IMO 2016 Solution Notes

Evan Chen《陳誼廷》
15 December 2024

This is a compilation of solutions for the 2016 IMO. The ideas of the
solution are a mix of my own work, the solutions provided by the competition
organizers, and solutions found by the community. However, all the writing
is maintained by me.
These notes will tend to be a bit more advanced and terse than the “official”
solutions from the organizers. In particular, if a theorem or technique is not
known to beginners but is still considered “standard”, then I often prefer to
use this theory anyways, rather than try to work around or conceal it. For
example, in geometry problems I typically use directed angles without further
comment, rather than awkwardly work around configuration issues. Similarly,
sentences like “let R denote the set of real numbers” are typically omitted
entirely.
Corrections and comments are welcome!

Contents
0 Problems 2

1 Solutions to Day 1 3
1.1 IMO 2016/1, proposed by Art Waeterschoot (BEL) . . . . . . . . . . . . . 3
1.2 IMO 2016/2, proposed by Trevor Tau (AUS) . . . . . . . . . . . . . . . . 5
1.3 IMO 2016/3, proposed by Aleksandr Gaifullin (RUS) . . . . . . . . . . . . 7

2 Solutions to Day 2 8
2.1 IMO 2016/4, proposed by Gerhard Woeginger (LUX) . . . . . . . . . . . . 8
2.2 IMO 2016/5, proposed by Russia . . . . . . . . . . . . . . . . . . . . . . . 9
2.3 IMO 2016/6, proposed by Josef Tkadlec (CZE) . . . . . . . . . . . . . . . 11

1
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

§0 Problems
1. In convex pentagon ABCDE with ∠B > 90◦ , let F be a point on AC such that
∠F BC = 90◦ . It is given that F A = F B, DA = DC, EA = ED, and rays AC
and AD trisect ∠BAE. Let M be the midpoint of CF . Let X be the point such
that AM XE is a parallelogram. Show that F X, EM , BD are concurrent.

2. Find all integers n for which each cell of n × n table can be filled with one of the
letters I, M and O in such a way that:
• In each row and column, one third of the entries are I, one third are M and
one third are O; and
• in any diagonal, if the number of entries on the diagonal is a multiple of three,
then one third of the entries are I, one third are M and one third are O.
Note that an n × n table has 4n − 2 diagonals.

3. Let P = A1 A2 . . . Ak be a convex polygon in the plane. The vertices A1 , A2 , . . . ,

Ak have integral coordinates and lie on a circle. Let S be the area of P . An odd
positive integer n is given such that the squares of the side lengths of P are integers
divisible by n. Prove that 2S is an integer divisible by n.

4. A set of positive integers is called fragrant if it contains at least two elements and
each of its elements has a prime factor in common with at least one of the other
elements. Let P (n) = n2 + n + 1. What is the smallest possible positive integer
value of b such that there exists a non-negative integer a for which the set

{P (a + 1), P (a + 2), . . . , P (a + b)}

is fragrant?

5. The equation

(x − 1)(x − 2) . . . (x − 2016) = (x − 1)(x − 2) . . . (x − 2016)

is written on the board, with 2016 linear factors on each side. What is the least
possible value of k for which it is possible to erase exactly k of these 4032 linear
factors so that at least one factor remains on each side and the resulting equation
has no real solutions?

6. There are n ≥ 2 line segments in the plane such that every two segments cross
and no three segments meet at a point. Geoff has to choose an endpoint of each
segment and place a frog on it facing the other endpoint. Then he will clap his
hands n − 1 times. Every time he claps, each frog will immediately jump forward
to the next intersection point on its segment. Frogs never change the direction of
their jumps. Geoff wishes to place the frogs in such a way that no two of them will
ever occupy the same intersection point at the same time.
(a) Prove that Geoff can always fulfill his wish if n is odd.
(b) Prove that Geoff can never fulfill his wish if n is even.

2
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

§1 Solutions to Day 1
§1.1 IMO 2016/1, proposed by Art Waeterschoot (BEL)
Available online at https://aops.com/community/p6637656.

Problem statement

In convex pentagon ABCDE with ∠B > 90◦ , let F be a point on AC such that
∠F BC = 90◦ . It is given that F A = F B, DA = DC, EA = ED, and rays AC and
AD trisect ∠BAE. Let M be the midpoint of CF . Let X be the point such that
AM XE is a parallelogram. Show that F X, EM , BD are concurrent.

