USAMO 2014 Solution Notes

Compiled by Evan Chen

November 23, 2019

This is an compilation of solutions for the 2014 USAMO. Some of the

solutions are my own work, but many are from the official solutions provided
by the organizers (for which they hold any copyrights), and others were found
on the Art of Problem Solving forums.
Corrections and comments are welcome!

Contents

0 Problems 2

1 USAMO 2014/1, proposed by Titu Andreescu 3

2 USAMO 2014/2, proposed by Titu Andreescu 4

3 USAMO 2014/3, proposed by Razvan Gelca 6

4 USAMO 2014/4, proposed by Palmer Mebane 7

5 USAMO 2014/5, proposed by Titu Andreescu and Cosmin Pohoata 9

6 USAMO 2014/6, proposed by Gabriel Dospinescu 11

1
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

§0 Problems
1. Let a, b, c, d be real numbers such that b − d ≥ 5 and all zeros x1 , x2 , x3 , and x4
of the polynomial P (x) = x4 + ax3 + bx2 + cx + d are real. Find the smallest value
the product (x21 + 1)(x22 + 1)(x23 + 1)(x24 + 1) can take.

2. Find all f : Z → Z such that

f (x)2
xf (2f (y) − x) + y 2 f (2x − f (y)) = + f (yf (y))
x
for all x, y ∈ Z such that x 6= 0.

3. Prove that there exists an infinite set of points

. . . , P−3 , P−2 , P−1 , P0 , P1 , P2 , P3 , . . .

in the plane with the following property: For any three distinct integers a, b, and c,
points Pa , Pb , and Pc are collinear if and only if a + b + c = 2014.

4. Let k be a positive integer. Two players A and B play a game on an infinite grid of
regular hexagons. Initially all the grid cells are empty. Then the players alternately
take turns with A moving first. In his move, A may choose two adjacent hexagons
in the grid which are empty and place a counter in both of them. In his move, B
may choose any counter on the board and remove it. If at any time there are k
consecutive grid cells in a line all of which contain a counter, A wins. Find the
minimum value of k for which A cannot win in a finite number of moves, or prove
that no such minimum value exists.

5. Let ABC be a triangle with orthocenter H and let P be the second intersection of
the circumcircle of triangle AHC with the internal bisector of ∠BAC. Let X be
the circumcenter of triangle AP B and let Y be the orthocenter of triangle AP C.
Prove that the length of segment XY is equal to the circumradius of triangle ABC.

6. Prove that there is a constant c > 0 with the following property: If a, b, n are
positive integers such that gcd(a + i, b + j) > 1 for all i, j ∈ {0, 1, . . . , n}, then

min{a, b} > (cn)n/2 .

2
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

§1 USAMO 2014/1, proposed by Titu Andreescu

Let a, b, c, d be real numbers such that b − d ≥ 5 and all zeros x1 , x2 , x3 , and x4 of the
polynomial P (x) = x4 + ax3 + bx2 + cx + d are real. Find the smallest value the product
(x21 + 1)(x22 + 1)(x23 + 1)(x24 + 1) can take.

The answer is 16 . This can be achieved by taking x1 = x2 = x3 = x4 = 1, whence

the product is 24 = 16, and b − d = 5. √
Now, we prove this is a lower bound. Let i = −1. The key observation is that
4
Y 4

Y
x2j + 1 = (xj − i)(xj + i) = P (i)P (−i).
j=1 j=1

Consequently, we have

x21 + 1 x22 + 1 x23 + 1 x21 + 1 = (b − d − 1)2 + (a − c)2

≥ (5 − 1)2 + 02 = 16.

This proves the lower bound.

3
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

§2 USAMO 2014/2, proposed by Titu Andreescu

Find all f : Z → Z such that

f (x)2
xf (2f (y) − x) + y 2 f (2x − f (y)) = + f (yf (y))
x
for all x, y ∈ Z such that x 6= 0.

The answer is f (x) ≡ 0 and f (x) ≡ x2 . Check that these work.

