PROBLEMS OF THE 52nd

AUSTRIAN MATHEMATICAL
OLYMPIAD 2021
About the Austrian Mathematical
Olympiad1

The story of the Austrian Mathematical Olympiad begins fairly late, as compared to other
countries, with an observer being sent to the 10th IMO in Moscow in 1968. The first Austrian
team participated in IMO 1970. Austria hosted the 18th IMO in 1976.
In 1977, the first Austrian-Polish Mathematics Competition was held. This competition was
superseded in 2007 by the Middle European Mathematical Olympiad (MEMO), where by now
10 countries take part.
Most students taking part in the Austrian Mathematical Olympiad take part in preparatory
courses prior to the competition. These courses are held in schools by school teachers. Currently,
around 70 such courses are offered in Austrian schools.
Students who are participating in the Austrian Mathematical Olympiad for the first time
(typically in grade 9) can participate on the Juniors’ Level. This competition is held in two
rounds. The first round is the Course Competition, which is held in the courses. The problems
are chosen individually by the teachers of each course. The best students from each course
competition (a maximum of five) then qualify for the second and final round, the Junior Regional
Competition. For this round, the students are invited to a central location in each region, where
the competition takes place.
The main competition of the Austrian Mathematical Olympiad is held in four rounds.
The first of these is the Course Competition. The highest-ranked participants of the course
competition qualify for the second round, the Regional Competition. There are three such
regional competitions, all of which are held simultaneously (end of March), and in all of which
the same problems are posed.
The best-ranked students from each of the regional competitions then qualify for the third
round of competition (National Competition—Preliminary Round ). Altogether, 40 students
qualify for this round. It is a one-day contest with four problems. Prior to the competition
itself, the students who have qualified are offered a ten-day seminar preparing them for the
competition. The seminar has been held for many years now in Raach am Hochgebirge, a tiny
mountain village in the province of Lower Austria. There is virtually nothing for the students
to do there other than work on mathematics.
In the third round, 24 students qualify for the fourth round (National Competition—Final
Round ). This is a two day contest with three problems per day. It is again preceded by a
ten-day seminar in Raach am Hochgebirge. This final round also serves to select the teams for
the IMO and the MEMO.
Prior to departure to IMO or MEMO, participants attend an additional three-day training.
Those students who participated in the IMO or the MEMO of the previous year are invited to
attend two additional three-day seminars during the school year.
More information can be found at https://www.oemo.at. In particular, English problems
of the last years are available at https://www.oemo.at/problems/.
1
This is an abbreviated and updated version of: R. Geretschläger, Th. Mühlgassner and G. Perz, 25 Years
(and more) of the Austrian Mathematical Olympiad, Mathematics Competitions, Vol. 10, Nr. 1, 1997, pp. 79–96.

1
Problems

Junior Regional Competition

15th June 2021

1. The pages of a notebook are numbered consecutively such that the first sheet contains
the numbers 1 and 2, the second sheet contains the numbers 3 and 4, and so on. One
sheet is torn out of the notebook. The page numbers on the remaining sheets are added.
The resulting sum equals 2021.

(a) How many pages can the notebook have had originally?
(b) Which page numbers could be found on the sheet that has been torn out?

(Walther Janous)

2. Let ABC be a triangle with circumcenter U such that ∠CBA = 60° and ∠CBU = 45°.
Let D be the point of intersection of the lines BU and AC.
Prove that AD = DU .
(Karl Czakler)

3. The eight points A, B, . . ., G and H are placed on five circles as in the figure below. Each
of these letters will be replaced with one of the numbers 1, 2, . . ., 7 and 8 such that the
following two conditions hold:

(i) Each of the eight numbers is used exactly once.

(ii) The sum of the numbers on each of the five circles is the same.

How many possibilities are there to replace the letters with numbers in this way?

D C

H F

A B

2
 (Walther Janous)

4. Let p be a prime and let m and n be positive integers such that p2 + m2 = n2 .

Prove that m > p.
(Karl Czakler)

Regional Competition
25th March 2021

5. Let a and b be positive integers and c be a positive real number satisfying

a+1 b
= .
b+c a

Prove that c ≥ 1 holds.

(Karl Czakler)

6. Let ABC be an isosceles triangle with AC = BC and circumcircle k. The point D lies
on the shorter arc of k over the chord BC and is different from B and C. Let E denote
the intersection of CD and AB.
Prove that the line through B and C is a tangent of the circumcircle of the triangle BDE.
(Karl Czakler)

7. The numbers 1, 2, . . . , 2020 and 2021 are written on a blackboard. The following operation
is executed:
Two numbers are chosen, both are erased and replaced by the absolute value of their
difference.
This operation is repeated until there is only one number left on the blackboard.

