Shakhmatny Vestnik' 1914

White to play and draw

White is a rook behind, but he can

create threats on the kingside.
1. f5-f6 tt:Jg3xh5
Kubbel, 'Shakhmaty' 1925
Not l . . . Wf8 on account of 2 .h6 ! .
White to play and draw
2. f6-f7+ �e8-f8
The black king is now bricked in Both sides have a dangerous passed
and, consequently, vulnerable. pawn. Due to certain circum­
White is going to exploit this in a stances, the black king turns out to
curious way. be the safer of the two.
3. ..ig5-d21 c6-c5 1. �e1 -f2+
Please take note: White still has Driving the king into the corner.
four squares for his king here. 1. ... '1t>g1 -h1
4. �d2-e3 l:la3-a5 2. h6-h7 c3-c2+
5. �e3-f4! tt:J h5xf4 Now the white king cannot move
Stalemate. upwards : 3 . Wg4 fails to

147
The A r t of t h e E n d g a m e

3 .. .<�:Jf6 + , while 3 .�h4 does not

work in view of 3 . . . tt::l e 7 , when
there is a family check in the po­
sition.
3. i£.f2-e31
White sacrifices his bishop in or­
der to obtain the f2-square for his
king.
3. ... .lla 3xe3+
4. 'it>g3-f2 .lle 3-h3
Rusinek, 'Problem' 1973
White to play and draw

White is two pieces behind, but his

passed pawn is not easy to stop.
1. b5-b6 .ll h5-f5
2. b6-b71
White sacrifices the pawn in order
to disturb the coordination in the
enemy camp.
2. ... .llf5-f7+
3. 'it>c7-d6 t2Je5-c4+
It looks as if White is heading for
Things look gloomy for White. Af­ disaster. However, there is a hidden
ter S .hxg8'ti' .l:th2 + 6 . �f3 c l 'ti' he rescue.
has nothing to hope for. 4. 'it>d6-e6 i.f3xb7
5. �a2-d5+11 5. .l:l.a3-h3+ 'it>h6-g7
But with this bishop sacrifice 6. .l:l.h3-g3+ 'it>g7-f8
White introduces a surprising
stalemate.
5. ... c6xd5
6. h7xg8'tW J:rh3-h2+
7. 'it>f2-f3 c2-c1 'tW
8. 'MHg8-g2+!!
The beautiful point of the bishop
sacrifice.
8. ... J:rh2xg2
Stalemate.
For this study, Kubbel received first
prize in the Shakhmaty tourney. 7. .l:l.g3-g8+! '>t'f8xg8
Stalemate.
Rusinek is a past master in this For this study, Rusinek received
genre. first prize in the Problem tourney.

1 48
 S t a l em a te P a t t erns

Five years later he made a new

variation on this theme, which
,\gain was awarded first prize!

7. �g3-f3! �e6-d5
8. a7-a8'ii' �d5xa8
Stalemate.

Rusinek, 'Szachy' 1978 Nice stalemate motifs can also oc­

White to play and draw
cur in practical play.
Again White is two pieces behind,
and again he has a dangerous
passed pawn.
1. a6-a7 J:!.b5·a5
2. e5-e6!
With this pawn sacrifice White
tries to disturb the coordination in
the enemy camp. Black has to take
the pawn.
2. ... ..td7xe6
3. .l:!.c4-e4
Vedder-Gagu nashvi I i
The point of the previous move.
Vlissingen 2004
White attacks the queen's bishop
and at the same time threatens to Black threatens 7 2 . . . .t!.h4. White
give check on e 1 . has several ways to prevent this.
3. ... �b2·c1 + 73. i.e5·d6?
4. �f4-g3 J:ta5·a3+ But this is not one of them.
5. �g3-f2! 73. J:!.h2-h4!
White must choose the squares for 74. ..td6·e7+ �g5-f4
his king with care. 75 . .ie7xh4
5. ... J:!.a3-a2+ Stalemate.
6. 'it>f2-g3 J:!.a2-g2+
Again it looks as if White is hope­ A bitter disappointment for Vedder,
lessly lost. And again, salvation is but it inspired me to make a nice
close at hand. study.

1 49
The Art of the Endgame

4. J::Id 4-d1 �b1 -c2

5. J:td1 -g 1 h5xg4

Timman 2011
White to play and draw

White cannot capture the rook,

since after l .�xe6 lLJd3 + 2 .We3 Now 5 .lhg4 is insufficient in view
lt:Jxc5 3 .�f7 �f5 ! Black obtains a of S . . . �d l ! , and the f-pawn can­
.

technically winning position. not advance.

1. J:tc5-c4+ 'it>a4-b3 6. f3-f4!!
2. J:lc4xf4 Highly surprising. White sacrifices
Here also, 2 .�xe6 was insufficient a pawn instead of capturing one.
in view of 2 . . . lt:Jxe6 3 . l:tc6 �f5 6. ... g5xf4
and now White cannot prevent 7. l::I g 1 xg4 .!:[e5-f5
Black bringing his king closer. 8. c;i;>f2-f31
2. l:te6xe5 Steering towards Gagunashvili's
3. J:tf4-d4 stalemate with reversed colours.
8. �c2-d1 +
9. <M3-e4 �d1 xg4

White doesn't have to fear the cap­

ture on g4 yet. After 3 . . . hxg4
4Jhg4 White would be able to Stalemate.
trade off the last black pawn.
3. ... �b3-c3 In January 1 984, I made the fol­
Forcing the rook to leave the 4th lowing study on a flight from Bel­
rank, since 4 . .l:.a4 is met by 4 . . . �c2 . grade to Amsterdam.

I SO
 S t a l e m a t e P a t t erns

whereupon Black deals the final

blow: 6 ....th6+!
Despite pointed play, White cannot
save himself in this way. So, this
variation is the temptation.
3J�tc7xc5+1 'it>c3-d4

Timman 1984
White to play and draw

White is a piece down, but it looks

as if he will have no trouble win­
ning it back.
1. .a.d8-d7 .ib7-c8
On closer inspection, this turns out
to be not so easy. On 2..I:I:xg7 Black This pos1t1on drove many strong
replies 2 . . JH6! and after 3 . .l:tg5 c4, players, to whom I showed the
followed by a general exchange on study, to despair. It is clear that
fS , he wins the pawn ending. Not, White cannot capture the queen's
however, 2 . . . ..txfS in view of bishop in any way, since after both
3 .l:lc7 ! with a draw. 4.lhf5 ..th6 + and 4.'�ixf5 .l:lf6 +
2. l:.d7-c7 �c8:xf5 h e loses his rook. S o h e will end up
Now White has two possibilities. two bishops down, and then what
The first is: does he have in this position?
3.�xf5 %:tf6+ 4.�gS Black now has 4. l:rc5-c71
to abandon either the bishop or the This is the answer. The rook re­
pawn. Of course, he opts for the turns to capture the black king's
first. 4 ... Wd4 Why not 4 . . . c4 ? That bishop, if the queen's bishop is re­
will soon become clear. 5.f4 Before moved from fS . After, for instance,
capturing the bishop, White first 4 . . . �h3 5 Jhg7 l:.xg7 . . .
saves his f-pawn. After 5 . . . �h8
6 . .l:tc8 Black can no longer save his
bishop. S . .l:i:xf4! Now the purpose
. .

of Black's 4th move becomes clear.

If White captures the rook, there
follows 6 . . . �e5 ; if he takes the
bishop he will end up in a lost
rook ending, and his situation is
not improved by 6.l:.d7+ <>t>e3 .
The last possibility is 6 Jhc5

151
The Art of t h e Endgame

White is stalemated.
4. ... i.f5·e4
5. f3xe4 �g7·e5+
6. 'Oiif4·f5 l:!.g6·f6+
7. c;t;>f5·g5 '1>- 112
The stalemate in the diagram posi­
tion was my starting point for this
study. In fact, it is not more special
than a stalemate on the edge of the
board, as the black king plays the
part of 'the edge' here. It deprives This seems like the final straw, but
the white king of all squares on the the position contains a hidden re­
e-file. source.
4. ... b6-b5
Chapter 3 centred on the famous Closing off the diagonal for the
combination from the game bishop. However, White is not dis­
Ortueta-Sanz. This fragment did couraged.
not contain a stalemate combina­ 5. �a6xb5! c6xb5
tion. Nevertheless, it is possible to 6. J:rf1 ·e1 !
conceive of a stalemate combina­ Preparing the stalemate.
tion against this background. 6. ... e3xd2
7. J:l.e1 xe2 d2-d 1 .J:I.
Promotion to queen amounts to
the same.
8. .J:I.e2-d2+1 l:td 1 xd2
Stalemate.

In the following study, the king

saves itself on the corresponding
square on the kingside.

The black e-pawn has to be stopped.

1. llg2·g1 i.g6·f7+
2. c;t;>a2·a31
The only square for the king. After
six moves, the reason will become
clear.
2. ... t2ig3·f1
Black starts the combination.
3. l:!.g1 xf1 e4-e3 Timman 2009
4. .ltb7·a6! White to play and draw

152
 S t al e m a t e P a t t erns

The white pieces are hanging. He 9. .!:!:g3-g2 l:!.c2-c1

will have to wriggle and squirm to 1 0. .!:!.g2-g1 !
avoid decisive material loss.
1. ti:Jd6-e4 .ia2-d5
2. �f1 -g2
Only move. On 2 .jt_d3 , 2 . . Jie6
would have decided immediately.
2. ... .id5xe4
3. .ig2xe4 .l:rg6-e6
Again White has to think of some­
t hing to avoid the loss of a piece.
He cannot capture on cS , as then
an endgame arises which is a theo­
retical win for Black. This stalemate trick saves White.
4. J::!. c 3-b3+! 1 0. ... l:!.c1 -c3+
Forcing the king to step aside. 11. l:tg1 -g3 .ie7-b4
4. ... �b4-c4 A final attempt. 1 1 .. Jhg3+ is met
5. .ie4-c2!1 by 1 2 .�xg 3 , followed by 1 3 .h4,
when Black is left with the bishop
of the wrong colour.
1 2. .!:!.g3xc3+ �b4xc3
Threatening 1 3 . . . �e 1 , after which
the bishop could nestle on h4. But
White pips him at the post.
1 3. 'iot>h3-g4 1/,- V,

The following study is much more

comphcated. In the various stages,
the variations are so complicated that
Introducing a stalemate combina­ this version is devalued by duals.
tion.
5. ... J:!.e6-e1 +
6. �h1 -g2 l:!.e1 -e2+
7. �g2-h3
Only on this square will the king
turn out to be safe.
7. ... .!:!.e2xc2
8. l:!.b3-g3 .ic5-e7
The only way to protect the
g-pawn. However, now the black
king and rook turn out to be un­ Timman 1987
fortunately placed. White to play and win

1 53
T h e Art of t h e E n d g a m e

Here the black king has barricaded I s .jlf2 ! jles 1 6.c5 ! tbe4 1 7 .�d4!
itself on the same square as the �xd4+ 1 8 .r:i.xd4 tbxcS 1 9.l:lf4
white king had in the previous and Black is powerless against the
study. White has a clear material penetration of the white rook.
plus and also a far-advanced passed 3. ... tt:lb3-d4!
pawn. However, the position con­ Only by sacrificing the knight
tains all kinds of stalemate motifs, again can Black continue the battle.
and moreover, the white king is After 3 . . . �xc3 4.lhc3 tbd4 S J:[e3,
not entirely safe. he would run out of possibilities.
1. Jld2-d31 4. l:td3xd4 �a1 xc3
A quiet move with a positional jus­
tification. White's plan is simple:
he wants to play 2 .c3 and then pro­
mote his h-pawn.
1. ..
. tt:lc7-b51
The first knight sacrifice, which
White has to accept.
2. a4xb5 tt:la5-b31
The point of the previous move.
White cannot accept the second
knight sacrifice, since after 3 .cxb3
�d4+! 4.l:hd4 Black is stalemated. Black's defence is based on the fact
3. c2-c31 that White cannot get a queen.
White continues the struggle for Now comes the climax of the
the d4-square. However, also suffi­ study.
cient for the win was 3 .h8'ii' , in 5. �a3-b211
order to drive the bishop to the ot­ As preparation for a bishop pro­
ter end of the diagonal. The main motion White first sacrifices his
line runs as follows: 3 . . .�xh8 4.c3 queen's bishop. The purpose of
tbcs S . .Ue3 tbe4 6.�b4! .ieS 7 .hS this is to force the black bishop to
tbgS 8 .h6 �f4 9 ..l:.d3 �es . This is move to a less favourable square on
as far as Mark Dvoretsky's variation the long diagonal. After immedi­
goes in his book Studies for Practical ately S .h8�? �e l he would even
Players, followed by the conclusion lose.
that Black can hold. However, with 5. ... �c3xb2
careful manoeuvres White can still 6. h7-h8�1
make progress: 1 O .�a3 �f4 Here we have the Phoenix Princi­
1 1.. �b 2 ! tbh7 1 2 .c4 ..txh6 ple, which I discussed in Chapter S .
1 3 .�d4 and White will collect the Now it looks as if White wins eas­
b-pawn. Here too, he has to keep ily, but Black can continue the
playing accurately. An important struggle since the white king is still
line is 1 3 . . . tbgS 1 4.�xb6 �g7 unsafe.

1 54
 S t a l em a t e P a t t erns

6. ... �b2-c3 11. ... e5-e4

Again threatening 7 . . . �e 1 . 1 2. h6-h7
7. i.h8-e5! Also here, 1 2 .d6 is a dual solution .
With the idea of capturing on g 3 . 1 2. ..tb2-h8
� - d6xe5 1 3. d5-d6 e4-e3
8. .!:!.d4-e4 14. d6-d7 e3-e2
In this phase of the study, White
must keep restraining the black
king's bishop with his rook. With
1 he text move he takes control of
t he square e 1 .
8. ... �c3-b2
The bishop is looking for new
squares.
9. h4-h5
Certainly not 9 .d6 in view of
9 . . �a3 with the double threat of
.

I O . . . �cS + and 1 O . . . .ild6. With the 1 5. �c4-h4+! 'it>h3xh4

text move White creates a new 16. d 7-d8�+
strong passed h-pawn. And wins.
9. ... �b2-a3 For this study I received second
1 0. l:le4-c4 prize in the Kaminer Memorial.
That was nice, of course, but the
duals in the final play continued to
nag at me, even though the first
dual, on move 3 , had gone unno­
ticed by the jury. I decided to make
a new version, with a correct
finale.

Preventing the deadly check on cS.

From here on, the study contains a
few more duals. 1 1 Jk2 is forced
only after 1 0 . . �f8 , and then it is
.

indeed a clear win. The real main

variation continues as follows:
1 0. ... �a3-b2
11. h5-h6 Timman 1987/2011
White to play and win
1 1 .d6 also wins.

1 55
The A r t of t h e E n d g a m e

Instead of a passed pawn on h4, Black threatens to bring his king to

White has one on fS . This makes f3 , after which White would not
no difference for the opening be able to win. Black would just
phase of the study, apart from the keep playing waiting moves with
joyful fact that the dual has been his bishop on the a3 -f8 diagonal.
eliminated. 1 3. .l::!.g 2-a2!
1. l:ld2-d3 tt::l c7-b5 With this attack on the bishop
2. a4xb5 tt::l a 5-b3 White vacates the g2-square for his
3. c2-c3 tt::l b 3-d4 king with gain of tempo.
4. .l::t d3xd4 �a 1 xc3 1 3. ... �a3-c5
5. �a3-b2 jLc3xb2 1 4. �g1 -g2
6. h 7-h8� �b2-c3 And wins.
7. �h8-e5 d6xe5 A satisfactory version of the study.
8. .l::!. d4-e4 �c3-b2 All the pieces take part.
9. f5-f6
From here on, things will be dif­ In the following study, the stale­
ferent. The white f-pawn is less mate is hidden.
dangerous for Black than the
h-pawn, since both White's passed
pawns can be stopped by the
bishop on the a3-f8 diagonal. This
means that White will have to play
very accurately in order to convert
his advantage to a win.
9. ... �b2-a3
1 0. .l::!. e4-c4 e5-e4
11. .l::!.c4-c2!
The only way to win. The foremost
Timman 2008
black g-pawn has to be eliminated.
White to play and win
11. ... e4-e3
1 2. l:Ic2xg2 �h3-g4 The white rook is under attack, and
so is the f-pawn. White makes a
virtue out of necessity by sacrific­
ing the rook.
1. llh4-h1 I .:lf1 xh1
2. �b3-d5 l:lh1 -f1
3. �d5-f3
The point of the rook sacrifice. The
black rook is held off and it looks
as if White will win without trou­
ble. However, there is a stalemate
combination for Black.

156
 S ta l e m a t e P a t t erns

3. ... tt:Jc8-d6! A curious position. White deliber­

4. c5xd6 J::tf1 -f2! ately puts both his minor pieces en
5. f6-f7 �e1 -f1 prise, thus eliminating the stale­
mate threat. Now the study enters a
new phase.
8. _ d7xe6
9. d6-d7
Here there is a dual. The alternative
9 . \:t>f3 would lead to the same re­
sult.
9. ... J:lb2-b8
This is why Black's 7th move was
best. The rook can go back and de­
fend.
Black's plan becomes clear. If 1 0. �g3-f3 'itg1 -h2
White queens, he plays his king to 11. �f3-e4 �h2-g3
g 1 , after which White cannot Now the kings dominate the bat­
evade the rook checks on g2 and tlefield.
h2. 1 2. �e4-e5 �g3xg4
6. f7-f8ttJI
Shadows of Chapter 5 . Only by this
knight promotion can White avoid
the stalemate. But the stalemate
threat continues to loom in the
background.
6. ... �f1 -g1
7. ttJf8-e6 J::tf 2-b21
The best defence, as we will see. If
White now plays his knight, Black
again has the check on g2.
8. .if3-c611 1 3. �e5-d6!
There's no time to take the e-pawn.
1 3. ... �g4xh5
1 4. �d6-c7 J::t b8-f8
1 5. d7-d8'iY J::tf8xd8
16. �c7xd8
From this point on, there are duals
in the main line. White is going to
collect the a-pawns, which can be
done in two different ways. Despite
this, the struggle remains sharp,
since it will be decided by a race.

157
T h e Art of t h e E n d g a m e

16. ... g5-g4 The black rook and bishop are boxed
1 7. ..t>d8-c7 ..t>h5-g5 in on the queenside. The bishop
1 8. ..t>c7-b7 h6-h5 cannot very well move, as then the
1 9. ..t>b7xa7 h5-h4 a-pawn falls. However, the black
20. ..t>a7xa6 h4-h3 knight threatens to liberate its col-
21 . ..t>a6-b61 leagues in a few moves. Therefore
The best king move. White must act very energetically.
21 . ... g4-g3 1. b2-b41 tt:le1 -d3
22. a5-a6 g3- g2 2. b4-b5 tt:ld3-b4
23. a6-a7 g2-g 1 � Now the situation looks critical for
24. a7-a8� �g1 -b1 + White.
3. b5-b6!
Not 3.l:ra5 in view of 3 ... �c7 4.b6
.l::lg 8 + , followed by 5 ... axb6, and
Black wins.
3. ... tt:lb4xa6
4. b6-b7 tt:la6-c7

Another important moment.

25. i.c6-b5
Protecting the d-pawn, after which
White wins easily.

In Chapter 6 , I introduced the Ex­

celsior theme. It also features in the Now all Black's pieces are boxed in,
following study. but he threatens to free himself
with 5 . . . a5 or 5 ... a6.
5. i.g 1 -c5!!
A spectacular move that poses
Black a difficult question.
5. ... d6xc5
6. d5-d6 a 7-a51
The only move that doesn't lose.
7. d6xc7 .l:!.a8-a6+
8. �g6-h5!
Preparing the stalemate.
Timman, 'The Problemist' 2009 8. ... �b8xc7
Draw (dedicated to Geertje) 9. b7-b8'i¥

1 58
 S ta l em a t e Po t terns

Excelsior! �d8 3 .h7 �e7 , followed by a de­

9. ... .ic7xb8 cisive intervention by the bishop.
Stalemate. 2. �h8xh7
3. .l:.g1 -g8
Shutting the black rook off from
play.
3. . .. �c8-d8
4. c5-c6!
Limiting the black bishop's free-
dom of movement.
4. ... b7xc6
5. c4-c5 �d8-e7
6. 'ito>h3-g4

In the following study, White

spends a number of moves on the
creation of a self-stalemate. Black
manages to dismantle this threat,
but then the white king finds a
new square on which to get stale­
mated.

Black is a full bishop to the good,

but he cannot free his pieces. The
white rook keeps both Black's rook
and his bishop under control at the
same time. Still, White has to keep
playing purposefully, because if he
keeps alternating between g3 and
Timman 1986
h3 with his king, he will lose. See:
White to play and draw
6.�g3 �d7 7 . �h3 1:1h6 8 . �g3
White is a piece behind and has to jLe6! and Black frees his bishop,
act courageously in order to prevent since in the pawn ending after
Black consolidating his position. 9.fxe6 l:.g6+ White doesn't stand a
1. h5-h6 .l:.f8-h8 chance.
2. h6-h7 6. ... .!::r h 7-g7+1
White must sacrifice this pawn as This rook sacrifice is Black's best
quickly as possible. Other moves chance. He has to make it right
will not do, for example: 2 . .l:!.g7 away, as after 6 . . . �d7 7 .hS it would

1 59
The Art o f t h e E n d g a m e

be too late. In that case White could Compare this posltlon with
never be brought into zugzwang, as Vukcevich's. White is about to cre­
he would always have the squares ate a self-stalemate situation.
g4 and h4 for his king. 1 0. ... �e8·d7
7. t%g8xg7 'it>e7-f8 11. J:lh6-h8!
By sacrificing his passive rook, The consequence of the preceding
Black has now managed to put the play. The enclosed rook is sacrificed.
white rook out of play. 11. ... 'it>g7xh8
8. .l:.g7-h7 1 2. 'it>h5-h6 'it;>h8-g8
Tempting is 8 .1::tg 5 . Black then The black king has to leave the cor­
wins in the nick of time: 8 .. .fxg5 ner in order to eliminate the stale­
(not 8 . . . �d7 9.h5 !) 9.�xg5 Wg7 mate threat.
I O . f6 + �h7 l l .�f5 �d7 + 1 3. h4-h5 Wg8-f8
1 2 . <>t>xe5 �e6 1 3 . Wd6 �b3 14. 'it;>h6-h7
1 4.<>t>xc6 �xa4+ 1 5 .Wb6 �c2 ! . Or
1 4.e5 Wg6 1 5 . Wxc6 it.xa4+
1 6. Wb6 ltc2 1 7 .�xa5 <>ttfS and
the rest is not difficult. Originally I
had made this study without the
a-pawns. In this variation the rea­
son for their presence becomes
clear. Since the pawn on a4 is vul­
nerable, White cannot achieve the
draw in this way.
8. ... 'it>f8-g8
9. J:::t h 7-h6! The white king is on its way to the
Only move. 9 J:thS is prospectless, corner in order to establish a new
as Black will then bring his bishop self-stalemate situation there.
outside the chain and attack the 1 4. ... �d7xf5+
white pawns. There was no time to play the
9. ... r;.t1g8-g7 bishop round to a better square. So
1 0. 'it>g4-h5 it is better to sacrifice it right away
in order to create a passed e-pawn.
1 5. e4xf5 e5-e4
1 6. h5-h6 e4-e3
1 7. 'it;>h7-h8 e3-e2
1 8. h6-h7
Forcing the black king to leave f8 .
1 8. ... 'it>f8-e7
1 9. 'it;>h8-g8
Not 1 9.'it>g7 in view of 1 9 ... e l 'iW
20.h8� �g3 + 2 I .Wh7 �h3+

1 60
 S t a l e m a t e P a t t erns

l VJ!tg7 �g4+ 23.�h6 'ifg5 +, and The white king is caught in a mat­
l he f-pawn will be lost with check. ing net. White has to keep the en­
1 9. ... e2-e1 'ir' emy bishop from the e4-square.
20. h7-h8� 1. J:l:f1 -f41
Fi nally a queen ending has arisen, The only way to control both the
w here White still has to find a few e4- and f3-squares.
accurate moves. 1. ... �h7-e4+!
20 . ... 'tlr'e1 -g3+ With this bishop sacrifice the knight
21 . �h8-g7 'tlr'g3-g5 gets access to the crucial f3-square.
The most interesting try. After 2. .!:tf4xe4 liJh2-f3
2 1 . . . �b8 + 2 2 .�h7 White would Now White will have to achieve
have nothing to fear. the aim by either sacrificing all his
22. l.t>g8-h8 'i:Yg5xf5 pieces or reaching the h-file with
his rook.
3. liJb1 -d2
First the knight.
3. ... J:b2xd2
4. l::!. e4-b4+
With the knight sacrifice White
has cleared the way to the b-file for
the rook.
4. ... <;t>b8-c8

23. 'tlr'g7-f8+1
The final point. 2 3 .'ihf7 + would
have led to the same result.

A simple example of stalemate in the

corner is seen in the following study.

A new sacrifice is hanging in the air.

5. �e8-d7+1
If the king now takes the bishop,
there follows a check on b 7 , after
which the white rook reaches the
h-file.
Timman, 'The Problemist' 2004 5. �c8-c7
White to play and draw 6. J:l:b4-b7+1

161
The A r t of t h e E n d g a m e

Black must always accept this rook Generally the queen doesn't get a
sacrifice. defensive role, but necessity knows
6. ... �c7xb7 no law. Now 5 . . . g5 would not yield
7. i.d7-c6+! �b7xc6 anything after 6 .'ifh7 + <;t>g4
Stalemate. 7 .�h2 .

