Polynomials Problem Set

piyush kumar jha

March 21, 2023

§1 First Problem

Example 1.1 (India TST 2001 )

For any positive integer n, show that there exists a polynomial P (x) of degree n
with integer coefficients such that P (0), P (1), . . . , P (n) are all distinct powers of 2.

Proof. We would like to prove a generalization for this problem which states as:

Claim 1.2 — ∃ a non-constant polynomial f ∈ Z[t] such that deg(f (x)) = n and
f (0), f (1), · · · f (n) are all distinct powers of a prime p
Proof:- The proof follows from proof by construction which goes as follows:

set n! = pα · β for some α, β ∈ Z+ such that gcd(β, p) = 1 and q = pφ(β)

x  
X
α x
so we see that f (x) = p · (q − 1)k works as
k
k=0

f (x) = pα+x·φ(β) which gives every distinct powers of p

now the stuff which is left to prove is that f ∈ Z[t]

so we make another claim that:

Claim 1.3 — f ∈ Z[t]

x pα ·(q−1)k Qk−1
Proof:- we notice in particular the coefficient is pα · ·(q−1)k =


k k! · i=0 (x−i)

now since q = pφ(β) and gcd(p, β) = 1 we get from euler’s theorem that (q − 1)k ≡ 0
(mod β)

which implies n!|pα · (q − 1)k =⇒ k!|pα · (q − 1)k

which gives f ∈ Z[t]

which proves the Subclaim and hence also finishes our main claim □

1
piyush kumar jha — March 21, 2023 Polynomials Problem Set

hence the original problem works for p = 2 ■

§2 Second Problem

Example 2.1 (IMO 1993 , Day 1 problem 1)

Let n > 1 be an integer and let f (x) = xn + 5 · xn−1 + 3. Prove that there do not
exist polynomials g(x), h(x), each having integer coefficients and degree at least one,
such that f (x) = g(x) · h(x).

Proof. FTSOC assume the contrary and set f (x) = h(x) · g(x), where h(x), g(x) ∈ Z[x],
plugging x = 0 and yeilds h(0) · g(0) = 3 , now observe f (x) = xn−1 (x + 5) + 3, plug
x = −5 and we get f (−5) = 3 =⇒ h(−5) · g(−5) = 3

Now by Rational Root Theorem, we can clearly observe that there is no integer root
of f (x) and h(x) and g(x) are not linear.

set h(x) = (x − z1 )(x − z2 ) · · · (x − zr ), deg((h(x)) = r and zi′ s are roots of h(x) for
1⩽i⩽r

now WLOG take h(0) = ±1( since h(x) ∈ Z[x])

Qr
| i=1 zi | =1

take | ri=1 (zi + 5)| = |(z1 + 5)(z2 + 5) · · · (zr + 5)|

Q

now notice for zl ∈ {z1 , z2 , · · · , zr }

we get zln−1 (zl + 5) = −3

Qr n−1
| l=1 zl (zl + 5)| = 3r

|h(−5)| = 3r

we know that h(−5)|3 =⇒ 3r |3, r > 1 this follows a beautiful contradiction! and
hence problem statement follows

§3 Third Problem

Example 3.1 (IMO 1976, Day 1, Problem 2)

Let P1 (x) = x2 − 2 and Pj (x) = P1 (Pj−1 (x)) for j= 2, . . . Prove that for any positive
integer n the roots of the equation Pn (x) = x are all real and distinct.

Solution.

Claim 3.2 — all real roots of Pn (x) = x lies in [−2, 2]

2
piyush kumar jha — March 21, 2023 Polynomials Problem Set

Proof:- clearly through induction we can show that Pn (x) > x ∀x > 2 and for x < −2
we have Pn (x) > 2 , hence we have all roots of Pn (x) = x lies in [−2, 2] , so claim follows
□
now we substitute x = 2 cos θ, we have P1 (x) = 2 cos 2θ, from again induction we have
Pn (x) = 2 cos 2n θ

we need to solve cos 2n θ = cos θ =⇒ θ = 22mπn −1 and θ =

2πm
2n +1 , where m ∈ Z∗ hence we
have all 2n roots of Pn (x) real and distinct □

§4 Fourth Problem

Example 4.1 (INMO 2020)

Suppose P (x) is a polynomial with real coefficients, satisfying the condition P (cos θ +
sin θ) = P (cos θ − sin θ), for every real θ. Prove that P (x) can be expressed in the
form
P (x) = a0 + a1 (1 − x2 )2 + a2 (1 − x2 )4 + · · · + an (1 − x2 )2n
for some real numbers a0 , a1 , . . . , an and non-negative integer n.
Proposed by C.R. Pranesacher

√ √
2 · sin(θ + π4 ) = P 2 · (cos θ + π4 )