(a) Show that 2021 can be the final number on the blackboard.
(b) Show that 2020 cannot be the final number on the blackboard.

(Karl Czakler)

8. Determine all triples (x, y, z) of positive integers satisfying

x | (y + 1), y | (z + 1) and z | (x + 1).

(Walther Janous)

3
National Competition—Preliminary Round
1st May 2021

9. Let a, b, c be positive real numbers with a + b + c = 1.

Prove that
a b c 1
+ + ≤ .
2a + 1 3b + 1 6c + 1 2
When does equality hold?
(Karl Czakler)

10. Let ABC denote a triangle. The point X lies on the extension of AC beyond A, such
that AX = AB. Similarly, the point Y lies on the extension of BC beyond B such that
BY = AB.
Prove that the circumcircles of ACY and BCX intersect a second time in a point different
from C that lies on the bisector of the angle ∠BCA.
(Theresia Eisenkölbl)

11. Let n ≥ 3 be an integer.

On a circle, there are n points. Each of them is labelled with a real number at most 1 such
that each number is the absolute value of the difference of the two numbers immediately
preceding it in clockwise order.
Determine the maximal possible value of the sum of all numbers as a function of n.
(Walther Janous)

12. On a blackboard, there are 17 integers not divisible by 17. Alice and Bob play a game.
Alice starts and they alternately play the following moves:

• Alice chooses a number a on the blackboard and replaces it with a2 .

• Bob chooses a number b on the blackboard and replaces it with b3 .

Alice wins if the sum of the numbers on the blackboard is a multiple of 17 after a finite
number of steps.
Prove that Alice has a winning strategy.
(Daniel Holmes)

National Competition—Final Round

4th/5th June 2021

First day
13. Let a, b, c be pairwise distinct natural numbers.
Prove that
a3 + b 3 + c 3
≥ abc + a + b + c.
3
When does equality hold?
(Karl Czakler)
4
 14. Mr. Precise wants to take his tea cup out of the microwave precisely at the front. The
microwave of Mr. Precise is not precisely cooperative.
More precisely, the two of them play the following game:
Let n be a positive integer. The rotating plate of the microwave takes n seconds for
a full turn. Each time the microwave is turned on, the plate is turned clockwise or
counterclockwise for an integer number of seconds such that the tea cup can end up in n
possible positions. One of these positions is marked „front“.
At the start of the game, the microwave rotates the tea cup in one of these positions.
Afterwards, for each move, Mr. Precise enters the integer number of seconds and the
microwave decides whether to turn clockwise or counterclockwise.
For which n can Mr. Precise ensure that after a finite number of moves, he can take out
the tea cup of the microwave precisely from the front position?
(Birgit Vera Schmidt)

15. Determine all triples (a, b, c) of integers a ≥ 0, b ≥ 0 und c ≥ 0 that satisfy the equation

ab+20 (c − 1) = cb+21 − 1.

(Walther Janous)

Second day
16. Let α be a real number.
Determine all functions f : R → R satisfying

f (f (x) + y) = f (x2 − y) + αf (x)y

for all x, y ∈ R.
(Walther Janous)

17. Let ABCD be an inscribed convex quadrilateral with diagonals AC and BD. Each of
the four vertices is reflected on the diagonal it does not lie on.
Prove that the resulting four points lie on a common circle or a common line.

(a) Investigate when the four resulting points lie on a common line and give a simple
equivalent condition for the quadrilateral ABCD.
(b) Prove that in all other cases, the four resulting points lie on a common circle.

(Theresia Eisenkölbl)
p2 +1
18. Suppose that p is an odd prime number and M a set of 2
integer squares.
Investigate if one can choose p elements of this set so that the arithmethic mean of these
p elements is an integer.
(Walther Janous)

5
Solutions

Junior Regional Competition

15th June 2021

Problem 1. The pages of a notebook are numbered consecutively such that the first sheet
contains the numbers 1 and 2, the second sheet contains the numbers 3 and 4, and so on.
One sheet is torn out of the notebook. The page numbers on the remaining sheets are added.
The resulting sum equals 2021.

(a) How many pages can the notebook have had originally?

(b) Which page numbers could be found on the sheet that has been torn out?

(Walther Janous)

Answer. There is exactly one solution. The notebook had 64 pages and the sheet with the page
numbers 29 and 30 is ripped out.