In the following study it is even

clearer that White aims for stale­
mate.

5. ... g7-g6!
An extremely level-headed reply.
Timman 1989 Black prevents the check on h7 and
Black to move - White draws
again puts the question to White:
White has an overwhelming mate­ how does he want to lose both of
rial surplus, but Black now sets up his pieces?
a lethal battery. 6. 'iYe4-f5+1
1. l:f.h2-g2+ Bull's eye. Black must take the
2. 'it>g1 -h 1 �e2-f3 queen, and then the stalemate is
3. .ilb7-f7 achieved.
Only by attacking the bishop with
the rook can White defend himself Stalemate in the corner can be
After 3 .'itb3 .l:(g3 + 4.'iixf3 .l:hf3 much more subtle.
5 J:Ixg7 l::rf l + Black would win the
pawn ending.
3. .. . �f3-d5
The only safe square on the long
diagonal. The bishop has to be on a
protected square.
4. 1:tf7-f3+
With clear intentions. After
4 . . .i1t.xf3 White forces stalemate
with 5 .'i¥h4+.
4. I!g2-g3 Timman 2010
5. �b4-e41 White to play and draw

1 62
 S ta l em a te P a t t e rn s

Th� whit� king is in danger. More­ The only square for the rook. Now
over, he cannot exchange on f6 as comes the climax of the study.
then the bishop on f7 will be lost.
1. �f7-d51 :r.:!.c6-c71
Threatening mate in one.
2. .ib2-e5!
A beautiful, problem-like move,
which serves to transfer the bishop
to the h-file.
2. ... .l:tc7-h7+
3. .ie5-h2
White seems to have warded off
the danger. However, now Black
starts an interesting combination.
3. ... l2Je1 -f3! 8. �h2-f41
With this knight sacrifice, the f-file With this second bishop sacrifice
is closed. White puts the black rook in a pre­
4. .id5xf3 dicament. Black cannot capture the
The only move. After 4.gxf3 �eS bishop, as that would mean stale­
White would be irrevocably mated. mate.
4. ... .if6-e5 8. ... :r.:!.g5xg4
9. �f4xe5 :r.:!.g4-g1 +
1 0. 'iti>h1 -h2 J:[g1 -g5
A last try.
11. �e5-d4+
Draw.

The construction of the following

study is even more complex.

5. .if3-h51
In turn, White sacrifices his king's
bishop in order to re-open the
f-file.
5. ... :r.:!.h7xh5
6. .l:!.f8xf5+1
This rook sacrifice is the conse­
quence of the bishop sacrifice.
6. .l:!.h5xf5 Timman 2010
Draw (dedicated to Geertje)
7. g2-g4 :r.:!.f5-g5

1 63
T h e Art of t h e E n d g a m e

Both sides have a far-advanced The stalemate trap is set. Black can­
passed pawn, so White's first move not promote to a queen.
speaks for itself. 4. ... l::.g 1 -g2!
1. h6-h7 �c1 -b2 The best try. After 4 . . . a l .:l'. 5 .l:!.d2+
But now things become compli­ White would have a 'rampant
cated. Not only has Black prevented rook'. This means that he can keep
promotion, but he also threatens to giving checks with his rook, even
give a devastating rook check. in it is en prise. If the opponent ac­
2. tt:Je6-c5+11 cepts the sacrifice, White is stale­
This knight sacrifice introduces a mated.
stalemate combination eight
moves deep.
Alternatives were:
A) 2 .l'hb2 �g i + 3 .Wf7 .trfl +
4.Wg6 (it looks as if Black cannot
win, as after the mutual promo­
tions, the white king would be safe
enough) 4 . . JH6+ ! ! (a tremendous
surprise) 5 . Wxf6 a I 'iV and wins;
B) 2.<1:Jf8+ Wc6 3 Jhb2 llg l +
4.Wf7 .:l'.g7 + ! (an echo of variation
A) 5 . Wxg7 a i 'ii' 6.h8'ii 'ifxb2+ 5. l:rb2-b1 !!
7 .Wg8 'ii'xh8 + 8 . Wxh8 and now A fantastic riposte. Firstly White
both 8 . . . d5 and 8 . . . Wd5 suffice for threatens 6 . .ild 1 + , when he would
the draw. again have a rampant rook. And on
The point of the text move will be 5 . . . l:rh2 , 6 .lla l suffices for the
revealed later on. Incidentally, White draw.
could not switch round the first and 5. .. .
J:[g2-g8+!
second moves. After I . tbc5 + dxc5 A stunning attempt. Black also sac­
2.h7 Black would win with 2 ... .:!.b l ! , rifices his rook, steering towards a
for example: 3Jh2 .l:b8+ 4.Wg7 queen ending where he has two
ii.e3 , followed by 5 ...�d4+. extra pawns.
2. ... d6xc5 6. h7xg8'if a2xb1 'if
On the alternative 2 . . . '0t>c6 White 7. 'fig8-g4+
had the following road to a draw: The only correct check. He must
3 Jhb2 l:rg l + 4.�f7 l:lfl + 5 .�g6 not let the black king escape via c8.
Itf6+ 6.Wxf6 a l 'ii' 7 .ltJd3 ! , fol­ 7. �d7-d8
lowed by promotion. The knight is 8. 'iYg4-g8+ �d8-d7
used to protect the rook. Now 9. 'ifg8-g4+ �d7-d6
stalemate motifs start to play a role. There's nothing to be done about
3. l:re2xb2 l:ra1 -g1 + it. Now White finally forces stale­
4. <;,tg8-h8! mate - with a queen sacrifice.

1 64
 S ta l em a te P a t t e rn s

1 0. 'irg4-g6+! �b1 xg6 On 2 . . . �e6 White would already

Stalemate. have 3 . tt:Jcs +.
For this study I received second 3. .:tf2-e2+
prize in the Nona 201 0 tourney. It is White has to keep giving checks.
dedicated to my wife Geertje, be­ After 3 .tbc3 hlb l ! 4.tlJd5 + We6
cause I made it just before her 3 6th 5 .tbxc7+ �d7 6.Rxa2 �xc7 Black
birthday. would be winning. The white
passed pawn is not strong enough.
The study was received enthusiasti­ 3. ... we7-d7
cally and published in many places. 4. ti:la4-c5+!
Unfortunately, later it was discov­ The well-known sacrifice - only
ered that it contained a dual. In the knight comes from the other
variation B, in the comments to side this time.
White's second move, 3 . .l:.c2 + 4. d6xc5
�dS 4.l:hb2 .!:tg l + 5 .tlJg6! .!:txg6 5. wh8-g8 �c1 -b2
6.�h8 leads to a draw. Here, too, 6. .!:.e2xb2
White gets a rampant rook; I had­ And we know the rest.
n't thought of that possibility at all. The echo variations are lost, but in­
I had to construct a new version. It stead a pretty knight sacrifice for
looks like this: Black has been added.

Besides a rampant rook, there is

also such a thing as a rampant
bishop.

Timman 2011
White to play and draw

White's king is stuck in the corner.

He has to give check.
Kubbel , 'Kolnische Vol kszeitung' 1926
1. llh1 -f1 + tLld1 -f2!
White to play and draw
With this knight sacrifice, Black
creates space for his rook, although The white king has so many
this cannot be seen at this point. squares that stalemate looks im­
After other moves White could probable. Nevertheless, the deci­
play 2 .�g8 without trouble. sive stalemate trick will appear
2. .l:tf1 xf2+ Wf7-e7 soon enough.

1 65
The A r t of t h e E n d g a m e

1. e6-e7 i.a7-b8 Threatening 2 .�g7 , s o Black's re­

The only way to stop the foremost ply is forced.
e-pawn. 2. .
. . �b4-f8
2. e4-e51 Now what is left for White to do?
An important pawn sacrifice. Black threatens to collect the
2. ... i.b8xe5 f-pawn with his king.
3. ..te8·d7 �e5·d6+ 3. b2-b4!
4. �a3-a41 Threatening 4-.�cS , so Black has to
The king barricades itself on the take en passant.
edge of the board. 3. ... c4xb3
4. �d6xe7 4. �c1 ·b2
5. i.d7·e6+ 'Ot>g8-g7 Now a new motif emerges. If
6. �e6-d51 White managed to exchange bish­
Black cannot capture the bishop in ops, the position would be a dead
view of stalemate. draw, even without the f-pawn,
6. .ig2-f1 since the black queen's bishop is
7. i.d5-c4 �f1 ·h 3 very unfortunately placed on a2. If
8. i.c4-e6 Black succeeds in avoiding this ex­
And the black bishop cannot es­ change, then he will bring his king
cape the attacks. to d 1 and give check on c 1 , driving
the white king into the comer.
Next he will play his king to c2 and
The same principle is also possible his queen's bishop to b 1 , and give
with the king in the comer. mate. In order to avoid this sce­
nario, White has a hidden ma­
noeuvre at his disposal that gives
him a rampant bishop.
4. ... 'Ot>h4-g5
5. 'it>b2·a1 'it>g5·g6

Timman 1983/2009
White to play and draw

This is an extended version of an

old study of mine. White must
push his f-pawn.
1. f5-f6 i.a5·b4
2. f6·f7 6. �d4·b2!

166
 Stalemate Patterns

The bishop retreats to make the The white king heads for the
drawing mechanism work . kingside. However, he couldn't
6. ... i.f8-c5 play 3.'it>b4? in view of 3...�a6!
7. f7-f8'tW �c5xf8 4.<;£;lc3 f2 S.�g2 �e2 with a draw.
8. i.b2-a3 �f8-g7+ 3. ... �g4-h5
9. �a3-b2 �g7-h6 Black must keep waiting. After the
10. �b2-c1 tempting 3 ...�d7+ 4.�c5 �c6
Draw. White would free his queen's
bishop with 5 .�b8! and win easily.
A follow-up study shows how ac­ 4. 'it>b5-c41 'it>b7-a6
curately the attacking side must 5. 'it>c4-d31
play in order to avoid the exchange As a consequence of these manoeuv­
of bishops. res, White has taken control of the
e2-square. This renders the march of
the black f-pawn harmless.
5. ... �h5-g4

Timman2011
White to play and win

In this miniature study, Black is

threatening to initiate the same
drawing mechanism that White 6. 'it>d3-d21
had in the previous study. So White makes a king triangulation,
White has to play sharply from the point of which will become
move 1 . clear at the end of the study.
1. .llf1-g2+ f4-f3 6. ... �g4-h5
2. �g2-h1! 7. 'it>d2-e3
White deliberately positions his 7.�e 1 is also good.
bishop in the corner in order to fix 7. ... �h5-g4
the black bishop on hSI g4. If Black 8. 'it>e3-f2 .i.g4-h5
reached the long diagonal, the 9. 'it>f2-g3
draw would be inevitable. Now everything becomes clear.
2. ... �h5-g4 Black has been put into fatal
3. 'it>a5-b5 zugzwang.

167
Chapter 9

Mutual Zugzwang
In pawn endings, mutual The king scorns the white pawns
zugzwang (also known as recipro­ in order to keep control of the
cal zugzwang) within the margins square d6.
of winning and losing is a quite 2. d4-d5+ 'it>e6-e5
common phenomenon. Take, for 3. b5-b6
instance, a white pawn on d4 and a The time has come for the pawn
black one on d5 , flanked by both push.
kings, a pattern known as trebuchet. 3. ... tt:Jg5xe4+
The obligation to move inescapably 4. 'it>g3-h4!
leads to one of the players' down­ Avoiding all knight checks.
fall. Outside pawn endings. mutual 4. tt:Je4-d6
zugzwang within the margins of 5. �a3xd6+ �e5xd6
winning and losing does not often 6. 'it>h4-g51
occur. Kubbel composed a famous
study on this theme.

A situation of mutual zugzwang

Kubbel 1928/1929 has arisen. If it were his move,
White to play and win
White would lose.
The white b-pawn is dangerous, 6. ... tt:Jg8-e7
but if White pushes it right away, What else?
Black can stop it with I . . . lt:Jxe4+ 7. b6-b7 <;1o>d6-c7
followed by 2 . . . li:Jd6. 8. d5-d6+
1. d2-d4+1 And wins.
The battle for the d6-square flares This study gained Kubbel first
up. prize in the 'Siberian Chess Section
1. ... 'it>e5-e6 Tourney' .

1 68
 M u t u a l Z u g zwang

The Russian endgame study com­ The white rook is under attack and
poser M.G.Kliatskin has made a has to be sacrificed.
draft study that has become very 1. .!':.c6-c8+1 �b8xc8
famous. 2. b6-b7+ �c8-b8
3. d4-d5
In Kliatskin's study, the mutual
zugzwang meant the difference be­
tween a win and a draw, but here
we have a win-or-lose situation.
Black's doubled d-pawn makes the
difference. These two pawns also
make a stalemate trap possible.
3. ... s.f.i>b8-c7

Kliatskin, 'Shakhmaty' 1924

White to play and win

The solution plays itself.

1. c6-c7! �d6xc7
2. a5xb6+ wc7xb8
3. b6-b7
And wins. The reciprocal
zugzwang is lethal.

More than a decade later, the It may be exaggerated to speak of a

well-known Russian player and trap, as this is Black's only move.
endgame study composer Now White has to take care which
A.S.Selesniev added an element to piece he promotes to.
Kliatskin's draft study. 4. b7xa8�!
Not 4.bxa8tb+ �b8 , and the win
is gone.
4. ... �c7-b8
5. .ia8-b7
And wins.
The final position is also a case of
mutual zugzwang.

It irritated me that Selesniev's

study, however pointed, lacked fit­
ting introductory play. I succeeded
Selesniev, '64' 1935 in making a study where White
White to play and win
starts with a double rook sacrifice.

1 69
T h e Art o f t h e E n d g a m e

An incredible situation. In spite of

his enormous surplus in material
Black is lost.
4. ... 'Wg5xd5
5. e4xd5
And the rest as in Selesniev's study.

In this study, the Phoenix Principle

made another appearance: the
white bishop, which had been cap­
Timman 2011 tured by the queen, rose from its
White wins (after Selesni ev)
ashes on a8. It's possible to build in
White starts with the same rook a knight promotion as well.
sacrifice.
1. J:!.c1 -c8+1 'it>b8xc8
2. b6-b7+ 'it>c8-b8
But now what? Without the rook it
would be simple - the bishop
would go to dS with deadly force.
With the rook on the board, White
must even take care that he does
not lose. On 3 . .l:f.bS there would
follow 3 . . . dS +.
3. J:[d5-g5!1
Absolutely the only move. Black Timman 2011
White to play and win (after Selesniev)
has to take the rook, since after
3 . . . d 5 + 4.l:rg6 White keeps a Here Black has counterplay be­
devastating check on g 8 in re- cause he has a check on fl .
serve. 1. b6-b7+ 'it>c8-b8
3. 'i'h6xg5 An important alternative is
4. �f7-d5 1 . . . Wc7 . If White then reacts
thoughtlessly with 2 .bxa8'i?i?,
Black has a perpetual check start­
ing with 2 . . . 'fifl + . The only way
to win is 2 .bxa8lt:l+ ! . After
2 . . . �b8 3 . .l::thxd6 'iffl + 4.l:.d3
't!Ya l + 5 .�b5 't!Yb2 + 6.Wc4 't!Yc2 +
7 . Wb4 't!Yb2 + 8 . .tlb3 ! Black has no
useful checks left.
After the text move White has a
problem. If he continues with
2.bxa8'i!V+ �xa8 3 J lhxd6, then

1 70
 M u t u al Z u g zwang

3 . . . 'iVfl + does give Black enough The main intention of this little
counterplay, for example: 4.J:I:d3 rook move is to block the e-file. Af­
iVa i + 5 .�b5 'ti'b2+ 6.�c5 'iYc2 + ter 2.exd7 exfl 'tiV 3 .d8'i¥+ 'Ot>b7
7 . �d4 �f2 + and there is no good 4.'i¥d7+ 'it>b6 5 .'i¥c6+ �aS White
way for White to escape the wouldn't be able to win.
checks. 2. ... f3-f2
2. .th5-f7!! Black has to bite the bullet, other­
Only with this problem move can wise White would block the black
White win. The queen's passage to pawns with 3 .�g3 .
f1 is cut off, and at the same time 3. e6xd7 wb8-c7
the bishop is ready to take up its An important alternative was
position. 3 .. .fxe 1 'tiV. Then the play continues
2. ... 1Wf8xh6 as follows: 4.d8'iY+ �b7 s .'it>e6 !
3. J:!.d5-h5 1Wh6xh5 (threatening mate) 5 . . . g3 6.�d5 +
4. �f7-d5 r;i>b6 7 . b4! and the black king is
And the rest is well-known. caught in a mating net.
4. �e5-e6+ c.t>c7-d8
The play in the following study is
quite sharp.

How can White continue his siege

Timman 2011
on the black king? He cannot move
White to play and win
his rook.
The black passed pawns are very 5. b5-b6!
dangerous. For example, after Threatening mate on c 7 .
l .l:re l exd6+, followed by 2 . . . f2 , 5. ... a7xb6
Black could look to the future with 6. ..th2-g1 1
confidence. The point of the previous move.
1. d6-d71 Black cannot take the rook in view
Opening the h2-b8 diagonal for of mate on b6. Now a field of ten­
the bishop. sian has been created that looks
1. ... ..te8xd7 like Nadareishvili's.
2. .!:r.f1 -e1 ! 6. ... g4-g3

171
The Art of the Endgome

Protecting the f-pawn. The situa­ The rooks and pawns on the
tion is extremely sharp. Paradoxi­ queenside form an unmovable
cally, this calls for a quiet little block. In order to maintain the op­
move. position, White has to keep his
7. b2-b3! king on the first rank.
All other moves lose. 1. Wh1-g11
7. b6-b5 After l.�g2 �g4 White would be
8. b3-b4 lost.
1. 'it>h5-g5
2. Wg1-f1 �g5-f5
3. �f1-e1 'it>f5-e5
4. we1-d1 �e5-d4
Other wise Black cannot make
progress.
5. 'it>d1-d2
Mutual zugzwang. Black cannot win .

As opposition and mutual zug­

zwang go together, so do corre­
Mutual zugzwang. Black is going to sponding squares. This is shown in
get mated. the following study.
8. g6-g5
9. h4xg5 f2xe1�
10. �g1-b6#

Besides in pawn endings, the op­

position motif can also occur in
rook endings. And with it, mutual
zugzwang.

Nadareishvili, 'Ceskoslovensky
Sach' 1953- White to play and draw

The first moves are obvious.

1. h6-h7 J:.c6-c8
Now White has to be on his guard.
After the obvious 2.�c4 a3 he
would end up in a fatal zugzwang
situation, for instance: 3.11Lf7 J::i.b8!
Timman 1993
4.�c4 r!b2 5.h8'i¥ l::lg2+ 6.�h7
White to play and draw
.:t.h2+ 7.�g7 l:!.xh8 8.�xh8 fS!

172
 Mutual Zugzwang

9 .exfS �f6 and the bishop cannot

stop the two black passed pawns.
2. �f1-b51
Forcing the pawn to move forward.
2. ... a4-a3
3. �b5-c4
The same position, with Black to
move. By carefully choosing the
hest squares for his bishop, White
prevents the black rook from turn­
ing up behind the white lines. The situation still looks precarious
3. ... J:!.c8-a8 for White. However, there is a way
4. �c4-a2 out.
Preventing the further advance of 9. �d5-g81
the a-pawn. Preparing a self-stalemate.
4. ... J:ra8-b8 9. f6-f5
10. e4xf5 e5-e4
11. f5-f6 e4-e3
12. f6-f7
And stalemate is unavoidable.
Summarizing: the corresponding
squares were a2-a8, f7-b8, c4-c8
and d5-d8.
For this study Nadareishvili re­
ceived third prize in the Prague
Problem Tourney.

An important moment. White does In the following endgame, mutual

not have the b3 -square at his dis­ zugzwang also plays a crucial role.
posal, because there his bishop
would simply be captured. There­
fore, he will have to find another
corresponding square.
5. �a2-f71
Now the rook cannot penetrate.
5. ... J:rb8-c8
6. �f7-c4
This is the corresponding square.
6. ... J:!.c8-d8
7. �c4-d5 J:rd8-h8
Timman 1984/2010
A last try.
Black to move- White wins
8. 'it>g7xh8 �e7-f8

173
The A r t o f the E n d g a m e

In order to properly understand

this position, it is necessary to real­
ize that Black's king aims for f6. Af­
ter he has swapped his bishop for
the pinned rook, he must try to
prevent White fixing the e-pawn
with d4-d5 .
1. ... �d8-e8
2. �e2-h5+!
Not L�f3 right away in view of
2 . . . . ..txb3 + 3 .�xb3 'it;>£7 4.Wc4 The critical moment in this study.
�f6 5 . �b5 d5 ! , followed by 7. �g4-h3!
5 . . . e5, when White's last pawn will Only this little bishop move wins.
be exchanged. White does not want to put his
2. ... 'Ot>e8-d7 bishop on e6 until Black has
After 2 . . . Wf8 3 . �b2 ..txb3 pushed his d-pawn. The alternative
4.Wxb3 �g7 5 .d 5 ! White would 7 .�c8 was not a good waiting
win effortlessly. move, as Black would then hold
3. .ih5-f3! the draw with 7 . . . e5 8 .�b7 + d5 .
Threatening 4.d5 , and so forcing 7. ... d6-d5
Black to capture the rook. Insuffi­ Now, 7 . . . e5 was insufficient in
cient for the win was 3 .�b2 on ac- view of 8 . ..tg2 + , followed by 9 .d5 ,
count of 3 . . . .ixb3 4.�xb3 d5 , fol- and White keeps his pawn.
lowed by 5 . . . 'it>d6 and 6 . . . e5, with 8. ii.h3-e6
a draw. Reciprocal zugzwang. Both sides
3. ... �e6xb3+ have taken up their optimum posi­
4. 'it;>a2xb3 �d7-e6! tions. Black has to release the pres­
Black has to play as actively as he sure on the white d-pawn.
can. Waiting too long would hand 8. ... 'lt>e4-f4
White a technically winning posi- 9. 'it;>c3-b4
tion. Of course not 9 .�xd5 because of
5. �f3-g4+ 9 . . . e5.
Not 5 . Wc4 in view of 5 ... �f6 , 9. ... 'it;>f4-e4
with the same position as in the 1 0. 'it;>b4-c5 'it;>e4-f4
comment to White's second move. 11. 'it;>c5xd5
The bishop check forces the black And wins.
king to move forward.
5. ... 'it;>e6-d5 In the following study too, the mu­
Not 5 . . . �f6 , as then White fixes tual zugzwang arises when both
the pawn formation by 6.d5 . sides have taken up their optimum
6. 'it;>b3-c3 'it;>d5-e4 positions.

1 74
 M u t u a l Z u g zwang

Kuriatnikov, 'Shakhmaty (Baku)' 1988 8. ... �a2-g8

White to play and win
9. �f5-b1
This looks like a simple exercise. Besides the mutual zugzwang,
The black knight is hemmed in and domination also plays an impor­
will be captured by the king. tant role in this study. The black
1. c;t>c3-b4 tt:Ja3-c4! bishop is systematically deprived
After this sharp reply, things turn of squares.
out to be more complicated. White 9. ... c;t>b7-a6
has to take the knight. 1 0. �b1 -e41
2. 'it;>b4xc4 �b7-c8 After this accurate move the black
With a double attack. Black wins bishop is irrevocably lost.
the a-pawn. 1 0. ii.g8-b3
3. tt:Jg4-e5 ii.c8-e6+ 11. ii.e4-d3+ 'ito>a6-b7
An important moment. Who 1 2. tt:Jc6-a5+
wouldn't automatically play 4.�c5 And wins.
here? But that is the wrong square.
4. 'it>c4-b5! Sometimes mutual zugzwang can
Anticipating the mutual zugzwang be overlooked by the best endgame
situation which will arise shortly. study composers.
4. �e6xa2
5. tt:Je5-c6+ 'it;lb8-b7
6. �h7-e4
After other bishop moves Black
would play 6 . . �d5 .
.

6. ... w b7-c8
The only move to prevent the im­
mediate loss of the bishop.
7. �e4-f5+ c;t>c8-b7
8. �b5-c5!
Mutual zugzwang. Black has only
Fritz, 'Svobodne Slovo' 1955
one move to prevent the immedi­
White to play and draw (incorrect)
ate loss of his bishop.

