Introduction

In 2010 the Dutch Mathematical Olympiad consisted of three rounds. The

first round was held on 29 January 2010 at the participating schools. The
paper consisted of eight multiple choice questions and four open-answer
questions, to be solved within 2 hours. In total 4150 students of 226 sec-
ondary schools participated in this first round.

In March we organised a new round at ten universities in the country. This

round contained five open-answer questions, and two problems for which
the students had to give extensive solutions and proofs. The contest lasted
2,5 hours. Those students from grade 5 (4, 6 3) that scored 14 (12, 10)
points or more on the first round (out of a maximum of 36 points) were
invited to this new second round.

From those 599 participants to the second round, the best students were
invited for the final round. Those students from grade 5 (4, 6 3) that scored
28 (18, 14) points or more on the second round (out of a maximum of 40
points) were invited to the final round. Also some outstanding participants
in the Kangaroo math contest or the Pythagoras Olympiad were invited.

We organised training sessions at the ten universities for the 154 students
who had been invited for the final round. Former Dutch IMO-participants
were involved in the training sessions at each of the universities.

Out of those 154, in total 147 participated in the final round on 17 Septem-
ber 2010 at Eindhoven University of Technology. This final round contained
five problems for which the students had to give extensive solutions and
proofs. They were allowed 3 hours for this round. After the prizes had been
awarded in the beginning of November, the Dutch Mathematical Olympiad
concluded its 49th edition 2010.

The 25 most outstanding candidates of the Dutch Mathematical Olympiad

2010 were invited to an intensive seven-month training programme, con-
sisting of weekly problem sets. Also, the students met twice for a three-day
training camp, three times for a day at the university, and finally for a
six-day training camp in the beginning of June.

On 18 March 2011 the first selection test was held. The best ten students
participated in the third Benelux Mathematical Olympiad (BxMO), held
in Mersch, Luxembourg.

1
In June, out of those 10 students and 1 reserve candidate, the team for
the International Mathematical Olympiad 2011 was selected by two team
selection tests on 8 and 11 June 2011. A seventh, young, promising student
was selected to accompany the team to the IMO. The team had a training
camp on Texel, one of the Dutch Frisian Islands, from 9 until 16 July,
together with the team from New Zealand.

For younger students we organised the third Junior Mathematical Olympiad

in October 2010 at the VU University Amsterdam. The students invited
to participate in this event were the 30 best students of grade 1, grade 2
and grade 3 of the popular Kangaroo math contest. The competition con-
sisted of two one-hour parts, one with fifteen multiple choice questions and
one with ten open-answer questions. The goal of this Junior Mathematical
Olympiad is to scout talent and to stimulate them to participate in the
first round of the Dutch Mathematical Olympiad.

The Dutch team for IMO 2011 Amsterdam consists of

• Ragnar Groot Koerkamp (16 y.o.)

• Jeroen Huijben (15 y.o., observer C at IMO 2010)
• Madelon de Kemp (18 y.o., bronze medal in IMO 2010)
• Daniël Kroes (17 y.o., honourable mention in IMO 2010)
• Merlijn Staps (16 y.o., bronze medal in IMO 2010)
• Jetze Zoethout (16 y.o.)

We bring as observer C the promising young student

• Jeroen Winkel (14 y.o.)

The team is coached by

• Johan Konter (team leader), Utrecht University

• Sietske Tacoma (deputy leader), Utrecht University

We are grateful to Jinbi Jin and Raymond van Bommel for the composition
of this booklet and the translation into English of most of the problems and
the solutions.

2
 First Round, January 2010
Problems
A-problems

A1. Consider figures consisting of three circles and two lines. What is the
maximal number of intersection points in such figures?
A) 15 B) 16 C) 18 D) 19 E) 20

A2. A test consists of six questions worth successively 1 to 6 points. If your

answer to a question is correct, the number of points that that question
is worth will be added to your score. If your answer is incorrect, that
number of points will be deducted from your score. So, if you only answered
questions 1, 3 and 4 correctly, your score will be 1 − 2 + 3 + 4 − 5 − 6 = −5.
What is the number of possible scores you can get for this test?
A) 20 B) 22 C) 41 D) 43 E) 64

A3. A regular hexagon ABCDEF has area 1. What is

the area of the kite ACDE?
√ √
A) 32 B) 12 3 C) 65 D) 34 E) 14 6

A4. Three players play a game with coins. Each round the player (or one of
the players) who has the largest amount of coins will put one coin in a vase
and then he will give each of the other players one coin. The vase is empty
when the game starts and the three players possess respectively 13, 14 and
15 coins. The game ends when one of the players has lost all his coins.
How many coins will be in the vase when the game ends?
A) 36 B) 37 C) 38 D) 39 E) 40

A5. What is the last digit of ((((76 )5 )4 )3 )2 ?

A) 1 B) 3 C) 5 D) 7 E) 9

 √
7 √
7  2  √
7 √
7  2
A6. Calculate 2+1 + 2−1 − 2+1 − 2−1 .
√
A) 2 B) 4 C) 8 2 D) 128 E) 512

3
A7. An odometer indicates that an car has driven 2010 km. The odometer
consists of six gears and there are no decimals; so the odometer displays
002010. However, each of the gears misses the digit 4 and will hop from
3 to 5 directly. What is the actual number of kilometres that the car has
driven?
A) 1409 B) 1467 C) 1647 D) 1787 E) 1809

A8. Thirty people with different lengths are positioned in a rectangle of six rows
each containing five persons. From each row we select the shortest person
and from these six shortest persons we select the tallest; that is Piet. We
also select the tallest person from each row and select the shortest from
these six tallest persons; that is Jan. Then we put all thirty people in one
line ordered by their length; the shortest person is standing on the left end
and the tallest on the right end of the line. On which position can Jan not
stand?
A) 21 positions left of Piet D) 19 positions right of Piet
B) 19 positions left of Piet E) 21 positions right of Piet
C) next to Piet

B-problems
The answer to each B-problem is a number.
?
B1. Seven equally long matches are situated on a table
as illustrated. How many degrees is the indicated
angle?

B2. How many positive integers a exist such that dividing 2216 by a gives
remainder 29?

B3. A figure consists of a square ABCD and a semicir-

cle with diameter AD outside of the square. The
square has side length 1. What is the radius of the
circumscribed circle of the figure?

4
B4. On a board containing 28 rows and 37 columns a number will be written in
red in each of the squares in the following way: in the top row the numbers
1 to 37 will be written from left to right, in the second row the numbers
38 to 74, etcetera.
In green a number will be written in each of the squares in the following
way: in the leftmost column the numbers 1 to 28 will be written from top
to bottom, in the column next to that column the numbers 29 to 56 will
be written, etcetera.
In the square in the top left corner the number 1 is written in red and
green. Add the red numbers of the squares in which the numbers written
in red and green are the same. What is the sum of these numbers?

Solutions
A-problems

A1. D) 19 Any two circles intersect in at most 2

points, a circle and a line intersect in at most 2 points
and two lines intersect in at most 1 point. Therefore,
the number of intersections cannot be more than 6 +
6 + 6 + 1 = 19. You can easily check that this number of intersections can
be realized, see for example the figure on the right.

A2. B) 22 The possible scores are exactly the odd numbers

−21, −19, . . . , 19, 21. To see that only odd numbers can occur, first con-
sider the perfect score: 21 points. Now for every wrong answer, an even
number must be subtracted from it, leaving an odd number.
Conversely, all odd numbers between −21 and 21 are possible scores. In-
deed, if none or exactly one question is answered incorrectly, the resulting
scores are 21, 19, 17, 15, 13, 11 and 9. Answering question 6 and one of the
first four questions incorrectly results in scores of 7, 5, 3 and 1. Similarly,
answering at most one question correctly, or answering question 6 and one
of the first four questions correctly, result in scores of −21, −19, . . . , −1.

5
 A

A3. B) 2
3 Connect the centre M to A, C and E. B F

Also connect C and E. Then the hexagon is split

M
into six equal triangles, four of which form the kite.
C E

A4. B) 37 After three rounds, every player has lost 1 chip, and the pool
has gained 3. After 12 · 3 rounds, the players have 1, 2 and 3 chips left
respectively and 36 chips are in the pool. In the next and final round, one
more chip is added to the pool.

A5. A) 1 If two numbers have a 1 as their last digit, then so does

their product. Since the last digit of 74 = 2401 equals 1, the same holds
for every power of 74 . In particular, the last digit of ((((76 )5 )4 )3 )2 =
76·5·4·3·2 = (74 )180 is equal to 1.

√ √
A6. B) 4 Let a = ( 2 + 1)7 and b = ( 2 − 1)7 . Then the given
expression is equal to
√ √
(a + b)2 − (a − b)2 = 4ab = 4( 2 + 1)7 ( 2 − 1)7
√ √
= 4(( 2 + 1)( 2 − 1))7 = 4 · 17 = 4.

A7. B) 1467 After 9 miles, the second digit from the right increases. After
nine increments, so after 9 · 9 miles, the third digit from the right increases.
After 9 · 9 · 9 = 729 miles the fourth digit from the right increases. As
the display shows 002010, the total distance travelled must be equal to
2 · 729 + 1 · 9 = 1467 miles.

A8. E) 21 places to the right of Paul Among 6 selected people, Paul is tallest.
Therefore at least 5 people are shorter than Paul. James is shortest among
6 selected people, so at least 5 people are taller than James. James can
therefore not be 21 places to the right of Paul, since then at least 5 + 1 +
20 + 1 + 5 = 32 people would be standing in line.
The other positions are indeed possible. Here we only show this for an-
swer C: James and Paul stand next to each other. Number the people in
increasing height from 1 to 30. Put numbers 1, 2, 3, 4 and 10 in one row
and put one of the numbers 5, 6, 7, 8 en 9 in each of the other rows. The
rest may be distributed freely over the remaining positions. The shortest

6
 people in the six rows are numbers 1, 5, 6, 7, 8 and 9. Therefore Paul has
number 9. James has number 10, since he is the tallest of his row and in
every other row, there is someone with a number bigger than 10.