Solution. Let b > 0 be the number of sheets. The number of pages will be 2b. We are looking
for a number 2b such that
(2b) · (2b + 1)
1 + 2 + · · · + (2b − 1) + 2b = > 2021.
2
2
Since 602 = 1800 has the right order of magnitude, we check the integers beginning with b = 30
and find:
62 · 63 64 · 65
= 1953 < 2021 < = 2080.
2 2
Therefore, the smallest possible number of pages is 64.
The torn out sheet h contains the page numbers 2h − 1 and 2h (first odd, then even).
This gives the equation

2h − 1 + 2h = 2080 − 2021 = 59

which implies h = 15.

So, one solution is that the book originally had 32 sheets and the 15th sheet with page
numbers 29 and 30 has been torn out.
It remains to explain why this is the only solution. If the book has 64 sheets, this is the only
possibility because h could be computed uniquely. Now, assume that the number b of sheets is
larger than 32 and the number of pages at least 66.
The sheet that has been torn out can have at most page numbers 2b − 1 and 2b, so the
remaining sum is at least 1 + 2 + · · · + 63 + 64 = 2080 > 2021. So there cannot be a solution
with more than 32 sheets.
(Lukas Donner)
6
Problem 2. Let ABC be a triangle with circumcenter U such that ∠CBA = 60° and ∠CBU =
45°. Let D be the point of intersection of the lines BU and AC.
Prove that AD = DU .
(Karl Czakler)

Solution.
In the isosceles triangle AU B, we have

∠BAU = ∠U BA = 60° − 45° = 15°,

and therefore
∠AU B = 180° − ∠BAU − ∠U BA = 150°.
The inscribed angle theorem implies
1
∠BCA = ∠BU A = 75°,
2
and therefore
∠BAC = 180° − 60° − 75° = 45°.
We can finally compute the two angles of interest:

∠U AD = ∠BAD − ∠BAU = ∠BAC − ∠BAU = 45° − 15° = 30°

∠DU A = 180° − ∠AU B = 180° − 150° = 30°

Therefore, the triangle AU D is isosceles with apex D and we have AD = DU as desired.

(Theresia Eisenkölbl)

Problem 3. The eight points A, B, . . ., G and H are placed on five circles as in the figure
below. Each of these letters will be replaced with one of the numbers 1, 2, . . ., 7 and 8 such that
the following two conditions hold:

(i) Each of the eight numbers is used exactly once.

7
 (ii) The sum of the numbers on each of the five circles is the same.
How many possibilities are there to replace the letters with numbers in this way?

D C

H F

A B

(Walther Janous)
Answer. There are eight possibilities.

Solution. Let a, b, . . . , g and h be the numbers that replace the letters A, B, . . . , G and H. Since
the sum on each circle has the value a + b + c + d, we immediately get e = c + d, f = d + a,
g = a + b and h = b + c. This implies e + f + g + h = 2(a + b + c + d). On the other hand,

36 = 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 = (a + b + c + d) + (e + f + g + h) = 3(a + b + c + d),

and therefore a + b + c + d = 12.

Consequently, we get e + g = (c + d) + (a + b) = 12 and f + h = (d + a) + (b + c) = 12.
Therefore, the pairwise distinct numbers e, f, g, h have to come from the set {4, 5, 7, 8} where
8 and 4, as well as 7 and 5 lie on opposite points in the figure. There are eight possibilities
to replace the letters E, F, G, H in this way with the numbers 4, 5, 7, 8, because we can freely
choose one of the four letters for 8 and then we have two choices for 7.
It remains to show that these choices of E, F, G, H determine the numbers in A, B, C, D
uniquely. Without loss of generality, let g = 8 and h = 7, and therefore e = 4 and f = 5.
From e = 4 = c + d and f = 5 = d + a we get that only b can take the value 6. The values
a = 2, c = 1, d = 3 are now a direct consequence of the circle sums. Therefore, there are 8
possibilities.
(Stefan Leopoldseder)

Problem 4. Let p be a prime and let m and n be positive integers such that p2 + m2 = n2 .
Prove that m > p.
(Karl Czakler)

Solution. We have p2 = n2 − m2 = (n − m)(n + m). Since p is a prime, the number p2 has the
divisors 1, p and p2 . Since the two factors n − m and n + m are distinct, they cannot be both
equal to p. Furthermore, n − m is smaller than n + m, therefore, n − m = 1, i. e. n = m + 1.
We find

p2 + m2 = (m + 1)2 ⇐⇒ p2 = 2m + 1.
8
 This immediately implies that p is odd, therefore p ≥ 3. We find 2m + 1 = p2 ≥ 3p > 2p + 1,
which gives m > p as desired.
(Reinhard Razen)

Regional Competition
25th March 2021

Problem 5. Let a and b be positive integers and c be a positive real number satisfying
a+1 b
= .
b+c a
Prove that c ≥ 1 holds.
(Karl Czakler)

Solution. The following equations are equivalent to the given equation:

a2 + a = b2 + bc
4a2 + 4a + 1 = 4b2 + 4bc + 1
(2a + 1)2 = 4b2 + 4bc + 1.