Here is a “long” solution which I think shows where the “power” in the configuration
comes from (it should be possible to come up with shorter solutions by cutting more
directly to the desired conclusion). Throughout the proof, we let
θ = ∠F AB = ∠F BA = ∠DAC = ∠DCA = ∠EAD = ∠EDA.
We begin by focusing just on ABCD with point F , ignoring for now the points
E and X (and to some extent even point M ). It turns out this is a very familiar
configuration.

Lemma (Central lemma)

The points F and C are the incenter and A-excenter of 4DAB. Moreover, 4DAB
is isosceles with DA = DB.

Proof. The proof uses three observations:

• We already know that F AC is the angle bisector of ∠ABD.
• We were given ∠F BC = 90◦ .
• Next, note that 4AF B ∼ 4ADC (they are similar isosceles triangles). From this
it follows that AF · AC = AB · AD.
These three facts, together with F lying inside 4ABD, are enough to imply the result.

B C

3
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

Corollary
The point M is the midpoint of arc BD
d of (DAB), and the center of cyclic quadri-
lateral F DCB.

Proof. Fact 5.

Using these observations as the anchor for everything that follows, we now prove several
claims about X and E in succession.

A X

B C

Claim — Point E is the midpoint of arc AD

d in (ABM D), and hence lies on ray
BF .

Proof. This follows from ∠EDA = θ = ∠EBA.

Claim — Points X is the second intersection of ray ED with (BF DC).

Proof. First, ED k AC already since ∠AED = 180◦ − 2θ and ∠CAE = 2θ.

Now since DB = DA, we get M B = M D = ED = EA. Thus, M X = AE = M B, so
X also lies on the circle (BF DC) centered at M .

Claim — The quadrilateral EXM F is an isosceles trapezoid.

Proof. We already know EX k F M . Since ∠EF A = 180◦ − ∠AF B = 2θ = ∠F AE, we

have EF = EA as well (and F 6= A). As EXM A was a parallelogram, it follows EXM F
is an isosceles trapezoid.

The problem then follows by radical axis theorem on the three circles (AEDM B),
(BF DXC) and (EXM F ).

4
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

§1.2 IMO 2016/2, proposed by Trevor Tau (AUS)

Available online at https://aops.com/community/p6637677.

Problem statement

Find all integers n for which each cell of n × n table can be filled with one of the
letters I, M and O in such a way that:

• In each row and column, one third of the entries are I, one third are M and
one third are O; and

• in any diagonal, if the number of entries on the diagonal is a multiple of three,

then one third of the entries are I, one third are M and one third are O.

Note that an n × n table has 4n − 2 diagonals.

The answer is n divisible by 9.

First we construct n = 9 and by extension every multiple of 9.

I I I M M M O O O
M M M O O O I I I
O O O I I I M M M
I I I M M M O O O
M M M O O O I I I
O O O I I I M M M
I I I M M M O O O
M M M O O O I I I
O O O I I I M M M

We now prove 9 | n is necessary.

Let n = 3k, which divides the given grid into k 2 sub-boxes (of size 3 × 3 each). We say
a multiset of squares S is clean if the letters distribute equally among them; note that
unions of clean multisets are clean.
Consider the following clean sets (given to us by problem statement):

• All columns indexed 2 (mod 3),

• All rows indexed 2 (mod 3), and

• All 4k − 2 diagonals mentioned in the problem.

Take their union. This covers the center of each box four times, and every other cell
exactly once. We conclude the set of k 2 center squares are clean, hence 3 | k 2 and so
9 | n, as desired.

5
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

Shown below is the sums over all diagonals only, and of the entire union.

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 2 2 1 4 1 1 4 1 1 4 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 2 2 1 4 1 1 4 1 1 4 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 2 2 1 4 1 1 4 1 1 4 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

6
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

§1.3 IMO 2016/3, proposed by Aleksandr Gaifullin (RUS)

Available online at https://aops.com/community/p6637660.

Problem statement

Let P = A1 A2 . . . Ak be a convex polygon in the plane. The vertices A1 , A2 , . . . ,

Ak have integral coordinates and lie on a circle. Let S be the area of P . An odd
positive integer n is given such that the squares of the side lengths of P are integers
divisible by n. Prove that 2S is an integer divisible by n.