Proof. P (cos θ +sin θ) = P (cos θ −sin θ) =⇒ P
∀θ∈R
√ √
Now this implies P ( 2 sin x) = P ( 2 cos x) ∀ x ∈ R
√ √
Take G(x) = P ( 2 sin x) − P ( 2 cos x)

Claim 4.2 — G(x) is identically 0 ∀ x ∈ R

√ √
Proof:- Since deg(P) is finite and P ( 2 sin x) − P ( 2 cos x) = 0 ∀ x ∈ R , we have
G(x) to be identically 0 and follows claim
√ √ √
also P ( 2 sin x) = P (− 2 sin x) = P ( 2 cos x)

we have P (x) to be even degree polynomial for all x

so we have P (1) = P (−1) = k, k ∈ R

so P (x) = Q(x)(x + 1)(x − 1) + k

also,

Claim 4.3 — (1 − x2 )|Q(x)

Proof:- Notice Q(cos θ + sin θ) + Q(sin θ − cos θ) = 0

for θ = π2 , Q(1) = 0

3
piyush kumar jha — March 21, 2023 Polynomials Problem Set

for θ = 0, Q(−1) = 0

so (1 − x2 )|Q(x)

Now this gives P (x) = (1 − x2 )2 R(x) + k

Notice R(x) is even degree polynomial we get R(cos θ + sin θ) = R(cos θ − sin θ)

Pn
Claim 4.4 — P (x) = a0 + i=1 a2i (1 − x2 )2i ∀ x ∈ R

base case:- n = 1 its trivially True

suppose it’s true for n, then we seek to show from the induction hypothesis that it
holds for (n + 1)

P (x) = J(x)(1 − x2 )2n + a0

for case of n + 1

P (x) = J(x)(1 − x2 )2(n+1) + a0

plugging back cos θ + sin θ and cos θ − sin θ we get J(x) as a even degree polynomial
and J(cos θ + sin θ) = J(cos θ − sin θ)

so we get P (x) as J(x)(1 − x2 )2(n+1) + a0 to be hold true hence from induction hy-
pothesis we get

P (x) = T (x)(1 − x2 )2n + c for some polynomial T (x) and constant c

§5 Fifth Problem

Example 5.1 (INMO 2021)

Find all pairs of integers (a, b) so that each of the two cubic polynomials

x3 + ax + b and x3 + bx + a

has all the roots to be integers.

Proposed by Prithwijit De and Sutanay Bhattacharya

Solution.

Claim 5.2 — only (a, b) = (0, 0) works

proof:- denote α1 , α2 , α3 to be the roots of the equation x3 + ax + b and β1 , β2 , β3 to be

the roots of the equation x3 + bx + a

4
piyush kumar jha — March 21, 2023 Polynomials Problem Set
P P Q P P Q
Notice αi = 0, i̸=j αi · αj = a, αi = −b, βi = 0, i̸=j βi · βj = a , βi = −a

αi2 = −2a, βi2 = −2b

P P
also

define ⊙(x, y, z) to be the parity of x, y, z

Claim 5.3 — ⊙(α1 , α2 , α3 ) = ⊙(β1 , β2 , β3 )=even

Proof:- FTSOC consider WLOG ⊙(α1 , α2 ) ̸= ⊙(α3 ), ⊙(β1 , β2 ) ̸= ⊙(β3 )

clearly, we first have 2|a and 2|b, also we observe a = −α22 + α · α3 =⇒ 2| − α22 + α1 · α3
which is from our assumption not true, hence a contradiction, the same way goes for
β1 , β2 , β3

so we get ⊙(α1 , α2 , α3 ) = ⊙(β1 , β2 , β3 )=even as claim follows ■

first of all there is a trivial solution pair of (a, b) = (0, 0) so ignoring this case for a
moment and considering P a ̸= 0 and bP ̸= 0 we observe:
8|a, 8|b and then 8| i̸=j αi · αj , 8| i̸=j βi · βj (♠)

set αi = 2qi and βi = 2pi

P P
so clearly we have i̸=j qi · qj ≡ 0 (mod 2) and qi = 0

which gives q1 , q2 , q3 to be even similarly p1 , p2 , p3 to be even

−b
· qj = a4 ,
P P Q
also we have qi = 0, i̸=j qi qi = 8 ,

−a
pi · pj = 4b ,
P P Q
pi = 0, i̸=j pi = 8

hence gives 8| a8 , 8| 8b =⇒ 8| a4 , 8| 4b =⇒ 8|
P P
i̸=j qi · qj , 8| i̸=j pi · pj

from (♠) we observe this creates an infinite descent, hence we only have (α1 , α2 , α3 ) = 0
and (β1 , β2 , β3 ) = 0

which gives (a, b) = (0, 0) as the only pair , hence claim follows □

§6 Sixth Problem

5
piyush kumar jha — March 21, 2023 Polynomials Problem Set

Example 6.1 (INMO 2023)

Suppose a0 , . . . , a100 are positive reals. Consider the following polynomial for each k
in {0, 1, . . . , 100}:

a100+k x100 + 100a99+k x99 + a98+k x98 + a97+k x97 + · · · + a2+k x2 + a1+k x + ak ,

where indices are taken modulo 101, i.e., a100+i = ai−1 for any i in {1, 2, . . . , 100}.
Show that it is impossible that each of these 101 polynomials has all its roots real.