B-problems

B1. 100 The outer four matches form a paral- 180−α 120−α
α α
lelogram. The four angles at the bases of the two 60

isosceles triangles are therefore all equal. This an- 120−α

gle, denoted by α, equals 180 degrees minus the 60
angle we are looking for. The inner three matches α
α
form an equilateral triangle (angles equal to 60 de-
grees). As the angles of any triangle add up to 180 degrees, we have that
(180 − α) + 2(120 − α) = 180. It follows that α = 80 and therefore the
angle in the question measures 100 degrees.

B2. 4 The statement that dividing 2216 by a leaves a remainder

of 29, is the same as saying that a divides 2216 − 29 = 2187 and is larger
than 29 (the remainder is always smaller than the divisor a). The divisors
of 2187 = 37 larger than 29 are 81, 243, 729 and 2187. There are 4 choices
for a.

5
B3. 6 Let O be the centre of the circum-
scribed circle, E the point where it touches the semi-
circle and let M be the midpoint of BC. Applying
the Pythagorean theorem gives |OC|2 = |M C|2 +
|OM |2 . Since |OM | = |EM | − |OE| = 32 − |OC| this
implies that |OC|2 = ( 12 )2 + ( 32 − |OC|)2 . Solving
this equation, we obtain |OC| = 65 .

B4. 5185 Number the rows of the array from top to bottom using 0 to
27 and the columns from left to right using 0 to 36. Consider the position
in row r and column k. The red number written equals 1 + k + 37r and
the green number equals 1 + r + 28k. These two numbers are equal exactly
when 36r = 27k, that is, if and only if 4r = 3k. As solutions we obtain for
r the multiples of three: 0, 3, . . . , 27 and for k the matching multiples of
four 0, 4, . . . , 36. The coloured numbers in the corresponding ten positions
are 1, 1 + 115, 1 + 2 · 115, . . . , 1 + 9 · 115. Adding these numbers we find the
solution: (1 + (1 + 9 · 115)) · 5 = 5185.

7
 Second Round, March 2010
Problems
B-problems
The answer to each B-problem is a number.

B1. Alice has got five real numbers a < b < c < d < e. She takes the sum
of each pair of numbers and writes down the ten sums. The three small-
est sums are 32, 36 and 37, while the two largest sums are 48 and 51.
Determine e.

B2. Let AB be a diameter of a circle. Point C is the point on segment AB

such that
2 · |AC| = |BC|. The points D and E lie on the circle such that CD is
perpendicular to AB and such that DE is also a diameter of the circle.
Write the areas of the triangles ABD and CDE as O(ABD) and O(CDE).
O(ABD)
Determine the value of O(CDE) .

B3. A 24-hour digital clock displays the times from 00:00:00 till 23:59:59 during
the day. You can add the digits of the time on every second of the day;
this will give you an integer. For example, at 13:07:14 you will get 1 + 3 +
0 + 7 + 1 + 4 = 16. When you write down this sum for every possible state
of the clock and then take the average of all these numbers, what will be
the result?

B4. For the infinite sequence of numbers

0, 1, 2, 2, 1, −1, −2, −1, 1, 3, . . .

the following rule holds. For each four consecutive numbers

. . . , a, b, c, d, . . . of the sequence the number d is equal to c minus the small-
est of the two numbers a and b. For example, the ninth number of the
sequence is equal to −1 − (−2) = 1 and the tenth number is equal to
1 − (−2) = 3. Calculate the 100th number of this sequence.

8
B5. Raymond has got five coins. On the heads side of each coin is the number 1.
On the tails sides of the coins are the fractions 12 , 31 , 14 , 15 and 16 respectively.
Because every coin has got either its heads side or its tails side facing up,
there are 32 ways to put the five coins on the table. Raymond multiplies
the five numbers facing up for each of these 32 situations and writes down
all results.
If Raymond adds up these 32 numbers, what will be the final result?

C-problems
For the C-problems not only the answer is important; you also have to describe the way you
solved the problem.

C1. Determine all positive integers n consisting of four digits for which it holds
that n plus the sum of the digits of n is equal to 2010.

C2. Line segment AB has length 10. A point C lies on line segment AB such
that |AC| = 6 and |CB| = 4. Two points X and Y lie on the same side
of the line AB, such that |Y B| = |Y C| = 3, |XA| = 8 and |XC| = 6.
Determine the length of line segment XY .

Solutions
B-problems

B1. Out of the ten sums 32, 36, 37, . . . , 48, 51 the largest one is d + e and the
second largest one is c+e. Therefore d+e = 51 and c+e = 48. Furthermore
a + b is the smallest sum and a + c the second smallest, so a + b = 32 and
a + c = 36.
The third smallest sum could be either a + d or b + c. However, we know
that
a + d = (a + c) + (d + e) − (c + e) = 36 + 51 − 48 = 39.

Apparently a + d is not the third smallest sum, so b + c = 37.

Combining the things we have found so far yields

2e = 2(c + e) − (a + c) − (b + c) + (a + b) = 2 · 48 − 36 − 37 + 32 = 55.
55
Thus, the answer is e = 2 .

9
B2. Let M be the midpoint of the circle. Tri-
angles CDM and CEM have equal areas,
because they have bases of the same length
|DM | = |EM | and the same height. This
yields

O(CDE) = 2 · O(CDM ).

Since |AC| = 31 |AB| and |AM | = 12 |AB|

are true, |CM | = |AM | − |AC| = 21 |AB| −
1 1
3 |AB| = 6 |AB|. Because triangles ABD
and CDM have the same height with respect to the bases AB and CM ,

O(ABD) = 6 · O(CDM ).
O(ABD) 6
Combining the two equalities yields O(CDE) = 2 = 3.

B3. Let S be the number of possible states of the clock (for this problem it is
not necessary to calculate S). If we would write down only the last digit for
1
each state, then we would write down each of the digits 10 S times. Thus,
1 45
the total sum of these last digits is 10 S (0 + 1 + 2 + . . . + 9) = 10 S = 92 S.
If we would do the same for the second last digits, then we would only
write down the digits 0 to 5 each exactly 16 S times. Thus, the total sum
of these digits is 16 S (0 + 1 + . . . + 5) = 15 5
6 S = 2 S.
Analogously the sum of the digits in the middle two positions is equal to
9 5
2 S + 2 S.
We have to pay special attention to the first two digits, representing the
hours of the time, because every digit does not appear evenly frequent.
1
However, each of the digit pairs 00, 01, . . . , 23 appears exactly 24 S times.
Thus, the total sum of the first to digits is
1
24 S ((0 + 0) + (0 + 1) + (0 + 2) + . . . + (2 + 3))
1
= 24 S (2(0 + 1 + . . . + 9) + 10 · 1 + 4 · 2 + (0 + 1 + 2 + 3))
= 114 19
24 S = 4 S.

The total sum of all digits of all possible states is 2· 92 S +2· 52 S + 19 75

4 S = 4 S.
75 3
The average sum is the total sum divided by S and that is 4 (= 18 4 ).

10
B4. It is easy to calculate some more numbers of the sequence:

0, 1, 2, 2, 1, −1, −2, −1, 1, 3, 4, 3, 0, −3, −3,

0, 3, 6, 6, 3, −3, . . .

We now see a clear pattern: after fifteen terms the sequence repeats itself,
but all the terms are three times as large. By considering the following
argument, we can verify that the sequence indeed does follow this pattern.
Each number of the sequence is uniquely determined by its three prede-
cessors. Assume that three consecutive numbers in the sequence a, b, c will
repeat fifteen positions further, multiplied by 3, however:

. . . , a, b, c, d, . . . , 3a, 3b, 3c, . . .

The successor of 3c is (by definition) equal to 3c minus the smallest of 3a

and 3b. That is exactly three times as much as: c minus the smallest of a
and b. So this is exactly three times d. If three consecutive numbers in the
sequence repeat themselves after fifteen positions (with an extra factor 3),
this also applies to the next number and therefore the number after the
next number, etcetera.
Because of this pattern we see that the hundredth number is equal to the
tenth number (3), but then multiplied 90 15 = 6 times by 3. This yields:
3 · 36 = 2187.

B5. The results that Raymond will get

correspond exactly to
the terms

of the
expansion of the product 1 + 21 1 + 13 1 + 14 1 + 15 1 + 16 .

Therefore, the total sum is

1 + 12 1 + 13 1 + 41 1 + 15 1 + 61 = 32 · 43 · 54 · 56 · 67 = 27 .

C-problems

C1. Suppose that n is an integer consisting of four digits. We use the notation
n = 1000a+100b+10c+d, where a, b, c and d are the digits of n. Therefore,
0 6 a, b, c, d 6 9 and because a number does not start with a zero, a > 0.
We now want to know for which n it is true that n + (a + b + c + d) =
1001a+101b+11c+2d = 2010. Because 1001·3 is already larger than 2010,
a has to be smaller than 3, so a = 1 or a = 2. We try both possibilities.
1. If a = 1, then 101b + 11c + 2d = 2010 − 1001 · 1 = 1009. We will
determine the possible values of b.