Assume to the contrary that c < 1 holds. This yields

(2b)2 = 4b2 < (2a + 1)2 = 4b2 + 4bc + 1 < 4b2 + 4b + 1 = (2b + 1)2 .

This is a contradiction as the square of an integer cannot lie strictly between two consecutive
square numbers. Therefore, c ≥ 1 holds (for instance, a = b yields c = 1 and therefore there is
a solution of the equation with c ≥ 1).
(Karl Czakler)

Problem 6. Let ABC be an isosceles triangle with AC = BC and circumcircle k. The point
D lies on the shorter arc of k over the chord BC and is different from B and C. Let E denote
the intersection of CD and AB.
Prove that the line through B and C is a tangent of the circumcircle of the triangle BDE.
(Karl Czakler)

Solution. We denote the center of the circumcircle of the triangle BDE by M and ∠BAC =
∠CBA by α. Since the quadrilateral ABDC is cyclic, we obain ∠BDE = α. By the inscribed
angle theorem, ∠BM E = 2α and thus ∠EBM = ∠M EB = 90◦ − α. Therefore,

180◦ = ∠CBA + ∠M BC + ∠EBM = α + ∠M BC + 90◦ − α = ∠M BC + 90◦

holds and we get

∠M BC = 90◦ ,
completing the proof.
(Karl Czakler)

9
 C

A B E

Figure 1: Problem 6

Problem 7. The numbers 1, 2, . . . , 2020 and 2021 are written on a blackboard. The following
operation is executed:
Two numbers are chosen, both are erased and replaced by the absolute value of their difference.
This operation is repeated until there is only one number left on the blackboard.

(a) Show that 2021 can be the final number on the blackboard.

(b) Show that 2020 cannot be the final number on the blackboard.

(Karl Czakler)

Solution. (a) Let us first choose the following 1010 pairs of numbers:

(1, 2); (3, 4); . . . ; (2019, 2020).

The absolute value of the difference within each of these pairs is 1. After applying the
operation for each of these pairs, the number 2021 and 1010 times the number 1 remain
on the blackboard. Now we execute the given operation 505 times with pairs of the form
(1, 1). Then the number 2021 and 505 times the number 0 remain on the blackboard. As
2021 − 0 = 2021 and 0 − 0 = 0, we end up with 2021 as the final number on the board
after additional 505 operations, regardless of the pairs we pick at each step.

(b) We prove a more general statement: The final remaining number on the blackboard
cannot be even.
As
a−b≡a+b (mod 2),
we obtain that the parity of the sum of all numbers on the board is an invariant throughout
the game. At the beginning, the sum of the numbers on the blackboard is
2021 · 2022
= 2021 · 1011,
2
an odd number. Therefore, the final number on the board must be odd as well. In
particular, 2020 cannot be the final number on the blackboard.
(Karl Czakler)
10
Problem 8. Determine all triples (x, y, z) of positive integers satisfying

x | (y + 1), y | (z + 1) and z | (x + 1).

(Walther Janous)

Answer. There are ten triples satisfying the three conditions. They are given by (1, 1, 1),
(1, 1, 2), (1, 3, 2), (3, 5, 4) and their cyclic permutations.

Solution. Without loss of generality, let x be the smallest of the three numbers (or one of the
smallest), i.e. x ≤ y and x ≤ z. From z | x + 1 we obtain x ≤ z ≤ x + 1. Thus we have to
consider two cases.

• Case 1. Let z = x. Then z = x | x + 1 leads to x = z = 1 and y | z + 1 = 2. Therefore

y = 1 or y = 2, and we get the two solutions (1, 1, 1) and (1, 2, 1).

• Case 2. Let z = x + 1. Then the two conditions x | y + 1 and y | x + 2 must be fulfilled.

In particular, we obtain x ≤ y + 1 and y ≤ x + 2. This yields x − 1 ≤ y ≤ x + 2 and we
have to examine the following cases for y.