1 75
The Art of t h e E n d g a m e

The intended solution runs as fol­ the other hand, White has a
lows: trump card that is not to be un­
1. tt:Jg8-f6+ �h5-g6 derestimated: inside the crum­
2. tt:Jf6-e8 �g6xf7 bled pawn structure on the
3. tt:Je8-d6+ �f7-e7 queenside - which the attentive
4. tt:Jd6xc8+ �e7-d7 reader will immediately associate
5. tt:Jc8xa7 �d7-c7 with Fritz's study - the black king
6. �a4-a51 �c7-b7 is caught in a mating net. First
7. tt:Ja7-b5 White has to clear a passage to c6
And whichever way Black takes, for the queen.
White is stalemated. 1. 'it>f6-f5
If the white king moves to h7 to
The stalemate appears out of the hide from the checks, then Black
blue here. In that sense, this is a can save himself:
nice study. However, it is not cor­ l .�f7 l:tf3 + 2 .�g8 l:ig3 + 3 .�h7
rect. Probably Fritz didn't realize .!::t g 4 4.'ii' c 6 l:re4 5 . a4 .l:th4+
that the position after White's 5 th 6.�g6 tt::lxa4 and White will not
move is a case of mutual zugzwang. manage to collect the black rook
By losing a tempo with his king with checks, for instance: 7 .'ii'a 8+
somewhere along the way, Black �b5 8 .'ii'e 8+ �a6 9.'1i'c8 + �a7
can win. With a revised study with and White cannot make any prog­
the same background, constructed ress. He constantly has to watch
by myself, I will explain this. out that Black doesn't parry a
check by interposing his knight on
b6.
1. ... l:tg3-f3+
After l . . . e2 2 .'ii'c 6 l:tf3 + 3 . �g4
l:rg3 + 4.�h4 Black would soon
run out of moves.
2. �f5-g4 l:rf3-g3+!
Black sacrifices a full rook in or­
der to be able to promote with
check.
3. �g4xg3
Timman 1984
Certainly not 3 . Wh4 .tf2. , and
White to play and win
Black wins by the force of his bat­
One of my colleagues, to whom I tery.
showed this position, inquired if 3. ... e3-e2
White was supposed to make a Now White has to withdraw his
draw here. Indeed, Black's three queen in order to prevent promo­
pieces and advanced passed pawn tion.
are superior to White's queen. On 4. �h6-c1

1 76
 Mutual Zugzwang

Now Black has two possibilities: With reversed colours we have

A) 4. . ttJd3 The
. most obvious now reached the same position as
move. Now White conquers the in Fritz's study. However, now
e-pawn with a series of checks: White decides to lose a tempo.
5.'ilfg5+ �c5 6.'it'd8+ <ot>a6 Other­ 7.We3!! This is the key. The black
wise the queen will give mate on knight remains boxed in, but
the a-file. White makes no haste to collect it.
7 ... ttJxa2 8. <ot>d2 Wa4 After
8 ...'it>b6 White has to be patient
again, because if he goes for the
knight right away, it is a draw:
9.Wc2 Wc5 10.Wb2 Wxc4
1l.�xa2 �xc3. Winning is 9.a4
WcS 10.aS or 9...WaS 10.cS. The
black knight is kept under control
by the white pawn on c3. 9.Wc2

7.'�'c8+! Only this check wins. Af­

ter 7 .'iia8+ �a7 White would
have no better than a perpetual.
7...Wa5 8.'i!Ya8+ <ot>b6 9.'iib8+
'Oita6 Or 9...<;,i;>c6 1 O.'t!Ye8+ and
wins. IO.'ii'bS+ Wa7 ll.'ti'd7+
followed by a check on e6 or e8,
winning the e-pawn.
B) 4... �£'2+ This is the main line.
After first having sacrificed a whole Black is in zugzwang. After 9...Wxa3
rook, he now gives the bishop. IO.c5, again, the hindmost pawn
White has no choice. 5.Wxf2 ttJd3 keeps the knight under control,
6.Wxe2 tZ:lxcl+ while the foremost pawn promotes.

177
T h e Art of t h e E n d g a m e

Sometimes one side has an enor­ 5 .�c5 �c7 6.b6+ �d7 7 . �b5
mous material surplus, but he can �d6 White will be out-tempoed.
still be put in zugzwang if he has But now comes a surprise.
too little space at his disposal. 4. �b4-c5!1
Gaining an important tempo by
leaving the black rook untouched.
4. ... .ll b8-a8
The alternative 4 . . . ita8 does not
lead to a win either, as is shown by
5 .axb8'iV+ �xb8 6.b6 �c8 n £?d6
�d8 8 . �e6 and the white king
cannot be driven from its active
position.
5. 'it>c5-b6
This is the moment of reciprocal
Timman 1984 zugzwang. Black has to give his rook.
White to play and draw
5. Ita8xa7
White is a rook behind, but Black 6. �b6xa7 �c8-c7
has trouble bringing his forces into 7. b5-b6+ �c7-c6
play. 8. 'Wt>a7-b8 �c6xb6
1. b4-b5 Stalemate.
Certainly not 1 . Wc3 �a6 and Here, at the end, it becomes clear
White doesn't stand a chance. why there had to be this seemingly
1. .. . �e6-d7 useless pawn on d3. Without this
Black wants to eliminate the white pawn, Black could have aimed for a
bishop as soon as possible. winning pawn ending with
2. �d2-c3 �d7-c8 8 . . . ..ta6 9 .�a7 Wb5 .
3. 'ito>c3-b4 J:ta8xb8
Mutual zugzwang can suddenly
appear in complicated positions.

It looks as if Black is heading

straight for a technically winning Timman 2008
position. After 4.axb8'iV+ Wxb8 White to play and draw

1 78
 Mutual Zugzwang

Black has an enormous material White cannot accept this sacrifice,

surplus, but still it will be hard for as Black then promotes with check.
him to stop the white passed 7. 'it'f2·e2 l:tf3·f8
pawns.
1. d6-d7
After l.ti:Jxe3 dl 'iY 2.ti:Jxdl .l:i.d2+!
Black would win.
1. ... .!:!:e3-e2+
2. 8g4-f21
White sacrifices his knight in order
to block the enemy f-pawn. After
2.'�hl �h4, mate would be inevi­
table.
2. ... l:te2xf2+
3. 'lt>h2·h3 Finally Black has managed to stop
Now Black has to sacrifice a lot of the dangerous white d-pawn. Now
material to stop the d-pawn. White has to be on his guard.
3. ..
. .!:::.f2·h2+ 8. b2·b31
The first sacrifice. White has to ac­ Only move. On any other move
cept. Black would play 8 ... b3. In order to
4. 'it'h3xh2 understand this, it is important to
realize which plan White has here.
The white d- and h-pawn will in­
evitably be swapped for the black
d-pawn. Next, White will have to
try to reach the square aS with his
king, and then push his a- and
b-pawns. We know this stalemate
construction from Chapter S. If the
black pawn were on b3, the white
king would need an extra tempo to
reach the square aS, as he would
4. ... �e1-g 3+ first have to capture that pawn.
The second sacrifice. 8. ... l:[f8·d8
5.
cJi>h2xg3 9. h6-h7 d5·d4
Forced, because after S.�h3 �c7! Again White must choose his
6.h7 f2+ there is no hope for move carefully.
White. 10. 'it>e2-d11
5. ... f3·f2+ Paradoxically, White must lose a
The f-pawn has to be sacrificed tempo in the heat of the battle.
too. Only after the trade of pawns will
6. �g3xf2 l:td3-f3+ speed become important again. I

179
The Art o f t h e E n d g a m e

will explain later why the alterna­ In the following study, White aims
tive 1 0. Wxd2 loses. for an ending of bishop versus
1 0. ... d4-d3 boxed-in rook. This ending is
11. �d1 xd 2 highly similar to a pawn ending,
where the motif of mutual
zugzwang fits in nicely.

Mutual zugzwang. If White had

played 1 O .Wxd2, it would now be
his move. Then the play would Timman 2010
White to play and win
continue as follows : 1 O . . . d3
1 1 .h 8'ii .l:txh8 1 2.< ;1;xd3 .C:d8 The first move is obvious.
1 3 . �c4 �xd7 1 4. �xb4 .C:d8 1. b5-b6
1 S .Wa5 Wc8! 1 6.a4 Wd7 , and Not only protecting the bishop,
1 7 .b4 fails to 1 7 . . . rl..a 8#. Now that but also creating a strong passed
it is Black's move, everything is dif­ pawn.
ferent. Black cannot move his rook. 1. ... �d4-e5
11. ... �b8-a8 An important alternative was
There is nothing else. Now White 1 . . . I:te7+ . Then White has to pur­
does have success with his plan. sue a different winning plan than
1 2. h7-h8'if in the main line. Play continues as
Not 1 2.'�r>xd3 in view of follows: V i tfl �eS 3 J�c8 ! <ot>g7
1 2 .. .lhd7 + 1 3 . Wc4 ld.d8 ! and 4.b 7 ! 1:hc7 S .b8"ti' I:c l + 6 Jhc 1
again Black is in time to give mate �xb8 7 . Wg 1 �d6 8 .l:tc6 and
on a8 with his rook. He does not White has a technically winning
have to collect the white h-pawn. position. In this variation it is im­
1 2. ... J::.d 8xh8 portant that White can protect the
1 3. Wd2xd3 J:ih8-d8 h-pawn with his king, which is a
14. �d3-c4 .U.d8xd7 consequence of the black rook
1 5. 'it>c4xb4 l:id7-d6 check.
1 6. 'it>b4-a5 2. �c7xe5
With a draw. Black cannot prevent Under these circumstances, 2 .l:!.c8
White continuing with 1 7 .a4 and was just insufficient for the win.
1 8 .b4. Black has the strong reply 2 . . . Wg6 ! .

1 80
 M u t u a l Z u g zwang

He does not have to fear rook end­ This rook sacrifice is the point of
ings, since after 3 .<iJfl (threaten­ the previous move.
ing 4.b7) 3 . . . jlxc7 4.lhc7 lH8 4. ... l:te7xb7
S .b7 llb8 , he will hold. After the 5. �e5-c7
text move, Black cannot enter the The bishop cuts off the black rook
rook ending with 2 . . . fxe5 on ac­ from further play, so that there is
count of 3 .b7, after which Black now virtually a pawn ending on
cannot prevent White from giving the board.
a rook check and then promoting 5. ... Wh7-g6
his b-pawn. 6. �e1 -f2 'it>g6-g5
2. ... J:!.f7-e71 Now White has to be careful. After
The chief aim of this little move is 7 . <iJxf3 <iJfS he would be in
to put the rook behind the white zugzwang, for example: 8 .�g3
passed pawn. For example, after �e4 9 . <it>xh3 <iJf3 ! and Black has
3 .<iJd2 l:txeS 4.b7 .l:tbS White nothing to fear.
would have no winning chances. 7. �f2-g31
So White has to try something else. Anticipating the mutual zugzwang.
3. f2-f41 7. �g5-f5
The introduction to a pretty com­ 8. �g3xf3
bination. The intention of this
move is to let Black take en passant,
after which his kingside pawn
structure is weakened.
3. ... g4xf3
The alternatives were:
A) 3 . . .fxe5 4.b7 exf4+ 5 . �d2
and wins;
B) 3 . . . g3 4.b7 gxh2 s J:Ih8+
�xh8 6.b8'iV+ '1t>g7 7.'Yi'a8 ! fxeS
8 .'il'e4 and wins.

Black is in zugzwang now. He can­

not prevent the white king reach­
ing the important centre square e4,
after which the winning process is
automatic.

Shortly after finishing this chapter,

I played a small open tournament
in Torre delle Sterre, a spa in the
south of Sardinia. Asphalted sand
4. l:tb8-b711 tracks, a bungalow at sea and Sar-

181
T h e Art o f t h e E n d g a m e

dinian red wine were ingredients 69. ... ..t>c4-d5

for a quite special time. 69 . . . Wd4 would be met by
In the fourth round I got the fol­ 7 0. Wf3 , with an immediate draw.
lowing endgame on the board: 70. ..t>e3

Timman-Shytaj
Torre delle Stelle 2011
The mutual zugzwang is estab­
It is clear that White is going to lished.
lose a piece. Where do his best 70. ... tt:lc6-d4
chances of salvation lie? White can Now White can force the draw by
try 66.f5 , but after 66 . . J:tf2 6 7 .f6 advancing his f-pawn. The best
cxd2 68Jhd2 .:txf6 69.l:t.d5 l:f.h6 winning try was 70 . . . 'iit e 6. After
there is no immediate draw. Fortu­ 7 l . We4 Wf6 7 2 . f5 li:JeS Black
nately, I had enough time on the would manage to block the white
clock. After fifteen minutes' pawns after all. However, here the
thought I discovered a hidden mu­ problem is that another mutual
tual zugzwang. zugzwang situation has arisen,
66. <oftc2-d1 1 .lle 2xd2+ for example : 7 3 . W f4 li:Jg4
67. .l:l.d5xd2 c3xd2 74. We4 ti:Jh6 7 S .Wf4. With a
68. <oftd1 xd2 <oftb4-c4 knight you cannot win (or lose) a
Black brings his king closer. It tempo.
looks as if he will manage to block 71 . f4-f5! <oftd5·e5
the white pawns. 72. f5-f6 tt:ld4-e6
69. �d2-e21! 73. f6-f7
The only move. Draw.

1 82
Chapter 10

Building a Fortress
When Smyslov was 8 3 , he created 3. ... '>t>f6xg7
following study.
1 he Black could play 3 . . . h4 here, to
maintain perspectives for his
bishop. It wouldn't have yielded
him any winning chances: White
just places his rook on gS and
waits. After the inevitable rook
swap he has nothing to fear.
4. h3-h4

Smyslov 2005
White to play and draw

In view of the closed structure of

the position, it is clear that Black
will have great problems convert­
ing his material advantage. Yet,
White will have to tread very care­
fully in order to actually build a
fortress. An unbelievable situation. Black is
1. tt:la6-b8 a rook and a bishop to the good,
Insufficient was 1 . tt::J c s on account but the position is so closed that
of 1 . . . �h8 ! and Black keeps all his the only try he has left is to pene­
trumps. trate with his rook.
1. ... l:c6-d6 4. ... J:!.d7-d6
Black tries to dominate the knight. Black is going to bring his rook to
2. tt:lb8-d7+! the g-file, hoping to penetrate the
Anyway! The aim of this knight white position. However, a sur­
sacrifice is to lure the black rook prise awaits him.
away from the 6th rank. 5. <ito>h1 -g2 l:d6-g6+
2. ... l:Id6xd7 6. <it>g2-h3!
3. J:!.g1 xg7! And the black rook has to leave the
With this exchange sacrifice White g-file again in order to prevent
builds up his fortress further. stalemate.

1 83
The A r t o f the E n d g a m e

This principle of making a line in­ A curious situation. In view of

accessible to an enemy piece can stalemate, the black queen cannot
also be realized on a diagonal. infiltrate along the fl -a6 diago­
nal.

Hasek's study looked to me like an

invitation to new adaptations. I
made two new versions.

Hasek, 'Narodni Listy' 1951

White to play and draw

Both sides still have all their pawns.

In order to avoid a breaking move
with the black h-pawn, White sac­
rifices his knight.
1. tt:Jf6-g4+ h5xg4 Timman 2010
White to play and draw (after Hasek)
Now how can White meet the threat
of a penetration by the queen? First of all White must concentrate
2. d3-d4+ �e5-f5 on closing the position.
3. llb1 -h1 1 1. d2-d3+ 'it>e4-d4
This mysterious rook move pro­ 2. �e1 -d2
vides the answer to the above ques­ Threatening mate in one, so Black
tion. White is weaving a stalemate is forced to allow the closure of the
construction into the position. position.
3. li'h6-f8 2. ... e5-e4
4. �d2-e1 �f8-a8 3. e2-e3+ <;t>d4-e5
5. �e1 -f1 �a8-a6 4. d3-d4+ <;t>e5-f6
6. �f1 -g1

1 84
 Bu ilding a F o r t ress

Now the position is almost com­ rook up after the queen sacrifice,
pletely closed. Just as in the two he must play precisely in order to
previous studies, White must now reach the safe haven of a draw.
sacrifice a knight, though for a dif­ 1 3. a4xb5 c6xb5
ferent reason this time. 14. 'iiig 1 -f1
5. ltJg7-e8+1 The king hurries back to the
With this knight sacrifice White threatened wing.
closes off the back rank for the 1 4. ... <;i;>d7-c6
black queen, thus winning a tempo. 1 5. 'iiif 1-e1 b5-b4
5. ... ..IW.f7xe8 1 6. <;i;>e1 -d2 <;i;>c6-b5
6. �g1 -h21
Otherwise Black would play
. . . h3-h2 .
6. ... 'it'h7-d7
7. l:.c1 -h1 1
The self-enclosing process is in full
swing.
7. ... 'tWd7-c8
8. �d2-e1 'iYc8-a8
9. <;i;>e1 -f1
Just in time.
1 7. l:th1 -c1 1
An accurate little move.
1 7. ... �e8-c6
1 8. �d2-c2
And there is no way for Black to
make any real progress.

My second adaptation contains a

new element.

9. ..
. �f6-e7
Before Black plays his queen to a6,
he first improves the position of
his king.
1 0. �f1 -g1 <J;e7-d7
11. <;i;>g1 -f1 'tWa8-a6+
It makes no sense to wait any longer.
1 2. �f1 -g1 li'a6-b5
This was the intention of the king Timman 2010
march. Although White remains a White to play and draw (after Hasek)

1 85
The Art of t h e E n d g a m e

Again, first of all White has to try 4. i.g6xh7

to close the position as much as 5. l:ta1 -h 1 �d7-c8
possible. 6. Wd2-e1 1fc8-a8
1. e2-e3 7. we1 ·f1
Otherwise Black would play With a draw, as in the previous
l ... e3 + . study
1. a6-a5!
Vacating the a6-square for the It can happen that certain strong
queen. grandmasters are not familiar with
2. tt:Jg7-e8! the notion of a fortress.
This knight sacrifice is familiar by
now. Black has to accept.
2. ... Wf7xe8
3. .Ua2-a 1 !
Heading for the other corner.
3. ... �h7-d7

Position after 52.'i!;>b2

Larsen-lvkov, Manila 1973

In those days, the time control was

after the 5 6th move. Doubtlessly,
Ivkov was in blood-curdling
time-trouble here, a shortcoming
which seemed to be an addiction
A crucial moment. If White for him. The rook ending is easily
thoughtlessly continues to build up tenable, as long as Black realizes
his self-stalemate construction, he that he has a fortress.
is lost; 4.J:lh l ? is met by 4 . . . �h7 ! . However, he produced the horri­
Black can then evade the white ble S 2 . . . g 6 ? , destroying his for­
stalemate trap by leaving the square tress with his own hands. After
h7 with the bishop at the right this, Larsen had no trouble win­
moment, for example: S .We l "ii'c 8 ning the endgame: S 3 .fxg6 a3+
6.�fl '*Va8 7 . �g l rJJ e 7 8 .�fl S 4 .Wb3 'it> g 7 S S .'it>a4 ..t>xg6
'tWa6+ 9 .Wg l �g6 1 0.h7 "ii'e 2 S 6 . 'it>bS 1Jf8 S 7 .'it>c6 Uf2
with mate. S 8 ..l:lxb6 .l:Z.xa2 S 9 . '1t>xd6 lib2
4. h6-h7! 6 0 J:ta6 1 -0 .
In view of the above, White has to As Hans Bouwmeester showed in
sacrifice the pawn right away Schaakbulletin, Black could safely

1 86
 B u i l d i n g a F o r tress

have given his b-pawn and pro­ 8 3 .h5 tt::lxe3 84.h6 Wf6 the game
tected his a-pawn from a7 with his was drawn.
rook, also keeping the 7th rank un­ In his book Hogeschool van het eind­
der control, for instance: S 2 . . . l:ta8 spel ('Endgame University') Max
S 3 . �a3 .l:ra7 5 4Jhb6 �e7 Euwe indicates that 70 .�f7 would
S S . .!:i:b8 �f7 5 6 . .l:i.h8 �e7 5 7 .l:!.hS have been 'much more convinc­
�f6, and White cannot make ing' . He considers this position
progress. The black king can always winning for White, substantiating
protect all the weak pawns, while this with the following variation:
the white king cannot penetrate 7 0 . . . tt::l c 3 + 7 l .�b3 tt::l d l 7 2 .ilxh5
anywhere. tt::l xf2 7 3 .�g6 f4 74.exf4 e3
7 S . �c2 e2 7 6 . �d2 and White
So far I have shown only positions wins.
with fixed pawn structures. There However, Black has a much stron­
are also fortresses possible in posi­ ger defence: 7 0 . . . tt::ld 6 ! . The knight
tions where the structure is not does not go out on a pawn hunt,
fixed. And then it can happen that but restricts itself to a purely de­
even a World Champion in his fensive task. After 7 l .�xh5 �c6
comments does not spot the 7 2..�. g 6 Wd7 the black king is in
fortress. the square of the white h-pawn.
White does not have a better plan
than to push his passed pawn. After
7 3 .h5 ri;e7 74.h6 Black has to
watch out.

Position after 69 . ..l2Jd6xb5

Taimanov-Bronstein, Riga 1958

Here, Taimanov continued

7 0 . .ixfS , after which Black saved
himself with the manoeuvre He cannot play 74 ... Wf6 ? in view
7 0 . . . tt::l c3 + 7 1 . � b3 tt::ld s ! . After of 7 S . .ths ! , and Black is in
7 2 . .ixe4 tt::l f6 7 3 .j_g6 �c6 zugzwang. However, 74 . . . �f8 is
7 4 . �c4 �d6 7 S . �d3 �e6 sufficient, after which the black
76.�e2 c4 7 7 . �d2 �eS 7 8 . �c3 king calmly awaits the proceed­
t2:le4+ 7 9 . �xc4 tt::lxf2 8 0.�xh5 ings. The range of the knight is re­
t2:l e4 8 l .�xg4 tt::lx g3 8 2 .�d3 tt::l fl markable. On the one hand, it has a

187
T h e Art o f t h e Endgame

defensive function on the kingside 40. h2-h4

by protecting fS . On the other Euwe gives this move an exclama­
hand it keeps the white king at bay tion mark, and rightly so. The posi­
on the queenside. The passed pawn tion is indefensible for Black and
on cS is of crucial importance White won after a few more
here: it deprives the white king of moves.
the squares b4 and d4. This looks like an instructive ex­
There is no way for the white king ample of a bishop outside the
to penetrate. If it goes to b6, then pawn chain becoming tremen­
the black c-pawn marches on. dously strong. Nevertheless, the di­
A little further on in Hogeschool van agram position is not at all win­
het eindspel , Euwe once again ning for White. What's more, he
misses the concept of a fortress isn't even better here! To under­
which can be built on the �+tb stand this, we will return to the po­
cooperation. sition after White's breakthrough
on the queenside.
Diagram after Black's 36th move.

Position after 39.axb5

Krylov-Runza
Here Runza made a terrible mis­
Prague 1951
take by decentralizing his king.
Euwe investigates this endgame Instead of 3 9 . . . Wb 7 ?? he should
under the theme 'The king pene­ have preferred 3 9 . . . \t>d7 ! , since
trates' . Black doesn't have to fear the fur­
The game continued as follows: ther advance of the b-pawn. After
37. a2-a4 'it.td7·c8 4 0 . b 6 cxb 6 4 l . cxb 6 Wc8
38. b4·b5 a6xb5+ 42.\t>b4 �b7 43 . c;t>as tbcs Black
39. a4xb5 'it.tc8·b7 can look to the future with confi­
After this move Euwe remarks: dence.
'White has carried through the If White does not push his b-pawn,
breakthrough on the queenside, Black plays 40 . . . cxb 5 + and after
and Black has positioned himself 4 l .Wxb5 c6+ he has a comfortable
accordingly.' fortress.

1 88
 B u i l d i n g a F o r t ress

Fortresses can also be built with 3. ... 'ti'e6xb6

heavily reduced material. As an il­ Black could not take the queen, as
lustration I give a study where after 3 . . . -txg S 4 . .l::r b 8 + 'i:Vg8
White just barely manages to pre­ 5 .�eS + he would get mated.
vent the building of a fortress. 4. .if4-e5

Timman 2010
White to play and win
What can Black think of now?
Actually, this is a middlegame posi­ 4. ... l:!.c3-h3+1
tion where White must play for an Black also sacrifices a rook. The
attack. point of the sacrifice is that Black
1. �c1 -f4 tries to build a fortress.
Insufficient for the win was l . .tb2 5. g2xh3
on account of l . . .td4. White wins
. White could not capture with the
an exchange, but loses his attack­ king on account of a queen check
ing chances. After the text move on e6. Now Black is going to try to
White threatens to administer a exploit the doubling of White's
devastating check on eS . pawns on the h-file.
1. ... �f2-d4 5. ... 'ti'b6-d61
Controlling the long diagonal. White
has to fmd new attacking ideas.
2. h5-h6
Threatening 3 . 'if d8 + 'if g 8
4.'i:Vxd4+ ! cxd4 S .�eS + , winning.
2. ... .id4-f6
The only defence. At first sight it is
unclear how White can breathe
new life into his attack.
3. l:!.b5xb61
A rook sacrifice which serves to
give the bishop control of the This was the plan. The white
square eS . bishop is pinned, and White does

189
T h e Art o f t h e E n d g a m e

not seem to have anything better 1 . e4-e5+1

than 6 .'ihf6 + . However, that leads White sacrifices a pawn in order to
to an ending he cannot win, due to drive the black king to a worse
the presence of the pawn on h6. square.
White cannot make progress be­ 1. ... �f6xe5
cause there are constantly stale­ 2. '>t>e1 -f2 �g2-e4
mates in the position. So he will 3. �f2-g3 �e5-f5
have to find a method to breach the Keeping the white king from f4.
fortress. 4. '>t>g3-h4 �f5-g6
6. 'i?Vg5-g7+11 5. a6-a7
White sacrifices his foremost
h-pawn, after which the endgame
is winning after all.
6. �f6xg7
7. h6xg7+ '>t>h8-g8
8. �e5xd6
And wins.

