Problem 1.

Find all triples (a, b, c) of real numbers such that the following system holds:
1 1 1

a + b + c = + +

a b c
a2 + b2 + c2 = 1 1 1
+ 2+ 2

a 2 b c
Solution. First of all if (a, b, c) is a solution of the system then also (−a, −b, −c) is a solution. Hence
we can suppose that abc > 0. From the first condition we have
ab + bc + ca
a+b+c= · (1)
abc
Now, from the first condition and the second condition we get
1 1 1 2
Å ã Å ã
2 2 2 2 1 1 1
(a + b + c) − (a + b + c ) = + + − + + .
a b c a2 b2 c2
The last one simplifies to
a+b+c
ab + bc + ca = · (2)
abc
First we show that a + b + c and ab + bc + ca are different from 0. Suppose on contrary then from
relation (1) or (2) we have a + b + c = ab + bc + ca = 0. But then we would have

a2 + b2 + c2 = (a + b + c)2 − 2(ab + bc + ca) = 0,

which means that a = b = c = 0. This is not possible since a, b, c should be different from 0.
Now multiplying (1) and (2) we have
(a + b + c)(ab + bc + ca)
(a + b + c)(ab + bc + ca) = .
(abc)2
Since a + b + c and ab + bc + ca are different from 0, we get (abc)2 = 1 and using the fact that abc > 0
we obtain that abc = 1. So relations (1) and (2) transform to

a + b + c = ab + bc + ca.

Therefore,
(a − 1)(b − 1)(c − 1) = abc − ab − bc − ca + a + b + c − 1 = 0.
This means that at least one of the numbers a, b, c is equal to 1. Suppose that c = 1 then relations
(1) and (2) transform to a + b + 1 = ab + a
+ b ⇒ ab = 1. Taking a = t then we have b = 1t . We can
now verify that any triple (a, b, c) = t, 1t
, 1 satisfies both conditions. t ∈ R \ {0}. From the initial
observation any triple (a, b, c) = t, 1t , −1 satisfies
conditions. t ∈ R \ {0}. So, all triples that
both
satisfy both conditions are (a, b, c) = t, 1t , 1 , t, 1t , −1 and all permutations for any t ∈ R \ {0}.

Comment by PSC. After finding that abc = 1 and

a + b + c = ab + bc + ca,

we can avoid the trick considering (a − 1)(b − 1)(c − 1) as follows. By the Vieta’s relations we have
that a, b, c are roots of the polynomial

P (x) = x3 − sx2 + sx − 1

which has one root equal to 1. Then, we can conclude as in the above solution.

1
Problem 2. Let 4ABC be a right-angled triangle with ∠BAC = 90◦ and let E be the foot of the
perpendicular from A on BC. Let Z 6= A be a point on the line AB with AB = BZ. Let (c) be the
circumcircle of the triangle 4AEZ. Let D be the second point of intersection of (c) with ZC and let
F be the antidiametric point of D with respect to (c). Let P be the point of intersection of the lines
F E and CZ. If the tangent to (c) at Z meets P A at T , prove that the points T, E, B, Z are concyclic.
Solution. We will first show that P A is tangent to (c) at A.
Since E, D, Z, A are concyclic, then ∠EDC = ∠EAZ = ∠EAB. Since also the triangles 4ABC
and 4EBA are similar, then ∠EAB = ∠BCA, therefore ∠EDC = ∠BCA.
Since ∠F ED = 90◦ , then ∠P ED = 90◦ and so
∠EP D = 90◦ − ∠EDC = 90◦ − ∠BCA = ∠EAC.
Therefore the points E, A, C, P are concyclic. It follows that ∠CP A = 90◦ and therefore the triangle
∠P AZ is right-angled. Since also B is the midpoint of AZ, then P B = AB = BZ and so ∠ZP B =
∠P ZB.

Furthermore, ∠EP D = ∠EAC = ∠CBA = ∠EBA from which it follows that the points P, E, B, Z
are also concyclic.
Now observe that
∠P AE = ∠P CE = ∠ZP B − ∠P BE = ∠P ZB − ∠P ZE = ∠EZB.

2
Therefore P A is tangent to (c) at A as claimed.
It now follows that T A = T Z. Therefore

∠P T Z = 180◦ − 2(∠T AB) = 180◦ − 2(∠P AE + ∠EAB) = 180◦ − 2(∠ECP + ∠ACB)

Thus T, P, B, Z are concyclic, and since P, E, B, Z are also concyclic then T, E, B, Z are concyclic as
required.