Proposed by Prithwijit De

Proof. denote Pk (x) = a100+k x100 + 100a99+k x99 + a98+k x98 + a97+k x97 + · · · + a2+k x2 +
a1+k x + ak ∀ k ∈ {0, 1, 2, 3, · · · , 100}

Claim 6.2 — all roots of Pk (x) are non-real

Proof:-FTSOC it’s all roots are real now if ai ∈ R+

we have roots of Pk (x) to be −αk,1 , −αk,2 , · · · − αk,100 where αk,i > 0

 
a99+k
now 100 and 100 ak
P Q
α
i=1 k,i = 100 · a100+k i=1 αk,i = a100+k

from AM ≥ GM we observe :
 100
a99+k ak
a100+k ≥ a100+k , now since k varies from 0 to 100 we observe:
Q100  a99+k 100 Q100 ak

a99
100
k=0 a100+k ≥ k=0 a100+k =⇒ a200 ≥1

but since we have a99 = a200 we get 1 ≥ 1 which forces equality

hence aj = aj+1 ∀j ∈ {k, k + 1, k + 2, · · · , k + 99}

so we get 100
P Q100
i=1 αk,i = 100 and i=1 αk,i = 1 =⇒ αk,1 = αk,2 = · · · αk,100 = β , for
some β ∈ R+

so we have Pk (x) = β(x + 1)100

also we had a100+k = a99+k = a98+k = · · · = ak which is not true for Pk (x) hence
contradiction and claim follows

§7 Seventh Problem

6
piyush kumar jha — March 21, 2023 Polynomials Problem Set

Example 7.1 (IMO 1975,Day 2, Problem 6)

Determine the polynomials P of two variables so that:

a.) for any real numbers t, x, y we have P (tx, ty) = tn P (x, y) where n is a positive
integer, the same for all t, x, y;

b.) for any real numbers a, b, c we have P (a + b, c) + P (b + c, a) + P (c + a, b) = 0;

c.) P (1, 0) = 1.

Solution. first of all we denote the assertion as Ω(x, y, z)

we have Ω(a, a, a, ) =⇒ P (2a, a) = 0 =⇒ P (x, y) + (x − 2y)J(x, y) for some polyno-

mial J(x, y) of degree n − 1

now we have P (1, 0) = Q(1, 0) = 1

Ω(a, b, b) gives P (2b, a)+2P (a+b, b) = 0 =⇒ (2b−2a)J(2b, a)+2(a−b)J(b+a, b) =⇒

J(a + b) = J(2b, a) ∀ a ̸= b

now set α = a + b, β = b, a = α − β

we have J(α, β) = J(2β, α − β) = J(2α − 2β, 3β − α) = · · ·

this gives J(α − ζ, β + ζ) − J(x, y) = 0 ∀ ζ, but this a polynomial of n − 1 degree hence

we have J(x, y) = c(x + y)n−1 , plugging J(1, 0) = 1

we have P (x, y) = (x − 2y)(x + y)n−1 ∀ n ∈ Z+

§8 Eighth Problem

Example 8.1 (INMO 2012)

Define a sequence {f0 (x), f1 (x), f2 (x), . . . } of functions by

f0 (x) = 1

f1 (x) = x
2
(fn (x)) − 1 = fn+1 (x)fn−1 (x)
for n ≥ 1. Prove that each fn (x) is a polynomial with integer coefficients.

Proof.

Claim 8.2 — fk (x) = xfk−1 (x) − fk−2 (x) ∀ k ⩾ 2

Proof:- Base case(k=2):

f2 (x) = x2 − 1 and xf1 (x) − f0 (x) = x2 − 1 hence f2 (x) = xfk−1 (x) − fk−2 (x) verified

7
piyush kumar jha — March 21, 2023 Polynomials Problem Set

now we assume that the Claim is true for k = k-th case

(fk )2 −1
Consider:- fk−1 (x) = fk+1 (x)
2
xfk (x)fk−1 (x)−fk (x)fk−2 (x)+fk (x)fk−2 (x)−fk−1 (x)
(fk )2 −1 (f +1)(xfk−1 (x)−fk−2 (x)−1)
fk−1 (x) = k fk−1 (x) = fk−1 (x) =
xfk (x) − fk−1 (x)

hence fk+1 (x) = xfk (x) − fk−1 (x), hence from induction hypothesis we have fn (x) =
xfn−1 (x) − fn−2 (x). hence claim follows □ , and now from induction again it suffices
that fn (x) ∈ Z[x] □