If, in the final position, Black had a

far advanced a-pawn instead of his
worthless c-pawns, he would have
been able to build a fortress after
all. This is shown in the following Mutual zugzwang. If White were
study. to move, he would be driven
back. For a good understanding
of the position, you have to real­
ize that White would have a for­
tress without the e-pawns. Then
he would move his king to and
fro on g3 and h3 , from where it
cannot be driven away. If Black
puts his king on gS and pushes
his pawn to h4, then White pro­
motes, with stalemate as a result.
With a pawn on e2 or e3 there is
Timman 2011 no stalemate and, consequently,
White to play and draw
no fortress.
In principle, the white pawns are 5. ... i.e4-b7
not strong enough to hope for a 6. e3-e4!
draw. For example, after 1 . 'itf2 Sacrificing the second e-pawn.
�xe4 2 .<it>g3 'itgS Black would be 6. ... i.b7xe4
winning. 7. e2-e3

1 90
 Building a F o rtress

And again, mutual zugzwang has 2. ... �f5-e41

arisen. White will also sacrifice his The point of the previous move.
t hird e-pawn, after which he has Obviously Black does not take the
completed his fortress. bishop, as then White will ex­
change on fS , after which there is
We have seen that there is no win if no winning potential left. Now
White has the queen's bishop, and Black has the nasty double threat of
if there is a white pawn on h6 and 3 . . . We5 and 3 . . . h3 .
a black one on h7. In that case 3. tUh6-f71
White cannot achieve anything, A deflecting sacrifice. Since 3 . . . h3
even with a king's bishop. And if fails to 4.tbg 5 + '.t>xeS S .tbxe4,
we move the pawns one file to the Black has to take the knight.
left, the result remains the same: 3. ... '1t>e6xf7
neither of the two bishops can 4. 'i.th1 -h2
hreach the fortress. Threatening S .g3 with an immedi­
ate draw.
4. ... jle4-f5

Timman 1989
White to play and draw

Material is equal, but due to the

double attack White cannot avoid The situation looks awkward for
losing material. What he has to do White. Black threatens to support
is give the piece in the most fa­ the h-pawn with his king, after
vourable circumstances. which the win will only be a mat­
1. tUf2-g4 .ih7-f5 ter of time.
Otherwise White will easily hold 5. jle5-g311
the draw, as both 1 . . . rJo>fS and An extraordinary way to attack the
l . ..hS are met by 2.tbf6, attacking black h-pawn.
the queen's bishop. 5. ... h4xg3+
2. tUg4xh6 6. ..t?h2-h1
Not 2..�.h2 in view of 2 . . . �xg4 An incredible situation. Black is a
3.�g l �xg l 4.rJo>xg l �fS , and full bishop pair ahead, but he can­
Black will stop the a-pawn. not win. To avoid stalemate he will

191
The A r t o f t h e E n d g a m e

have to sacrifice one of his bishops. 3. Wc1 xd 1

After 6 . . . �h3 7 .gxh3 �g6 8 .'it>g2 4. tt:lg4-e5+
he cannot maintain his g-pawn,
while after 6 . . .�g 1 White takes the
king 's bishop and obtains a
fortress.

In Chapter 8 I gave a study with

which I won second prize in the
Nona 2010 tourney.
First prize in this tourney went to
the following study:

This discovered check makes

things difficult for Black. He has no
direct win, since 4 . . . 'it>e 1 5 . tLld3 +
'Ot>d2 6.tLlxb2 c2 is met by 7 .tLld3 ,
and Black cannot promote. His
next king moves are intended to
rule out this trick.
4. 'lt>d1 -c2
5. ..ih5-g6+ 'it>c2-b3
6. ..ig6-f7+ �b3-a4
Zakharov 2010 7. ..if7-e8+
White to play and draw

If the black c-pawn advances fur­

ther, it will soon be over. With tac­
tical tricks White remains on his
feet for a while.
1. llc5-e5+
Not I...t g6 on account of l .tLle3 !
2 . tLle3 'iiV f2 + , followed by
3 . . . 'iY e3 + , when White is helpless.
1. ... 'it>e1 -d2
2. l:.e5-d5+ 'it>d2-c1
The black king has walked out of It looks as if the black king cannot
the checks. But White has another escape. However, he still has one
trump card up his sleeve. little move to keep the struggle go­
3. �d5xd 1 + ing.
With this exchange sacrifice White 7. ... c7-c6!
realizes optimum cooperation be­ This pawn sacrifice wrongfoots the
tween his bishop and knight. white bishop.

1 92
 Building a Fortress

8. �e8xc6+ �a4-b3
9. �c6-d5+ 'it'b3-c2
1 0. �d5-e4+ wc2-d1
11. �e4-f3+ �d1 -e 1
Now he can. White cannot prevent
t h e c-pawn's promotion after he
has captured the black queen.
1 2. ttJe5-d3+ 'it'e 1 -d 2
1 3. ttJd3xb2 c3-c2

Certainly not 1 8 . Wh3 ?, since after

1 8 . . . �xaS there is no fortress.
1 8. ... 'tWc7-a7+
1 9. �2-g3
Again, White has to choose the
square for his king with care. If
Black now takes the pawn, White
brings his king to h 2.
1 9. ... li'a7-g1
A last try. Without the a-pawn this
position would be lost for White.
From here on, White's play is fo­ 20. a5-a6 �b4-b5
cused on building a fortress. 21 . a6-a7!
1 4. ltJb2-c4+ �d2-c3 Forcing the queen to leave her post.
Not 1 4 ... <Jold3 in view of 1 S .�e4+ ! 21 . 'it'g1 xa7
c;i;>xe4 1 6 .lt:'Jd 2 + , followed by 22. �g3-h2 '12-'12
1 7 .lL'lb3 , after which the c-pawn is
stopped after all. This study inspired me to make an
1 5. ttJc4-d6 adaptation where White has to pre­
On its way to the central square e4. vent the building of a fortress.
1 5. ... c2-c1 'tW
1 6. ttJd6-e4+ wc3-b4
1 7. �f3-g2
Protecting the knight, and aiming
for a type of drawn position which
was discovered by Max Karstedt in
1 90 3 .
1 7. ... "iic 1 -c7+
Preventing 1 8.'ibh2. If the white
king reaches that square, then
Karstedt's fortress is realized. Timman 2010
1 8. wg3-f21 White to play and win

1 93
The Art of the Endgame

Black has a great material advan­ 4 . . . Wh6! ! The point of this king
tage, but he will have to give his move is that Black wants to capture
rook for the e-pawn. And then the white bishop with the knight.
White still has a strong passed It is striking how little weight the
a-pawn. white bishop carries.
1. e6-e7 �f4-f8 S .aS tLle2 6 .a6 . Otherwise the
Now the trouble starts. After knight can stop the pawn.
2.exf8Wk' �xf8 3 .a5 �h6 the black 6 . . . ti:Jxf4+ 7 ::Ji;c6 lLJxg6 , and again
bishop stops the a-pawn. Black is in time to build Karstedt's
2. g4-g51 fortress.
White sacrifices the g-pawn in or­ The intention of the text move is to
der to block the c l -h6 diagonal for force the black king to go to an­
the bishop. other square than h6. Curiously,
2. ... �h4xg5 first the black king was forced to
3. f3-f4+1 block the bishop's diagonal, and
In order to understand this second now the bishop blocks a square for
pawn sacrifice it is necessary to the king.
analyse what would happen if 3. �g5xg6
White took on f8. After 3 .exf8'fi 4. e7xf8'iW i.h6xf8
itxf8 White has the following pos­ 5. a4-a5
sibilities: Again the a-pawn has a free pas­
A) 4.a5 ti:Jxf3 S .a6 �xg6 6.a7 sage. But Black can try to build
i!.g7 7 .a8'li' �h7 ! and White can­ Karstedt's fortress.
not prevent the knight reaching the 5. ... tLlg1 -e2
crucial square eS , when Black has
built the Karstedt fortress;
B) With the pawn check 4.f4+
White tries to disturb the coordi­
nation in the black camp. But now
comes a tremendous surprise.

6. f4-f5+!
The third pawn sacrifice, which is
intended to drive the black king to
a worse square.
6. �g6xf5
7. a5-a6 tLle2-f4+
8. �d5-c6

1 94
 Bui lding a F o r t r ess

And, wins. The a-pawn will reach It seems impossible to stop the
the promotion square unhindered, white passed pawns. However,
whereas Black just misses the deci­ Black has a clever defence.
sive tempo that is needed for 3. . .. �h7-g8!
building the fortress. On 4 . e 7 Black has planned
4 .. Jid 1 +. Therefore White must
With an extra pawn there are more close off the d-file for the rook
chances for bishop + knight versus with a combination.
queen. At the end of the following
study, White must choose the right
square for his king in order to pre­
vent the construction of a fortress.

4. �c5-b6+11
A second bishop sacrifice. Black
has to accept.
4. ... �a5xb6
Timman 2011 5. tt:Jf4-d5+
White to play and win
The point of the previous move.
White has two powerful passed The knight check closes the d-file
pawns, but Black is not without for the black rook, after which the
counterplay. It is vital for White to e-pawn can march on. But that still
try and promote his pawn. doesn't decide the argument.
1. �d3-h71 5. ... '>t>b6-c5
A bishop sacrifice, in order to be The best square for the king.
able to march on with the f-pawn. 6. e6-e7
1. ... �g8xh7 Nadareishvili's field of tension has
An important alternative is l ... �c4. appeared on the board, only now
Now White has to play very accu­ with a black rook instead of a
rately in order to secure the win: queen.
2 .e6 I:td l + 3 .�h2 I:td2+ 4.Wg3 6. ... �g8xf7
tLle2+ 5 . Wf3 ! tLlxf4 6.e7 �[7 The only defence. Black has to al­
7.Wxf4 l:re2 8 .�f5 ! and the white low promotion, and tries to get as
passed pawns decide the battle. much material for the queen as
2. f6-f7 J:td2-d8 possible.
3. e5-e6 7. e7xd8'if i.f7xd5+

195
T h e Art o f t h e E n d g a m e

move leads to mate in 64 moves,

whereas with the strongest queen
check, 9 .�e7 + , it would take 6
more moves. It is important to
bring the white king to the battle
scene as soon as possible.
9. ... �d5-c4
Now White has to pay attention.
After 1 0 .'iYe 7 + �b5 1 l .Wf4,
Black would be able to build a for­
Which square should White tress after all with 1 1 . . . tt:la5 ! , fol­
choose for the king? lowed by 1 2 . . . tt:lc6 .
8. 'ii;> h 1 -h2! 1 0. 'ii;> g 3-f4
The alternative 8 .Wg 1 was insuffi­ Only move. Black has no time to
cient for the win, since Black re­ bring his knight to c6. The white
plies 8 . . . tt:le2+ and then brings his king will penetrate, causing death
knight to c3. Even then the draw and destruction.
requires a delicate touch, as after
9 .Wf2 tt:lc3 the fortress is not en­ In 1 9 84 it was established that two
tirely realized yet. White can try bishops win against a knight. The
the following moves: situation where Black has an addi­
A) 1 0 .We3 tt:lb5 ! 1 l .'it'a5 .tb3 , tional pawn is interesting. Some­
and the white king is held off; times he can use this to realize pro­
B) 1 0.'it'a5+ Wd4! and Black has motion, if the white king is far off
taken up a new fortress. Here also, It is also possible to use the pawn
the white king is kept at a distance. to build a fortress.
Insufficient was 1 0 .. .'i.illc 4? in view
of 1 1 . We3 ! , after which there is no
fortress.
8. ... ll:lc1 -b3
Under these circumstances the
knight could not reach the square
c3 , since 8 . . . tt:la2 fails to 9 .'i¥a5 + ,
winning the a-pawn. That is why
Black tries to build a different for­
tress.
9. 'ii;> h 2-g3
Not the only move; six queen Timman 2010
White to play and win
checks would also eventually
achieve the aim. Yet, we cannot Black threatens to bring his king to
speak of real duals here. The b 7 , after which he would build an
tablebases indicate that the text impregnable fortress. White has to

1 96
 B u i l d ing a F o r t ress

play very precisely in order to pre­ 4. ... 'it>c4-b5

vent this plan. The tablebases indi­
cate that it takes 99 moves to give
mate.
1. �a1 -d4
Driving the knight forward.
1. ... lba7-b5
2. �d4-c5!
The only way. To L�f6 Black
would react with 2 . . . c5 ! . Then the
hlack c-pawn cannot well be
stopped, as the white king is too
fiu away. 5. �f8-g71
2. ... Wb3-c4 Preventing the knight reaching the
3. �c5-f81 b5 -square.
Not 3 .�e7 ? on account of 3 . . . tLlc3 ! 5. ... lbd4-e6
and the knight will reach the cen­ Threatening to give check on f4.
tral square d5 with tempo. 6. �g7-e5 Wb5-b6
3. ... lbb5-d4 The black king has almost reached
Vacating the b5-square for the the safe haven. The white bishop
king. It is important to understand pair has to frantically continue ma­
what Black's fortress is going to noeuvring in order to prevent the
look like exactly. When he has construction of a fortress.
brought his king to b7, his knight 7. i.h5-g4
should be either on a7 or b5 . Then The king's bishop also comes into
White cannot make progress, even action.
if his king is close by on d7. The 7. ... lbe6-c7
black king can oscillate between
b6, b7 and b8 , and thus will not
end up in zugzwang. So after the
text move Black threatens 4 . . . �b5 ,
followed by 5 . . . �b6 and 6 . . . tt:Jb5 ,
after which he will be safe. After
the alternative 3 . . . d4, both 4.�g8
and 4. �g6 are winning, as White
can meet 4 . . . Wb5 with 5 .�g 7 ! .
Only not 4.�g7 ? , since then White
cannot prevent the construction of
a fortress after 4 . . . Wb5 . 8. �g4-c8!
4. 'it>h7-g6 Keeping the black king from b7.
The moment has come to bring the 8. ... lbc7-b5
king one step closer. 9. �e5-g3

1 97
T h e Art o f t h e E n d g a m e

9 .�h2 was also possible. However, Threatening 1 3 . . . tbc7, when the

9 .�f4 was insufficient for the win. king will finally reach the square
The reason will not become appar­ b 7 . The white king is just close
ent until much later. enough to frustrate this plan.
9. ... tt::l b 5-a7 1 3. jLf2-g3+ tt::l b 5-c7
Black does all he can to drive the 1 4. 'it>f6-e7 'it>b8-a7
king's bishop from its dominant 15. �a6-c81
position.
1 0. jLg3-f2+ �b6-c7
11. jLc8-a61
Still keeping the black king from b7.
11. ... tt::la 7-b5
1 2. ..t>g6-f6
Finally White has time to bring his
king closer. 1 2 .�£7 was also good.
1 2. ... Wc7-b8

Now it becomes clear why 9 .�f4

was not good. In that case Black
would have a check on dS here.
After the text move Black runs out
of options, since after 1 5 . . .'�b8
White can protect the king's
bishop with his king.

1 98
Chapter 11

Systematic Manoeuvres
One of the characteristics of a sys­ 3. c4-c5!
tematic manoeuvre is the lack of Mutual zugzwang as a precursor of
captures. the systematic manoeuvre. Black
The pieces on both sides play has to move his rook.
moves that adhere to a certain pat­ 3. ... �f8-f7
tern, which is repeated. Only at 4. 'ifg8-g7!
moments where the manoeuvres And the queen keeps pursuing it.
are interrupted, in between events,
a pawn can be snatched away.
One possible aim of a systematic
manoeuvre can be stalemate.

4. ... .l:f.f7-f6
5. 1i'g7-g61
The pattern becomes clear.
Kotov/Mitrofanov 1976 5. J:[f6-f5
White to play and draw
6. 'ii'g 6-g5! .l:f.f5-f4
Black threatens mate in one, which 7. 'ii'g 5-g4! .l:f.f4xg4
is why White does not have time to Black resigns himself to the inevi­
take the rook on c8. table: stalemate.
1. .t:[g8-f8+ .ig1 -f2 For this study Kotov and
Renewing the mate threat, but now Mitrofanov received first prize in
White can promote his pawn. the 'Leninska Smena' tourney.
2. g 7-g8'i¥ .Uc8xf8
The situation remains critical for Leopold Mitrofanov was a brilliant
White, as mate is threatened from Russian endgame study composer.
both sides. If White removes the The systematic manoeuvre was one
queen, a deadly check on h8 fol­ of his many specialties. Alexander
lows. Kotov - not to be confused with

1 99
T h e Art of the E n d g a m e

the grandmaster of the same name There are no more squares. But
- specializes in studies containing White has another string to his bow.
systematic manoeuvres. 7. b6-b7+!
With the white rook on the 5th
In the following study, the rook is rank, this pawn sacrifice is decisive.
successful in the fight against the Black has to take with the queen.
queen. 7. ... 'ii'b 5xb7+
8. �c8-d8
The pawn on c7 is much stronger
than the bishop in Mitrofanov's
study. Black has to pin the pawn.
8. ... 'it'b7-b6
9. �d8-d7!
White doesn't have to bother about
the rook.
9. ... 'f!fb6xc5
1 0. c7-c8'li'+ 'ii'c5xc8+
11. �d7xc8
Kotov, 'Vecherny Peterburg' 1995 And White wins the pawn ending.
White to play and win

The pawn promotion cannot stop In the two studies I have presented,
White; he's going to play for mate. the systematic manoeuvres oc­
1. b5-b6 b2-b1 'fr' curred on lines. Major pieces can
Defending against the mate threat. also pursue each other along adja­
2. l:!.f1 xc1 f5-f4 cent diagonals, but only queens
And this time it is Black who plays can do this. The studies with this
for stalemate. Rook and queen are theme are spectacular.
dribbling upwards together.
3. l:!.c1 -c2 'ifb1 -b2
4. l:tc2-c3 'ii' b 2-b3
5. l:l.c3-c4 'ii' b 3-b4
6. l:!.c4-c5 'it'b4-b5

BondarenkojKakovin , 'Shakhmaty' 1954

White to play and draw

White has to force the pawn pro­

motion as quickly as possible, but

200
 S ys t em a t i c M a n o e u vres

that does not put an end to his In this situation, White cannot cap­
problems. ture the queen on account of stale­
1. .l:!.c1 ·c8 ltJg6·f4 mate.
Threatening mate. 1. kd8·f6+
2. b7·b8'ilf 'ir'e8xc8 Threatening mate on b8. Black has
3. 'i!Vb8-c7! to screen off the b-file.
Keeping the knight pinned, and 1. ... �a1 ·b2!
preventing the passage of the black The start of a fascinating dance of
queen to c l . the queens. Black is threatening
3. ... �c8·d7 check on g2.
4. �c7·d6! 2. �b1 ·c2!
The queen screens off the second
rank.
2. ... �b2·c3!
Black has to follow suit, and so the
ritual is continued.
3. 'ikc2-d3!

The pattern repeats itself

4. ... 'ikd7·e6
5. �d6-e51 'ike6xe5
Stalemate.

Kotov used this study as a basis for

adaptations with a record number
of queen sacrifices on both sides. While the black queen is screening
off the files, the white queen is do­
ing the same with the ranks.
3. ... �c3·d4!
4. 'ir'd3·e4! �d4·e5
5. 'ir'e4·a8+
And mate.

In a follow-up study, Kotov in­

creased the number of queen sacri­
fices. Here also, the kings are con­
Kotov, 'Shakhmaty v SSSR' 1977 demned to hold each other in
White to play and win check.

201
T h e Art o f the E n d g am e

Preventing the check on a?. How­

ever, White has another way out.
7. 'fifd4xb4+ �c3xb4
8. J:lb2xa2+
The final sacrifice.
8. ... b3xa2
Stalemate.

Later that year the Russian end­

game study composer Leonard
Kotov, 'Shakhmaty v SSSR' 1977 Katsnelson made a new adaptation.
White to play and draw

In order to prevent direct mate,

White has to promote his pawn.
1. h7-h8'fif h2-h1 �!
Black also promotes. I . . . .iit.x h8
would have been met by 2 .'ii'xb3 +
with stalemate.
2. ..Wd1 xh1 <1Jg3xh1
3. 'ii' h 8-g71
The systematic manoeuvres start.
3. 'ifg8-f71
Katsnelson 1977
4. -.g7-f61 -.f7-e61
White to play and draw
5. 'it'f6-e51 'it'e6-d51
6. 'ife5-d41 No immediate mate is threatened,
so White has a little time to give
material; in this study he will have
to play for stalemate, too.
1. b7-b8..W
Threatening to give an annoying
check on the a-file, which is why
Black has to take the queen.
1. ... �d8xb8
2. f7-f8� 'ff b8xf8
3. h7-h8'it'
This promoted queen cannot be
captured. Since no check is threat­
Herein lies the difference with the ened on the a-file, Black has time
previous study. As the h 1 -square is to bring his bishop into position.
blocked by the knight, Black does 3. �e1 -c3
not have a winning check. 4. 'iVh8-f61 'fff 8-d61
6. ... �d5-c5 5. 'iVf6-d41

202
 S y s te m a t i c Manoeuvres

In 1 9 8 0 , Katsnelson published an
article in EG entitled 'Both sides
sacrifice queens - development of
an idea'. After presenting the stud­
ies by Kotov and himself he
concluded:
'A general shortcoming can be
seen in static initial positions
where the Royal prisons are al­
ready prepared and the introduc­
White seems to be safe, but now tory play is forced.'
the systematic manoeuvres are This is indeed a blot that is not easy
shifted to another diagonal. to avoid. I succeeded in construct­
5. ... 'iWd6-a6! ing a version where both kings
Suddenly threatening to give a dev­ move, although this is restricted to
astating check on fl . White can a single move.
only just save himself.
6. 'ifd4-d31
Pinning the bishop.
6. ... 'iWa6-b5!
The queen sacrifices continue.
7. 'ifd3-e21

Timman 2011
White to play and win

By moving the construction one

file to the left, I have created some
space for a modest introductory
play. The shift has yet another
Protecting the knight. The white point, which will be revealed later
queen always has two squares on.
available on the fl -a6 diagonal. 1. g6-g7+ �h8-g8
7. ... 'iWb5-c4 2. �b1 -a2+
8. 'itfe2-d3 V2- 'l2 Forcing the rook to step into a pin.
For this study Katsnelson received 2. �f8-f7
second prize in the 'All Union Team 3. <;t>h6-g6
Championship' . It is hard to imag­ On the kingside everything is
ine a better study on this theme. ready for systematic manoeuvring.

203
T h e Art of t h e E n d g a m e

3. ... l:ld6-e6 the construction has been shifted

one file to the left, White has no
immediate mate. In Kotov's second
study this mate wasn't available ei­
ther, as the corner square was oc­
cupied by a knight. In the dia­
grammed position, two further
queen sacrifices are necessary to
give mate with the bishop in the
end.
9. 'tlt'd4-e4! �d5-a2
1 0. 'tlt'e4-a8+1
The final queen sacrifice.
Black puts his second rook on the 1 0. ... �a2xa8
a2-g8 diagonal as well. After the al­ 11. ke6xf7#
ternative 3 . . :ds White would
.

quickly give mate by 4.�xd5 Earlier on, I had already made a

�xfS + S .gxfS 'il"a2 6.'iih l ! . There version where White first makes a
are no stalemate traps. double rook sacrifice.
4. �a2xe6 �c2xf5+
Black also has to sacrifice his
bishop in order to play for stale-
mate.
5. g4xf5 �a6-a21
The systematic manoeuvres start.
6. 'it'a1 ·b21 �a2-b31
7. 'it'b2-c3! �b3-c4!
8. 'tlt'c3·d4! 1i'c4·d5!

Timman, 'New in Chess' 2008

White to play and win

For the time being it is mainly the

white king that is in danger. Mate
in three is threatened, starting with
1 . . Jhg 7 + . A nice detail is that
both queens are hanging, but this
does not play a role during the first
moves of the study. The battle
scene is elsewhere.
Here the difference with Kotov's 1. J:!.c7-c8+1 kd7xc8
first study becomes clear. Because 2. ttJc5-e6

2 04
 S y s t e m a t i c Manoeuvres

Protecting the g-pawn. The black king is caught in a mat­

2. ... �c8xe6 ing net. To close the net, White
3. .l:!.c4-c8+! must first move aside with his king.
The second rook sacrifice serves to 1. �b3-a3 �h 1 -c1
open up the a2-g8 diagonal for the The only defence against the
king's bishop. threats of 2 .b3 /b4 mate.
3. ... �e6xc8 2. �c3-d41
4. �b1 -a2 Preventing a side attack by the
Finally Black's threats have been rook. After 2 ..�.. eS ? l::i. c 5 there
neutralized. would be no win.
4. ... �c8xf5+ 2. ... J:tc1 -d1
The only chance. Black is playing 3. �d4-e51
for stalemate. The same principle.
5. g4xf5 e7-e6 3. l:[d1-e1
6. �a2xe6 'ti'a5-a21 4. �e5-f61 l:[e1 -f1
And now the queen dance starts.
White wins in the same way as be­
fore.

So far I have only given studies

where the systematic manoeuvres
were carried out by major pieces.
The fight between rook and bishop
is also interesting. In the following
study the white bishop is strutting
step-by-step up the long diagonal
to escape from the attacks by the 5. �f6-g71
enemy rook. Patiently, White keeps moving the
bishop one square at a time.
5. ... l:[f1 -g1
6. �g7-h8
The corner square is reached, and
since the black knight obstructs the
h-file, no further rook attacks are
possible.
6. ... l:[g 1 -g3+
7. b2-b3+
With mate next move.

Liburkin, '64' 1940 The duel between rook and bishop

White to play and win can also revolve around the theme
of line control.

205
The A r t of t h e E n d g a m e

5. �e4xh1 J:!:a8-e8
The start of the systematic ma­
noeuvres. Black threatens to give
check on e l .
6. �f1 -e2
Screening off the e-file.
6. ... J:!:e8-d8
A new danger: Black threatens to
give mate in one.
7. �e2-d3 J:!:d8-e8
Korolkov, 'Shakhmaty v SSSR' 1934
White to play and win

The first move is obvious.