11
 It is true that 0 6 11c + 2d 6 13 · 9 = 117. Because 101b =
1009 − (11c + 2d), this yields 1009 > 101b > 1009 − 117 = 892.
We are looking for multiples of 101 between 892 and 1009, because
101b is a multiple of 101. The only multiple of 101 between 892 and
1009 is 909 and therefore b must be equal to 9.
Now 11c + 2d = 1009 − 909 = 100 has to be true. We will determine
the possible values of b.
It is true that 0 6 2d 6 2 · 9 = 18. Because 11c = 100 − 2d, this also
yields 100 > 11c > 100 − 18 = 82. Since 11c = 100 − 2d is even, 11c
has to be equal to an even multiple of 11 between 82 and 100. The
only even multiple of 11 between 82 and 100 is 88. Therefore, c = 8
and d = 6 have to be true.
We find n = 1986 as a possible solution.
2. If a = 2, then 101b + 11c + 2d = 2010 − 1001 · 2 = 8.
Now we see that b = c = 0 and d = 4.
We find n = 2004 as possible solution.
Finally, we check if the solutions 1986 and 2004 really suffice. Indeed,
1986 + 1 + 9 + 8 + 6 = 2010 and 2004 + 2 + 0 + 0 + 4 = 2010.
So, n = 1986 and n = 2004 are the only two solutions. 

C2. Let D be the midpoint of BC and M

the midpoint of AX.
Since |AM | = |M X|, |AC| = |CX|
and |M C| = |M C|, the triangles AM C
and XM C are congruent (SSS). This
yields ∠CM A = ∠CM X. Because
∠CM A + ∠CM X = 180◦ , this yields
∠CM A = ∠CM X = 90◦ .
Analogously, the triangles CDY and
BDY are congruent. This yields ∠CDY = ∠BDY = 90◦ .
|CD| |CY | 1 ◦
It is true that |AM | = |AC| = 2 . Combined with ∠AM C = ∠CDY = 90
this yields that triangles AM C and CDY are similar (ssr).
We now know that ∠ACM = ∠CY D and because of that ∠ACM +
∠DCY = ∠CY D + ∠DCY = 180◦ − ∠CDY = 90◦ . This yields that
∠M CY = 180◦ − ∠ACM − ∠DCY = 90◦ .
Draw the line through Y perpendicular to AX. Let F be the intersection
point of this line with AX. We see that M CY F is a rectangle.
The Pythagorean theorem yields |M C|2 = |XC|2 − |M X|2 = 62 − 42 = 20.
It is also true that |XF | = |M X| − |M F | = |M X| − |CY | = 1 and
2
|F Y | = |M C|. Applying the Pythagorean theorem
√ again yields |XY | =
2 2 2
|F Y | + |F X| = |M C| + 1 = 21 and |XY | = 21. 

12
 Final Round, September 2010
Problems
For these problems not only the answer is important; you also have to describe the way you
solved the problem.

1. Consider a triangle ABC such that ∠A = 90◦ , ∠C =

60◦ and |AC| = 6. Three circles with centers A, B B
and C are pairwise tangent in points on the three
sides of the triangle.
Determine the area of the region enclosed by the
three circles (the grey area in the figure).
C A

2. A number is called polite if it can be written as m+(m+1)+· · ·+(n−1)+n,

for certain positive integers m < n. For example: 18 is polite, since 18 =
5 + 6 + 7. A number is called a power of two if it can be written as 2` for
some integer ` > 0.
(a) Show that no number is both polite and a power of two.
(b) Show that every positive integer is polite or a power of two.

3. Consider a triangle XY Z and a point O in its X

interior. Three lines through O are drawn, par- F

A f E
allel to the respective sides of the triangle. The a
e
intersections with the sides of the triangle de- B b O
d Z
c
termine six line segments from O to the sides of D
the triangle. C
The lengths of these segments are integer num- Y
bers a, b, c, d, e and f (see figure).
Prove that the product a · b · c · d · e · f is a perfect square.

4. (a) Determine all pairs (x, y) of (real) numbers with 0 < x < 1 en 0 <
y < 1 for which
x + 3y and 3x + y are both integer. An example is
(x, y) = 83 , 78 , because x + 3y = 83 + 21 24 9 7
8 = 8 = 3 en 3x + y = 8 + 8 =
16
8 = 2.
(b) Determine the integer m > 2 for which there are exactly 119 pairs
(x, y) with 0 < x < 1 en 0 < y < 1 such that x + my en mx + y are
integer.
Remark: if u 6= v, the pairs (u, v) and (v, u) are different.

13
5. Amber and Brian are playing a game using 2010 coins. Throughout the
game, the coins are divided into a number of piles of at least 1 coin each. A
move consists of choosing one or more piles and dividing each of them into
two smaller piles. (So piles consisting of only 1 coin cannot be chosen.)
Initially, there is only one pile containing all 2010 coins. Amber and Brian
alternatingly take turns to make a move, starting with Amber. The winner
is the one achieving the situation where all piles have only one coin.
Show that Amber can win the game, no matter which moves Brian makes.

Solutions

1. We recognize triangle ABC to be half an equilat-

eral triangle. This implies that |BC| = 2|AC| =
12.
p The Pythagorean √ theorem√ yields: |AB| =
|BC|2 − |AC|2 = 108 = 6 3. B
Denote the pairwise tangent points of the three
circles by D, E and F (see figure) and the radii of
D
the three circles by rA , rB and rC . The strategy F
will be to determine the area of the three circu-
lar sectors and subtract them from the area of C E A
triangle ABC.
We √ |AC|+|AB|−|BC|
√ see that 2rA = (r√A +rC )+(rA +rB )−(rB +rC ) = √ =
6 3 − 6, so rA = 3 √3 − 3. It follows that rB = 6 3 − rA = 3 3 + 3 en
rC = 6 − rA = 9 − 3 3.
2
The area of a circle of radius r equals πr√ . Therefore,√the area of circular
sector AF E equals 360 ·πrA , or 4 π(36−18 3) = 9π− 92 3π. For the area of
90 2 1
30
√
circular sectors BDF and CED we find, respectively, 360 2
πrB = 3π + 32 3π
60 2
√
and 360 πrC = 18π − 9 3π.
√
Since ABC has an area of 21 · |AB| · |AC| = 18 3, we obtain a value of
√ √ √ √ √ √
18 3 − (9π − 92 3π) − (3π + 32 3π) − (18π − 9 3π) = 18 3 − 30π + 12 3π
for the area of the grey region. 

2. (a) Suppose that k = m+(m+1)+· · ·+(n−1)+n is a polite number. The

sum formula for arithmetic sequences gives k = 12 (m + n)(n − m + 1).
As m and n are different positive numbers, m+n > 3 and (n−m)+1 >
2 must hold.
Since (m + n) + (n − m + 1) = 2n + 1 is odd, one of the numbers
m + n and n − m + 1 is odd. Hence 2k = (m + n)(n − m + 1) has

14
 an odd divisor (greater than 1) and is therefore not a power of two.
This implies that k is not a power of two either.
We conclude that no number can be both polite and a power of two.
(b) Suppose that k is a positive integer, not a power of two. We will
show k to be a polite number. Collecting all factors 2, we can write
k = c · 2d , where c is odd and d > 0 is a nonnegative integer. The
assumption that k is not a power of two, means that c > 1. We need
to find n > m such that m + · · · + n = 12 (m + n)(n − m + 1) = c · 2d , or
(m + n) · (n − m + 1) = c · 2d+1 . We can achieve this by choosing m en
n in such a way that m + n = c and n − m + 1 = 2d+1 , or conversely:
m + n = 2d+1 and n − m + 1 = c. To ensure that m will be positive,
we consider two cases.
For c > 2d+1 we solve: m + n = c, n − m + 1 = 2d+1 . This gives
m = (c − 2d+1 + 1)/2 and n = (c + 2d+1 − 1)/2. Obviously, n > m
(since 2d+1 > 2). Both m and n are integers (the numerators are even
since c is odd) and positive by the assumption c > 2d+1 .
For c < 2d+1 we solve: m + n = 2d+1 , n − m + 1 = c. This gives
m = (2d+1 − c + 1)/2 and n = (2d+1 + c − 1)/2. Clearly, n > m holds
(since c > 1) and both m and n are positive integers. 

3. Since AO and XZ are parallel, ∠OAB = ∠ZXY X

are corresponding angles. Similarly, since BO F
A f E
and Y Z are parallel, ∠ABO = ∠XY Z holds. a
e
We deduce that 4OAB ∼ 4ZXY (equal an- B b O
d Z
gles). Hence there is a scaling factor u such c
that a = u|XZ| and b = u|Y Z|. Using simi- D
lar arguments we find that 4OCD ∼ 4XY Z Y C
and 4OEF ∼ 4Y ZX. So there are scal-
ing factors v and w such that c = v|XY |, d = v|XZ|, e = w|Y Z| and
f = w|XY |.
We now see that a · c · e = uvw · |XY | · |Y Z| · |ZX| = b · d · f . This implies
that a · b · c · d · e · f = (a · c · e)2 , which is a perfect square since a, c and e
are integers. 

4. (a) Suppose that (x, y) is such a pair and consider the integers a = x + 3y
and b = 3x + y. From 0 < x, y < 1 it follows that 0 < a, b < 4, or:
1 6 a, b 6 3.
Conversely, let a and b be integers such that 1 6 a, b 6 3. There is a
unique pair of numbers (x, y) that satisfies a = x + 3y en b = 3x + y.
Indeed, combining the two equations, we get 3b − a = 3(3x + y) −

15
 (x + 3y) = 8x and 3a − b = 8y. In other words x = (3b − a)/8 and
y = (3a − b)/8 (and these x and y do satisfy the two equations). If we
substitute 1, 2, 3 for a and b, we obtain the following nine pairs (x, y):

( 82 , 28 ), ( 58 , 18 ), ( 88 , 08 ), ( 18 , 58 ), ( 48 , 84 ), ( 78 , 38 ), ( 08 , 88 ), ( 38 , 78 ), ( 86 , 68 ).