3
Problem 3. Alice and Bob play the following game: Alice picks a set A = {1, 2, . . . , n} for some
natural number n > 2. Then starting with Bob, they alternatively choose one number from the set
A, according to the following conditions: initially Bob chooses any number he wants, afterwards the
number chosen at each step should be distinct from all the already chosen numbers, and should differ
by 1 from an already chosen number. The game ends when all numbers from the set A are chosen.
Alice wins if the sum of all of the numbers that she has chosen is composite. Otherwise Bob wins.
Decide which player has a winning strategy.

Solution. To say that Alice has a winning strategy means that she can find a number n to form
the set A, so that she can respond appropriately to all choices of Bob and always get at the end a
composite number for the sum of her choices. If such n does not exist, this would mean that Bob has
a winning strategy instead.
Alice can try first to check the small values of n. Indeed, this gives the following winning strategy
for her: she initially picks n = 8 and responds to all possible choices made by Bob as in the list below
(in each row the choices of Bob and Alice are given alternatively, starting with Bob):
1 2 3 4 5 6 7 8
2 3 1 4 5 6 7 8
2 3 4 1 5 6 7 8
3 2 1 4 5 6 7 8
3 2 4 5 1 6 7 8
3 2 4 5 6 1 7 8
4 5 3 6 2 1 7 8
4 5 3 6 7 8 2 1
4 5 6 7 3 2 1 8
4 5 6 7 3 2 8 1
4 5 6 7 8 3 2 1
5 4 3 2 1 6 7 8
5 4 3 2 6 7 1 8
5 4 3 2 6 7 8 1
5 4 6 3 2 1 7 8
5 4 6 3 7 8 2 1
6 7 5 4 3 8 2 1
6 7 5 4 8 3 2 1
6 7 8 5 4 3 2 1
7 6 8 5 4 3 2 1
7 6 5 8 4 3 2 1
8 7 6 5 4 3 2 1
In all cases, Alice’s sum is either an even number greater than 2, or else 15 or 21, thus Alice always
wins.

4
Problem 4. Find all pairs (p, q) of prime numbers such that
pq − q p
1+
p+q
is a prime number.
Solution. It is clear that p 6= q. We set
pq − q p
1+ =r
p+q
and we have that
pq − q p = (r − 1)(p + q). (3)
From Fermat’s Little Theorem we have
pq − q p ≡ −q (mod p).
Since we also have that
(r − 1)(p + q) ≡ −rq − q (mod p),
from (3) we get that
rq ≡ 0 (mod p) ⇒ p | qr,
hence p | r, which means that p = r. Therefore, (3) takes the form
pq − q p = (p − 1)(p + q). (4)
We will prove that p = 2. Indeed, if p is odd, then from Fermat’s Little Theorem we have
pq − q p ≡ p (mod q)
and since
(p − 1)(p + q) ≡ p(p − 1) (mod q),
we have
p(p − 2) ≡ 0 (mod q) ⇒ q | p(p − 2) ⇒ q | p − 2 ⇒ q ≤ p − 2 < p.
Now, from (4) we have
pq − q p ≡ 0 (mod p − 1) ⇒ 1 − q p ≡ 0 (mod p − 1) ⇒ q p ≡ 1 (mod p − 1).
Clearly gcd(q, p − 1) = 1 and if we set k = ordp−1 (q), it is well-known that k | p and k < p, therefore
k = 1. It follows that
q≡1 (mod p − 1) ⇒ p − 1|q − 1 ⇒ p − 1 ≤ q − 1 ⇒ p ≤ q
a contradiction.
Therefore, p = 2 and (4) transforms to
2q = q 2 + q + 2.
We can easily check by induction that for every positive integer n ≥ 6 we have 2n > n2 + n + 2. This
means that q ≤ 5 and the only solution is for q = 5. Hence the only pair which satisfy the condition
is (p, q) = (2, 5).

Comment by the PSC. From the problem condition, we get that pq should be bigger than q p , which
gives
ln p ln q
q ln p > p ln q ⇐⇒ > ·
p q
ln x
The function is decreasing for x > e, thus if p and q are odd primes, we obtain q > p.
x