1. 'ifa7-b7
Stopping the dangerous h-pawn
and making way for the a-pawn.
1. ... e5-e4
A pawn sacrifice to open the e-file.
After 1 . . . .!:l'.d8 L �.d3 , Black would
have no chance.
2. 'ifb7xe4 J:!:c8-e8
Play becomes sharp.
3. a6-a71 8. �d3-e41
Both sides strive for promotion. I n zigzag fashion the bishop finds
its way upwards, screening off the
d- and the e-file in turn.
8. .I:Xe8-d8
9. �e4-d51 J:!:d8-e8
1 0. �d5-e6! .I:Xe8-d8
11. �e6-d7
Completing the ladder manoeuvre.
The black rook is now dominated.
Vladimir Korolkov was one of the
great Russian endgame study com­
posers. For this study he received
shared first/ second prize in the
3. ... h2-h1 �! Shakhmaty v SSSR tourney.
4. a7-a8'if+l
The appropriate reply. Now both One year later Kasparyan made an
promoted queens will be elimi­ adaptation of this theme, only, in
nated. this study, the bishop zigzags
4. ... .I:Xe8xa8 downwards.

206
 S y s t e m a t i c Manoeu vres

A knight can dominate an enemy

rook in similar fashion.

Kasparyan, 'Shakhmaty v SSSR' 1935

White to play and win

Both kings are imprisoned. White

Gorgiev, 'Schackvarlden ' 1937
must prevent Black from getting a
White to play and win
rampant rook.
1. i.. d 8-e7 .!:t.e8-d8 In Gorgiev's book Izbrannye etiudy
Now danger looms on d 1 . this study can be found in a chapter
2. �e7-d6 .!:t.d8-e8 called 'Geometric manoeuvres' .
3. i.. d 6-e5 l:t.e8-d8 This is a nice alternative term.
4. i.e5-d4 .!:t.d8-e8 White has an enormous material
5. i.. d 4-e3 .U.e8-d8 advantage, but as both kings are
6. i.e3-d2 .U.d8-e8 boxed in, the motif of the rampant
rook plays a role.
1. tLlf5-g3
The h-pawn has to be stopped.
1. ... h2-h1 'if+!
The only chance. After 1 . . . .!:th8
2 . tLlh 1 the black rook would not
have a single file at its disposal.
2. tLlg3xh1 .!:t.d8-h8
Now, if the knight budges, Black
sacrifices his rook on h 1 .
3. 'ifa3-f8+!!
White sacrifices his queen in order
7. i.d1 -e21 to win a tempo.
Other moves don't suffice. 3. ... .!:t.h8xf8
7. f3xe2 4. tLlh1 -f2
8. i.d2-e1 .U.e8-h8 This is the start of the systematic
9. tLla2-c1 l:!.h8xh7 manoeuvres. The knight will
1 0. tLlc1 xe2 .!:t.h 7-e 7 screen off the f- and the h-file in
11. tLle2-g1 # turn.

207
The Art o f the Endgame

M itrofanov/Korolkov/Dolgov 1983
White to play and win
4. ... .l:!.f8-h8
5. tLlf2-h3 J:.h8-f8 Korolkov and Mitrofanov had
6. tLlh3-f4 J:.f8-h8 worked on this theme before. This
7. tLlf4-h5 J::t h8-f8 is the final version.
1. J:.a2-c2 tLlc1 -d3
2. J:.b4-d4 tLld3-e1
The scheme becomes clear. On
each move, the knight attacks one
of the rooks, who are at a knight
jump's distance from each other.
3. J:.c2-e2 tLle1 ·f3
4. J:.d4-f4 tLlf3-g1
5. J:.e2-g2 tL\g1 -h3
6. J:.f4-h4

8. tLlh5-f6!
Now, 8 . . . gxf6 will be met by 9.g7
l:1g8 I O.�b l , after which White
calmly brings his king into play.
The black rook does not have
enough freedom of movement.
8. ... J::tf 8-h8
9. d7-d8'tW+ .!:!.h8xd8
1 0. tLlf6-d7+ 1 -0
The end of the first phase. The
Almost half a century later, at the black knight has to give check, and
age of 7 5 , Korolkov worked on a then it will harass the white rooks
very long study, where two rooks from the other side.
eventually manage to triumph over 6. ... tLlh3-g5+
an enemy knight. 7. wh7·h6!

208
 S y s tema t i c Manoeu vres

The only time that the king attacks The end of the third phase. Black
the klli ght. must give his g-pawn to gain con­
7. ... tt:Jg5·f3 trol of the square e 1 for the knight.
8. .!:.h4·f4 tt:Jf3·e1 And then the knight will travel all
9. .!:.g2·e2 tt:Je1 ·d3 the way back to the queenside for
1 0. l:if4·d4 tt:Jd3·c1 the last time.
11. .!:.e2·c2 1 7. ... g3·g2
1 8. J:te2xg2 tt:Jf3·e1
1 9. l:rg2·e2 tt:Je1 ·d3
20. J::i:f4·d4 tt:Jd3·c1
21 . .!:.e2·c2 tt:Jc1 ·b3
22. .l:!:d4·b4
And wins.
This study was awarded the special
prize of the 'Friendship Tourney'.

So far I have exclusively dealt with

studies where both kings largely
11. ... tt:Jc1 xb3 refused to take part in the battle.
The first capture. A few skirmishes Only the last study, the long one,
follow, and then the third phase contained a king move. In the oth­
starts: the knight is driven towards ers, the kings were on squares
the kingside again. where they could be mated or
1 2. J:td4·b4 a3·a2 stalemated.
With this pawn sacrifice Black Play becomes more lively when
again safeguards the c 1 -square for one of the kings is involved in the
the knight. systematic manoeuvres.
1 3. J:tc2xa2 tt:Jb3·c1
1 4. J:ta2·c2 tt:Jc1 ·d3
1 5. J:tb4-d4 tt:Jd3-e1
1 6. J:tc2-e2 tt:Je1 ·f3
1 7. .tid4·f4

Gurgenidze, 'Shakhmaty v SSSR 1974'

White to play and win

In his book Chess Curiosities Tim

Krabbe categorizes this study un-

209
The Art o f the E n d g a m e

der the heading ' Prokes Ma­ 8. ... J:.a2-b2+

noeuvre' . He had good reasons for 9. 'it>b1 -c1 l:lb2-c2+
this, but it seems to me that this is 1 0. wc1 -d1 J:tc2-d2+
also a striking example of system­ 11. 'it>d1 -e1 l:rd2-e2+
atic manoeuvring. 1 2. we1 -f1 l:.e2-f2+
1. h6xg7 13. 'it>f1 -g1
Now Black has to try to sacrifice
his rook on h8. This can only be
done by first offering the other
rook as a gift.
1. ... J:rf2-g2+
2. Wg1 -f1 I
White does not accept the rook
sacrifice. Only at the end of the
study will we see why.
2. ... l:.g2-f2+
3. 'it>f1 -e1
The king continues its march to the
queenside. The same position is reached as af­
3. ... l:l.f2-e2+ ter White's first move, but now
4. we1 -d1 r:te2-d2+ without the white a-pawn.
5. 'it>d1 -c1 J:.d2-c2+ 1 3. ... .!:f2-g2+
6. 'it>c1 -b1 ttc2-b2+ 1 4. �g 1 xg2!
7. 'it>b1-a1 Now he can.
1 4. ... .!:h8xh2+
Finally Black can sacrifice his
' Prokes rook' .
1 5. �g2xh2 i.g8xf7
1 6. g6xf7 'it>e7xf7
1 7. ti:lb8xd7
The tension rises. Can the white
knight stop the enemy passed
pawns?
1 7. ... b4-b3
In the event of 1 7 . . a3 , White
.

blocks the pawns by 1 8 .lt:JcS a2

The corner square is reached. Black 1 9.tt:lb3 .
has to take the a-pawn. 1 8. ti:ld7-e5+ 'it>f7xg7
7. ... J:tb2xa2+ 1 9. ti:le5-c4
8. �a 1 -b1 Only now does the sense of
And now - back to the kingside White's long king manoeuvre be­
again. come dear.

210
 S y s te m a t i c Mano euvres

2. 'it>c5·c4 .l:.b5-b4+
3. 'it>c4-c3!
The time still isn't right to capture
the rook.
3. ... J:tb4-b3+
4. c;t>c3·c2! .l:.b3-b2+
5. <;i;>c2-c1 nb2·b1 +
6. �c1 xb1
Only now.
6. ... J:ra8xb8+
If the white pawn were still on a2,
then Black would capture on a2
here and win. By sacrificing the
pawn, White has ultimately suc­
ceeded in blocking the black
passed pawns. Amazingly, White
would even have lost if he had
played 2 .' �xg2?.
1 9. ... b3·b2
20. ti:lc4·a3 1 ·0

In the following study, the white 7. 'it>b1 ·a1 !

king eventually captures the rook The king has to go to the corner
for another reason. square. The reason for this will be­
come clear later.
7. ... 'it>e8-d8
8. d5·d6!
Thus White wins a rook. But that
does not mean the game is over.
8. _ e7xd6
9. .l:.h1 -h8+ 'it>d8-e7
1 0. J:rh8xb8 g4·g3

Timman 2011
White to play and win

Black is threatening to castle

queenside.
1. �h2-b8!!
The only way to prevent this.
1. ... .l:.b7-b5+

211
The A r t o f the E n d g a m e

The black passed pawns are dan­ The battle between king and rook
gerous. White has to play ex­ can also take place from a distance,
tremely accurately in order to safe­ as we can see in the following
guard the win. study.
11. d4-d5!
Creating a dangerous passed
pawn for himself. Now it be­
comes clear why the white king
has to be on a 1 . If White had
played 7 . Wa2 , there would have
followed 1 1 . . . g 2 1 2. .:f.b 1 tt::l b S
here. Because of the possibility of
a check on c3 Black would have
sufficient counterplay.
11. ... tba7-b5
1 2. l:rb8-b7+ We7-f8 Dolgov, 'Shakhmaty v SSSR' 1966
White to play and win
After 1 2 . . . �f6 1 3 .d4! the black
king would land up in a mating The pivotal question in this minia­
net. ture is whether the black rook will
1 3. l:tb7xb5! be able to stop the g-pawn.
The a-pawn will decide the issue. 1. g6-g7 l:a1 -b1 +
13. _ a6xb5 2. 'it>b3-a4!
14. 'it>a1 -a2 The king has to move up. After
Obviously, White could also push 2.<iii>c 2 l:rb2 + ! 3 . �d3 J:[b3 + the
the a-pawn first. black rook will reach the
14. ... g3-g2 g3-square.
1 5. a5-a6 g2-g1 'ii' 2. ... l:r b 1 -a 1 +
1 6. a6-a7 3. �a4-b5 l:ra1 -b 1 +
4. wb5-a6
The king strolls on in zigzag fash­
ion. For the time being, the pur­
pose of this manoeuvre is totally
unclear.
4. ... l:rb1-a1 +
5. wa6-b7 l:ra 1 -b1 +
6. �b7-a7
Suddenly the king moves to a dark
square. Now, on 6 . . . .:ra l + 7 .�a6+
wins, so the black bishop has to
give check.
The promotion cannot be pre­ 6. �h2-g 1 +
vented. White wins. 7. '>t>a7-a8!

212
 S y s t em a t i c Manoeuvres

The corner square is reached. Now The black bishop blocks the
White's plan will become clear g3-square.
soon. 1 6. ... l:!.b2-b3+
7. ... l:!.b1 -a1 + 1 7. �d3-e4 l:!.b3-b4+
8. �a8-b8 1 8. �e4-f5 l:!.b4-b5+
Again forcing Black to check with 1 9. �f5-g4
the bishop. And wins.
8. ... �g1 -h2+
9. g2-g311 There are also studies possible
This little move is what it was all where both kings are involved in
about. The long manoeuvre with the systematic manoeuvre.
the king was necessary to make it
possible. Black has to accept the
pawn sacrifice.
9. ... �h2xg3+
1 0. �b8-b7 .!:!.a1 -b1 +
11. �b7-a6
Now that the g3-square is blocked
for the black rook, the king travels
the long way back.
11. ... .l:l.b1-a 1 +
1 2. �a6-b5 l:!.a1 -b1 +
.l:l.b1-a 1 + DolgovjKuznetsov, 'Burevestnik' 1974
1 3. �b5-a4
White to play and win
1 4. �a4-b3
Step-by-step the king has arrived In general, two rooks are stronger
downstairs, and now it is ready to than a queen. However, here White
move to the kingside. has the initiative, and he has to
1 4. ... .r:l.a1 -b1 + hold onto it.
1 5. �b3-c2 l:!.b1 -b2+ 1. 'ifh4-c4+ �c6-d6
1 6. �c2-d3 2. �b8-c8
Now he can. Forcing the rook to step aside.

213
T h e Art o f the E n d g a m e

2. �d7-e7 The systematic manoeuvre now

3. �c4-d4+ 'it>d6-e6 continues in the opposite direc-
4. <Jo>cB-dB tion.
The queen and king are cooperat­ 11. 'ti'g4-f4+ �f6-e6
ing effectively. The enemy king is 1 2. 'it>g8-f8 l:.e7-d7
driven to the kingside. 1 3. 'ikf4-e4+ 'lt>e6-d6
4. ... �e7-f7 14. 'it>f8-e8 1ld 7-c7
5. �d4-e4+ <Jole6-f6
6. �d8-e8

How to go on from here?

15. 'ti'e4-f4+1 'ito>d6-c6
1 6. 'iff4-c1 +
All very systematic. This is what it was all about. White
6. ... J:rf7-g7 wins the other rook.
7. 'ii'e4-f4+ 'it>f6-g6
8. <;t>eB-fB J:rg7-h7 Especially beautiful are those stud­
9. 1i'f4-g4+ ies where king, rook and bishop
9 .'ife4+ was also sufficient, but it work together.
would slow down the winning
process.
9. ... 'it>g6-f6
1 0. 'lt>f8-g8 �h7-e7

Korolkov, 'Shachmaty v SSSR' 1964

White to play and win

White is a rook up, but how can he

prevent the promotion of the en­
emy h-pawn?

214
 S y s t em a t i c Manoeu vres

1. e4-e5 Now this is possible. The black

Clearing the e4-square for the king is drawn outside the square of
bishop. the white h-pawn.
1. �e3-c5+ 6. ... h2-h 1 'if
2. �a3-a2! 7. �a4-c6+ �e4-d4
The only square for the king. The 8. jg,c6xh 1 �d4-c5
square b3 must be kept free for the 9. h3-h4
bishop. And wins.
2. ... �c5-d4! Korolkov received 4th prize for this
A fantastic defence. 3 Jhd4 is met study in the tourney of the Russian
by 3 . . . h l 'iV 4.�e4+ 'it>e3 s . ..txh l chess magazine.
Wxd4, when White is left with the
wrong bishop. The study is impressive, but I felt
3. �c2-b3 that it could be improved.
The bishop is crawling up. Now
Black has to protect the dS-
square.
3. ... �f3-e4
4. .I:Ia4-a5
The rook also crawls up.
4. ... �d4-c5!
Again Black tries to lure the rook to
a critical square.

Korolkov 1964/Timman 2009

White to play and win

In this setting, too, the white

bishop has to try to reach the long
diagonal. However, White cannot
give check, since after l .cS+? �d4
he cannot win.
1. �d1 -c2+ �e4-f3
2. c4-c5
Now he can.
5. �b3-a4! 2. �e3xc5+
White carries on with the system­ 3. �a3-a2 �c5-d4
atic manoeuvre regardless. His 4. �c2-b3
bishop stills aims for a square on And the rest is well-known. In this
the long diagonal. version the white bishop travels the
5. �c5-b6 entire route along the short diago­
6. .I:Ia5-a6! nal d l -a4.

215
The Art o f the Endgame

There is a study by Bran that 4. l:th8-e8 we2-d3

greatly impressed me. Unfortu­ 5. �d4-c5 �e5-d6
nately it turned out to be incorrect. 6. l:te8-d8 �d3-c4
So far, the systematic manoeuvre is
correct. But now 7 .�e3 suffices for
the win.
7. �c5-b6
This is also possible.
7. �d6-c7
8. .ll d8-c8 wc4-b5
9. �b6-d4 �b5-c4
1 0. i.d4-e3 wc4-d3
11. �e3-c5 .tta 2-a 1 +
1 2. �g1 -g2 .tta 1 -a2+
Bron, Galitzky Memorial 1964
13. �g2-f3 l:ta2-f2+
White to play and win (incorrect)
14. �c5xf2 g3xf2
The intended solution runs as fol­ 15. �f3-g2 �d3-e2
lows: 1 6. l:tc8-e8+
1. �a7-d4 .l:tb2-a2 And wins.
2. a6-a7 �e1 xe2
3. l:th3-h8 Of course it is a great pity that such
a study should be incorrect. I
delved deeper into the matter, try­
ing to find a way to save this fantas­
tic systematic manoeuvre. In the
end I found the solution: the struc­
ture must be shifted one file to the
right.

3. ... �f4-e5
With this move Bran starts a sys­
tematic manoeuvre. However,
Black could have achieved a draw
with 3 . . . c.t>d3 ! . The point is that af­
ter 4 . ..tc5 �c4 5 . .l:!.c8 he has time
to improve the position of his
Timman 2011
bishop with 5 . . . �g5 ! . White can
White to play and win
try a few things, but there is no (correction of a study by Bron)
win.

2 16
 S y s t em a t i c Manoeuvres

The sit"\).ation of the white king And for the third time, the coun­
looks a little precarious. However, terattack.
the white passed pawn is strong 8. .i.c6-e41 �c5-d4
enough. Black must attack the bishop with
1. b6-b7 .:ta2-b2 the king , since after 8 . . . �f5
2. l:rh6-h8 9 .l::!.d 5 + he would lose immedi­
Threatening to promote the pawn. ately.
Black must attack the white bishop. 9. �e4-f3 �d4-e3
2. ... �g4-f5 1 0. �f3-d5
Neither was 2 . . . �e3 sufficient. Af­
ter 3 .�d5 �d4 4J:td8 White wins
effortlessly.
3. .l:th8-f8
Pinning the bishop, and so Black,
in turn, must attack the white
bishop again.
3. ... 'iitf2-e3
4. .i.e4-d51 jLf5-e6
5. .:tf8-e8!
The pattern is repeated.
5. ... 'it>e3-d4
Finally Black is outmanoeuvred. He
can still try a bishop move.
1 0. ... jLd7-e6
11. �d5-c6
And now what's left is two spite
checks.
11. ... l:rb2-b1 +
1 2. �h 1 -h 2 t!.b1 -b2+
1 3. \t>h2-g3
And wins.

When Benko turned 8 0 , he orga­

6. �d5-c61 nized an endgame study tourney,
In these circumstances this is the where the condition was that no
only move. On 6 .tf3 Black has the
. captures should be made during
saving move 6 . . . g4. the first 8 moves of the main line.
6. ... �e6-d7 This study has a phase of 1 3
7. l:re8-d8! moves, where there is only ma­
For the third time, the bishop is noeuvring. It seems to me that all
pinned. the elements of Bran's study have
7. ... �d4-c5 been preserved.

217
T h e A r t of t h e E n d g a m e

Also interesting is the struggle be­

tween king and minor piece.

The bishop will only find safe

squares on the g- and h-files.
5. ... �h5-g4
Katsnelson, Herbstman Centennial 2001
6. �g3-h2 cot>g4-h3
White to play and win
7. �h2-g1 ..t>h3-g2
Both sides' bishops are hanging. 8. �g1 -d4
White must not only preserve his After the bishop has climbed down
bishop, but he also has to win an the entire ladder, it sets off for a cen­
enemy piece - otherwise he won't tre square where it is relatively safe.
have enough winning potential. 8. ... tt:Jc4-a5
1. �f8-h6 Another resource. After 9. tt::lxa 6
Forcing Black to give check. tt::lc 6+ I O.c;t;>c3 tt::lxd4 I l .'it>xd4 bS !
1. ... tt:Jd2-c4+ White can't win since the black b­
2. 'iit;>a 3-b4 pawn is not blocked on the right
Now Black cannot move away his square. Then the knight pair is pow­
bishop, as then the knight on c4 erless in the fight against the king.
will be lost. Therefore he has no 9. �d4xb6 tt:Ja5-c4
choice but to attack the bishop 1 0. �b6-d4 1 -0
with his king. A struggle between king and two
2. ... 'it;>g8-h7 bishops is also possible.
3. �h6-g5!
The only square. 3 .�f4 would be
met by 3 . . . tt::l b 2, whereas after
3 .�c I Black would have a way out
with 3 . . . tt::l e 5 . The bishop must
stay outside the range of knight
checks.
3. ... �h7-g6
4. �g5-h4
The downward ladder movement.
4. ... Wg6-h5 Liburkin, 'Sachove Umeni ' 1946
5. i.h4-g3 White to play and win

218
 S y s t em a t i c Manoeuvres

Whitt; wants to promote his pawn,

but he must watch out that the
newly-born queen is not captured
by the black rook after a bishop
check.
1. b7-b8� �d8-h4+
2. <t>e1 -d2
The king cannot go to a light
square.
2. ... �h4-g5+
3. <t>d2-c3 .ig5-f6+ 11. ... �d8-b6+
4. �c3-b4 �f6-e7+ 1 2. �b8xb6 nhaxa8
5. Wb4-a5 �e7-d8+ 1 3. �b6-b7+ 1 -0
6. wa5-b51
The struggle between king and
knight is fascinating.

Only now that the king's bishop is

Liburkin, 'Shakhmaty v SSSR' 1938
blocking the 8th rank again, the
White to play and win
light squares are available for the
king. The white knight and the black
6. ... �c8-d7+ bishop are both under attack. By
The duel between the king and giving checks, White hopes to save
the queen's bishop starts. Now his knight, and then to capture the
White has to avoid the dark bishop.
squares. 1. tbb3-d4+ cJtc2-c3
7. �b5-c4 �d7-e6+ 2. tbd4-b5+ �c3-c4!
8. �c4-d3 �e6-f5+ Not 2 . .'\to>b4 in view of 3 J:tb l +
.

9. �d3-e2 �f5-g4+ and 4.rJilxe l .

1 0. �e2-f1 �g4-h3+ 3. tbb5-d6+ �c4-c5
11. �f1 -g1 ! Once more Black has to allow a
Finally the white king can move to knight check, as otherwise a rook
a dark square again. check would follow.

219
The Art o f the E n d g a m e

4. tt:ld6-b7+
The knight has to move up further.
With 4.tbe4+? �d5 5 . tbf6+ �e5
6.tLld7 + �e6 White would
achieve nothing.
4. ... �c5-c6
5. tt:lb7-d8+
The king-and-knight dance contin-
ues.
5. ... �c6-c7
The finish contains a few more at­
tractive points.
11. .l:.e1 -h1 g4-g3
1 2. �e2-e31
Now the white king is going to
collect the g-pawn via f4.
1 2. ... �h8-g7
The king moves out, just too late to
save the knight.
13. �e3-f4 g3-g2
14. l::th 1 -g 1 tt:lh2-f1
1 5. l::tg 1 xg2+
The knight and king have reached And wins. On the next move White
the edge of the board. Now the plays 1 6 .nf2 , after which the
systematic manoeuvre is continued knight is lost.
in the direction of the kingside. For this study Liburkin received
6. tt:ld8-e6+ cJ;;c 7-d7 4th prize; a scant reward for such
7. tt:le6·f8+ �d7-e7 an extraordinary idea. I started in­
Black must keep avoiding rook vestigating in what circumstances
checks. the duel between the king and
8. tt:lf8-g6+ �e7-f7 knight could take place. The knight
9. tt:lg6-h8+1 must be supported by a rook. Only
The ninth and last check! White then is the enemy king forced to
allows his knight to be caught in attack the knight diagonally in or­
the corner in order to conquer der to avoid rook checks. The
the enemy knight with tactical board has to be almost empty. This
tricks. is in fact a necessary condition for
9. .. . 'ft;f7-g7 studies with systematic manoeuv­
1 0. l:1a1 xe1 res - space is needed.
Finally White finds time to capture Moreover, it seemed a good idea to
the bishop. add a theoretical endgame to the
1 0. ... Wg7xh8 manoeuvres as a finale.

220
 Systematic Manoeuvres

The first series of checks is fin­

ished, and now the king and
knight move up.
6. ... lt:Jh3-f4+
7. Wg2-g3
Again White has to attack the
knight diagonally. After 7. <oitf3?
l:tf8 Black would have a winning
discovered check at his disposal.
7. ... tLlf4-h5+
Timman 2011 8. Wg3-h41
White to play and draw (after Liburkin)
For the first time White avoids a
White is a rook behind. He has to knight check.
try to promote his c-pawn. 8. ... na8-h8
1. ..ta6-b7 4Jb4-d3+ 9. c6-c7
The knight checks start. Insuffi­ Now the discovered check is less
cient for the win was 1...l:h2+ dangerous. But Black can still make
2.Wb3 dxc6 3 . .txc6 l:i.c2 4.�b5 a serious winning attempt.
and the knight is lost. 9. 4Jh5-f6+
2. 'Wt;>b2-c21 10 . ..t>h4-g5
The diagonal attack. After 2. �c3
l:ta3 + Black would win.
2. ... 4Jd3-e1 +
3. 'it>c2-d2
The king-and-knight dance starts.
First they move towards the
kingside.
3. 4Je1-f3+
4. �d2-e2 4Jf3-g1+
5. <;t>e2-f2 lt:Jg1-h3+
6. wf2-g2
White must put his king in as ac­
tive a position as possible.
10. ... tLlf6-e4+!
With this fine point Black reaches a
rook versus bishop ending with
one pawn on each side.
11. ..tb7xe4 J:rh8-c8
A continuation of Chapter 2. White
must push his passed pawn quickly,
as otherwise the ending is a theo­
retical loss.