◦ Case 2a. Let 0 < y = x. The conditions x | x + 1 and x | x + 2 can only hold
simultaneously for x = 1, giving the solution (1, 1, 2).
◦ Case 2b. Let y = x + 1. Then the two conditions are x | x + 2 and x + 1 | x + 2.
They cannot hold simultaneously.
◦ Case 2c. Let y = x + 2. The condition y = x + 2 | x + 2 = z + 1 is trivially fulfilled.
The requirement x | y + 1 = x + 3 can only hold for x | 3. And, indeed, for either
x = 1 or x = 3 the condition is fulfilled and we obtain the solutions (1, 3, 2) and
(3, 5, 4).

Summing up, the triples (1, 1, 1), (1, 2, 1), (1, 1, 2), (1, 3, 2) and (3, 5, 4) fulfill all three conditions.
As each of the three numbers can be the minimum, every cyclic permutation of these triples
is a solution as well.
(Lukas Donner)

National Competition—Preliminary Round

1st May 2021

Problem 9. Let a, b, c be positive real numbers with a + b + c = 1.

Prove that
a b c 1
+ + ≤ .
2a + 1 3b + 1 6c + 1 2
When does equality hold?
(Karl Czakler)

Solution. We will use the following inequalities:

a 2a + 1 b 3b + 1 c 6c + 1
≤ , ≤ and ≤ .
2a + 1 8 3b + 1 12 6c + 1 24

11
 They are an immediate consequence of the arithmetic-geometric mean inequality with the
pairs of values 1 and 2a, 1 and 3b, and 1 and 6c, so that equality holds for a = 1/2, b = 1/3
and c = 1/6.
From these inequalities and the condition a + b + c = 1, we obtain
a b c 2a + 1 3b + 1 6c + 1 1
+ + ≤ + + = .
2a + 1 3b + 1 6c + 1 8 12 24 2
Therefore, the given inequality is true and equality holds exactly for a = 1/2, b = 1/3 and
c = 1/6.
(Karl Czakler)

Problem 10. Let ABC denote a triangle. The point X lies on the extension of AC beyond A,
such that AX = AB. Similarly, the point Y lies on the extension of BC beyond B such that
BY = AB.
Prove that the circumcircles of ACY and BCX intersect a second time in a point different
from C that lies on the bisector of the angle ∠BCA.
(Theresia Eisenkölbl)

α β
B
A

Ic
X
wγ

Figure 2: Problem 10

Solution. As usual, we denote the angles of the triangle at A, B and C with α, β and γ.
It is sufficient to show that the center Ic of the excircle touching the line AB lies on the two
circles. To do this, we look at the respective inscribed angles.
Since the triangle AY B is isosceles, the following holds:
∠CY A = ∠BY A = 90◦ − ∠Y BA/2 = 90◦ − 90◦ + β/2 = β/2.
12
But it is also true that

∠CIc A = 180◦ − (180◦ − α)/2 − α − γ/2 = 90◦ − α/2 − γ/2 = β/2.

So Ic lies on the circumcircle of ACY by the inverse of the inscribed angle theorem. In the
same way, one also obtains that Ic lies on the circumcircle of BCX. So Ic is the second point
of intersection, which therefore lies on the angle bisector through C as required.
(Theresia Eisenkölbl)

Problem 11. Let n ≥ 3 be an integer.

On a circle, there are n points. Each of them is labelled with a real number at most 1
such that each number is the absolute value of the difference of the two numbers immediately
preceding it in clockwise order.
Determine the maximal possible value of the sum of all numbers as a function of n.
(Walther Janous)

Answer. If n is a multiple of 3, the maximal sum is 2n/3. Otherwise, the sum is always 0.

Solution. All the numbers are absolute values, so they are positive or zero. Either all of them
are zero, then their sum is also zero, or there is a maximal positive number. If we scale all
numbers such that this maximum is 1, the sum can only get larger, therefore, we may assume
that the maximum is 1 in this case.
We observe that the difference of two such neighboring numbers smaller than 1 is also
smaller than 1. If we iterate around the circle, we see that all numbers have to be smaller than
1 which is impossible if the maximum is 1.
Therefore, in the case of maximum 1, at least one of each pair of neighbors must equal 1.
We can now distinguish two cases for the two numbers after the maximum. Either the number
immediately after the maximum is also 1 which means that the list of differences continues as
1, 1, 0, 1, 1, 0, 1, 1, 0, . . . or the next numbers are 1, x, 1. However this means that |1 − x| = 1 so
that x = 0 and we get the list 1, 0, 1, 1, 0, 1, 1, 0, 1, . . . .
In both these subcases, n has to be divisible by 3 to wrap correctly around the circle with
these patterns. Then, we get a sum of 2n/3.
In conclusion, we get maximal sum 0 if n is not divisible by 3 and 2n/3 if n is divisible by
3.
(Theresia Eisenkölbl)

Problem 12. On a blackboard, there are 17 integers not divisible by 17. Alice and Bob play a
game. Alice starts and they alternately play the following moves:

• Alice chooses a number a on the blackboard and replaces it with a2 .