Now let’s prove these are the only solutions. Put y = 0 to obtain

f (x)2
xf (2f (0) − x) = + f (0).
x

Now we claim f (0) = 0 . If not, select a prime p - f (0) and put x = p 6= 0. In the above,
f (p)2
we find that p | f (p)2 , so p | f (p) and hence p | p . From here we derive p | f (0),
contradiction. Hence
f (0) = 0.
The above then implies that

x2 f (−x) = f (x)2

holds for all nonzero x, but also for x = 0. Let us now check that f is an even function.
In the above, we may also derive f (−x)2 = x2 f (x). If f (x) 6= f (−x) (and hence x 6= 0),
then subtracting the above and factoring implies that f (x) + f (−x) = −x2 ; we can then
obtain by substituting the relation
 2
1 3
f (x) + x2 = − x4 < 0
2 4

which is impossible. This means f (x)2 = x2 f (x), thus

f (x) ∈ {0, x2 } ∀x.

Now suppose there exists a nonzero integer t with f (t) = 0. We will prove that
f (x) ≡ 0. Put y = t in the given to obtain that

t2 f (2x) = 0

for any integer x 6= 0, and hence conclude that f (2Z) ≡ 0. Then selecting x = 2k 6= 0 in
the given implies that
y 2 f (4k − f (y)) = f (yf (y)).
Assume for contradiction that f (m) = m2 now for some odd m 6= 0. Evidently

m2 f (4k − m2 ) = f (m3 ).

If f (m3 ) 6= 0 this forces f (4k − m2 ) 6= 0, and hence m2 (4k − m2 )2 = m6 for arbitrary

k 6= 0, which is clearly absurd. That means

f (4k − m2 ) = f (m2 − 4k) = f (m3 ) = 0

for each k 6= 0. Since m is odd, m2 ≡ 1 (mod 4), and so f (n) = 0 for all n other than
±m2 (since we cannot select k = 0).

4
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

Now f (m) = m2 means that m = ±1. Hence either f (x) ≡ 0 or

(
1 x = ±1
f (x) =
0 otherwise.

To show that the latter fails, we simply take x = 5 and y = 1 in the given.
Hence, the only solutions are f (x) ≡ 0 and f (x) ≡ x2 .

5
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

§3 USAMO 2014/3, proposed by Razvan Gelca

Prove that there exists an infinite set of points

. . . , P−3 , P−2 , P−1 , P0 , P1 , P2 , P3 , . . .

in the plane with the following property: For any three distinct integers a, b, and c, points Pa ,
Pb , and Pc are collinear if and only if a + b + c = 2014.

The construction  !
2014 3

2014
Pn = n− , n−
3 3
works fine, and follows from the following claim:

Claim — If x, y, z are distinct real numbers then the points (x, x3 ), (y, y 3 ), (z, z 3 )
are collinear if and only if x + y + z = 0.

Proof. Note that by the “shoelace formula”, the collinearity is equivalent to

x x3 1
 

0 = det y y 3 1
z z3 1

But the determinant equals

X
x(y 3 − z 3 ) = (x − y)(y − z)(z − x)(x + y + z).
cyc

6
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

§4 USAMO 2014/4, proposed by Palmer Mebane

Let k be a positive integer. Two players A and B play a game on an infinite grid of regular
hexagons. Initially all the grid cells are empty. Then the players alternately take turns with A
moving first. In his move, A may choose two adjacent hexagons in the grid which are empty
and place a counter in both of them. In his move, B may choose any counter on the board and
remove it. If at any time there are k consecutive grid cells in a line all of which contain a counter,
A wins. Find the minimum value of k for which A cannot win in a finite number of moves, or
prove that no such minimum value exists.

The answer is k ≥ 6.
To show k ≥ 6 can’t be done, mark every third cell red, as shown in the figure below.