221
The Art of t h e Endgame

12. h2-h4 llc8xc7

1 3. h4-h5 .!::tc7-c5+
1 4. �e4-f5
White must maintain his active
set-up.
1 4. ... d7-d6
15. h5-h6+
With a draw.

In the end, just as in Liburkin's

study, the main line contained nine 6. tt:Jf6-g8+
knight checks; however, the final The knight has to travel to the far
two were not thematic. I started edges of the board again.
looking for a study version with 6. 'it>e7-f7
even more interesting content and 7. t2Jg8-h6+ 'fi;;f7-g7
more thematic knight checks. 8. t2J h6-f5+ 'ft;g7-g6
Not 8 .. .'�0lf6 on account of 9 JHl ,
followed by a devastating discov­
ered check by the knight.
9. tt:Jf5-h4+ 'ft;g6-g5
The h-pawn is in the way. Again,
Black has to allow a thematic check.
1 0. t2Jh4-f3+ 'ot>g5·g4
An important moment. Now the
thematic temptation is l l .tLlh2+.
But with this ninth thematic check
White would let the win slip
Timman 2011
through his fingers. Black would
Wh ite to play and win
reply l l . . .�h3 , when the white
Before the ritual starts, White has rook cannot intervene decisively.
to take some precautions. For the win, an extremely subtle
1. 'iitf 1 -g2 h2-h1 'iY+I move is needed.
Black sacrifices a pawn in order to 11. tla1 -a211
have more space for his king later Highly unexpected. Now Black
on, when it has to travel the long cannot take the knight, since after
road to the kingside. l l . . .'�xf3 1 2 .dxc3 he would not
2. 'it>g2xh1 �c1 -b2 have any good squares for his
3. tt:Jb5-d6+ bishop.
Starting the king-and-knight dance. 11. ... c3-c2
3. ... 'it>b7-c7 The best chance. Now White has
4. t2Jd6-e8+ 'ftJc7-d7 the same resource as in the previ­
5. tt:Je8-f6+ 'it>d7-e7 ous study.

222
 System a t i c Manoeu vres

1 2. tt:lf3-e5+! .ltb2xe5
Again, his best chance. If the white
knight reaches the d3-square,
Black has nothing left to hope for.
13. .l:!.a2xc2 �g4-f3
14 . .l:!.c2-c4!
In this final phase too, White has to
keep playing accurately. With the
text he denies the black king access
to the e-file.
14. ... �e5-g3
Threatening 1 5 . . h4, when White
. 1 6. d2-d4!
cannot win. You can also start a game with this
1 5. .l:!.c4-c3+ move, but in these circumstances,
Again the black king is cut off from on an almost empty board, it can
the e-file. be quite effective. Black cannot
15. ... �f3-g4 stop the passed d-pawn any more.

223
Chapter 12

The Disappearing Trick

In Chapter 7 I already touched on 2. d7-d8'i¥+
the theme of the present chapter: First the pawn is sacrificed.
in a study variation one of White's 2. ... �a5xd8
pieces disappears - in other words: After 2 . .'ifxd8 3 .tLld7 + , followed
.

the same position arises as before, by 4J1f8, White would immedi­

but now without a certain piece. ately attain the draw.
The absence of this piece gives 3. �f7-b7+
White extra possibilities. Introducing the disappearing trick.
Such piece sacrifices are always 3. �b8-c8
spectacular. I have named this 4. �b7-b8+1 'i!tc8xb8
theme 'the disappearing trick'
since it made me think of the work
of an illusionist: you make some­
thing disappear and then a brilliant
point follows.

Kubbel has made three studies

where he used the disappearing
trick to force stalemate.

On the board we have the same po­

sition as two moves ago, only now
without the white rook. The stale­
mate is not far offnow.
5. tt:lf8-d7+ Wb8-c8
6. tt:ld7-b6+1
The third and last sacrifice. If Black
does not take the knight, White
Kubbel, 'Novoye Vremya ' 1914
will give perpetual check with it.
White to play and d raw
6. ... ..td8xb6
This one is quite straightforward. Stalemate.
1. d6-d7+ This was a study without great sur­
Forcing the king to step aside. prises. More or less the same goes
1. ..
. 'i!tc8-b8 for the following study.

224
 The Disappearing Trick

Kubbel, 'Bohemia' 1909 Kubbel, 'Deutsche Schachzeitung' 1907

White to play and draw White to play and draw

Again White has to try to sacrifice The white king is already in a stale­
all his pieces. mate position. He has to sacrifice
1. lbb6·d7+ his pieces in an accurate way.
First the knight. 1. �f8·d6 'iib8xd6
1. ... 'ti'g4xd7 2. b7-b8'i¥+
2. l:ta6·c6+ 'it>c5·b5 Black has two ways to take back:
Without the rook and bishop this A) 2 .'.t>xb8
•. 3.�h3+! 9iic 7
would be stalemate. How does 4.l:!.xh7+ Wb6 sJ:th6! 'ifxh6 stale-
White get rid of these pieces? mate;
3. �e4·d3+1 B) 2 ••. 'ifxb8 In this variation our
Black has to capture the rook, as theme is featured. 3.l::txh7+ 9iia8
otherwise he will be mated. 4.ii.b7+ Wa7 s...tc8+ with either
3. \tb5xc6 perpetual check or stalemate.
4. �d3·b5+1 \tc6xb5
There are also present-day varia­
tions on this interpretation of our
theme.

The rook and bishop have disap­

peared from the stage and White is
stalemated. KralinjPervakov
'Uralski Problemist' 2003
White to play and draw
The next study is of a higher calibre.

225
T h e Art o f t h e E n d g a m e

Both kings are in danger, but the 7. d7·d8�+ Wb8-a7

white king more so. Moreover, it is 8. �d8xd4+ \t>a7xa6
almost stalemated. For this reason 9. tt:lb7-c5+ wa6-b5
White will now sacrifice all his 1 0. �d4-a4+
pieces. Forcing the king to take the knight.
1. llf4·f8+ 1 0. ... wb5xc5
White must first sacrifice the rook 11. 'iVa4-a7+
- otherwise he will not lose every­ The third queen sacrifice. This
thing in the end. time Black does not have to accept.
1. ... wb8-a7 11. ... wc5-c4!
2. l:U8-a8+ '.t>a7xa8 Sacrificing his own queen. Black
Not yet a disappearing trick; the plays for mate.
black king has landed on another 1 2. 'iVa7xh7
square.
3. d7-d8'ii' + wa8-a7
4. 'ti'd8·b8+1
With this queen sacrifice White
makes new checks possible.
4. ... wa7xb8

1 2. ... tlc2·b2+
13. �b1 ·c1 tlb2·b1 + 1
Now it is Black's turn to take refuge
in a disappearing trick.
14. 'i!Yh7xb1
Now the starting position is reached, Certainly not 1 4.'it;xb l in view of
without the rook and the d7 pawn. 1 4 . . . �d3 + and Black wins.
5. d6·d7+ Wb8·a7 14. ... �f1 ·d3
6. 'ti'g3·b8+1 The situation is precarious for
The second queen is sacrificed on White. However, he still has a stale­
the same square. mate trap.
6. ... wa7xb8 15. 'i!Yb 1 -c2!
And now the starting position ap­ The final queen sacrifice. Which­
pears on the board without the ever way Black takes, it is stalemate.
white queen, rook, and pawn on For this study, Kralin and Pervakov
d6! The depletion of the white received second prize in the tour­
forces is very swift. ney of the Russian magazine.

226
 The Disappearing Trick

Nikolay Kralin and Oleg Pervakov The only way to escape the checks.
are both grandmasters in the field 10. 'ti'b1-e4+ �d2-g2
of the endgame study. Together they 11. lt:ld5-f4!!
have made magnificent studies. A magnificent denouement.
Pervakov has also used our theme in 11. ... �g2xe4
a study he composed on his own. Stalemate.

Pervakov 2005 For this study Pervakov received

Black to move- White draws
first prize in the World Cup.
As far as material is concerned,
White has nothing to complain There is a difference between stale­
about, but his king is unsafe. mate studies from the beginning of
1. ... 'ii'c8-c1+ the 2Oth century and those from
2. 'it>h6-h5 the beginning of this century; the
The only square. present-day adaptations are more
2. ... ..te6-f7+ complicated, and play takes place
3. 'it>h5-h4 J:!.f2xg2 over the entire board. By the way,
After a few introductory checks for studies with a great number of
Black has created dangerous disappearing pieces, the stalemate
threats. motif is quite suitable.
4. �b4-d2!!
With this bishop sacrifice White
clears the way for his rook.
4. ... li'c1xd2
5. l:.b5-b1+ 'it>h1-h2
6. .l:lb1-h1+!
Here is the theme. White clears the
way for his queen.
6. ... Wh2xh1
7. �h7-b1+ �h1-h2
8. �b1-b8+ 'it>h2-h1 Timman 2011
9. �b8-b1+ l:.g2-g1 White to play and draw

227
T h e A r t o f the E n d g a m e

The white king is caught in a 5. �a7-b8+!

mating net. He could give his This fourth bishop sacrifice marks
queen with I .'it'b l .!:l:h8 + 2. .�h7 , the beginning of a sacrificial
but the ensuing ending is slugfest.
prospectless. 5. ... 'it>a8xb8
1. .ia3-c5+! �a7-b8 Now, all white pieces will be sacri­
Otherwise White would trade ficed on the corner square.
queens on a I . 6. Wk'a1 -a8+!
2. .ic5-a7+! '.it>b8-a8 First the queen.
Again, the only square. After 6. ... 'it>b8xa8
2 . . . '>itc8 3 .�f5 + White would win. 7. :g1 -a1 +
But what can he do now against the The first rook presents itself.
threatened mate? 7. 'it>a8-b8
3. �d3-g611 8. :a1 -a8+1 'it>b8xa8
A true problem move. The enemy 9. .l:!:h 1 -a 1 +
queen is driven to g6, after which And the second.
the white queen can control the 9. ... '1t>a8-b8
h8-square. 1 0. :a1 -a8+1 '1t>b8xa8
3. ... �g7xg6 Stalemate.
An important alternative was
3 . . .'iYf6 + 4.Wxg4 .r;:[ e l ! . With
5 . .tel ! White can just hold.
4. 'ti'f1 -a1 1 f7-f51
The best attempt. After 4 . . . l:t.g8
s . .te3 + ! �b8 6 . 'ifa 7 + �c8
7 .�a8 + Wd7 8 .'iYxg8 'ifxg8 9.f5 !
White would have the advantage.
After the text move he must again
arm himself against the threat of
mate in one.

A real deforestation has taken place.

The previous diagram position is
again on the board, only now
without the four white pieces.

In almost all the other studies that I

will present, White uses the disap­
pearing trick to win. Sometimes
White wants to lose a piece in or­
der to create a passage for a pawn.

228
 T h e D i sappearing Tri ck

Introducing a stalemate trap into

the position.
5. g7xh8�1
The carousel of minor promotions
continues. After 5 .gxh8'i¥? d4!
6.'i¥xd4 .l:ra 1 + Black would force
stalemate.
5. ... d5-d41
Under these circumstances, too,
Black continues to play for stale­
Bron, 'Magyar Sakkelet' 1948
mate.
White to play and win
6. �h8xd4 .!:i.a3-d3
Black threatens mate in one. Al­ 7. .ltf7-h5
though three of his pawns are about Covering the d 1 -square.
to promote, White cannot solve this 7. ... .!:i.d3-b3
problem. He will have to create a Continuing to harass White.
new candidate for promotion. 8. �h5-g6 .l:[b3-c3!
1. .!:i.b7-a7+ �a8-b8 Making use of the stalemate trap.
2. .t:: a 7-a8+! Now White has to put his queen's
Thus White gets rid ofhis rook. bishop on another diagonal.
2. ... �b8xa8 9. �d4-e3 .!:i.c3-a3
3. b6-b7+ 1 0. h7-h8�1
The point of the rook sacrifice. The The second bishop promotion on
b-pawn gets free passage. the same promotion square.
3. ... �a8-a7 1 0. ... .t:!a3xe3
The only move to prevent immedi­
ate mate.
4. b7xc8tt:l+!
With this minor promotion, the
important black defender of the
back rank is eliminated with tempo.
4. ... �a7-a6

The last attempt.

11. e7-e8.!:i.l
And wins.
For this study, Bron received first
prize in the tourney of the Hun­
garian magazine.

229
T h e Art of the E n d g a m e

Incidentally, the disappearing trick 3. ... lUg5-f3+

was not central in this study: it 4. �e1 -d1 lUf3-d4
served to make several minor pro­ 5. lUb3-a1 lUd4-c2
motions possible.

The Romanian endgame study

composer Emilian Dobrescu has
made a study where the disappear­
ing trick serves via an ingenious
manoeuvre to grant a passed pawn
free passage.

Black's counterplay materializes.

He steers towards a repetition of
moves.
6. lUd8xe6!
The introduction to the disappearing
trick. It is important that the black
e-pawn disappears from the board.
Dobrescu, 'L:Italia Scacchistica' 1968 6. ... lUc2xa1
White to play and win
7. lUe6-d4 lUa1 -c2
A coincidence: just as in the previ­ The ritual continues; the two
ous study by Bron, White has knights will change places.
passed pawns on b6 and c6. 8. lUd4-b3 lUc2-d4
1. c6-c7+ �b8-b7 9. lUb3-a1 lUd4-c2
His best chance.
2. c7xd8lU+I
A second coincidence: here, too,
White captures a rook and pro­
motes to a knight with check.
The three white knights that are
now on the board will all be sacri­
ficed on the same comer square in
the end.
2. ... <;;t> b 7xb6
3. lUd2-b3
The first episode is concluded. In four moves the same pos1tJ.on
White has a great material advan­ will occur, only without the knight
tage, but the a-pawn remains very on h6.
dangerous. 1 0. lUh6-f5!

230
 The D isappearing T r i c k

This. is now possible because there White has three extra pawns, but
is no longer a black pawn on e6. his knight is offside and also the
1 0. ... tt:lc2xa1 safety of his king is not guaran­
11. tt:lf5-d4 teed.
The third knight enters the battle. 1. tt:lb8-d7!
11. ... tt:la1 -c2 Black has to accept this knight sac­
1 2. tt:ld4-b3 rifice, otherwise the knight would
The ritual is played out for the reach the vital square f6.
third and last time. 1. ... i¥.e6xd7
12. ... tt:lc2-d4 2. e5-e61
1 3. tt:lb3-a1 tt:ld4-c2 The point of the knight sacrifice.
1 4. h5-h6! With a further pawn sacrifice,
This was what it was all about. The White opens up the long diagonal
white kingside pawns decide the for the queen.
outcome of the battle. 2. ... ii.d7xe6
1 4. ... tt:lc2xa1 The starting position, without the
1 5. h6-h7 tt:la1 -c2 white knight and e-pawn.
16. h7-h8� a2-a1 1i'+ 3. l:rc3-c8+1
1 7. 'ifh8xa1 tt:lc2xa1 The third consecutive sacrifice.
1 8. g4-g5 1 -0 White's attacking play is based on
For this study Dobrescu received the strong f-pawn, which paralyses
first prize in the tourney of the Ital­ Black's defences.
ian magazine. 3. .ie6xc8
4. 'ii' d3-c3+ l:rh7-g7
Things are a little more difficult if 5. '1Wc3xc8+
the disappearing trick serves to en­ White has regained material with
able a pawn to march, which is check, while his attack rages on.
then also sacrificed in order to 5. ... wh8-h7
open up a diagonal for the queen. 6. 'ifc8-f5+ wh7-h6

Bazlov 2009
White to play and win
Setting a stalemate trap.

23 1
T h e Art of t h e E n d g a m e

7. f7-f81:!.!
Certainly not 7 .f8'if? on account of
7 . . . 'ifxe 2 + ! with stalemate.
7. ... .l:!.g7-g6
The white attack is unstoppable.
8. 'iVf5-h3+ �h6-g7
9. 'iVh3-h8
Mate.
For this study Bazlov received first
prize in the Sarychev Memorial
Tourney. The king has to move up further.
5. ... tt::l d5xc3+
Dobrescu has made another inter­ 6. we4-f4
esting study with the disappearing The king moves to the kingside, in
trick. Although only a single pawn order to conquer the black bishop,
disappears, the motif behind this is which deprives it of the square e I .
worth seeing. 6. ... tt::l c 3-d5+
7. 'it>f4-g4 tt:Jd5-e3+
8. <t>g4xh4 tt:Je3-f5+
9. �h4-g4
And now, back again.
9. tt::lf 5-e3+
1 0. �g4-f4 tt:Je3-d5+
11. �f4-e4 tt:Jd5-c3+
How can White escape from the
checks now?
1 2. �e4-e3
Down again.
Dobrescu, 'Themes 64' 1981
1 2. ... tt:Jc3-d5+
White to play and win
1 3. 'it>e3-e2 tt::l d 5-f4+
Black cannot prevent the promo­ 1 4. 'it>e2-e1 !
tion of the white c-pawn, but he
gets counterplay because the
white king has few squares at its
disposal.
1. c6-c7 tt:Jd8-e6
2. c7-c8if �g3-g2+
The start of a long series of
checks.
3. <it>e1 -e2 tt:Je6-f4+
4. We2-e3 tt:Jf4-d5+
5. �e3-e4

232
 T h e D i sappearing T r i c k

The . theme becomes apparent. It looks as if Black has already run

White forces Black to take the out of moves. However, he still has
d-pawn. a surprising resource.
1 4. ... lt:lf4xd3+ 3. ... \jWe6-g4+!!
1 5. �e1 -e2 lt:ld3-f4+ An astonishing queen sacrifice, and
1 6. we2-e3 at the same time Black's contribu­
And wins, as White now gets the tion to the theme.
square d3 for the king. 4. �h4xg4 lt:ld7-e5+
For this study Dobrescu received The knight takes the place of the
first prize from the French maga- queen.
zine.

In the previous study, White made

the pawn disappear in order to
avoid perpetual check. For this
purpose, we can also make two
white pieces disappear.

5. 'Ot>g4-h4!
White sacrifices a rook in order to
create an escape square for the
king.
5. ... lt:le5xf3+
6. �h4-g4
The same position as two moves
Timman 2011 earlier, only now without the rook
Black to move - White wins on f3 .
6. ... lt:lf3-e5+
White threatens to give mate. Black 7. �g4-f4 lt:le5-d3+
has to take drastic measures to con­ 8. 'Ot>f4-f3 lt:ld3-e5+
tinue the struggle. The alternative was 8 . . . tt::l xe 1 +. Af­
1. ... 'iVb2-e5+ ter 9.Wg4 dxe4 1 O.tt::lc 3 a3 1 1 .Wf4
2. �f4-g4 tt::l c 2 1 2 .�xe4 Wg3 1 3 .-tc l ! Black
Now the queen is in the way of a cannot prevent the white king's de­
knight check. For this reason, Black cisive penetration via eS . After the
is the first to resort to a disappear­ text move White has to continue
ing trick. manoeuvring.
2. ... '1We5-e6+ 9. �f3-f2 lt:le5-g4+
3. 'it>g4-h4 1 0. wf2-f1 1

233
The Art o f the E n d g a m e

1 6. ttle2-c1 I
It all fits nicely.
1 6. ... ttlf3xe1
1 7. �d1 xe1 �h2-g3
1 8. e4-e5
And the e-pawn decides.

Square clearance for the king as a

motif for the disappearing trick
can also occur in a miniature
As in Dobrescu's study, White sac­ study.
rifices his bishop in order to clear
the e3-square for his king.
1 0. ... ttlg4xe3+
11. �f1 -f2 ttle3-g4+
1 2. �f2-f3 ttlg4-e5+
1 3. �f3-e3
Finally the white king has some
space. But the enemy checks are
not exhausted yet.
1 3. ... d5-d4+
1 4. �e3-d21
Bazlov, Nunn 50-JT 2005
Not 1 4.lbxd4 cxd4+ 1 S .�e2 on
White to play and draw
account of 1 S ... Wg3 1 6.l:[fl a3 ! ,
and thanks to his strong knight Without the knight, White could
Black can just hold. play 1 . �g6. If the white king takes
14. ... ttle5-f3+ a detour, it will be too late: l .�g7
1 5. �d2-d1 lbd6 2 .lbes g 3 3 .�g6 �d8 !
The conclusion of a long king 4.�hS lbe4 S .�g4 �c7 and the
march. But Black has yet another black g-pawn is solidly defended.
string to his bow. White has no chance to exchange
1 5. ... a4-a3 the knights.
1. ttlg6-h81
Offering the exchange of knights.
If Black complies, the white king
will be back just in time, as be­
comes clear after 1 . . . lbxh8 2 . �xh8
�c6 3 .�g7 �dS 4.�g6 �d8
s .�fs g3 6.�f4! �c7 + 7 .�e3 and
the king reaches the safe square e2,
after which White has a fortress.
1. ... ttlf7-e5

234
 T h e D i sappearing T r i c k

Ce:n,tralization. Now what can The disappearing trick can also

White do? serve to clear a square for other
2. lbh8-f71! pieces than the king. The following
The magnificent point. White sac­ miniature study is famous.
rifices his knight in order to obtain
the square g6 for the king.
2. ... lbe5xf7
3. �h7-g6
The king takes the place of the
knight.
3. . . . ttJf7-e5+

Kozlowski, 'Swiat Szachowy' 1931

White to play and win

The solution is not difficult. White

forces the decision with:
1. �h7-g7+! 'ot>g8xh8
2. .!:tg7-h7+ 'it>h8-g8
3. g6-g7
The bishop sacrifice has cleared the
The best try Now, after 4.�xg5 h8-square for the rook.
�c6 5 .�f4 �d6 ! , White would It is harder to prove that the alter­
not stand a chance. native l .g7 does not win. After all,
But he does not have to take the if White managed to exchange
bishop. rooks, it would no longer matter
4. �g6-f51! that the bishop on h8 is bad. The
A fantastic move. White would defensive scheme that Kozlowski
rather take the knight than the indicated implies that in the end
bishop and therefore he puts his Black sacrifices his f-pawn and puts
king in between the two pieces. his king on h7 and his rook on the
4. .. . lbe5-f7 g-flie. Indeed, it looks as if Black
There is nothing better. 4 ... l2Jf3 has a fortress in that case. By the
would be met by 5 .�xg4. way, I wouldn't enjoy defending
5. �f5-g6 such an endgame in practice.
And Black cannot make progress.
In my opinion, this is the best A bishop can also be sacrificed in
study with the disappearing trick order to clear a square for a
theme. knight.

235
T h e Art o f t h e E n d g a m e

The g-pawn has also disappeared

from the board, and stalemate is in
the air. However, White has an­
other trick.
5. h3-h4!
From stalemate to mate. The deadly
check on gS cannot be prevented.
5. ... 'ii'g 2·g1
6. l::tg 7-g5+
And wins.
Timman, WCCT-7 2004
White to play and win
Line clearance can also be a motif
Materially White clearly has the for the disappearing trick.
upper hand, but Black threatens to
bring his queen to c2.
1. tt:Jf5-g3+!
Forcing Black to take the bishop.
On l . . 't\Vxg 3 , the zwischenschach
.

2 . .!:i.h7 + would win instantly.

1. . .. 'it>h5xh4
2. tt::lg3-f5+ 'i.t;lh4-h5
The starting position, only without
the white bishop.
3. .!:[g7-h7+
Wotawa, 'Deutsche Schachzeitung' 1954
The point of the bishop sacrifice.
White to play and win
Now that square h4 has been
cleared for the knight, the black It looks as if Black has managed to
king cannot go to g6. keep command of the situation.
3. ..
. �h5xg5 Without the white bishop and
Playing for stalemate. b-pawn, a check on b2 would de­
4. l:th7·g7+ �g5-h5 cide the issue. But how can White
lose the bishop and the pawn?
1. wg2-g3!
Forces Black to take on b3 .
1. ... .l:rb1 xb3+
2. 'ii? g 3-g2 l::t b 3-b1
The b-pawn has been removed.
Now the bishop.
3. ..tb4-a5+! Wb6-a6
After 3 . . . �xa5 4 . .!:i.a2+ followed
by 5 . l:lb2+ White would have
achieved his goal.

236
 T h e D i sappea ring Trick

Black's counterplay is not to be un­

derestimated. White has to protect
his bishop; all other moves lose.
1. .'bh5-f4 'ifa6-b6
An important alternative was 1 . . . a2
with the threat of a queen sortie to
a3. In that case White has an ele­
gant, hidden way to win: 2 . l:th6+
�gS 3 .l:rh5 + ! �xf4 4 . .!:f.f5 + '1t>e4
s . ..tg2 + �d4 (s . . . �d3 6 . ..tfl +)
5. �a5-b6!! 6 . .l:rd5 mate.
A fantastic move. As the bishop
controls the square g 1 , it has to be
captured.
5. ... <it>a6xb6
6. l:!.d2-b2+1 l:!.b1 xb2+
7. <it>g2xh1 l:tb2-b1 +
8. �h1 -g2
The white king marches to f6, and
Black will run out of rook checks.