Solution by Jeck Lim: We will prove the result just for n = pe where p is an odd
prime pand e ≥ 1. The case k = 3 is resolved by Heron’s formula directly: we have
S = 14 2(a2 b2 + b2 c2 + c2 a2 ) − a4 − b4 − c4 , so if pe | gcd(a2 , b2 , c2 ) then p2e | S 2 .
Now we show we can pick a diagonal and induct down on k by using inversion.
Let the polygon be A1 A2 . . . Ak+1 and suppose for contradiction that all sides are
divisible by pe but no diagonals are. Let O = Ak+1 for notational convenience. By
applying inversion around O with radius 1, we get the “generalized Ptolemy theorem”
A1 A2 A2 A3 Ak−1 Ak A1 Ak
+ + ··· + =
OA1 · OA2 OA2 · OA3 OAk−1 · OAk OA1 · OAk

or, making use of square roots,

s s s s
A1 A22 A2 A23 Ak−1 A2k A1 A2k
+ + · · · + =
OA21 · OA22 OA22 · OA23 OA2k−1 · OA2k OA21 · OA2k

Suppose νp of all diagonals is strictly less than e. Then the relation becomes
√ √ √
q1 + · · · + qk−1 = q

where qi are positive rational numbers. Since there are no nontrivial relations between
square roots
p (see this link) there is a positive rational number b such that ri = qi /b
p

and r = q/b are all rational numbers. Then

X
ri = r.

However, the condition implies that νp (qi2 ) > νp (q 2 ) for all i (check this for i = 1, i = k −1
and 2 ≤ i ≤ k − 2), and hence νp (ri ) > νp (r). This is absurd.

Remark. I think you basically have to use some Ptolemy-like geometric property, and also
all correct solutions I know of for n = pe depend on finding a diagonal and inducting down.
(Actually, the case k = 4 is pretty motivating; Ptolemy implies one can cut in two.)

7
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

§2 Solutions to Day 2
§2.1 IMO 2016/4, proposed by Gerhard Woeginger (LUX)
Available online at https://aops.com/community/p6642559.

Problem statement

A set of positive integers is called fragrant if it contains at least two elements and
each of its elements has a prime factor in common with at least one of the other
elements. Let P (n) = n2 + n + 1. What is the smallest possible positive integer
value of b such that there exists a non-negative integer a for which the set

{P (a + 1), P (a + 2), . . . , P (a + b)}

is fragrant?

The answer is b = 6.
First, we prove b ≥ 6 must hold. It is not hard to prove the following divisibilities by
Euclid:

gcd(P (n), P (n + 1)) | 1

gcd(P (n), P (n + 2)) | 7
gcd(P (n), P (n + 3)) | 3
gcd(P (n), P (n + 4)) | 19.

Now assume for contradiction b ≤ 5. Then any GCD’s among P (a + 1), . . . , P (a + b)

must be among {3, 7, 19}. Consider a multi-graph on {a + 1, . . . , a + b} where we join two
elements with nontrivial GCD and label the edge with the corresponding prime. Then we
readily see there is at most one edge each of {3, 7, 19}: id est at most one edge of gap 2, 3,
4 (and no edges of gap 1). (By the gap of an edge e = {u, v} we mean |u − v|.) But one
can see that it’s now impossible for every vertex to have nonzero degree, contradiction.
To construct b = 6 we use the Chinese remainder theorem: select a with

a+1≡7 (mod 19)

a + 5 ≡ 11 (mod 19)
a+2≡2 (mod 7)
a+4≡4 (mod 7)
a+3≡1 (mod 3)
a+6≡1 (mod 3)

which does the trick.

8
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

§2.2 IMO 2016/5, proposed by Russia

Available online at https://aops.com/community/p6642565.

Problem statement

The equation

(x − 1)(x − 2) . . . (x − 2016) = (x − 1)(x − 2) . . . (x − 2016)

is written on the board, with 2016 linear factors on each side. What is the least
possible value of k for which it is possible to erase exactly k of these 4032 linear
factors so that at least one factor remains on each side and the resulting equation
has no real solutions?