§9 Ninth Problem

Example 9.1 (APMO 2009)

Let a1 , a2 , a3 , a4 , a5 be real numbers satisfying the following equations:

a1 a2 a3 a4 a5 1
k2 +1
+ k2 +2
+ k2 +3
+ k2 +4
+ k2 +5
= k2
for k = 1, 2, 3, 4, 5

a1 a2 a3 a4 a5
Find the value of 37 + 38 + 39 + 40 + 41 (Express the value in a single fraction.)

 
a1 a2 a5 1
Solution. we set P (x) = x(x + 1)(x + 2) · · · (x + 5) x+1 + x+2 + ··· + x+5 − x

clearly we can observe from problem statement that 1, 4, 9, 16, 25 are roots hence we
have P (x) = k(x − 1)(x − 4)(x − 9)(x − 16)(x − 25)
 
a1 a2 a5
now we can re write P (x) as P (x) = x(x+1)(x+2) · · · (x+5) x+1 + x+2 + · · · + x+5 −
(x + 1)(x + 2) · · · (x + 5)

1
plugging x = 0 we get k = 120

a1 a2 a5
hence we want to evaluate P (x) at 36 so doing so we get 37 + 38 + ··· + 41 =
35·32·27·20·11 1 187465
120·36·37·38·39·40·41 + 36 = 6744582 .

§10 Tenth Problem

Example 10.1 (Stanford Mathematics Tournament 2022)

Let a, b, and c be the roots of the polynomial x3 − 3x2 − 4x + 5.

a4 +b4 b4 +c4 c4 +a4

Compute a+b + b+c + c+a

Solution. so we first do some side calculations before solving the actual one. so we have
a2 + b2 + c2 = (a + b + c)2 − 2ab − 2bc − 2ca = 17

a3 + b3 + c3 = (a + b + c)(a2 + b2 + c2 − ab − bc − ca) + 3abc = 48

8
piyush kumar jha — March 21, 2023 Polynomials Problem Set

and a4 + b4 + c4 = (a2 + b2 + c2 )2 − 2a2 b2 − 2b2 c2 − 2c2 a2 = 289 − 92 = 197

now we get back to our problem:

4 +b4 4 +c4 4 +a4
we denote Ω = aa+b + b b+c + c c+a
P a(a3 −1)+b(b3 −1)
Ω−3= a,b,c a+b
P a(3a2 +4a−6) P b(3b2 +4b−6)
Ω−3= a,b,c a+b + a,b,c a+b
P 3a3 +4a2 −6a P 3b3 +4b2 −6b
Ω−3= a,b,c a+b + a,b,c a+b
 
P 3 2 1 1
Ω−3= a,b,c (3a + 4a − 6a) · 3−c + 3−b
 
P 3 2 6−(b+c)
Ω−3= a,b,c (3a + 4a − 6a) · (3−c)(3−b)
 
(a+3)(a−3)
(3a3 4a2
P
Ω−3= a,b,c + − 6a) · 7
P (13a2 +6a−15)(a2 −9)
Ω−3= a,b,c 7
P 13a4 +6a3 −132a2 −54a+135
Ω−3= a,b,c 7

13(197)+6(48)−132(17)−54(3)+3(135)
Ω−3= 7

869
Ω=
7

§11 Eleventh Problem

Example 11.1 (I.S.I 2019)

Let f be a polynomial with integer coefficients. Define

a1 = f (0) , a2 = f (a1 ) = f (f (0)) ,

and an = f (an−1 ) for n ⩾ 3.

If there exists a natural number k ⩾ 3 such that ak = 0, then prove that either
a1 = 0 or a2 = 0.

Proof. first of all we have f (ak−1 ) = 0 so f (ak ) = f (0) so ak+1 = a1 and similarly
ak+2 = a2

so we have f (x) = (x − ak−1 )g(x) for some g(x) ∈ Z[x]. Then using the fact that
a1 |a2 − a1 =⇒ a1 |a2 , then by induction we can show a1 |ak ∀ k ∈ Z+ . so we have
f (al ) = a1 (rl − rk−1 )g(al ), so if a1 = 0 we get that f (al ) = 0 ∀ l ∈ Z+ . sps that a1 ̸= 0 ,
then we proceed like this:

a2 − a1 |a3 − a2 | · · · |an − an−1 | · · · |ak+1 − ak |a1 since ak = 0 and ak+1 = a1

also we notice that a2 = f (a1 ) ≡ f (0) (mod a1 ) which gives a2 − a1 ≡ 0 (mod a1 ),

hence we have a1 |a2 − a1 |a3 − a2 | · · · |an − an−1 | · · · |ak+1 − ak |a1 , which gives a2 − a1 =

9
piyush kumar jha — March 21, 2023 Polynomials Problem Set

±a1 , a3 − a2 = ±a1 , · · · , ak+1 − ak = ±a1

now we observe that ki=1 ai − ai−1 = ak = 0 , which gives that we must have at least
P
a ℓ s.t. aℓ − aℓ−1 = −(aℓ+1 − aℓ ) hence we get that aℓ−1 = aℓ+1 , take over f until we get
=⇒ ak = ak+2 = a2 which proves a2 = 0 and done

§12 Twelfth Problem

Example 12.1 (Russia 1995)

Can the equation f (g(h(x))) = 0, where f , g, h are quadratic polynomials, have the
solutions 1, 2, 3, 4, 5, 6, 7, 8? Proposed by S. Tokarev

Solution.