For my 2 004 study I received 1 1 th

prize in the 2003 World Champi­
onships, which had the disappear­ A clean mate, in the middle of the
ing trick as its theme. board.
With hindsight I wasn't satisfied 2. 'it>a1 -a2!
with this study. The play only oc­ A quiet move, after which the
curs on the kingside, and the white black queen is dominated.
king plays no active role. A better 2. ... 'iYb6-g1
version turned out to be possible. 2 . . . �d4 would be met by 3 .l:th6+
�g5 4.lLle6+ and wins. Also after
2 . . . 'iVf2 3 .l:th6+ \t>gs 4.l:ths +
�f6 S . l:!.fS + the queen is lost.
3. .'bf4·g2+
Preparing the disappearing trick.
3. ... <;i;>h4xh3
4. tiJg2·f4+ �h3·h4
5. J:g6·h6+ Wh4xg4
6. J:h6·g6+ 'lt;>g4-h4
7. h2·h31
Timman 2011 Again this little move. Black can
White to play and win still resist and play for stalemate.

237
The Art o f the Endgame

7. ... �g1 -b1 + White has to capture the knight

8. �a2xb1 a3-a2+ in order to create winning poten­
tial.
1. �c2-e4+ �a2-b3
2. �e4xd5+ �b3-c3
3. l:[d2-d1 <J;>c3-c2
Black's plan of defence consists of
simply pursuing the rook with the
king.
4. l:[d1 -d4 �c2-c3
5. l:!:d4-c4+ �c3-b3
How to go on from here?

9. �b1 -c21
The only good square for the king.
White has to avoid the a-pawn pro­
moting with check.
9. ... a2-a1 �
1 0. .l:tg6-g4#

After the bishop I made the

g-pawn disappear, with the aim of
opening the g-file. It is only logical
that this is achieved by means of a
pawn sacrifice. After all, the pres­ 6. J:lc4-c6+11
ence (or absence) of pawns deter­ The climax of the study. White
mines if a file is closed, half-open forces the king to capture the
or open. This motif also features in b-pawn.
a youthful study by Kralin. Insufficient for the win would have
been 6 . .!:tc5 + . White has the same
idea: 6 . .'�xb4 is met by 7 . .l:f.c4+.
.

But with 6 . . .'�ia4! 7 . .l:f.a5 + Wxb4

Black could wrongfoot the white
rook.
6. ... �b3xb4
7. l:.c6-c4+
Now the rook will travel the whole
way back.
7. . .. 'it'b4-b3
8. .l:!.c4-d4+ �b3-c3
Kralin, 'Shakhmatnaya Moskva' 1968
9. l:.d4-d1 'it'c3-c2
White to play and win

238
 T h e D i s appearing Trick

The win is not easy for White. For

instance, he cannot play 4 . .U.g5 , as
then the black passed pawns be­
come too strong.
4. <>t>h5-g4 J:lf5-f4+
5. �g4-h31

1 0. jLd5·b3+!
The final point. Now that the b-file
is open, White has this tactical
trick.
1 0. ... 'Ot>c2xb3
11. .l:!:d1 ·b1 +
And wins.
This way White forces Black to take
In the following study the f-file has the f-pawn.
to be opened. 5. ... l:.f4xf3+
6. �h3-g4 l:.f3-f4+
7. '>tg4-h5 J:lf4-f5+
8. '>tr>h5-h6 l:.f5-f6+
9. '>th6-h7 J:lf6-f5
Via a systematic manoeuvre we are
back in the position of seven
moves ago, but the white f-pawn
has fallen.
1 0. l:.g1 -f1 !
That's why White wanted to open
the f-file. With the rook sacrifice
Sochniev, Gurgenidze JT 2004
he gains time to promote his
White to play and win
pawns.
Black is threatening mate in one. 1 0. ... .:tf5xf1
Before White prevents this, he first 11. c6-c7 J:lf1 -f5
sacrifices his knight in order to Black is defending tenaciously.
create a mighty, far-advanced pawn Again, mate is threatened.
pair. 1 2. d7·d8tLl+l
1. tt:Jb5·d6+1 c7xd6 The only way.
2. cJo>h7-h6 .l:f.f5-f6+ 1 2. ... l:.b8xd8
3. �h6-h5 .l:f.f6-f5+ 1 3. c7xd8tt:J+I wf7·f6

239
T h e Art o f t h e E n d g a m e

2 . f8'ti' l:[xf8 3 .�xf8 h3 , and the

h-pawn decides the issue. There­
fore White must prepare the pro­
motion of his b-pawn.
1. tba8-c7! l::tf2xd2+
2. \t>a2-a3!
The only good square for the king.
Here is a long variation which il­
lustrates why 2 .'1t>b3 is not good:
2.'1t>b3 .l:tb2+ 3 . <;t>c4 trxb6 4.f7
Black maintains his threat. �d6 s .tt:le8 �e7 6.�f6 .l:rb4+
1 4. g 7-g8tbl 7 . '1t>c3 �f8 8 .�g7 iie7 9 . ..tes
Mate. l:txb7 1 0 .�d6 h3 ! and in the end
For this study Sochniev received Black draws the long end of the
shared first prize in the Gurgenidze stick.
Jubilee Tourney. 2. ... J:td2xd3+
3. �a3-a2
Back to the 2003 World Champi­ The theme is clear again: White has
onships. Second prize went to given up the d-pawn in order to
Pervakov, who incorporated a reach the same position. The ab­
highly complicated square clear­ sence of the pawn provides him
ance in a study that makes a with an extra square for the king
slightly chaotic impression. later on.
3. ... l::td 3-d2+
4. 'it>a2-b3
Now he can play this.
4. . .
. J:td2-b2+
5. �b3-c4 .!:[b2xb6
6. f6-f7 ..lte5-d6

Pervakov, WCCT· 7 2004

White to play and win

Many things are going on here.

White has two dangerous passed
pawns, but the black h-pawn
should not be underestimated ei­
ther. For example, l .f7 ? would be
insufficient in view of 1 . �b8 ! . . 7. tbc7-e8!

240
 T h e D i sappearing T r i c k

So far the variation is the same as in 3. .l:tb7xc7 �e8xh5

the comments to White's second Perpetual check is threatened.
move. Now 7 . . . �e7 would fail to White now also sacrifices his sec­
8 .ii.f6 �b4+ 9.'it>d3 �f8 1 0.�c3 , ond rook.
when White gains a decisive 4. .l:tc7-c8+ �a8-a7
tempo by attacking the rook. 5. .!::!:c8-a8+ �a7xa8
7. ... .itd6-a3 6. c6-c7+
The best try. A familiar motif The rook has dis­
8. tt:Je8-d61 appeared in order to provide pas­
The thematic knight sacrifice. After sage for the c-pawn.
8 .ii.e5 ? .ti.xb 7 9 .�d6 �xd6 6. ... �a8-a7
1 0.tt::lxd6 .ki.b8 l l .tLle8 .l:f.b4+ , fol­ 7. c7-c8�
lowed by 1 2 . . . l:!.f4, Black would
win.
8. ... �a3xd6
9. .itg7-e5 .itd6-a3
1 0. �e5-c7
The bishop goes to the vacated
square. White wins.

And the first prize study? That was

a quite exceptional piece of work.

On 7 . . . 'ifdl + White has prepared

8.\i'c l . Black has to play for stale-
mate.
7. ... �h5xd5
8. �c8-c7+ 'ita7-a8
The stalemate is still not fore­
stalled.
9. �c7-a7+1
But now it is!
Benn6, WCCT-7 2004
9. ..
. 'it>a8xa7
White to play and win
The same position, now without
If White does not play carefully, the white queen.
Black will have a kamikaze queen. 1 0. e4xd5
Therefore he must first get rid of And wins.
his foremost a-pawn.
1. .l:tb7-b8+ �a8xa7 Avoiding stalemate is a useful mo­
2. .l:tb8-b7+ �a7-a8 tif for a disappearing trick.

241
The Art of t h e E n d g a m e

The knight on d 1 was sacrificed in

order to clear its square for the
rook - the rook in its turn has
made its original square available
for the bishop.
6. ... l:!:d5-e5
For the third time the rook returns
to the centre square.
7. �c1 -g51
The battle scene is shifted to the
Korolkov, 'Shakhmaty v SSSR ' 1938
g-file.
White to play and win
7. ... .Ue5xg5
Black is threatening to sacrifice his 8. J:[h1 -g1
rook on e8. There is only one way Forcing the rook to stay on the
to avoid the stalemate. g-file. Now there follow several
1. t2Je3-d51 l:!:e5xd5 systematic manoeuvres.
Now White has a move's respite 8. J:tg5-g2
before Black has a concrete stale­ 9. h3-h41
mate threat again.
2. t2Jd1 -e3
The second knight enters the stage. . .
2. ... l:td5-e5
3. t2Je3-d5!
and is sacrificed on the same
square.
3. ... J:te5xd5
4. J:[c1 -d1 l:td5-e5
5. J:td1-d51
The rook also chooses the central
square on which to be sacrificed. Reciprocal zugzwang. The black
5. ... .!:!e5xd5 rook has to move up.
6. �b2-c1 9. ... l:!:g2-g3
10 . .Ug1 -g21
And the white rook follows suit.
1 0. ... l:tg3-g4
11. .:tg2-g31 .Ug4xg3
There's nothing to be done about
it. There now follows a finale in
which White manages to neutral­
ize Black's rampant rook.
1 2. <j;lf8-e7 .Ug3-e3+
1 3. ctJe7-d7 J:te3-d3+

242
 The D i sappearing T r i ck

1 4. Wd7xc6 J:l:d3-d6+ Both white pieces are under attack.

15. 'it>c6-b5 .!:i.d6·b6+ It looks as if White can win easily
16. <;t>b5-c4 with I ..t a4 cxb4 2.c5, but with
2 . . .b3 3 .�xb3 tbxb3 4.c6 tbaS !
5 .c7 tbc6+ 6 . �d6 tba7 Black suc­
ceeds in stopping the c-pawn.
1. tt:l b4-c2!
White has no objections to playing
a pawn ending.
1. ... tt:la1 -b3
The only way to maintain the
knight. Now it appears that White
has not achieved anything special,
since after 2 .tbe3 tbas Black easily
The checks are exhausted. holds the draw.
16. ... .!:i.b6xf6 2. tll c 2·a1 !I
1 7. <;t>c4-b5 A tremendous surprise. White sac­
And wins. An impressive study. rifices his knight in order to domi­
There is also a version with a white nate the enemy knight. This minia­
queen on f3 instead of a knight on ture looks like Bazlov's, only there
e3 in the starting position. There the white knight was first played to
the first move is I . 'ifdS ! . the corner square and then sacri­
ficed. Here the corner square is
An interesting occasion for a disap­ where the sacrifice takes place.
pearing trick is when a piece is at­ 2. ... tt:lb3xa1
tacked by an enemy pawn. White Otherwise 3 .tbb3 + follows.
wants to prevent this capture by 3. �d1 ·a4 <;t>d2-c3
sacrificing the piece in question on 4. we5-d5 wc3·b4
another square. 5. �a4·d 1 !

Afek, 'Themes 64' 1977

White to play and win
Reciprocal zugzwang. White wins.

243
T h e Art o f the Endgame

My own contribution looks slightly The picture has changed dramati­

different, as the piece that will be cally. The ensuing bishop ending
sacrificed is not under attack yet. does not seem so bad for Black.
White sacrifices his bishop in antic­ However, he has a problem: his re­
ipation of that attack. mote king. White is going to ex­
ploit this factor.
6. �d4-f2
A 'switchback' by the bishop.
6. ... �g3-h2
Black should not consent to enter­
ing the pawn ending. The text move
seems a safe option. If White sets
his majority in motion, Black has
no problems. After 7.g4 fxg3 , each
move by the white bishop is an­
swered by 8 . . g2+, after which the
.

Timman 1999/2011
black bishop is again free to move.
White to play and win
7. �f2-g1 �h2-g3
To the old version I have added a Again he avoids the bishop trade.
starting move. 8. �g1 -h21!
1. d4-d5! An amazing version of the disap­
A thematic pawn sacrifice with pearing trick.
which White increases the action 8. ..
. �g3xh2
range of his bishop.
1. ... 'i*'c6xd5
2. l:te2-a2+!!
The beautiful point of the previous
move. White wins the queen.
2. 'it'd5xa2
3. �f2-d4+ 'it>a1 -b1
4. t2Jd1 -c3+ <ot>b1 -b2
5. t2Jc3xa2+ 'itb2xa2

9. g2-g41!
Recently I read about someone
who wanted to abolish the en
passant rule. The game of chess
would lose a lot of its beauty in
that case.
9. ..
. f4xg3
1 0. 'itf1 -g2

244
 T h e D i sapp earing T r i c k

The ppint. Without the white His best chance. Black allows the
bishop, the black bishop will re­ queen exchange and tries to obtain
main cut off from play, and the counterplay with his rooks.
white f-pawn has free passage. 2. 'ii'a 4-a6+ 'ii'e6xa6
1 0. ... �a2-b3 3. b5xa6+ Wb7-a81
11. f3-f4 �b3-c4 A clever move. The alternative
1 2. f4-f5 �c4-d5 3 . . . <;;t>xa6 was insufficient on ac­
1 3. f5-f6 count of 4 . ..ta3 (preventing
And wins. 4 . . . llbf8) , 4 . . J:I:h8 5 .tLlc3 l:!xh7
6.0-0-0! and wins.
Is it possible to make a study where 4. �c1 -a3
in the starting position no fewer Now too, White has to prevent the
than five white pieces are undevel­ black rook occupying the f8-square.
oped on the first rank? Costeff suc­
ceeded in composing such a study,
which is rich in content and strik­
ing in its originality.

4. ... .:bsxb1 +1
The magnificent point of Black's
previous move. Now, after 5 .l:!xb 1
Black forces stalemate by 5 . . . l:tfl + .
Costeff, WCCT-7 2004 5. �a3-c1 11
White to play and win
The ultimate disappearing trick.
Materially, White has a large advan­ The black rook has to go back.
tage, but particularly the far-ad­ 5. ... .l:!.b1 -b8
vanced black d-pawn creates dan­ 6. �c1 -b2
ger for his king. Therefore he will Preventing the rook check.
have to do everything he can to 6. ... .l:.b8-f8
achieve the exchange of queens. 7. 0-0-0
1. �d 1 -a41 The final point. With the transac­
After other moves White will have tion that started with 4.�a3 ,
a draw at best, for example: l .'i!fb3 White has lost his knight, eventu­
'ifd6 ! 2 .exf5 'i¥g3 + with a draw ally gaining time to get his king
by perpetual check. out of the danger zone by castling
1. ... .l:!.f5xf3 queenside. White wins.

245
The Art o f the Endgame

I succeeded in making a study Again, White has the same prob­

where both sides take refuge in lem. But there is an elegant way
the disappearing trick. Actually out:
this happened by coincidence. I 4. �g8-g2!
had been occupied with the fol­ A second queen sacrifice, which
lowing study by the Hungarian Black again has to accept.
endgame study composer Attila 4. ... f3xg2+
Koranyi. 5. Wf1 xg2 1/z-Vz

A modest study with an attractive

theme. Something more could be
gotten out of it.

Koranyi , match Hu ngary-Finland 1979

White to play and d raw

Black threatens to give mate. It's

clear that White has to queen.
Timman 2011
1. c7-c8'if+ 'it>c5-b4
Black to move - White draws
Now what? Only by sacrificing his
queen can White prevent mate in It looks hard to come up with
one. something for Black. His king is
2. 'i!Yc8·c4+! 'lt>b4xc4 caught in the crossfire of white
3. g7-g8�+ 'it>c4·b4 pieces. With the disappearing trick
he manages to turn the tables.
1. ..
. ..g5-g2+1
2. Wf1 xg2 f2·f1 �+
3. 'it>g2xf1 f4-f3
Suddenly White has great trouble
averting the mate threat. He cannot
play 4.a8Vi'+ on account of
4 . . . �a7 mate. So he has to start sac­
rificing on a grand scale.
4. J:!.c4-a4+ Wa5xa4
Now White follows up with a dis­
appearing trick.

246
 The D i sapp e a r i n g T ri c k

9. a7·a8'i!V+ 'it>a5xb6

5. �c6·c4+ <;t>a4·b3
6. J::[c4·b4+ wb3xa3
7. l:!.b4·a4+!
Not 7 J�b3 + ? Wa2 and Black wins. And now for the queen sacrifices:
7. ... \t>a3xa4 1 0. �a8xb7+ <;t>b6xb7
The same position as two moves 11. �e8-c6+!
earlier, only now without the A bishop sacrifice in between
white rook and knight. By the way, events.
Black had to capture the rook. After 11. ..
. ..ti>b7xc6
7 . . .';i(b2 8 .l::ta 2+ ! �b l 9 Jhh2 ! the 1 2. c7-c8'i¥+ \t>c6·b5
mate threat would have been elimi­ Now 1 3 .llt'e8+ fails to 1 3 . . . 'it>aS !
nated, and White would have won and White has no good checks left.
with his material surplus. 13. 'ik'c8·c4+!
8. b5·b6+ Koranyi's queen sacrifice.
The point. This discovered check 1 3. ... '1tb5xc4
renders the a7 -square inaccessible 1 4. g7-g8'i¥+ lt>c4xc3
for the black bishop. 1 5. 'ti' g8-g2!
8. ... <;t>a4-a5 White's third queen sacrifice.
Not 8 . . . 'it>a3 9 .a8"ii + �b2 1 5. ... f3xg2+
1 0 ."iia 2+ ! Wxa2 l l .g8'i¥+ fol­ 1 6. Wf1xg2
lowed by 1 2 .'ifg2 , and White wins. Draw.

247
Chapter 13

Three Themes

1. Troltzky/Amelung position White is a piece behind, and taking

An interesting question is: can we on g7 will not save him.
compose a position where one side 1. h5-h6!
is a rook, bishop and pawn up, it is This pawn sacrifice has a special in­
his move and yet he cannot win? I tention.
shall give two clues: stalemate is 1. ... g7xh6
not a theme, and without that 2. �e2-f3+ �a2·b1
pawn it isn't possible. 3. .l:!.g2-b2+1
In 1 89 6 this position was discov­ A rook sacrifice that serves to va­
ered twice: by the Baltic/German cate the g2-square for the king.
endgame theoretician Friedrich 3. ... 'ito>b1 xb2
Amelung in Riga, and by Alexey 4. �f3·g2
Troitzky in Smolensk. Each of them And the black rook has no squares.
published a modest study in that
year which showed the discovery: In Troitzky's study, the play unfolds
Amelung in the Diina Zeitung, and on the queenside.
Troitzky in Novoye Vremya. In those
times it was normal that new stud­
ies were published in daily news­
papers.
Amelung's study looked as fol­
lows:

Troitzky, 'Novoye Vremya' 1896

White to play and draw

Here also, a crucial square must be

cleared for the white king.
1. b2·b4+1 �c5xb5
2. l:!:a6xb6+1 cJtb5xb6
Amelung, 'Dun a Zeitung' 1896
3. �c3·b2
White to play and draw
And again the black rook is caught.

248
 T h ree Themes

Which of these studies appeared

first? We don't know the publica­
tion date of Amelung's study in the
Diina Zeitung, but we do have
something to go on: his study was
published again later, on May 1 7 ,
in Deu tsches Wochenschach.
Troitzky's study is not dated. In fact
this is a little strange.
In 1 89 6 , Troitzky published 5 1
Herbstman 1934
studies in Novoye Vremya. The
White to play and d raw
greater part of these were dated,
but precisely this one isn't. Because Black threatens to promote his
of this, we cannot say with cer­ e-pawn, and at the same time
tainty who was first with this dis­ White's rook is under attack.
covery. Somewhere in Russia there The only solution to both these
may be someone who knows; it's problems is a pin on the black
unlikely that all the newspapers e-pawn.
from before the Revolution have 1. l:lg6-g2 l:tf6xf4+
been lost. This seems to guarantee an easy
It does seem certain that Troitzky win. However, White can just save
and Amelung did not know each himself thanks to the Troitzky I
other's studies. News did not travel Amelung position.
so fast in those days, but there does 2. 'it>c4-d31
seem to have been an exchange of On the road to g2.
studies between Russia and the 2. ... l:tf4xf1
Baltic countries. Nevertheless, it 3. 'it>d3xe2
seems improbable to me that Driving the rook to the corner.
Troitzky and Amelung knew each 3. J:[f1 -h1
other's studies. Both composers 4. <;1;;> e2-f3+ <;1;;>a 2-b1
had a reputation to lose. If the 5. l:tg2-b2+!
drawing position had been already Amelung's sacrifice.
known, they would have had to 5. �b1 xb2
make an effort to compose a spe­ 6. 'it>f3-g2
cial study. As it was, it sufficed for Draw.
them to show the idea in a modest
form. For this study Herbstman received
third prize in the 'Communist
A fair number of studies have been Party Congress Tourney' published
made on this theme. Four decades in 64.
later, Herbstman came up with the This adaptation is not entirely con­
following adaptation. vincing. There are actually few sur-

249
T h e Art of t h e E n d g a m e

prises, and the black bishop is al­

ready poised to block the rook in
the final position.
It turns out to be not at all easy to
conceive of a fitting introduction.
The Finnish study composer Pauli
Perkonoja found this out when he
made the following study.

5. l:i.a2·a 1 !
An attractive rook sacrifice in order
to reach the famous position.
5. ... J:!.b1 xa1
6. 'i.itc3·b2
Draw.

I had the same trouble myself As

Perkonoja, 'Tidskrift for Schack' 1963
an illustration of the disappearing
White to play and d raw
trick I had invented the following
White will first have to take care of mini-study.
the black h-pawn. If he takes it
with the knight, Black has a pretty
win: l .ltJxh3 �dl ! (threatening
mate on b3) 2 . .l:!.f4 l:tb3 + 3 . �c4
a4 and the a-pawn cannot be
stopped, for example 4.llf7 �e2 +
followed by 5 .. a 3 and the black
.

king will assist in the pawn promo-

tion.
1. l:tc4·h4 �c2-d1
2. J:!.h4xh3+ �e3·f2
Timman, 'New In Chess' 2004
So far everything is clear. However
White to play and d raw
White now has two roads to a
draw. The intended solution was:
3. J:!.h3·h2+ 1. J:!.b3·b2+ �g2xg1
The thematic way. The alternative 2. J:!.b2·g2+1 'iti>g1 xg2
3 . �d2 was sufficient too. Two white pieces have disap­
3. ..
. 'iti>f2xg1 peared.
4. J:!.h2·a2 a5-a4 3. 'iti>c3-b2 'l2- 'l2

250
 T hree T h emes

l .l:f.oS also does the trick. In order

to exclude this solution, a black
pawn on a6 can be added.

In the following mini-study, White

manages to avoid the Troitzky I
Amelung position.

Timman 2010
White to play and d raw

First of all White has to stop the

h-pawn.
1. jLa4-c6
Now, on 1 . . . l:.b6 White has planned
2 .�d5 , and l . ..�bS 2..�e4 also
leads to a problem-free draw. How­
Timman 2002
ever, Black has a stronger move.
Black to move - White wins
1. . .. �e2-d1
It is tempting to force a rook Now what? This question brought
exchange by l . . .l:i.b7 , but after many strong players to despair when
2 . l:.xb7 Wxb 7 3 .�d8 ! White I confronted them with this posi­
wins. So Black has to try some­ tion. If White captures on h2 , there
thing else. follows a decisive check on c 1 .
1. ... a7-a5 2. �c6-d7+11
2. b4xa5 .l:rd7-a7 Simply amazing! White sacrifices
If White now plays the careless his bishop in order to drive the
3 .�d8 , there follows 3 . . . l:i.a8 ! with black king to a less favourable
a draw. square.
3. .l:rb8-c8+! 2. ... �c8xd7
An important zwischenschach. 3. .!:lc2xh2
3. ... �c6-d6
Black has to keep control of the
square cS .
4. i.h4-d8 �d6-d7
5. .l:!.c8-b8
And wins.

So far my contribution to this

theme had been modest. The fol­
lowing study is more ambitious.

25 1
T h e Art o f t h e E n d g a m e

At first glance it doesn't look as if White has to give his knight for the
White will be able to save himsel£ d-pawn.
However, Black has trouble keep­ 1. tt:Je2-c3 d3-d2
ing the coordination of his forces 2. .!:!h8-h2! d2-d 1 �
intact. The a-pawn is indirectly at­ 3. tt:Jc3xd1 .ib3xd1 +
tacked and 3 . . . �c6 is satisfactorily 4. 'it>b4-c3!
met by 4J:th 1 , when, due to the And the position has been reached.
pin along the first rank, Black can­ For this study Matous received sec­
not avoid the exchange of rooks. ond prize in the Reti Memorial. I
3. ... a7-a5 could console myself with the
Now 4 . .l::rh 1 can be met by 4 . . . a4. thought that my introductory play
Then S .Wd2 is decisively answered was more surprising and probably
by 5 . . a3 , since the a-pawn be­
. also better.
comes too strong. Therefore White
shuffles the cards. 2. Valladao Task
4. l:Ih2-a21 In Chapter 5 I already mentioned
Forcing the a-pawn further forward. this task. Like the Novotny and
4. ... a5-a4 Plachutta themes, the Valladao Task
5. .!:!a2-a1 11 stems from the world of problem
Perkonoja's rook sacrifice. chess, only it is of a much more re­
5. ... �b1 xa 1 cent date. In 1 964, the Brazilian
6. �c3-b2 '1>- '1, Joaquim Valladao Monteiro com­
posed the following problem.
I was very pleased with this study.
The play is lively, and the position
after White's third move is intrigu­
ing. Great was my disappointment
when later I saw that this position
had already been discovered by
someone else, 3 1 years earlier.