The condition 0 < x, y < 1 rules out the two candidates (x, y) = ( 88 , 80 )
and (x, y) = ( 08 , 88 ). This leaves the 7 pairs we were asked to find.
(b) Suppose that 0 < x, y < 1 holds and that a = x + my and b = mx + y
are integers. Then 1 6 a, b 6 m holds.
Given integers a and b with 1 6 a, b 6 m, there is a unique pair (x, y)
for which x + my = a and mx + y = b hold. Indeed, combining the
two equalities gives : mb − a = (m2 − 1)x and ma − b = (m2 − 1)y,
or: x = (mb − a)/(m2 − 1) and y = (ma − b)/(m2 − 1). These x and
y indeed satisfy the two equations.
For given a and b, we determine whether the corresponding numbers
x and y satisfy 0 < x, y < 1. From 1 6 a, b 6 m it follows that
x > (m · 1 − m)/(m2 − 1) = 0 and x 6 (m · m − 1)/(m2 − 1) = 1.
The cases x = 0 and x = 1 exactly correspond to (a, b) = (m, 1)
and (a, b) = (1, m) respectively. Similarly, 0 < y < 1 holds, unless
(a, b) = (1, m) or (a, b) = (m, 1). Among the m2 possible pairs (a, b),
there are exactly two for which (x, y) is not a solution. In total there
are m2 − 2 solutions (x, y).
From m2 − 2 = 119, we see that m = 11. 

5. A strategy that guarantees a win for Amber is as follows. In Amber’s turn,

she splits every pile with an even number of coins (say 2k) in two piles with
an odd number of coins: 1 coin and 2k − 1 coin respectively. The piles
having an odd number of coins, she leaves untouched. So in her first turn,
she created one pile of 1 coin and one of 2009 coins.
When Brian gets to make a move, all piles will have an odd number of
coins. He is therefore forced to split an odd pile, creating a new pile with
an even number of coins. This implies that Amber, in the next turn, can
continue her strategy, since there will be at least one even pile.
With each turn, the number of piles increases, so after at most 2009 turns,
the game is over. Since Brian always creates an even pile, the game cannot
end during his turn. Therefore, it will be Amber who wins the game. 

16
 BxMO Team Selection Test, March 2011
Problems

1. All positive integers are coloured either red or green, such that the following
conditions are satisfied:
• There are equally many red as green integers.
• The sum of three (not necessarily distinct) red integers is red.
• The sum of three (not necessarily distinct) green integers is green.
Find all colourings that satisfy these conditions.

2. In an acute triangle ABC the angle ∠C is greater than ∠A. Let E be such
that AE is a diameter of the circumscribed circle Γ of 4ABC. Let K be
the intersection of AC and the tangent line at B to Γ. Let L be the the
orthogonal projection of K on AE and let D be the intersection of KL
and AB.
Prove that CE is the bisector of ∠BCD.

3. Find all triples (x, y, z) of real numbers that satisfy

x2 + y 2 + z 2 + 1 = xy + yz + zx + |x − 2y + z|.

4. Let n > 2 be an integer. Let a be the greatest positive integer such that
2a | 5n − 3n . Let b be the greatest positive integer such that 2b 6 n.
Prove that a 6 b + 3.

5. A trapezoid ABCD is given with BC k AD. Assume that the bisectors of

the angles BAD and CDA intersect on the perpendicular bisector of the
line segment BC.
Prove that |AB| = |CD| or |AB| + |CD| = |AD|.

17
 Solutions

1. Suppose that we colour the integer k red. We prove by induction to n that

(2n + 1)k then also is coloured red for all n > 0. For n = 0 this is trivial.
Suppose that (2n − 1)k is red for a certain n, then k + k + (2n − 1)k =
(2n + 1)k is also red. This concludes the induction. Analogously if k is
coloured green, then also (2n + 1)k is coloured green for all n > 0.
Without loss of generality we may assume that 1 is red. Then all odd
numbers are red. Now suppose that there is also an even number 2m that
is red. Because both 1 and 2m are red, we can easily show by induction
that 2m + 2n is red for all n > 0. Then only finitely many numbers
are left that could be green, namely the even numbers smaller than 2m.
However, if one of these numbers is green, then also all odd multiples of
that number are green and these are infinitely many numbers, yielding a
contradiction. If on the other hand none of these numbers is green, then
we have a contradiction with the first condition. We conclude that no even
number is red. Therefore all even numbers are green.
This colouring also satisfies all conditions: the sum of three odd numbers
is always odd and the sum of three even numbers is always even, which
shows that the second and third condition are satisfied. The first condition
is also clearly satisfied.
Hence the only possible colourings are: all even numbers are coloured green
and all odd numbers red, or all odd numbers green and all even numbers
red. 

2. The conditions in the problem fix the configuration. Because ∠LAD =

∠EAB = ∠ECB by the inscribed angle theorem on chord EB of circle Γ:

∠BDK = ∠ADL = 180◦ − ∠DLA − ∠LAD = 90◦ − ∠LAD

= 90◦ − ∠ECB = ∠ECA − ∠ECB = ∠BCA = 180◦ − ∠BCK.

This yields that BDKC is a cyclic quadrilateral. Therefore ∠BCD =

∠BKD. Now we will prove that ∠BKD = 2∠BCE, from which the
problem follows.
By the inscribed angle theorem, ∠EBK = ∠BCE, hence ∠ABK =
∠ABE + ∠EBK = 90◦ + ∠BCE. Together with the sum of the angles of
triangle ABC this yields

∠AKB = 180◦ − (90◦ + ∠BCE) − ∠BAK = 90◦ − ∠BCE − ∠BAK.

18
 The inscribed angle theorem gives ∠BAK = ∠EAK + ∠BAE = ∠EAK +
∠BCE, therefore we find

∠AKB = 90◦ − 2∠BCE − ∠EAK = 90◦ − 2∠BCE − ∠LAK.

From the sum of the angles of triangle AKL follows ∠AKL = 90◦ −∠LAK,
hence
∠AKB = ∠AKL − 2∠BCE.
This yields
∠BKD = ∠AKL − ∠AKB = 2∠BCE,
what we wanted to prove. 

3. We rewrite the given equation as

1 2
2x − xy + 21 y 2 + 12 y 2 − yz + 12 z 2 + 12 z 2 − zx + 12 x2 + 1 = |x − 2y + z|,

or, equivalently,
1
2 (x − y)2 + 12 (y − z)2 + 12 (z − x)2 + 1 = |(x − y) + (z − y)|. (1)

Now substitute a = x − y and b = z − y. Then x − z = a − b, thus we get

1 2
2a + 12 b2 + 12 (a − b)2 + 1 = |a + b|. (2)

From (a − b)2 > 0, it follows that a2 − 2ab + b2 > 0, hence 2a2 + 2b2 >
2
a2 + b2 + 2ab, which means that a2 + b2 > (a+b)
2 with equality if and only
if a = b. Furthermore, also (a − b)2 > 0. Hence

(a + b)2
|a + b| = 12 a2 + 12 b2 + 12 (a − b)2 + 1 > + 1.
4
Now write c = |a + b|, then the expression becomes

c2
c> + 1.
4
We can rewrite this to c2 − 4c + 4 6 0, or, equivalently (c − 2)2 6 0.
Because the left hand side is a square, equality must hold, so c = 2.
Furthermore, in our previous inequalities equality also has to hold, so a = b.
Substituting this in (2) gives a2 + 1 = 2, so a = ±1. Thus we find the
triples (y + 1, y, y + 1) and (y − 1, y, y − 1) for arbitrary y ∈ R. Substituting
this in the equation (1) (that is equivalent to the original equation) shows
that these triples are indeed solutions for all y ∈ R. Hence, all solutions
are given by (y + 1, y, y + 1) and (y − 1, y, y − 1) with y ∈ R. 

19
4. First we prove the statement for all odd numbers n. In this case modulo
4 we have 5n ≡ 1n = 1 and 3n ≡ (−1)n ≡ −1, hence 5n − 3n ≡ 2 mod 4.
Therefore, if n is odd, then a = 1. Because b > 1, the inequality a 6 b + 3
is satisfied.
Now suppose that n ≡ 2 mod 4. Write n = 2k with k an odd positive
integer. Notice that 52k − 32k = (5k − 3k )(5k + 3k ). We just showed that
5k − 3k has precisely one factor 2, since k is odd. Now consider 5k + 3k
modulo 16. For m = 1, 2, 3, 4 we have that 5m modulo 16 is congruent
to respectively 5, 9, 13, 1. Because 54 ≡ 1 mod 16, we have 5k ≡ 5 for all
k ≡ 1 mod 4 and 5k ≡ 13 for all k ≡ 3 mod 4. For m = 1, 2, 3, 4 we
have that 3m modulo 16 is congruent to respectively 3, 9, 11, 1. Because
34 ≡ 1 mod 16, we have 3k ≡ 3 for all k ≡ 1 mod 4 and 3k ≡ 11 for all
k ≡ 3 mod 4. Altogether 5k + 3k ≡ 5 + 3 ≡ 8 mod 16 if k ≡ 1 mod 4
and 5k + 3k ≡ 13 + 11 = 24 ≡ 8 mod 16 if k ≡ 3 mod 4. In both cases
5k + 3k contains precisely 3 factors 2.
We conclude that for n ≡ 2 mod 4 we have: a = 4. Because b > 1, the
inequality a 6 b + 3 is now satisfied.
We will prove by induction to m that a 6 b + 3 for all positive numbers n
with precisely m > 1 factors 2. The induction basis is m = 1, that is the
case n ≡ 2 mod 4, which we just solved.
Now let m > 1 and assume as induction hypothesis that we already showed
that a 6 b + 3 for all numbers n with precisely m factors 2. Now consider
a number n with m + 1 factors 2. We write n = 2k, where k has precisely
m factors 2. For clarity let a(k) and b(k) be the a and the b corresponding
to k, and a(n) and b(n) the a and b corresponding to n. The induction
hypothesis yields a(k) 6 b(k) + 3. Now we will prove that a(n) 6 b(n) + 3.
We have 5n − 3n = 52k − 32k = (5k − 3k )(5k + 3k ). Because k is even (it
contains m > 1 factors 2) modulo 4 we have 5k + 3k ≡ 1k + (−1)k ≡ 2
mod 4. Hence 5k +3k contains precisely one factor 2. Furthermore, 5k −3k
contains precisely a(k) factors 2. Thus a(n) = a(k) + 1. We already know
that 2b(k) 6 k and 2b(k)+1 > k, which yields 2b(k)+1 6 2k and 2b(k)+2 > 2k.
Hence b(n) = b(k) + 1. Now we conclude: a(n) = a(k) + 1 6 b(k) + 3 + 1 =
b(n) + 3, which concludes the induction.
This proves that a 6 b + 3 for all integers n > 2. 