The answer is 2016. Obviously this is necessary in order to delete duplicated factors. We
now prove it suffices to deleted 2 (mod 4) and 3 (mod 4) guys from the left-hand side,
and 0 (mod 4), 1 (mod 4) from the right-hand side.
Consider the 1008 inequalities

(x − 1)(x − 4) < (x − 2)(x − 3)

(x − 5)(x − 8) < (x − 6)(x − 7)
(x − 9)(x − 12) < (x − 10)(x − 11)
..
.
(x − 2013)(x − 2016) < (x − 2014)(x − 2015).

Notice that in all these inequalities, at most one of them has non-positive numbers
in it, and we never have both zero. If there is exactly one negative term among the
1008 · 2 = 2016 sides, it is on the left and we can multiply all together. Thus the only
case that remains is if x ∈ (4m − 2, 4m − 1) for some m, say the mth inequality. In that
case, the two sides of that inequality differ by a factor of at least 9.

Claim — We have
Y (4k + 2)(4k + 3)
< e.
(4k + 1)(4k + 4)
k≥0

Proof of claim using logarithms. It’s equivalent to prove

 
2
log 1 +
X
< 1.
(4k + 1)(4k + 4)
k≥0

To
P this end, we use the deep fact that log(1 + t) ≤ t, and thus it follows from
k≥0 (4k+1)(4k+4) < 2 , which one can obtain for example by noticing it’s less than
1 1

4 6 .
1 π2

Remark (Elementary proof of claim, given by Espen Slettnes). To avoid calculus as above,

9
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

for each N ≥ 0, note the partial product is bounded by

N    
Y (4k + 2)(4k + 3) 2 3 6 7 10 4N + 3
= · · · · · ··· ·
(4k + 1)(4k + 4) 1 4 5 8 9 4N + 4
k=0
4N + 3
< 2 · 1 · 1 · ··· · < 2 < e.
4N + 4

This solves the problem, because then the factors being multiplied on by the positive
inequalities before the mth one are both less than e, and e2 < 9. In symbols, for
4m − 2 < x < 4m − 1 we should have
(x − (4m − 6))(x − (4m − 5)) (x − 2)(x − 3)
× ··· × <e
(x − (4m − 7))(x − (4m − 4)) (x − 1)(x − 4)

and
(x − (4m + 2))(x − (4m + 3)) (x − 2014)(x − 2015)
× ··· × <e
(x − (4m + 1))(x − (4m + 4)) (x − 2013)(x − 2016)
because the (k + 1)st term of each left-hand side is at most (4k+1)(4k+4) ,
(4k+2)(4k+3)
for k ≥ 0. As
e2 < 9, we’re okay.

10
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

§2.3 IMO 2016/6, proposed by Josef Tkadlec (CZE)

Available online at https://aops.com/community/p6642576.

Problem statement

There are n ≥ 2 line segments in the plane such that every two segments cross and
no three segments meet at a point. Geoff has to choose an endpoint of each segment
and place a frog on it facing the other endpoint. Then he will clap his hands n − 1
times. Every time he claps, each frog will immediately jump forward to the next
intersection point on its segment. Frogs never change the direction of their jumps.
Geoff wishes to place the frogs in such a way that no two of them will ever occupy
the same intersection point at the same time.

(a) Prove that Geoff can always fulfill his wish if n is odd.

(b) Prove that Geoff can never fulfill his wish if n is even.

The following solution was communicated to me
by Yang Liu.

Imagine taking a larger circle ω encasing all n2 intersection points. Denote by P1 , P2 ,
. . . , P2n the order of the points on ω in clockwise order; we imagine placing the frogs on
Pi instead. Observe that, in order for every pair of segments to meet, each line segment
must be of the form Pi Pi+n .
1
8

3
6
4

Then:

(a) Place the frogs on P1 , P3 , . . . , P2n−1 . A simple parity arguments shows this works.

(b) Observe that we cannot place frogs on consecutive Pi , so the n frogs must be
placed on alternating points. But since we also are supposed to not place frogs on
diametrically opposite points, for even n we immediately get a contradiction.

11
IMO 2016 Solution Notes web.evanchen.cc, updated 15 December 2024

Remark. Yang says: this is easy to guess if you just do a few small cases and notice that
the pairs of “violating points” just forms a large cycle around.

12