Claim 12.2 — There do not exist such f, g and h

Proof:- FTSOC consider there exist such f, g and h

so we get that f (g(h(1))) = 0, f (g(h(2))) = 0, · · · , f (g(h(7))) = 0 and f (g(h(8))) = 0

consider set S = {h(1), h(2), · · · , h(7), h(8)}

since we have f (g(x)) to be a biquadratic we must have 0 < |S| ⩽ 4 and since h(x) is
a quadratic polynomial we must have that h(x) can have equal values at atmost two points

so from this, we set:

WLOG

h(1) = h(8)

h(2) = h(7)

h(3) = h(6)

h(4) = h(5)

now since we also had that f (x) is a quadratic polynomial we get that:

WLOG

g(h(1)) = g(h(4)) and g(h(2)) = g(h(3))

now consider g(x) = c(x − a)(x − b) for some c ̸= 0

we get that h(1)2 − h(1)(a + b) + ab = h(4)2 − h(4)(a + b) + ab , now since h(1) ̸= h(4)
we get:

a + b = h(1) + h(4) and similarly a + b = h(2) + h(3)

10
piyush kumar jha — March 21, 2023 Polynomials Problem Set

now setting h(x) = γ(x − α)(x − β) for some γ ̸= 0 we get:

h(1) + h(4) = 17 − 5(α + β) + αβ = 13 − 5(α + β) + αβ which gives a contradiction!

hence our assumption was wrong and hence there do not exist any such quadratic
polynomials f, g and h as the claim follows □

§13 Thirteenth Problem

Example 13.1 (Greek National 2020 )

Find all non constant polynomials P (x), Q(x) with real coefficients such that:
P ((Q(x))3 ) = xP (x)(Q(x))3

Solution.

Claim 13.2 — P (x) = ax2 ∀ a ∈ R∗ , Q(x) = x are the only polynomial satisfying
the problem statement
Proof:- denote Deg(P(x)) = m, Deg(Q(x)) = n

so we get 3mn = 3n + m + 1 =⇒ (3n − 1)(m − 1) = 2 =⇒ n = 1, m = 2

so we set P (x) = ax2 + b and Q(x) = jx + k where a, j ∈ R∗ and b, k ∈ R

now we notice a(jx + k)6 + b = x(ax2 + b)(jx + k)3

now plugging x = 0 we get:

ak 6 + b = 0

also comparing coefficient of x6 we get that aj 6 = aj 3 =⇒ j = 1

also comparing coefficient of x we get that ak 5 j = k 3 b =⇒ ak 5 = k 3 b

Case1:- k = 0

clearly, this gives b = 0

which gives P (x) = ax2 and Q(x) = x

Case2:- k ̸= 0

then we have ak 2 = b but we also had ak 6 = −b so we get k 4 = −1 which is clearly

not possible

hence we get P (x) = ax2 ∀ a ∈ R∗ , Q(x) = x as the only solutions , as claim follows
□

11
piyush kumar jha — March 21, 2023 Polynomials Problem Set

§14 Fourteenth Problem

Example 14.1 (Belarusian National Olympiad 2007)

Find all polynomials with degree ≤ n and nonnegative coefficients, such that
P (x)P ( x1 ) ≤ P (1)2 for every positive x
Pj ia
Solution. we set P (x) = i=1 x i for some j ≤ n

1

Pj ai
so we get P x = i=0 xi
P
j
2 P
j √ √ √ √ 2
i=0 ai = i=0 ai x · ai · x−i
i

so from C-S, we get:

P 2 P  P 
j j i j 1
i=0 ai ≤ i=0 ai x · i=0 ai xi

1 1
so we get P (1)2 ≤ P · P (x) ≤ P (1)2 =⇒ P (1)2 = P (x) · P



x x

now we equate the degrees to get that P (x) = aj xj for some aj ∈ R+ □

§15 Fifteenth Problem

Example 15.1 (Putnam 1940 A1)

Prove that if f (x) is a polynomial with integer coefficients and there exists an integer
k such that none of f (1), . . . , f (k) is divisible by k, then f (x) has no integral root.