• Bob chooses a number b on the blackboard and replaces it with b3 .

Alice wins if the sum of the numbers on the blackboard is a multiple of 17 after a finite
number of steps.
Prove that Alice has a winning strategy.
(Daniel Holmes)

13
Solution. Since both the problem statement and the winning condition are given in terms of
divisibility by 17, it is sufficient to consider the numbers modulo 17. In the beginning, all the
remainders are different from zero and Alice wins if the sum modulo 17 becomes zero.
The moves a 7→ a2 and b 7→ b3 turn remainders into powers of the original nonzero values.
Therefore, Fermat’s little theorem can be applied. For a ̸≡ 0 (mod 17) and the prime number
17, one has
a16 ≡ 1 (mod 17).
So if Alice squares the same number a four times in a row, then the remainder 1 modulo 17
is always obtained. Bob cannot do anything about it, because if Bob raises this number to the
third power k times, we get a result of
4 ·3k k k
a2 = 163 ≡ 13 = 1 (mod 17).
The timing of Bob’s moves does not matter, as the order of the factors in the exponent does
not change anything.
Therefore, Alice can make all the remainders equal to 1 by squaring each number four times.
Then of course the sum is 17 · 1 ≡ 0 (mod 17) and Alice has won.
(Theresia Eisenkölbl)

National Competition—Final Round

4th/5th June 2021

Problem 13. Let a, b, c be pairwise distinct natural numbers.

Prove that
a3 + b 3 + c 3
≥ abc + a + b + c.
3
When does equality hold?
(Karl Czakler)

Solution. It is well-known and easily verified that

1
a3 + b3 + c3 − 3abc = (a + b + c)((a − b)2 + (b − c)2 + (c − a)2 ). (1)
2
Assume without loss of generality that a > b > c ≥ 0. Since the numbers are integers, we
obtain a − b ≥ 1, b − c ≥ 1 and a − c ≥ 2.
Equation (1) now implies
1
a3 + b3 + c3 − 3abc ≥ (a + b + c)(1 + 1 + 4) = 3(a + b + c)
2
as desired.
Equality holds for a = b + 1, b = c + 1 and a = c + 2, which are exactly the triples
(t + 2, t + 1, t) where t ≥ 0 is an integer, and for all their permutations.
(Karl Czakler)

Problem 14. Mr. Precise wants to take his tea cup out of the microwave precisely at the front.
The microwave of Mr. Precise is not precisely cooperative.
More precisely, the two of them play the following game:
Let n be a positive integer. The rotating plate of the microwave takes n seconds for a full
turn. Each time the microwave is turned on, the plate is turned clockwise or counterclockwise
14
for an integer number of seconds such that the tea cup can end up in n possible positions. One
of these positions is marked „front“.
At the start of the game, the microwave rotates the tea cup in one of these positions.
Afterwards, for each move, Mr. Precise enters the integer number of seconds and the microwave
decides whether to turn clockwise or counterclockwise.
For which n can Mr. Precise ensure that after a finite number of moves, he can take out
the tea cup of the microwave precisely from the front position?
(Birgit Vera Schmidt)
Answer. Mr. Precise can ensure his victory when n is a power of 2.

Solution. We label the positions consecutively 0, 1, . . . , n − 1 where 0 is the front position.

If n is a power of 2, say n = 2k , Mr. Precise can simply always put in the current position
as number of seconds. If the microwave turns the plate backwards, the tea cup will end up
front immediately. Otherwise, the number of the position will be doubled and reduced modulo
2k at each turn and therefore be divisible by 2k after at most k turns. This means that the tea
cup ends up front.
Now, let n = 2k · m, where m > 1 is an odd number.
We will show that the microwave can always choose a number not divisible by m.
This is clearly true for the first position, for example by choosing the position 1. After that,
the tea cup is in a certain position p not divisible by m and Mr. Precise puts in s seconds. If
both p + s and p − s were divisible by m, this would also be true for the sum, so that m | 2p.
Since m is odd, this implies m | p which is wrong.
(Stephan Pfannerer)

Problem 15. Determine all triples (a, b, c) of integers a ≥ 0, b ≥ 0 und c ≥ 0 that satisfy the
equation
ab+20 (c − 1) = cb+21 − 1.