20
5. Let M be the midpoint of BC and let P be the intersection of the per-
pendicular bisector of BC and AD. Let K be the intersection of M P and
the two angle bisectors. Let L and N be the feet of the lines through K
perpendicular to sides AB and DC, respectively. Because AK and DK
are bisectors, |KL| = |KP | = |KN |. Furthermore K also lies on the per-
pendicular bisector of BC, thus |KB| = |KC|. Because triangles BLK
and CN K both have a right angle, they are congruent by (RHS).
Now we distinguish between four cases. First consider the case that L
and N lie on the interior of respectively sides AB and DC. Notice that
triangle KBC is isosceles, yielding ∠KBC = ∠BCK. Hence by 4BLK ∼ =
4CN K:

∠ABC = ∠LBK + ∠KBC = ∠N CK + ∠BCK = ∠BCD.

From this follows that ABCD is a isosceles trapezoid and hence |AB| =
|CD|.
In the case that L and N both lie on the exterior of sides AB and CD we
can show analogously that |AB| = |CD|.
Now consider the case that L lies on the interior of side AB, but N on the
exterior of side DC. Because AK and DK are bisectors, |AL| = |AP | and
|DN | = |DP |. Therefore by 4BLK ∼ = 4CN K:

|AB| + |CD| = (|AL| + |LB|) + (|DN | − |N C|)

= |AP | + |LB| + |DP | − |LB| = |AD|.

In the case that L lies on the exterior of AB and N on the interior of DC,
we analogously show that |AB| + |CD| = |AD|. 

21
 Benelux Mathematical Olympiad, May 2011
Mersch, Luxembourg
Problems

1. An ordered pair of integers (m, n) with 1 < m < n is said to be a Benelux

couple if the following two conditions hold: m has the same prime divisors
as n, and m + 1 has the same prime divisors as n + 1.

(a) Find three Benelux couples (m, n) with m 6 14.

(b) Prove that there exist infinitely many Benelux couples.

2. Let ABC be a triangle with incentre I. The angle bisectors AI, BI and CI
meet [BC], [CA] and [AB] at D, E and F , respectively. The perpendicular
bisector of [AD] intersects the lines BI and CI at M and N , respectively.
Show that A, I, M and N lie on a circle.

3. If k is an integer, let c(k) denote the largest cube that is less than or
equal to k. Find all positive integers p for which the following sequence is
bounded:

a0 = p and an+1 = 3an − 2c(an ) for n > 0.

(A sequence a0 , a1 , . . . of reals is said to be bounded if there exists an

M ∈ R such that, for all n > 0, |an | 6 M .)

4. Abby and Brian play the following game: They first choose a positive
integer N . Then they write numbers on a blackboard in turn. Abby starts
by writing a 1. Thereafter, when one of them has written the number n,
the other writes down either n + 1 or 2n, provided that the number is not
greater than N . The player who writes N on the blackboard wins.
(a) Determine which player has a winning strategy if N = 2011.
(b) Find the number of positive integers N 6 2011 for which Brian has a
winning strategy.

22
 Solutions

1. (a) It is possible to see that (2, 8), (6, 48) and (14, 224) are Benelux cou-
ples.
(b) Let k > 2 be an integer and m = 2k − 2. Define n = m(m + 2) =
2k (2k − 2). Since m is even, m and n have the same prime factors.
Also, n + 1 = m(m + 2) + 1 = (m + 1)2 , so m + 1 and n + 1 have the
same prime factors
as well. We have thus obtained a Benelux couple
2k − 2, 2k (2k − 2) for each k > 2. 

2. The quadrilateral AM DB is cyclic. Indeed, M is the intersection of the

line BI, which bisects the angle ∠ABD in ABD and the perpendicular
bisector of [AD]. By uniqueness of this intersection point, it follows that M
lies on the circumcircle of ABD, and hence AM DB is cyclic. Analogously,
AN DC is cyclic.
Since AM DB and AN DC are cyclic, ∠AM I+∠AN I = ∠AM B+∠AN C =
◦
∠ADB + ∠ADC = 180 , because B and M , and C and N lie on either
side of AD. Hence AM IN is cyclic, for M and N lie on opposite sides of
AD. 

3. Since c(an ) 6 an for all n ∈ N, an+1 > an , where equality holds if and only
if c(an ) = an . Hence the sequence is bounded if and only if it is eventually
constant, which is if and only if an is a perfect cube, for some n > 0. In
particular, the sequence is bounded if p is a perfect cube.
We now claim that, if an is not a cube for some n, then neither is an+1 .
Indeed, if an is not a cube, q 3 < an < (q + 1)3 for some q ∈ N, so that
c(an ) = q 3 . Suppose to the contrary that an+1 is a cube. Then

an+1 = 3an − 2c(an ) < 3(q + 1)3 − 2q 3 = q 3 + 9q 2 + 9q + 3

< q 3 + 9q 2 + 27q + 27 = (q + 3)3 .

Also, since c(an ) < an , an+1 > an > q 3 , so q 3 < an+1 < (q + 3)3 . It follows
that the only possible values of an+1 are (q + 1)3 and (q + 2)3 . However,
in both of these cases,

3an − 2q 3 = an+1 = (q + 1)3 ⇐⇒ 3an = 3(q 3 + q 2 + 1) + 1

3an − 2q 3 = an+1 = (q + 2)3 ⇐⇒ 3an = 3(q 3 + 2q 2 + 4q) + 8

a contradiction modulo 3. This proves that, if an is not a cube, then

neither is an+1 . Hence, if p is not a perfect cube, an is not a cube for any

23
 n ∈ N, and the sequence is not bounded. We conclude that the sequence
is bounded if and only if p is a perfect cube. 

4. (a) Abby has a winning strategy for odd N : Observe that, whenever
any player writes down an odd number, the other player has to write
down an even number. By adding 1 to that number, the first player
can write down another odd number. Since Abby starts the game by
writing down an odd number, she can force Brian to write down even
numbers only. Since N is odd, Abby will win the game. In particular,
Abby has a winning strategy if N = 2011.
(b) - Let N = 4k. If any player is forced to write down a number
m ∈ {k + 1, k + 2, ..., 2k}, the other player wins the game by
writing down 2m ∈ {2k + 2, 2k + 4, ..., 4k}, for the players will
have to write down the remaining numbers one after the other.
Since there is an even number of numbers remaining, the latter
player wins. This implies that the player who can write down k,
i.e. has a winning strategy for N = k, wins the game for N = 4k.
- Similarly, let N = 4k + 2. If any player is forced to write down a
number m ∈ {k + 1, k + 2, ..., 2k + 1}, the other player wins the
game by writing down 2m ∈ {2k + 2, 2k + 4, ..., 4k + 2}, as in the
previous case. Analogously, this implies that the player who has
a winning strategy for N = k wins the game for N = 4k + 2.
Since Abby wins the game for N = 1, 3, while Brian wins the game
for N = 2, Brian wins the game for N = 8, 10 as well, and thus for
N = 32, 34, 40, 42 too. Then Brian wins the game for a further 8
values of N between 128 and 170, and thence for a further 16 values
between 512 and 682, and for no other values with N 6 2011. Hence
Brian has a winning strategy for precisely 31 values of N with N 6
2011. 

24
 IMO Team Selection Test 1, June 2011
Problems

1. Find all pairs (x, y) of integers that satisfy

√
x2 + y 2 + 33 = 456 x − y.

2. We consider tilings of a rectangular m × n-board with 1 × 2-tiles. The tiles

can be placed either horizontally, or vertically, but they aren’t allowed to
overlap and to be placed partially outside of the board. All squares on the
board must be covered by a tile.

(a) Prove that for every tiling of a 4 × 2010-board with 1 × 2-tiles there is
a straight line cutting the board into two pieces such that every tile
completely lies within one of the pieces.
(b) Prove that there exists a tiling of a 5 × 2010-board with 1 × 2-tiles
such that there is no straight line cutting the board into two pieces
such that every tile completely lies within one of the pieces.

3. The circles Γ1 and Γ2 intersect at D and P . The common tangent line of

the two circles closest to point D touches Γ1 in A and Γ2 in B. The line
AD intersects Γ2 for the second time in C. Let M be the middle of line
segment BC.
Prove that ∠DP M = ∠BDC.

4. Determine all integers n for which the polynomial P (x) = 3x3 − nx − n − 2

can be written as the product of two non-constant polynomials with integer
coefficients.

5. Let ABC be a triangle with |AB| > |BC|. Let D be the midpoint of AC.
Let E be the intersection of the angular bisector of ∠ABC and the line
AC. Let F be the point on BE such that CF is perpendicular to BE.
Finally, let G be the intersection of CF and BD.
Prove that DF divides the line segment EG into two equal parts.