Valladao Monteiro, '0 Globo' 1964

M ate i n 2

The key move is

1. g2-g4
Threatening to give mate immedi­
ately by promoting on a8. Black
Matous 1979 has only two moves to postpone
White to play and draw the mate.

252
 T h r e e T h emes

A) . l . . . hxg3 2 .h8'ii' mate; 6. f3-f4 �f2-f3

B) l . . ..ta2 2 . 0-0-0+ and Black is
. 7. f4-f5 wf3-f4
mated as well. 8. f5-f6 wf4-f5
Not a very impressive problem. 9. f6-f7 wf5-g6
Van der Heijden claimed that this
task is better suited for endgame
studies. Moreover, he specified the
composer's task further: to the cas­
tling move and the en passant cap­
ture, a minor promotion should be
added.

Not many studies have been made

with this criterion so far. It isn't
easy to add new points if all the
'Valladao ingredients' are already If White now queens, Black is
present. A study by Costeff is rela­ stalemated; after a bishop promo­
tively highly valued. tion, Black has a fortress.
1 0. f7-f8.U!
The only correct promotion. White
wins.

Afek has also fulfilled the Valladao

Task in a nice way.

Costeff, 'The Problemist' 2006

White to play and win

The first move is obvious.

1. 0-0-0+ g2-g1 'iV
2. .Ud1 xg 1 + Wh1 xg1
3. �c1 -b2 d5xe4
Afek 2011
4. f2-f4!
White to play and win
Not 4.f3 in view of 4 . . . e3 , with a
draw. The black pawns look threatening.
4. ... e4xf3 White will have to play for mate.
5. e2xf3 Wg1 -f2 1. g2-g4+ f4xg3
Now the black king is going to In order to prevent mate, Black has
pursue the f-pawn. to take en passant.

253
T h e A r t o f the E n d g a m e

2. 0-0+ We know the construction in the

Two-thirds of the task has been lower left-hand corner from Chap­
completed. ter 5 . It is clear that White has to
2. ... �f5-g6 castle.
3. l:l:e7-g7+ �g6-h5 1. 0-0 c7-c5
4. f6-f7 c2-c1 'fV 2. e2-e4!
The alternative 2 .e3 is too slow. Af­
ter 2 . . .f3 ! White would not pro­
mote in time at the end of the line.
2. ... f4xe3
After 2 . . .f3 3 .e5 c4 4.e6 h5 5 .e7 h4
6.e8if h3 7 .'ilie5 h5 8 .�d2 h4
9 .�c 1 Black would also be mated.
3. f2-f41
The f-pawn has free passage.
3. ... e3-e2
4. .!:!:f1 -e1 c5-c4
Now for the final touch. 5. f4-f5 h6-h5
5. l:lg7-h7+ �h5-g6 6. f5-f6 h5-h4
6. f7-f8l2J 7. f6-f7 h7-h5
Mate.
When I worked on my own inter­
pretation of the task I decided, in
imitation of Zinar, to focus on the
trusted territory of bishop and
knight promotions. The castling
move and the en passant capture
could then be fitted in, so I
thought.

8. f7-f8..t!l
Excelsior.
8. ... h4-h3
9. �f8-h6 h5-h4
1 0. �h6-c1 1
The well-known point.
1 0. ... �b1 -e4
11. �c1 xb2
Mate.
Timman 2011
White to play and win
I also made a livelier version.

254
 T h ree T h emes

On its way to f8 .
4. ... g3·g2
5. .!:tf1 -e1 d6xe5
There's nothing for it.
6. f4-f5 e5-e4
7. f5-f6 e4-e3
8. f6·f7 e3·e2

Timman 2011
White to play and win

Black threatens to win with a

queen check on b 1 .
1. 0·0+ �g6-b1
Black has to brick himself in, since
after l .. .'ifb l 2. .�xb2+ he cannot
recapture on account of mate on 9. f7·f8j_!
a6. Again this combination of an Ex­
2. 'ife2xe5 d7·d6 celsior and a bishop promotion.
If White now moves the queen, he 9. ... c5-c4
will be too late, since his plan to 1 0. �f8-h6 c7·c5
move his bishop to c 1 is prevented 11. �h6·c1
by Black's pawn mass. Therefore he Just in time.
has to sacrifice the queen. 11. ... �b1 ·d3
3. g2-g4!! f4xg3 1 2. �c1xb2#
The alternative 3 . . . dxe5 wouldn't
help. After 4.g5 e4 5 .g6 f3 6.g7 e3 3. Uquldatlon to a pawn ending
7 .g8'if exf2+ 8 .'it>h2 ! c4 9.'ifg7 c5 This is a theme that has been sub­
1 O .�d2 ! c3 1 1 .'iixc3 c4 1 2...�.c 1 limely adapted by Shaya Kozlowski.
he would be mated.
4. f2·f4!

Kozlowski , 'Gios Poranny' 1931

White to play and win

255
The Art of the E n d g a m e

The black bishop is boxed in on its Both sides have promoted their
starting square, and it will never pawns. Now White forces the ex­
leave there. change of queens by a series of
1. tt:Je5-g6+1 checks.
A positional knight sacrifice. After 1 0. �e8-e4+ c.th7-h8
other moves Black would free him­ 11. 'ife4-e5+ 'it>h8-h 7
selfby pushing his g-pawn. 1 2. �e5-f5+
1. ... h7xg6+ Step by step the queen approaches.
2. c.tf5xg6 c;t>h8-g8 1 2. ... <ot>h7-h8
3. h4-h5 13. l\ff5-f6+ �h8-h 7
Mutual zugzwang. If it were his 14. �f6-g6+
move, White would lose. And White wins.
3. ... c;t>g8-h8
4. 'it>g6-f7 Even more interesting is the fol­
lowing miniature study.

4. ... g7-g5
Kozlowski, 'Gios Poranny' 1931
His best chance. A pawn race follows.
White to play and win
5. 'it>f7xf8 g5-g4
6. \it?f8xe7 g4-g3 Without the black f-pawn this po­
7. 'it>e7-f7 g3-g2 sition would be drawn. As in the
8. e6-e7 g2-g1 'iY previous study, White is going to
9. e7-e8�+ Wh8-h7 cut off the enemy bishop from
play.
1. tt:Je5-d7+ 'iit> b8-c7
2. tt:Jd7-f811
Amazing. White locks in his own
knight in order to deprive the
black bishop of the h7 -square.
2. ... 'it>c7-d8
The alternative was 2 . . . Wd6. Then
White wins with 3 . �g4 �d5
4.Wh5 , and White has the opposi-

256
 T h r e e Themes

tion, Since the two minor pieces White is threatened with the loss
are condemned to keep each other of his extra piece.
imprisoned, this position can be 1. tba7-c8
treated as a pawn ending. The only move to avoid this.
3. �f3-f4 1. ... �b6-a5!
Also good is 3 . 'it>g 4. A subtle defence. Before he cap­
3. ... 'it>d8-e8 tures the king's bishop, Black
4. Wf4-g5 cJI>e8xf8 drives the other bishop to a less fa­
5. 'lt>g5-h6 vourable square.
2. �b4xa5+ 'it>d8xd7
Now White has a problem. If he
plays 3 . tt:Ja 7 , there follows 3 . . .f6 !
4.tLlb5 �h7 5 .tLld4 <Jtd6, and he
cannot maintain his f-pawn.
3. tbc8-b6+1
Preparing a bishop sacrifice.
3. ... 'it>d7-c6

Mutual zugzwang. White liqui­

dates into a winning pawn ending.
Without the black bishop it would
be a draw.

I reckoned that it ought to be pos­

sible to elaborate on this study.
With a little extra material, some
new points can be added. 4. f5-f6!
Fixing the black f-pawn.
4. ... 'it>c6-b5
5. tt:lb6-d71
The point of the previous move.
The knight is on its way to f8 .
5. ... 'it>b5xa5
6. tbd7-f81
Here also, with both kings further
removed, Kozlowski's concept
works.
6. ... 'it>a5-b6
Timman 2011
7. �e1 -f2
White to play and win (after Kozlowski)
The only good king move.

257
The Art of t h e E n d g a m e

7. ... wb6-c5 1. d5-d6 �c7-d8

The alternative 7 . . . �c7 would lead Forced, since all rook endings are
to Kozlowski's main line. winning for White.
2. wa6-a7
Not immediately 2 JH4+ 'it>e3
3 .l::rf8 , as then Black can save him­
self just in time with 3 . . . �c7 .
2. ..
. J::t b8-c8
3. .!:td4-f4+
Insufficient for the win was 3 .'i.tOJb7
in view of 3 . . . .l:!.c4.
3. ... 'it>f2-e3
4. J:[f4-f8
It looks as if the matter is settled.
8. 'it>f2-g31 What can Black do against the
Again the only move. The king is threat of 5 .�b7 ?
on its way to h4. 4. ... b6-b5
8. ... ..t>c5-d5 5. 'iita 7-b7 �d8-c7!
9. 'it>g3-h4!
Thus White gains the opposition.
9. �d5-e4
1 0. 'it>h4-g4 ..t>e4-d4
11. 'it>g4-h51 �d4-e5
12. 'iith 5-g5 1 -0

Obviously we can conceive of

countless studies where White liq­
uidates into a pawn ending.

The point of the previous move. If

White takes the rook, he cannot
win, and after 6.dxc7 .l:f.xf8 he even
has to be careful not to lose.
6. l::tf8-f3+11
Terrific. White drives the black king
further away from the b-pawns.
6. ... �e3xf3
7. d6xc7
Timman 2011 Now the pawn endgame is won by
White to play and win
the narrowest of margins.
White must aim for maximum 7. ... l:.c8-f8
domination of the enemy pieces. 8. c7-c8'ii' J:rf8xc8

258
 Th ree Themes

9. 'it>b7xc8 �f3-e4 1. ... h7xg6

1 0. �c8-c7 2. �h2-b8!!
Also good is I 0. 'it>b7 . The same problem move we saw
1 0. ... 'it>e4-d5 earlier on. Insufficient was 2 .Jl.c 7 ?
11. 'it>c7-b6 'it>d5-c4 in view of 2 . . . '1l!Yd4! 3 . cxd4 l:txc7 +
1 2. 'it>b6-a5 and Black wins.
2. ... 1Wa7-d41
Here, too, the only defence.
3. c3xd4 �a5-d2+
Black has to keep sacrificing, or
else mate is unavoidable.
4. �c1 xd2 tt:Ja3-c4+

An elementary mutual zugzwang

situation. White wins.

Even with a board full of pieces,

eventually a liquidation into a
pawn ending may be possible.
White has to take the knight, as
otherwise Black will play his king
to f8. If Black controls the e5-
square, his king can escape via g 7 .
5. b3xc4 tt:Je3xf1 +
6. 'it>d2-c3!
The best square for the king.
6. ... l:!.b7-b3+
Black has to keep sacrificing in or­
der to avoid mate.
Timman 2011
7. ..t>c3xb3 l:ra8xb8+
White to play and win
8. ..t>b3-c3 �e8-d8
The same construction as we saw Finally Black has managed to bring
in Chapter I I Black is threatening
. his king into safety. Now, however,
to bring his king to safety by cas­ White coolly liquidates.
tling queenside. 9. l:r h 1 xf1 d5xc4
1. 'ifb1 -g6+! 1 0. l:[f1 -f8+ �d8-c7
White is willing to sacrifice a full 11. l:[f8xb8 ..t>c7xb8
queen just to open the h-file. 1 2. 'it>c3xc4 1 -0

259
The Art of t h e E n d g a m e

Chapter 14

Various Endgame Studies

1. Studies with others

Some endgame study composers
regularly create studies in collabo­
ration with colleagues. I have done
this only very rarely.
One example of such a collabora­
tion was a new, elaborated version
that Jan van Reek made of one of
my miniatures.

Now the lines divide into two:

A) 4 . . . �g6 S . ..ihs tbxeS
6. �a 1 , and we reach a mirror
version of the final position from
the first chapter, with an extra
black pawn;
B) 4 . . c2+ S .'iitx c2 ..ig6 6.�h5
.

tbxeS 7 . <bc3 ! and wins. Again, the

white king is unassailable.
TimmanjVan Reek 1994
For this study Van Reek and I re­
White to play and win
ceived first prize in the Wola
In order to take control of the situ­ Gulowska Tourney.
ation, White has to block the
c-pawn. Israeli Grandmaster Emil Sutovsky
1. �d2-c3 �c7xe5 is known to be well-disposed to­
Black sacrifices a bishop in order to wards endgame studies. Together
create a passage for the c-pawn with Boris Gelfand, he has com­
anyway. posed one magnificent study, pub­
2. �c3xe5 c4-c3+ lished in New In Chess Magazine.
3. �a2-b1 During the closing ceremony of
The only good square for the king. the third Inventi tournament in
3. ... tt:Jb2·d3 Antwerp he showed me an inter­
4. l:lf4xf5 esting idea for a finish.

260
 Various Endgame Studies

�xf3, and White does not have

enough pawns.
1. ... g6-g5
With I ...�g2 Black could have
aimed for further simplifications.
But then White liquidates to a win­
ning rook ending: 2.g4+ �h4
3 ..txg2 lhg2 4Jhg6 �xh3
5..l:.xg7, and the white f-pawn de­
cides the battle.
Sutovsky
2. .ltf1-d3 l:Ic2-c5+
Black to move achieves a surprising In order to avoid mate, Black has to
draw with l .. g6! 2..1:hg6 �xf5+!
. give the exchange.
3.�xf5 g4, and in order to prevent 3. 'it>e5-d4 J::ic5-d5+
stalemate White has to consent to 4. '.td4-c4 J::id5xd3
playing a pawn ending that offers 5. �c4xd3
him no chances.

Working backwards, I made the

following study, which shows the
idea in a slightly more pronounced
way.

5. ... �c6-g2
The bishop harasses the white
pawns from behind .
Inadequate was 5 ...�d7 in view of
6.f5 g4 7 .h4 g5 8 .l:.e5!
. gxh4
Sutovsky(Timman 2009 9 .gxh4 and again, the f-pawn de­
White to play and win
cides.
Material is depleted. White must 6. f4-f51
exploit the exposed position of the White always has to strive to ob­
black king. tain a strong f-pawn.
1. f3-f41 6. ... �g2xh3
Insufficient for the win was l.�d3 7. 'it>d3-e4 g7-g61
on account of 1....l:f.c5+ 2. �d4 Here we have Sutovsky's clever dis­
.:l.d5+ 3 . �c4 l:!.xd3! 4. �xd3 covery.

261
The Art o f the Endgame

1. ... �h8·h7
2. �b3-f71
By preparing a pin on the black
bishop, White manages to keep his
piece.
2. ... wh7xh6
3. .l:ta4-h4 J::l.d 8·h8
An interesting defence. Black could
also have given a series of checks,
but this wouldn't have brought
However, now that the black him anything in the end, for in­
bishop is on h3 instead of g4, stance: 3 . . . I:td3 + 4.'1t>g2 .l:i.d2+
White has something up his sleeve. S .'�g l .!:ld l + 6 . �f2 .!:rd6 7 .llx h5 +
.

8. g3·g4+! c;t>g? 8 .�b3 l:!h6 9 J:tg5+ Wf6

This little check eliminates the I O.l:.g2 l:!.g6 l l .l:!.g3 ! and if Black
stalemate threat. exchanges rooks, White takes back
8. ... �h3xg4 with the h-pawn.
9. .l:te6xg6 1 ·0 4. �f7xh5 .l:th8·h7!

2. Studies with no
specific theme
In the following study, the concen­
tration of pieces at the edge of the
board is striking. As if by a magnet,
the remaining pieces are also
drawn to the h-file.

A curious situation. White has a

hard time avoiding the exchange
of the rooks. For the time being the
pieces, crammed together on the
edge of the board, are condemned
to each other's company. Black's
threat is 4 . . . ®gs .
5. 'Wth3·g4 .l:th7-g7+
Timman 1990
6. 'it>g4-f31
Black to move - White wins
The only winning move. Obvi­
As Black will not win anything ously, 6.®f5 failed to 6 . . . l:Ig5 + ,
back after l . . . .::l: d 3+ 2.'it?h4, he has but 6 .'1t>f4 was harder to refute.
to attack the knight first. Black achieves the draw with

262
 Va r i o u s E n d g a m e S t u d i e s

6.,.e5+ 7 .< �e4 Wg5 s J:tg4+ Wh6 White cannot take on c7 immedi­
9 .h3 b5 ! 1 0 .�e8 1:1xg4+ l l .hxg4 ately, since after l .bxc7 �h4+ he
'it>g5 1 2 .Wf3 e4+ 1 3 .'lt>g3 b4 and cannot protect his king from rook
the black pawns are too far ad­ checks.
vanced for the white king and 1. �c3-b4+ Wd6-d5
bishop to block them effectively. The only square for the king.
6. ... l:tg7-h7 2. �b5-c4+!
After 6 .. .'�g5 7 J �g4+ Wh6 8 .h3 ! With this bishop sacrifice, White
White would win without trouble. blocks the c-file for the black rook.
7. wf3-f4! 2. ... Wd5xc4
3. b6xc7

Now this is possible. Insufficient was

7 ..ie8+ <;t>g? 8.1:1g4+ Wf8 9 . .id7 Now Black has two bishop checks,
e5. with a draw soon to follow. which White has to play accurately
7. ... e7-e5+ to parry.
8. wf4-f5 3. ... i.g3-h4+
And wins. On 3 . . . .te5 + White has to prevent
Black from giving a winning rook
In the following study, many check. The play continues as fol­
themes are combined. lows: 4.We7 ! �d6+ ! s . <;t>f6 �e7 +
6.'iitf5 ! 1:1d5 + 7 .Wf4! , with a draw.
The black rook blocks the long di­
agonal for the black queen's
bishop, and so the c-pawn cannot
be stopped.
4. wf6-f5 �f3-e4+
Without this bishop sacrifice Black
cannot keep any winning chances.
After 4 . . . 1:1d5 + 5 .'lt>f4 it would be
an immediate draw.
Timman 2010 5. 'ot>f5xe4 l:td2-e2+
White to play and d raw 6. We4-f5 J:te2-e8

263
The Art o f the Endgame

1. �g4-f3+ wg2-g3
2. .ia3·d6+1
Before the bishop goes to cS , the
black king has to be driven to a less
favourable square. Later we will see
why.
2. ... Wg3-h4
3. �d6-e7+1
This check is necessary too.
3. ... Wh4-h3
4. �e7-c5
It looks as if Black will triumph.
However, White has a beautiful fi­
nesse up his sleeve.
7. �b4·e71
By sacrificing his second bishop,
White conquers the e6-square for
his king.
7. �h4xe7
8. wf5·e6
Draw.

Sometimes a strong passed pawn

can almost compensate for a rook. Only now. The king is not well
placed on h3 .
4. . . . �e8·b5+1
Without this surprising finesse,
White would attain the draw with
ease.
5. wc4xb5 l:tg8-g5
An unpleasant pin. Now, however,
it is White's turn to sacrifice a
bishop.
6. �f3·d51 l:tg5xd5
7. Wb5-c6
Timman 2011
The point of the bishop sacrifice.
White to play and draw
Black is forced to take on cS with
White will have to lift the blockade the rook, closing in the bishop on
on a7 by playing his queen's a7.
bishop to cS . But first he has to do 7. ... l:d5xc5+
something about the unprotected 8. Wc6-b7 .!:!.c5-b5+
position of his other bishop. 9. �b7xa7

264
 Various Endgame Studies

9. ... 'i.t>h3-g4
1 0. e2-e4!
Cutting off the king. After 1 O .�a8
'lt>f5 1 1 .e4+ We5 1 2 .a7 c;t>d6
1 3 .e5+ �c7 Black would have
time to catch the white king in a
mating net.
1 0. .. . �g4-f4
11. e4-e51
According to the same principle.
Without the white e-pawn it 11. ... �f4-f5
would be a draw without further 12. e5-e6
ado. But the extra pawn may play The last accurate move. After
tricks on White. Thus, Black would 1 2 .�a8? �e6 1 3 .a7 'ltd? 1 4.e6+
be winning if his king were on g3 rtlc7 1 5 .e7 .ti.e5 White would get
or g4. Then he would succeed in mated after all.
blocking the e-pawn. Draw.

265
Explanation of Sym bols

0 White to move
• Black to move
Good move
!! Excellent move
? Bad move
?? Blunder
!? Interesting move
?! Dubious move
\t> King
'ii Queen
: Rook
� Bishop
ttJ Knight

The chess board with its coordinates :

8
7
6
5
4
3
2
1

a b c d e f g h

266
 Index of Composers
Numbers refer to pages

A Koranyi 246 s
Afek 243, 2 5 3 Korolkov 2 0 6 , 2 1 4-2 1 5 , Selesniev 1 69
Amelung 248 242 Shmulenson 93
Kotov 200-202 Smyslov 1 03 - 1 05, 1 8 3
B Kotov/Mitrofanov 1 99 Sochniev 1 23 , 2 3 9
Bazlov 2 3 1 , 234 Kozlowski 235, 255-256 Sutovsky/Timman 261
Benko 39, 70, 1 02 Krabbe 36
Benno 241 Kralin 238 T
Bondarenko/Kakovin 200 Kralin/Pervakov 225 Timman 1 2 , 1 4- 1 9 , 2 3 ,
Bron 52, 2 1 6, 229 Kubbel 147, 1 65 , 1 68, 3 2 - 3 3 , 3 5 , 3 7 , 40, 4 1 , 44,
224-225 48, 50, 5 1 , 5 3 , 5 5 - 6 0 , 6 2 ,
C Kuriatnikov 1 75 64- 6 7 , 7 1 , 7 2 , 74, 7 5 , 7 9 ,
Costeff 1 09 , 1 1 3 , 245, 2 5 3 Kuznetsov,AG/ 8 1 , 9 7 , 9 9 , 1 04, 1 0 8 , I l l ,
Sakharov 1 43 1 1 3 - 1 1 8 , 1 2 1 , 1 2 2 , 1 24,
D 1 2 6, 1 2 7 , 1 3 0 , 1 3 1 , 1 3 3 ,
De Feijter 1 39 L 1 34, 1 3 6 , 1 3 7 , 1 4 1 , 1 42 ,
Dobrescu 230, 232 liburkin I I , 8 7- 8 8 , 2 0 5 , 1 44, 1 45 , 1 5 0- 1 5 3 , 1 5 5 ,
Dolgov 212 2 1 8-2 1 9 1 5 6, 1 5 8 , 1 59, 1 6 1 , 1 63 ,
Dolgov/Kuznetsov,AP 2 1 3 Lommer 8 5 - 8 6 , 89-90, 1 6 5 - 1 6 7 , 1 7 0- 1 7 3 , 1 7 6,
1 0 6, 1 1 2 1 7 8 , 1 8 0 , 1 84, 1 8 5 , 1 8 9 ,
E 1 9 1 , 1 93 , 1 95 , 1 96, 203,
Evreinov 1 2 8- 1 2 9 M 2 04, 2 1 1 , 2 1 5 , 2 1 6 , 2 2 1 ,
Manella 78 2 2 2 , 2 2 7 , 2 3 3 , 2 3 6, 2 3 7 ,
F Matous 1 3 , 1 23 , 252 244, 246, 2 5 0, 2 5 1 , 2 5 4,
Fernhout 136 Mattison 63 2 5 5 , 2 5 7- 2 5 9 , 2 6 2-264
Fritz,F 83 Mitrofanov /KorolkovI Timman/Krabbe 36
Fritz,] 77, I 75 Dolgov 208 Timman/Van Reek 260
Troitzky 84, 248
G N Troitzky /Korolkov I 07
Gorgiev 207 Nadareishvili 38, 1 72
Grinblat 54 Novikov 85 u
Gunst 139 Novotny 69 Ulrichsen 91
Gurgenidze 209
Gurvich 1 00, 1 2 6 p v
Perkonoja 250 Valladao Monteiro 252
H Pervakov 2 2 7 , 240
Hasek 1 84 Pogosyants 44 w
Herbstman 49 , 7 3 , 249 Prokes 34 Wotawa 6 1 , 63, 1 20, 236
Herland 82
Hurtig 1 00 R z
Ratner 138 Zakharov 1 92
K Rinck 75 Zinar 9 1 -9 2 , 94-96, I 0 6 ,
Kasparyan 1 2 0, 207 Rokirovkin 101 1 10
Katsnelson 202, 2 1 8 Rusinek 8 0 , 99, Zinar/Didukh 93
Kliatskin 1 69 1 48 - 1 49 Zoltan 77

267
Acknowledgements
The studies by Timman ( 1 983) and Matous 1 984 (Chapter 1 ) , Timman
1 984 and Timman 1 990 (Chapter 7 ) , Timman 1 9 84 and Timman 1 98 6
(Chapter 8 ) and Fritz 1 95 5 and Timman 1 9 84 (2x) have been published
before in Schaakwerk II (Timman, 1 9 9 1 ) .
In Chapter 2 , the study by Timman (20 1 0) was published before in
New In Chess 2 0 1 0 / 8 , the rest of this chapter was published before in
New In Chess 2 0 1 0 / 6 .
I n Chapter 3 , the study Timman 1 99 2 was published before in New In
Chess 1 994/ 3 , the Ortueta-Sanz analysis was published in New In Chess
1 99 7 / 3 , and Timman 1 99 7 and Timman 1 999 in New in Chess 1 999/ 5 .
In Chapter 6 , the article 'Smyslov as an Endgame Composer' up to the
study Smyslov 1 9 7 6 was published before in New In Chess 2 0 1 0 I 3 .
Here and there, slight alterations have been made.

269