(Walther Janous)
Answer. {(1, b, 0) : b ∈ Z>0 } ∪ {(a, b, 1) : a, b ∈ Z>0 }

Solution. One can first see that the right side factors:

ab+20 (c − 1) = (cb+20 + cb+19 + · · · + c + 1)(c − 1).

The case c = 1 will be handled separately (and is very simple). For c ̸= 1 the equation simplifies
to
ab+20 = cb+20 + cb+19 + · · · + c + 1.
We therefore distinguish the two cases for c.
• c = 1 leads to
0 = 0.
Therefore, in this case, arbitrary natural numbers a and b are solutions.

• For c ̸= 1 we can divide by c −1 (see the above equations) and get the equivalent equation

ab+20 = cb+20 + cb+19 + · · · + c + 1.

Obviously,
cb+20 + cb+19 + · · · + c + 1 > cb+20 .
15
 Therefore a ≥ c + 1 must hold. Because of the binomial theorem we, thus, obtain

ab+20 ≥ (c + 1)b+20
   
b+20 b + 20 b+19 b + 20
=c + c + ··· + c+1
1 b + 19
≥ cb+20 + cb+19 + · · · + c + 1
= ab+20 .

Hence, both inequalities must be equations.

We consider the second inequality in particular. Because of
 
b + 20
= b + 20 > 1
1

this can only be an equation if c = 0. In the case of c > 0, the second inequality is strict
and therefore leads to a contradiction and there is no solution.
In the remaining case c = 0, the resulting equation

ab+20 = 1

is easy to solve. Since a is a natural number, a = 1. (This also follows from the necessary
relationship a = c + 1.) Hence, in thic case b may be any natural number.

Alternatively, one can see in the case c > 0 that

cb+20 < cb+20 + cb+19 + · · · + c + 1

   
b+20 b + 20 b+19 b + 20
<c + c + ··· + c+1
1 b + 19
= (c + 1)b+20 .

So ab+20 is in this case strictly between cb+20 and (c + 1)b+20 , which is impossible for natural
numbers. (The case c = 0 must then be treated separately as above.)
(Michael Drmota)

Problem 16. Let α be a real number.

Determine all functions f : R → R satisfying

f (f (x) + y) = f (x2 − y) + αf (x)y

for all x, y ∈ R.
(Walther Janous)

Solution. First, we set y = (x2 − f (x))/2 which gives αf (x)(x2 − f (x))/2 = 0. Now, we
distinguish the cases α ̸= 0 und α = 0.

(a) In the case α ̸= 0, this equation implies f (x) = 0 or f (x) = x2 for each x separately.
In particular, f (0) = 0.

◦ It is easily verified that f (x) = 0, x ∈ R, is a solution.

16
 ◦ Next, we investigate the function f (x) = x2 , x ∈ R. The functional equation becomes

(x2 + y)2 = (x2 − y)2 + αx2 y ⇐⇒ 2x2 y = −2x2 y + αx2 y ⇐⇒ (α − 4)x2 y = 0

which holds for all x and y exactly when α = 4. Therefore, for α = 4, there is an
additional solution f (x) = x2 , x ∈ R.
◦ It remains to investigate the case, where there are numbers x, y ∈ R \ {0} with
f (y) = y 2 and f (x) = 0. Suppose that x and y were two such numbers.
Then the original functional equation becomes f (y) = f (x2 − y). Because of f (y) =
y 2 ̸= 0, we have f (x2 − y) ̸= 0 and therefore f (x2 − y) = (x2 − y)2 . This implies

y 2 = (x2 − y)2 = x4 − 2x2 y + y 2 ,

i.e.. y = x2 /2, so that y is the only number with f (y) = y 2 and f (z) = 0 for all
z ∈ R \ {y}. Repeating this argument, we obtain y = z 2 /2 for all z ∈ R \ {y}, a
contradiction.

(b) For α = 0, the functional equation becomes

f (f (x) + y) = f (x2 − y). (2)

It is easy to check that constant functions and the function f (x) = −x2 are solutions.
Now, assume that there is a real number a with f (a) = b ̸= −a2 . We define d = b+a2 ̸= 0.
Putting x = a in the functional equation (2) gives f (b + y) = f (a2 − y) for all y ∈ R.
With y = z − b, we obtain f (z) = f (d − z) for all z ∈ R. Using x = z and x = d − z in
the functional equation (2), we get

f (z 2 − y) = f (f (z) + y) = f (f (d − y) + y) = f ((d − z)2 − y).