25
 Solutions

1. Since the left hand side is an integer, the right hand side must also be an
integer. The square root of an integer is either an integer or irrational (but
√
never a rational number that is not an integer). Hence x − y must be an
integer. The right hand side is divisible by 3, so the left hand side must
also be divisible by 3. We deduce that 3 | x2 + y 2 . Since squares are always
congruent to 0 or 1 modulo 3, it follows that x2 ≡ y 2 ≡ 0 mod 3, hence
both x and y are divisible by 3. Writing and substituting x = 3a, y = 3b:
√
9a2 + 9b2 + 33 = 456 3a − 3b.
√
Square roots of integers are either integer or irrational. But since 3a − 3b =
2 2 3
9a +9b +3
456 is rational, it must be an integer as well. Hence 3a − 3b is a
square, and a multiple of 3, so it must also be a multiple of 9. Now we can
divide both sides by 9, which yields
r
2 2 a−b
a + b + 3 = 152 .
3
Writing a − b = 3c2 with c > 0, and substituting a = b + 3c2 gives

9c4 + 6c2 b + 2b2 + 3 = 152c.

Considering this as a quadratic equation in b and knowing that we are

looking for real solutions, we can deduce that the discrimant must be non-
negative. Hence 36c4 −8(9c4 −152c+3) > 0 and thus 36c4 +24 6 8·152c. If
c > 4 we have 36c4 +24 > 36·64c > 8·152c, this is a contradiction. Hence we
may conclude c 6 3. Furthermore, 152c is even, and so is 6c2 b + 2b2 , hence
9c4 + 3 must be even as well. We deduce that c is odd, and consequently
that the only possibilities are c = 1 and c = 3. For c = 3 the discriminant
is 36 · 34 − 8(9 · 34 − 152 · 3 + 3) ≡ 1 · 3(1 − 152) ≡ −1 · 3 · 151 ≡ 6 mod 9.
Hence it is not a square and the solutions of the quadratic equation will
not be integers. Substituting c = 1 gives

9 + 6b + 2b2 + 3 = 152,

or, equivalently,
b2 + 3b − 70 = 0.
This can also be written as (b − 7)(b + 10) = 0. Hence either b = 7 or
b = −10. In the first case, we have a = b + 3c2 = 10, so x = 30 and
y = 21. In the second case, we have a = b + 3c2 = −7, hence x = −21
and y = −30. Note that both pairs do satisfy the equation. Hence the
solutions are (x, y) = (30, 21) and (x, y) = (−21, −30). 

26
2a. Let a dividing line be a straight line that divides the board in two parts in
such a way that every tile lies entirely in one of these parts. Suppose that
there exists a tiling without dividing lines. Consider the columns k and
k+1, with 1 6 k 6 2009. Then there is a tile lying horizontally in these two
columns; otherwise, the vertical line between these two columns would have
been a dividing one. There are 4k squares in the first k columns, an even
number. Since every tile lying entirely inside the first k columns, covers
exactly 2 squares, it follows that the number of tiles lying horizontally in
the columns k and k + 1 is even. We’ve shown earlier that this number is
at least 1, so it must be at least 2.
So for every k with 1 6 k 6 2009, there are two tiles lying horizontally
in columns k and k + 1. These tiles together cover 2 · 2009 · 2 squares.
Furthermore, for every i with 1 6 i 6 3, there must be a tile lying vertically
in columns i and i + 1. These tiles together cover 3 · 2 squares. So the total
number of squares covered by these tiles is (2 · 2009 + 3) · 2 > 2 · 2010 · 2.
But the board only contains 4 · 2010 squares. This is a contradiction.

2b. We use induction on n to show that for all n > 3, a

5 × 2n-board can be tiled without dividing line. For
n = 3, this can be done as indicated in the picture.
Suppose that we have a tiling of a 5×2n-board with-
out dividing lines. Then there are at least 4 tiles that
lie vertically. Hence there is a k with 1 6 k 6 2n − 1
such that there is a tile lying vertically in column k. Now add two columns,
called k1 and k2 , between columns k and k+1. Every tile lying horizontally
in columns k and k + 1, is replaced by two horizontal tiles, one in columns
k and k1 , one in columns k2 and k + 1. Since column k contains a vertical
tile, not all squares in columns k1 and k2 are covered yet. Also note that if
in a certain row, the square in k1 is not covered, then neither is the square
in k2 in the same row. Hence we can put a horizontal tile there. In that
way, all the squares of the new columns are covered.
Now it’s clear that no horizontal dividing lines are created in this process,
and no vertical ones between columns that are unchanged. Moreover, there
is at least one tile lying horizontally in columns k1 and k2 , so there is no
dividing line between these two columns. In the original board, there was
no dividing line between columns k and k + 1. That implies that there are
tiles lying horizontally in columns k and k1 , and in columns k2 and k + 1.
So there are no dividing lines between these pairs of columns. Hence the
board constructed contains no dividing lines at all.
This completes the induction. We conclude that there is a tiling of a
5 × 2010-board without dividing line. 

27
3. Let S be the intersection of P D and AB. Then S lies on the radical axis
of the two circles, so |SA| = |SB|. So P S is a median in triangle P AB.
By the inscribed angle theorem on Γ1 with chord AP , we have ∠BAP =
180◦ − ∠ADP = ∠CDP = ∠CBP . By the inscribed angle theorem on
Γ2 with chord BP , we have ∠ABP = ∠BCP . Hence 4P AB ∼ 4P BC
(aa). Since P M is a median in triangle P BC, we have ∠SP B = ∠M P C.
It follows that

∠DP M = ∠DP B + ∠BP M = ∠SP B + ∠BP M = ∠M P C + ∠BP M

= ∠BP C = ∠BDC. 

4. Suppose that P (x) can be written as P (x) = A(x)B(x) with A and B are
non-constant polynomials with integer coefficients. Since A and B are non-
constant, they both have degree at least 1. The sum of the degrees is equal
to the degree of P , hence equal to 3. This implies that the two degrees are 1
and 2. So we can write, without loss of generality, that A(x) = ax2 + bx + c
and B(x) = dx + e, where a, b, c, d and e are integers. The product of the
leading coefficients a and d is equal to the leading coefficient of P , hence
equal to 3. By multiplying both A and B if necessary, we may assume that
both a and d are positive, so that they are equal to 1 and 3 in some order.
First suppose that d = 1. Substituting x = −1 yields

P (−1) = 3 · (−1)3 + n − n − 2 = −5,

so
−5 = P (−1) = A(−1)B(−1) = A(−1) · (−1 + e).
Note that −1 + e is a divisor of −5, hence equal to −5, −1, 1 or 5, giving
four possible values for e, namely −4, 0, 2 or 6. Moreover, x = −e is a
zero of B, hence also of P .
If e = −4, then

0 = P (4) = 3 · 43 − 4n − n − 2 = 190 − 5n,

so n = 38. So we can indeed factor P (x):

3x3 − 38x − 40 = (3x2 + 12x + 10)(x − 4).

If e = 0, then
0 = P (0) = −n − 2,
so n = −2. So we can indeed factor P (x):

3x3 + 2x = (3x2 + 2)x.

28
 If e = 2, then

0 = P (−2) = 3 · (−2)3 + 2n − n − 2 = −26 + n,

so n = 26. So we can indeed factor P (x):

3x3 − 26x − 28 = (3x2 − 6x − 14)(x + 2).

If e = 6, then

0 = P (−6) = 3 · (−6)3 + 6n − n − 2 = −650 + 5n,

so n = 130. So we can indeed factor P (x):

3x3 − 130x − 132 = (3x2 − 18x − 22)(x + 6).

Now suppose that d = 3. Then we have:

−5 = P (−1) = A(−1)B(−1) = A(−1) · (−3 + e).

Note that −3 + e is a divisor of −5, hence equal to −5, −1, 1 or 5. This

gives four possible values for e, namely −2, 2, 4 or 8. Furthermore, x = −e
3
is a zero of B, hence also of P . Note that e is never divisible by 3. We
have
e3
0 = P ( −e −e 3 e e−3
3 ) = 3 · ( 3 ) + 3 n − n − 2 = − 9 + 3 n − 2,
3 3
so e−3 e e
3 n = 9 + 2, hence (e − 3)n = 3 + 6. But this is a contradiction,
since the left hand side is an integer, whereas the right hand side is not,
since 3 does not divide e.
Finally, we deduce that the solutions are: n = 38, n = −2, n = 26 and
n = 130. 

5. By |AB| > |BC|, the points D and E are distinct, and the order of the
points on the line AC is: A, D, E, C. Moreover, ∠BCA > ∠CAB, so
∠BCA + 12 ∠ABC > 90◦ , which implies that F lies on the interior of BE.
Consequently, G lies on the interior of BD.
Now let K be the intersection of CF and AB. Since BE is the angle
bisector of ∠ABC, we have ∠KBF = ∠F BC. Furthermore, ∠BF K =
90◦ = ∠CF B and |BF | = |BF |, hence by (ASA), we have 4KBF ∼ =
4CBF . This implies that |BK| = |BC| and |KF | = |CF |. In particular,
F is the midpoint of CK. Since D is the midpoint of AC, it follows that
DF is a midparallel in 4AKC, so DF k AK and |DF | = 12 |AK|.