Proof. f (x) ∈ Z[X]

FTSOC f has an integer root n.

set n − 1 = km + r where 0 ≤ r < k

since f (n) = 0 =⇒ f (mk + r + 1) = 0

reducing (mod k) both sides we observe:

f (r + 1) ≡ 0 (mod k)

but since r + 1 ∈ {1, 2, · · · , k} we can’t have f (r + 1) ≡ 0 (mod k)

hence our assumption was wrong and hence we get that f has no integral roots

§16 Sixteenth Problem

12
piyush kumar jha — March 21, 2023 Polynomials Problem Set

Example 16.1 (India TST 2001)

Complex numbers α , β , γ have the property that αk + β k + γ k is an integer for
every natural number k. Prove that the polynomial

(x − α)(x − β)(x − γ)

has integer coefficients.

Proof.

Claim 16.2 — αβγ is an integer

Proof:- for k = 3

α3 + β 3 + γ 3 ∈ Z

we have α3 + β 3 + γ 3 =
( α)P α2
− α · β + 3αβγ =⇒ 2(α3 + β 3 + γ 3 ) =

P P P

(α + β + γ) 2 α2 + β 2 + γ 2 − 2 α · β + 6αβγ

which gives 6αβγ ∈ Z

similarly for k = 6 we get

6α2 β 2 γ 2 ∈ Z

a
now we have αβγ = 6 for a ∈ Z

a2
6α2 β 2 γ 2 = 6 =⇒ a2 ≡ 0 (mod 6) =⇒ a ≡ 0 (mod 6) =⇒ αβγ ∈ Z. Hence claim
follows ■

P
Claim 16.3 — α·β ∈Z
P P b
Proof:- Clearly 2 α · β ∈ Z , so take α·β = 2 where b ∈ Z

2
α2 α4 + 2 α2 · β 2 =⇒ 2 α2 · β 2 ∈ Z
P P P P
and we have =

b2
α · β)2 =
P
so ( 4

b2 b2 b2
α2 · β 2 + 2αβγ(α + β + γ) = α2 · β 2 + 4αβγ · (α + β + γ) =
P P
4 =⇒ 2 2 =⇒ 2 ∈Z

which gives b2 ≡ 0 (mod 2) =⇒ b ≡ 0 (mod 2) which gives

P
α·β ∈Z

hence claim follows ■

hence (x − α)(x − β)(x − γ) = x3 − αx2 +

P P
α · βx − α · β · γ = 0 has integral
coefficients □

§17 Seventeenth Problem

13
piyush kumar jha — March 21, 2023 Polynomials Problem Set

Example 17.1 (Hungary-Israel MC, 1991)

Suppose f (x) is a polynomial with integer coefficients such that f (0) = 11 and
f (x1 ) = f (x2 ) = ... = f (xn ) = 2002 for some distinct integers x1 , x2 , ..., xn . Find the
largest possible value of n.

Qn
Solution. we can set f (x) = i=1 (x − xi )Q(x) + 2002 for some Q(x) ∈ Z[X]

f (0) = (−1)n · ( ni xi ) Q(0) + 2002

Q

clearly this gives (−1)n · ( ni xi ) Q(0) = −1991

Q

Claim 17.2 — n ≤ 4
Qn
Proof:- since we have xi ̸= xj we observe we can’t have (−1)n · ( i xi ) = 1 and
Q(0) = −1991

so we have either (−1)n · ( ni xi ) = −1, Q(0) = 1991 or (−1)n · ( ni xi ) = −1991 and

Q Q
Q(0) = 1 Q
or (−1)n · ( ni xi ) = 1991 and Q(0) = −1 , and some more cases

clearly from above cases we get only n = 4 to be the maximal case working ( for
(x1 , x2 , x3 , x4 ) ∈ {−1, 1, 11, 181})

hence we have n ≤ 4 as claim follows ■

hence maximal n that works is 4

§18 Eighteenth Problem

Example 18.1 (CentroAmerican 2013)

Determine all pairs of non-constant polynomials p(x) and q(x), each with leading
coefficient 1, degree n, and n roots which are non-negative integers, that satisfy
p(x) − q(x) = 1.

Solution. first of all p(x) = 1 + q(x)

set

p(x) = (x − α1 )(x − α2 )(x − α3 ) · · · (x − αn )

q(x) = (x − β1 )(x − β2 )(x − β3 ) · · · (x − βn )

s.t. α1 ≤ α2 ≤ α3 · · · ≤ αn and β1 ≤ β2 ≤ β3 · · · ≤ βn

where αi , βi ∈ Z∗
Case1:- αn ≥ βi ∀ 1 ≤ i ≤ n

14
piyush kumar jha — March 21, 2023 Polynomials Problem Set

Claim 18.2 — only n = 1 works

Proof:- q(α1 ) = ni=1 (α1 − βi )
Q

Qn
clearly q(αn ) = −1 =⇒ i=1 (αn − βi ) = −1 =⇒ β1 = β2 = β3 = · · · = βn := β ∈ Z∗

so q(x) = (x − β)n and p(x) = (x − β)n + 1 =⇒ n has to be odd

n(n−1)β 2 P 2
= nβ 2
P
now i̸=j αi · αj = 2 , αi

from C-S we get:

P 2
2
αi2 =⇒ (n − 1)2 ≤ 4 =⇒ n ≤ 3
P
i̸=j αi · αj ≤

so clearly n = 1 or 3 but clearly n = 3 don’t works as (x − β)3 + 1 has not all roots to
be integer

so we get n = 1 to be working , hence claim follows ■

p(x) = x − β + 1 and q(x) = x − β

Case 2:- βn ≥ αi ∀ 1 ≤ i ≤ n ∈ Z∗

Claim 18.3 — only n = 2 works

Proof:- clearly p(βn ) = ni=1 (βn − αi )
Q

Qn
so p(βn ) = q(βn ) + 1 =⇒ p(βn ) = 1 =⇒ i=1 (βn − αi ) = 1 =⇒ α1 = α2 = α3 =
· · · αn := α ∈ Z∗

so p(x) = (x − α)n and q(x) = (x − α)n − 1 =⇒ n has to be even

n(n−1)α2
βi2 = nα2
P P
now again i̸=j βi · βj = 2 and

again from C-S we get:

n ≤ 3 , so n = 2 works as claim follows ■

hence p(x) = (x − α)2 and q(x) = (x − α)2 − 1

so clubbing all solutions we got we get:

(p(x) = x − β + 1, q(x) = x − β), (p(x) = (x − α)2 , q(x) = (x − α)2 − 1) to be the poly-

nomials satisfying given condition for some α, β ∈ Z∗

§19 Nineteenth Problem

15
piyush kumar jha — March 21, 2023 Polynomials Problem Set

Example 19.1 (STEMS 2023 Maths CAT A Part A)

Consider a polynomial P (x) ∈ R[x], with degree 2023, such that P (sin2 (x)) +
p
P (cos2 (x)) = 1 for all x ∈ R. If the sum of all roots of P is equal to with p, q
q
coprime, then what is the product pq?

Solution. so first of all as we know that sin2 (x) + cos2 (x) = 1 =⇒ sin2 x = 1 − cos2 x

now we get P (y) + P (1 − y) = 1 ∀y

now just take p(y) = a2023 x2023 + a2022 x2022 + · · · + a0

we get a2023 + 2a2022 = 0

−a2022 2023
so we get a2023 = 2

hence we get pq = 4046

§20 Twentieth Problem

Example 20.1 (Austrian MO 2017)

Determine all polynomial P (x) ∈ R[x] satisfying the following two conditions: (a)
P (2017) = 2016 and (b) (P (x) + 1)2 = P (x2 + 1) for all real number x.

Solution. well trivially 2016 and 2017 really don’t play a significant role here so we go
for this : P (n) = n − 1 for some n ∈ R

so we have P (x) = (x − n)Q(x) + x − 1 for some Q(x) ∈ R[x] (♠)

now we have (P (x) + 1)2 = (x − n)2 · Q2 (x) + x2 + 2(x − n)Q(x)

P (x2 + 1) = (x2 + 1 − n)(Q(x2 + 1)) + x2 (⋄)

as we have (P (x) + 1)2 = P (x2 + 1) for all real x we get:

(x2 + 1 − n)Q(x2 + 1) = (x − n)Q(x)((x − n)Q(x) + 2)

clearly Q(n2 + 1) = 0 (♣)

so plugging (♣) in (⋄) we get P (n2 + 1) = n2 (♡)

so from (♡) we get:

P (x) = (x − (n2 + 1))J(x) + x − 1 for some J(x) ∈ R[x] =⇒ P (x) = (x − i)J(x) + x − 1

for i = n2 + 1

which is similar as (♠)

so clearly we continue like this to get that Q(x) has infinite degree but clearly this is
not possible hence Q(x) ≡ 0

16
piyush kumar jha — March 21, 2023 Polynomials Problem Set

plugging this in (♠) to get:

P (x) = x − 1

§21 Twenty-first Problem

Example 21.1 (Pan African 2000)

Define the polynomials P0 , P1 , P2 · · · by: P0 (x) = x3 + 213x2 − 67x − 2000 Pn (x) =
Pn−1 (x − n), n ∈ N Find the coefficient of x in P21 (x).

Solution.
Pn
Claim 21.2 — Pn (x) = P0 (x − i=1 i)

Proof:-

Base case:-

n = 1 ( trivial)

assume the claim is true for n

from definition we have Pn+1 = Pn (x − (n + 1)) (♠)

Pn
=⇒ Pn (x−(n+1)) = P0 (x − (n + 1) − ni=1 i)
P
we also have Pn (x) = P0 (x − i=1 i)
 
= P0 x − n+1
P
from (♠) we conclude Pn+1 i=1 i , which makes the claim true ■

so from claim it follows P21 (x) = P0 (x − 231)

coefficient of x in P21 (x) = 3(231)2 − 213(231)(2) − 67 = 61610 □

§22 Twenty-second Problem

Example 22.1 (I.S.I 2012)

Prove that the polynomial equation x8 − x7 + x2 − x + 15 = 0 has no real solution.