Therefore,
f (z 2 − y) = f ((d − z)2 − y)
for all real numbers y and z. With y = z 2 , we obtain f (0) = f (d2 − 2dz) for all z ∈ R.
Because of d ̸= 0 the second argument attains all real numbers, so that f is constant.
This proves that there are no other solutions.

(Walther Janous)

Problem 17. Let ABCD be an inscribed convex quadrilateral with diagonals AC and BD.
Each of the four vertices is reflected on the diagonal it does not lie on.
Prove that the resulting four points lie on a common circle or a common line.

(a) Investigate when the four resulting points lie on a common line and give a simple equiv-
alent condition for the quadrilateral ABCD.

(b) Prove that in all other cases, the four resulting points lie on a common circle.

(Theresia Eisenkölbl)

17
Solution. (a) We denote the reflections of A, B, C and D with A′ , B ′ , C ′ resp. D′ and we
denote the intersection of the diagonals with S. Since the points A and C are reflected in
the same line BD and the point S remains invariant under this reflection, the whole line
ASC becomes A′ SC ′ after reflection in BD. Analogously, the line BSD becomes B ′ SD′
after reflection in AC.
If we denote the smaller angle between the two diagonals by φ, these two actions on the
lines correspond to a rotation of the line AC with center S in direction BD with rotation
angle 2φ and a rotation of the line BD with center S in direction AC with rotation angle
2φ.
Therefore, the angle between the lines A′ SC ′ and B ′ SD′ is the angle 3φ. This has to be
a multiple of 180°, so that the original angle has to be 0° or 60°. The first case is not
possible since the points of the inscribed quadrilateral cannot lie on a line.
We obtain that the four new points lie on a line if and only if the diagonals of the given
inscribed quadrilateral make an angle of 60°.
(b) Since the reflections do not only preserve the collinearity of ASC and BSD, but also the
position of S between the two points and the distances to the two points, we want to use
the power of S with respect to the circle ABCD.
B′
A′

C′

B
′
D

Figure 3: Problem 17

Because of the reflections, we have

SA = SA′ , SB = SB ′ , SC = SC ′ , SD = SD′ ,
18
 and since ABCD is an inscribed quadrilateral, we have

SA · SC = SB · SD.

Therefore, we obtain

SA′ · SC ′ = SA · SC = SB · SD = SB ′ · SD′ .

Since the two lines A′ SC ′ and B ′ SD′ do not coincide in this case, we can apply the
properties of the power of a point in reverse, and we get that A′ , B ′ , C ′ and D′ lie on a
circle.
(Theresia Eisenkölbl)

2
Problem 18. Suppose that p is an odd prime number and M a set of p 2+1 integer squares.
Investigate if one can choose p elements of this set so that the arithmethic mean of these p
elements is an integer.
(Walther Janous)

Answer. Yes.
2
Solution. The idea is to choose from the p 2+1 square numbers p numbers that are in the same
residue class modulo p. Obviously, the sum of these p numbers is then divisible by p and thus
the arithmetic mean is an integer.
It is known that the square numbers do not run through all residue classes modulo p, but
only through 1 + p−1
2
= p+1
2
ones. (On the one hand, this is the residue class 0 if one squares a
number divisible by p. Because of a2 ≡ (p − a)2 mod p, the squares of numbers a that are not
divided by p run through a maximum of half of the p − 1 nonzero residue classes. On the other
hand, x2 ≡ y 2 mod p gives the relation p | (x − y)(x + y) and so x ≡ y mod p or x ≡ −y mod p.
Therefore, the squares of numbers a, which are not divisible by p, run through exactly half of
the p − 1 residue classes different from zero.)
2
We now divide the p 2+1 square numbers into the p+1 2
residue classes that correspond to
square numbers. Because of the pigeon hole principle, there is therefore a residue class, that
contains at least  2 
(p + 1)/2
(p + 1)/2
numbers.
Because of
(p2 + 1)/2 p2 + 1 p2 + p p − 1 p−1
= = − =p−
(p + 1)/2 p+1 p+1 p+1 p+1
and 0 < p−1
p+1
< 1 it follows that
(p2 + 1)/2
 
= p,
(p + 1)/2
what was to be shown.
(Michael Drmota)

19