29
From DF k AB, it follows that 4KGB ∼ 4F GD (aa), hence
1 1 1
|DG| |F D| |AK| 2 (|AB| − |KB|) 2 (|AB| − |BC|)
= = 2 = = . (3)
|BG| |KB| |KB| |KB| |BC|

|AE| |AB|
The angle bisector theorem now implies that |CE| = |CB| , or, equivalently
|AC|−|CE| |AB|
|CE| = |BC| .This yields |BC| · (|AC| − |CE|) = |AB| · |CE|, so
|BC| · |AC| = |CE| · (|AB| + |BC|), hence

|BC| · |AC|
|CE| = .
|AB| + |BC|

Continuing the calculation:

1
|DE| = |DC| − |CE| = |AC| − |CE|
2
1 1
|AC| · (|AB| + |BC|) − |AC| · |BC| 2 |AC| · (|AB| − |BC|)
= 2 = .
|AB| + |BC| |AB| + |BC|

Thus 1
2 |AC|·(|AB|−|BC|) 1
|DE| |AB|+|BC| 2 (|AB|− |BC|)
= |BC|·|AC|
= .
|CE| |BC|
|AB|+|BC|

Combined with (3) this gives

|DE| |DG|
= .
|CE| |BG|

In triangle DBC this implies that EG k BC. Now let S be the intersection
of DF and EG. Then we have

∠SGF = ∠EGF = ∠F CB = ∠F KB = ∠GF D = ∠GF S,

so 4SF G is isosceles, with |SF | = |SG|. Moreover,

∠SEF = ∠GEF = 90◦ − ∠EGF = 90◦ − ∠GF S = ∠SF E,

so 4SF E is also isosceles, with |SF | = |SE|. We deduce that |SG| = |SE|,
hence that S is the midpoint of EG. 

30
 IMO Team Selection Test 2, June 2011
Problems

1. Let n > 2 and k > 1 be positive integers. In a country there are n cities and
between each pair of cities there is a bus connection in both directions. Let
A and B be two different cities. Prove that the number of ways in which
you can travel from A to B by using exactly k buses is equal to
(n − 1)k − (−1)k
.
n

2. Find all functions f : R → R satisfying

xf (x + xy) = xf (x) + f (x2 )f (y)
for all x, y ∈ R.

3. Let Γ1 and Γ2 be two intersecting circles with midpoints respectively O1

and O2 , such that Γ2 intersects the line segment O1 O2 in a point A. The
intersection points of Γ1 and Γ2 are C and D. The line AD intersects Γ1
a second time in S. The line CS intersects O1 O2 in F . Let Γ3 be the
circumcircle of triangle ADF . Let E be the second intersection point of
Γ1 and Γ3 .
Prove that O1 E is tangent to Γ3 .

4. Prove that there exists no infinite sequence of prime numbers p0 , p1 , p2 , . . .

such that for all positive integers k:
pk = 2pk−1 + 1 or pk = 2pk−1 − 1.

5. Find all triples (a, b, c) of positive integers with a+b+c = 10 such that there
are a red, b blue and c green points (all different) in the plane satisfying
the following properties:
• for each red point and each blue point we consider the distance be-
tween these two points; the sum of these distances is 37;
• for each green point and each red point we consider the distance
between these two points; the sum of these distances is 30;
• for each blue point and each green point we consider the distance
between these two points; the sum of these distances is 1.

31
 Solutions

1. Let α(k) be the number of ways to travel from city A to city B 6= A by

using k buses. Let β(k) be the number of ways to travel from city A to
city A by using k buses. If we start in city A and then take a bus k times,
this can be done in (n − 1)k ways. In β(k) of the cases we end at city A
and in (n − 1)α(k) of these cases we end at a city different from A. Hence

(n − 1)α(k) + β(k) = (n − 1)k . (4)

Now suppose that k > 2. To travel from city A to city A by using exactly
k buses, we take a bus from A to an arbitrary city (this can be done in
(n − 1) ways); then we must travel from a city different from A to city A
by using k − 1 buses, which can be done in α(k − 1) ways. Therefore

β(k) = (n − 1)α(k − 1) for k > 2. (5)

Now we substitute this expression for β(k) in (4). This yields for k > 2
that
(n − 1)α(k) + (n − 1)α(k − 1) = (n − 1)k
and hence
α(k) = (n − 1)k−1 − α(k − 1). (6)
Now we prove by induction to k that for n > 2 and k > 1:

(n − 1)k − (−1)k
α(k) = .
n

For k = 1 this is equivalent to α(1) = (n−1)+1

n = 1 and that is true, because
there is exactly one way to travel from city A to city B 6= A by using one
bus. Now let m > 1 be an integer and suppose that the expression for α(k)
has been proven for k = m. Then, using (6), for k = m + 1 > 2:

(n − 1)m − (−1)m
α(m + 1) = (n − 1)m − α(m) = (n − 1)m −
n
n(n − 1)m − (n − 1)m + (−1)m (n − 1)m+1 − (−1)m+1
= =
n n
and that is exactly the expression we wanted to prove for k = m + 1. This
completes the induction. 

32
2. Substituting x = 0 and y = 0 yields 0 = f (0)2 , hence f (0) = 0. Sub-
stituting x = 1 and
y = −1 yields f (0) = f (1) + f (1)f (−1), hence
0 = f (1) 1 + f (−1) , therefore f (1) = 0 or f (−1) = −1. Substituting
x = −1 yields
−f (−1 − y) = −f (−1) + f (1)f (y) for all y ∈ R. (7)
Suppose that f (1) = 0, then this is equivalent to −f (−1 − y) = −f (−1)
and because −1 − y takes all values in R, this means that f is constant.
Because f (0) = 0, it must be true that f (x) = 0 for all x. It is clear that
this function satisfies the original equation. We found our first solution:
f (x) = 0 for all x ∈ R.
Now suppose that f (1) 6= 0, then f (−1) = −1. Now substitute y = −1
in (7), then we get −f (0) = −f (−1) + f (1)f (−1), hence 0 = 1 − f (1),
therefore f (1) = 1. Substituting x = 1 in the original equation yields
f (1 + y) = 1 + f (y) for all y ∈ R. (8)
Furthermore substituting y = −1 in the original equation yields xf (0) =
xf (x) − f (x2 ), hence
xf (x) = f (x2 ) for all x ∈ R. (9)
The original equation now can be written as
xf (x + xy) = xf (x) + xf (x)f (y) = xf (x)(1 + f (y)).
If x 6= 0, we may divide the left and right hand side by x and this yields
together with (8):
f (x + xy) = f (x)f (1 + y) for x 6= 0.
Notice that this is also true for x = 0. Now define z = 1 + y. Because this
takes all values in R, we get
f (xz) = f (x)f (z) for all x, z ∈ R.
Applying this to (9), we find
xf (x) = f (x2 ) = f (x)f (x),
hence for all x we have f (x) = 0 or f (x) = x. Now suppose there is a
x 6= 0 with f (x) = 0, then
1 = f (1) = f (x · x1 ) = f (x)f ( x1 ) = 0,
which is a contradiction. Hence for all x 6= 0 we have f (x) = x. This is
also true for x = 0. Substituting it in the original equation shows that
this is indeed a solution. The two functions satisfying the condition are
f (x) = 0 for all x ∈ R and f (x) = x for all x ∈ R. 

33
3. The intersection point of O1 O2 and the arch SD of Γ1 containing C, is
named T . Because A lies in the interior of Γ1 , we now know

∠O1 AS = ∠AT S + ∠T SA exterior angle theorem in 4AT S

= ∠O1 T S + ∠T SD
= ∠T SO1 + ∠T SD 4O1 ST is isosceles (|O1 S| = |O1 T |).

The line O1 O2 is perpendicular to CD and splits the line segment CD into

two equal parts, hence it is the perpendicular bisector of CD. Therefore
T lies on the perpendicular bisector of CD, which yields that arches T C
and T D are equally long. Hence by the inscribed angle theorem ∠T SD =
∠CST . Thus

∠O1 AS = ∠T SO1 + ∠T SD = ∠T SO1 + ∠CST = ∠CSO1 = ∠F SO1 .

This means that 4O1 AS ∼ 4O1 SF (hh). This yields

|O1 A| |O1 S|
= ,
|O1 S| |O1 F |

hence |O1 A| · |O1 F | = |O1 S|2 = |O1 E|2 . Because A and F lie on the
same side of O1 , it is even true that O1 A · O1 F = O1 E 2 . Using the power
theorem we now see that O1 E is tangent to the circumcircle of 4AF E and
that is Γ3 . 

4. Suppose that such an infinite sequence exists. By eventually leaving out

the first two elements, we can make sure that the first prime number of
the sequence is at least 5. Now we assume without loss of generality that
p0 > 5. Then we know that p0 6≡ 0 mod 3.
Suppose that p0 ≡ 1 mod 3. We prove by induction to k that for all
positive integers k: pk ≡ 1 mod 3 and pk = 2pk−1 − 1. Namely, suppose
that k > 1 and pk−1 ≡ 1 mod 3. Then 2pk−1 ≡ 2 mod 3, hence 2pk−1 + 1
is divisible by 3. Because pk is a prime number greater than 3, we have
pk = 2pk−1 − 1 and pk ≡ 2 − 1 ≡ 1 mod 3. This completes the induction.
In the case p0 ≡ 2 mod 3, we can prove analogously that pk ≡ 2 mod 3
and pk = 2pk−1 + 1 for all positive integers k.
Now we can compose a direct formula for the sequence. If p0 ≡ 1 mod 3,
then we get pk = (p0 − 1)2k + 1 for all k > 0. If p0 ≡ 2 mod 3, then we
get pk = (p0 + 1)2k − 1. We can prove these formulas by induction again.
By Fermat’s little theorem 2p0 −1 ≡ 1 mod p0 , hence also (−p0 +1)2p0 −1 ≡
1 mod p0 and (p0 +1)2p0 −1 ≡ 1 mod p0 . This yields p0 | (p0 −1)2p0 −1 +1

34
 and p0 | (p0 + 1)2p0 −1 − 1. We see that p0 always is a divisor of pp0 −1 .
Because it is also clear that pp0 −1 is greater than p0 , this yields that pp0 −1
is not a prime number. This is a contradiction. 