If B always removes a counter on a red cell, then he can prevent A from ever getting six
consecutive, since A can never cover two red cells simultaneously on her turn.
We now show A can achieve k = 5 in a finite number of moves by describing a strategy
which ensures her victory. Note that after B’s first turn there is one counter, so then
A may create an equilateral triangle, and hence after B’s second turn there are two
consecutive counters. Then, A places a pair of counters two spaces away on the same
line, as shown below.

x y

Now it is B’s turn to move; in order to avoid losing immediately, he must remove either
x or y. Then on any subsequent turn, A can replace x or y (whichever was removed) and
add one more adjacent counter. This continues until either x or y has all its neighbors
filled; we ask A to do so in such a way that she avoids filling in the two cells between x
and y until one hexagon is complete.
So, WLOG the x hexagon is filled completely. At that moment, B must choose to
remove x (or A wins on the next turn). Then A marks the two cells colored light green
below; then A clearly wins one turn later, because B cannot simultaneously block all the
threats.

7
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

x y

8
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

§5 USAMO 2014/5, proposed by Titu Andreescu and Cosmin

Pohoata
Let ABC be a triangle with orthocenter H and let P be the second intersection of the circumcircle
of triangle AHC with the internal bisector of ∠BAC. Let X be the circumcenter of triangle
AP B and let Y be the orthocenter of triangle AP C. Prove that the length of segment XY is
equal to the circumradius of triangle ABC.

B′
X′
A Y′

X
O
H

B C

We eliminate the floating orthocenter by reflecting P across AC to Q. Then Q lies

on (ABC) and moreover ∠QAC = 12 ∠BAC. This motivates us to reflect B, X, Y to B 0 ,
X 0 , Y 0 and complex bash with respect to 4AQC. Obviously

y 0 = a + q + c.

Now we need to compute x0 . You can get this using the formula

(b0 − a)(q − a) q − a − b0 − a


0
x =a+ .
(b0 − a)(q − a) − (b0 − a)(q − a)
c3
Using the angle condition we know b = q2
, and then that

aq 2
b0 = a + c − acb = a + c − .
c2

9
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

Therefore
   
aq 2 c2
c− (q − a) 1q − a1 − 1c + aq
c2 2
x0 = a +  2
   
c2
c − aq c2
1
q − 1
a − 1
c − aq 2
(q − a)
 
c3 −aq 2 c2
c2
(q − a) 1q − a1 − 1c + aq 2
=a+ c3 −aq 2 q−a c3 −aq 2
− c2 qa + aq2 c (q − a)
1 1 1 c2
q − a − c + aq 2
=a+ 1
− qa + aqc 2
aq 2
c2 − q 2 + aq − c
=a+
c−q
aq
=a+c+q+
c
whence
x − y 0  =  aq  = 0.
 0   
c

10
USAMO 2014 Solution Notes web.evanchen.cc, updated November 23, 2019

§6 USAMO 2014/6, proposed by Gabriel Dospinescu

Prove that there is a constant c > 0 with the following property: If a, b, n are positive integers
such that gcd(a + i, b + j) > 1 for all i, j ∈ {0, 1, . . . , n}, then

min{a, b} > (cn)n/2 .

Let N = n + 1 and assume N is (very) large. We construct an N × N with cells (i, j)

where 0 ≤ i, j ≤ n and in each cell place a prime p dividing gcd(a + i, b + j).
The central claim is at least 50% of the primes in this table exceed 0.001n2 . We count
the maximum number of squares they could occupy:
X  N 2 X N 2 X 1 X1 X
≤ +1 = N2 + 2N + 1.
p
p p
p p
p2 p
p p

Here the summation runs over primes p ≤ 0.001n2 .

Let r = π(0.001n2 ) denote the number of such primes. Now we consider the following
three estimates. First,
X 1 1
2
<
p
p 2

which follows by adding all the primes directly with some computation. Moreover,

X1 r
X 1
< = O(log r) < o(N )
p
p k
k=1

using the harmonic series bound, and

N2
X  
1<r∼O < o(N 2 )
p
ln N

via Prime Number Theorem. Hence the sum in question is certainly less than 12 N 2 for N
large enough, establishing the central claim.
Hence some column a + i has at least one half of its primes greater than 0.001n2 .
Because this is greater than n for large n, these primes must all be distinct, so a + i
exceeds their product, which is larger than

N/2
0.001n2 > cn · nn

where c is some constant (better than the requested bound).

11