Proof. so from above part (xp − 1)(x) < 0,hence we get (1 − x)(x) > 0 , so applying am
gm on (1 − x), x we get 12 ≥ (1 − x)(x) , so we get 14 ≥ (1 − x)(x) , so −1 −15
4 ≤ x6 +1 , so
we get x6 ≥ 59 , clearly x ∈ (0, 1) hence we get no solution for x , and hence no real
value of x is there for which polynomial is 0

§23 Twenty-third Problem

17
piyush kumar jha — March 21, 2023 Polynomials Problem Set

Example 23.1 (2006 iTest)

Compute the number of ordered quadruples (w, x, y, z) of complex numbers (not
necessarily nonreal) such that the following system is satisfied:

wxyz = 1
wxy + wx z + w yz + xyz 2 = 2
2 2 2

wx2 y + w2 y 2 + w2 xz + xy 2 z + x2 z 2 + ywz 2 = −3
w2 xy + x2 yz + wy 2 z + wxz 2 = −1

Solution. As we are given xyzw = 1 , so from this we get second equation as yz + xy + wx + wz =

2. so say a = yz , b = xy , wx = c, wz = d.

so we get a + b + c + d = 2.

1 1 1 1
From fourth equation we get a + b + c + d = −1.

So we get abc + abd + acd + bcd = −1.

Also from the third equation, we get ab + bc + cd + ad + w2 y 2 + x2 z 2 = −3.

1
Notice we want ac and bd. so ac = x2 z 2
. so this gives ab + bc + cd + ad + ac + bd = −3.
and abcd = 1.

So we get a equation α4 − 2α3 − 3α2 + α + 1 = 0 whose roots are a, b, c, d. so we get

(α + 1)(α3 − 3α2 + 1) = 0.

This gives α = −1. and three distinct complex ( not necessarily non-real) solutions.

So as α = −1. we get any one pair say xy = −1. so x = −y = k for some k ∈ C. so as

z, w, will be distinct we will get 4 quadruples from −k, k, w, z solution so we can have
such 4 · 4 = 16 quadruples.

§24 Twenty-Fourth Problem

Example 24.1 (Romania TST 2014)

Let p be an odd prime number. Determine all pairs of polynomials f and g from
Z[x] such that
p−1
X
f (g(X)) = X k = Φp (X).
k=0

Solution. The answer is g(x) ≡ ±x + c for any c (and f easily extractable). These clearly
work, so we show they are the only solutions.
We may verify by Eisenstein criterion that
   
d p−2 p p−3 p
f (g(x + 1)) = (p − 1)x + (p − 2) x + ··· +
dx 1 p−2

18
piyush kumar jha — March 21, 2023 Polynomials Problem Set

is irreducible.
But then dxd
f (g(x)) = g ′ (x) · f ′ (g(x)) must also be irreducible, so g is linear. Of course,
g has leading coefficient ±1, so we are done.

§25 Twenty-Fifth Problem

Example 25.1 (Romania TST 2013 )

Given an integer n ≥ 2, determine all non-constant polynomials f with complex
coefficients satisfying the condition 1 + f (xn + 1) = f (x)n

Solution. We claim that if n is odd, f (x) = −x is the only solution, and if n is even,
then there is no solution for f.
Lemma There exists a nonnegative integer k < n such that fx(x) n
k is a polynomial in x .

Proof:-

Notice that since f (x)n = f (xn + 1) + 1 is a polynomial in xn by writing out the terms
of f and expanding we can bash out the result.

Now, if k = 0 we can write f (x) = g(xn ) and g will satisfy the same condition as f
so by infinite descent we find that f (x) = cxn for some constant c ∈ C. This is clearly
impossible so we must have k > 0. In other words, we have that f (0) = 0. We now
proceed with casework on the parity of n.

Case 1: n is even.

In this case by letting x = 0 in the condition we find that f (1) = −1. Then by letting
x = 1 in the condition we find that f (2) = 0. Proceeding in this fashion, we find that
f (ak ) = 0 for all positive integers k where a1 = 0 and ak = (ank−1 + 1)n + 1. This means
that f (x) has infinitely many roots, contradiction.

Case 2: n is odd.

In this case by letting x = 0 in the condition we find that f (1) = −1. Then by letting
x = 1 in the condition we find that f (2) = −2. Proceeding in this fashion, we find that
f (bk ) = −bk for all positive integers k where a1 = 0 and ak = ank−1 + 1. Therefore the
polynomial f (x) + x has infinitely many roots, so it must be identically 0, as desired

19