5. We construct triangles consisting of a blue, a red and a green point. These

triangles may also be degenerated. In each of these triangles the (non-
strict) triangle inequality holds: the distance between the blue and the
red point is at most the sum of the distances between the blue and the
green and between the red and the green point. We add all these triangle
inequalities (one for each triangle we can form with a blue, a red and a
green point). We now count each distance between a red and a blue point c
times (because with a fixed blue and red point you may choose c points as
third green point), each distance between a green and a red point b times
and each distance between a blue and a green point a times. Thus, we get

37c 6 30b + a.

Because a + b + c = 10, this yields 37c 6 30b + (10 − b − c) = 10 + 29b − c,

hence 38c 6 10 + 29b. Otherwise stated
38c − 10
6 b. (10)
29
We can also apply the triangle inequality differently: the distance between
the green and the red point is at most the sum of the distances between
the red and the blue and between the blue and the green point. If we add
these inequalities, we get
30b 6 37c + a.
This yields 30b 6 37c + (10 − b − c) = 10 + 36c − b, so 31b 6 10 + 36c,
otherwise stated
10 + 36c
b6 . (11)
31
Combining (10) and (11) yields
38c − 10 10 + 36c
6 ,
29 31
hence 31(38c − 10) 6 29(10 + 36c). Expanding yields 134c 6 600, therefore
c 6 4. Now we consider one by one all possibilities for c.
46
Suppose c = 1. Then (11) yields b 6 31 < 2, which yields b = 1. Now we
get (a, b, c) = (8, 1, 1).
66 82
Suppose c = 2. Then (10) and (11) yield 2 < 29 6b6 31 < 3, hence there
is no integer b that satisfies the conditions.

35
Suppose c = 3. Then (10) and (11) yield 3 < 104 29 6 b 6
118
31 < 4, hence
there is no integer b that satisfies the conditions.
Suppose c = 4. Then (10) and (11) yield 4 < 142 29 6 b 6
154
31 < 5, hence
there is no integer b that satisfied the conditions.
The only triple that possibly could satisfy the conditions is (8, 1, 1). Now
we show that there are indeed 8 red points, 1 blue point and 1 green point in
the plane that satisfy all conditions. We use a standard coordinate system.
Choose for the blue point (0, 0) and for the green point (1, 0). Choose red
points (i, 0) with 2 6 i 6 8. Also choose a red point such that that point
together with the green and blue point forms an isosceles triangle with two
sides of length 2. The sum of the distances between the red points and
the blue point now is 2 + 2 + 3 + · · · + 8 = 37. The sum of the distances
between the red points and the green point is 2 + 1 + 2 + · · · + 7 = 30. The
distance between the blue point and the green point is 1.
We conclude that the only possible solution is: (8, 1, 1). 

36
 Junior Mathematical Olympiad, October 2010
Problems
Part 1

1. The letters A, B, C and D represent digits. The following holds:

A B A B
C A + and C A −
D A A

What is D?
A) 5 B) 6 C) 7 D) 8 E) 9

2. Peter is constructing a sequence of seven integers (they can be negative or

zero as well), such that the sum of four consecutive integers is always 1.
He wants his sequence to contain as many integers as possible, which are
greater than 13. What is the maximal number of such integers?
A) 0 B) 2 C) 3 D) 5 E) 6

3. An ant is walking on the surface of a cuboid with edges of length 3, 4

and 5. It starts walking on a vertex, and it wants to visit all the other
seven vertices. It doesn’t need to return to its starting vertex. What is
the length of the shortest possible route that accomplishes this?
A) 24 B) 25 C) 26 D) 27 E) 28

1
4. The integers 1 up to 5 are placed to form a circle.
We add every pair of neighbours together. The five ? ?
sums that we obtain, turn out to be consecutive in-
tegers. What are the neighbours of 1?
A) 2 and 4 B) 2 and 5 C) 3 and 4 D) 3 and 5 E) 4 and 5

5. Bert and Ernie both have 64 candies. Every day, one of them gives half of
his candies to the other. After six days, Bert has 61 candies, so Ernie has
67 candies. How many of the six days has Ernie shared his candies?
A) 1 B) 2 C) 3 D) 4 E) 5

37
 6. If you add 36 to 37, you obtain 73. When its digits are written in reverse
order, this becomes 37 again. How many two-digit integers are there with
the property that if you add 36 to it, then write its digits in reverse order,
you obtain the integer you started with?
A) 4 B) 5 C) 6 D) 9 E) 10

7. One hundred students participate in a mathematical olympiad. Problem

1 was solved by 90 participants. Problem 2 was solved by 80 participants
and problem 3 has been solved by 75 participants. What is the minimal
number of participants that solved all of the problems?
A) 35 B) 45 C) 54 D) 55 E) 60

8. From 125 small cubes, a 5 × 5 × 5-cube is made. In every direction, the

cubes are coloured white and black alternatingly; the cubes on the vertices
are black. We only consider the cubes that are visible from the outside.
What is the difference between the number of black cubes and the number
of white ones?
A) 2 more black cubes B) 1 more black cube C) no difference
D) 1 more white cube E) 2 more white cubes

9. We start by drawing an equilateral triangle, and then

its circumcircle. Around this circle, we draw a per-
fectly fitting square. We then draw its circumcircle,
and draw a perfectly fitting pentagon around it, and
so on, up to a regular 16-gon. In the figure, you can
see that the area inside of the pentagon is divided in
17 pieces. In how many pieces is the area inside the
regular 16-gon divided?
A) 134 B) 136 C) 248 D) 264 E) 267

10. How many integers between 1 and 1000 do not contain the digit 1?
A) 700 B) 728 C) 729 D) 880 E) 900

38
11. A square with edges of length 2010 is divided into
nine rectangles by four lines which are parallel to the
edges. You can see an example in the figure. You can
divide this square in such a way that the perimeter of
the rectangles obtained are nine consecutive integers.
What is the perimeter of the largest rectangle?
A) 671 B) 1340 C) 1790 D) 2684 E) 3577

12. A farmer has a stack of hay to feed his horse, cow and goat. With it,
he can feed his cow and horse for 12 months, or his cow and goat for 15
months, or his horse and goat for 20 months. For how many months can
he feed all three of his animals?
A) 7 56 B) 9 C) 10 D) 15 23 E) 47

13. There are twenty-five towers: five of height 1, five of 2

height 2, five of height 3, five of height 4 and five of 1
height 5. The towers have to be places on a 5 × 5- 3
board with on each square a tower, such that in every
row and every column, every height occurs exactly 2
once. Moreover, in the direction of each of the ar- 4
rows in the figure, you have to be able to see exactly
the number of towers mentioned at that arrow. You 3 5 ? 2 1
can’t see a tower if it’s behind a higher tower. What number should be
filled in the place of the question mark?
A) 1 B) 2 C) 3 D) 4 E) 5

14. Tania has played 10 basketball matches this season. In the sixth up to the
ninth match, she scored 23, 14, 11 and 20 points, respectively. As a result,
her average score per match became higher after nine matches, than it was
after five matches. After ten matches, she has an average score of more
than 18 points. What is the minimal number of points that Tania scored
in the tenth match?
A) 19 B) 27 C) 28 D) 29 E) 31

39
15. Lucas’ sequence starts with the integers 1 and 3. After that, the next
integer in the sequence is found by adding the two preceding integers. So
you obtain 1, 3, 4, 7, 11, . . .. What is the last digit of the hundredth
integer in Lucas’ sequence?
A) 1 B) 3 C) 4 D) 7 E) 8

Part 2
The answer to each problem is a number.

1. The length of the side of the small square is 5, and

the length of the side of the large square is 10. What
is the total area of all the black pieces?

2. The following sum is incorrect:

7 4 2 5 8 6
8 2 9 4 3 0 +
1 2 1 2 0 1 6

This sum can made correct by picking two distinct digits and replace every
occurrence of one of them with the other one. Which two digits should
you pick?

3. A rhombus is cut from a pattern of grey and white

squares. What fraction of the rhombus is grey?

4. Consider all four-digit integer in which each of the digits 1, 2, 3 and 4

occur exactly once. What’s the average of all these integers?

5. Three equilateral triangles with edges of length 21

are placed on a white table in such a way that we
obtain an equilateral triangle with edges of length
36. In the centre, a small triangular piece of table is
visible. How long are the edges of that piece?

40
 6. A rectangle is divided into eight squares. The area
of the grey square is 1. What is the area of the
rectangle?

7. Anton and Ben are competing in a race. Ben runs three times as fast
as Anton, but Anton gets a 30 meter lead. The arrive simultaneously at
the finish line. What is the distance in meters that Ben has run when he
crosses the finish line?

8. A square is divided into seven pieces by its diago- 160

nal and by two lines parallel to the edges. Of three
pieces, the area is given. What is the area of the 50
square? 90

9. Anne answered 33 questions, most of them well and the rest even very well.
Both grades are rewarded with a fixed number of points per questions,
very well answered questions being rewarded with more points than well
answered questions. Both number of points are integers 1, 2, 3, . . ., or
10, but we don’t know exactly which two integers they are. When Anne
calculated her average number of points per question, it turned out to be
an integer. How many questions did Anne answer very well?

10. On a circle, there are eleven points, numbered from 1 9

1 up to 11. From each point except point 11, a 5
6
line segment is drawn between that point, and the
point with the next number. Another line segment 10
is drawn between point 11 and point 1. What is the 2
maximal number of intersections of these 11 line seg- 4
ments? In the figure, there’s an example with only 7 8
16 intersections. 3 11

41
Solutions
Part 1

1. E) 9 6. B) 5 11. D) 2684

2. E) 6 7. B) 45 12. C) 10

3. B) 25 8. A) 2 more black 13. B) 2

4. C) 3 and 4 9. C) 248 14. D) 29

5. D) 4 10. B) 728 15. D) 7

Part 2
595
1. 25 6. 8 = 74 83

2. 2 and 6 7. 45

2
3. 3 8. 484

4. 2777 12 9. 11

5. 9 10. 44

42