Professional Documents
Culture Documents
Dutch
Mathematical
Olympiad
2007
DSE
W
IS
N
K
RLA
U
N
D
E
E
NED
P I AD E
OLYM
Contents
Introduction i
Solutions 3
Solutions 6
Solutions 9
Solutions 15
The Dutch Mathematical Olympiad consists of two rounds. The first round is held on the
participating schools and consists of eight multiple choice questions and four open questions
(see page 1–4). Students get two hours to work on the paper. On January 26, 2007, in total
2742 students of 185 secondary schools participated in this first round.
Those students that scored 20 points or more on the first round (out of a maximum of
36 points) were invited to the second round, which took place at University of Technology
Eindhoven after the summer holidays. Also some outstanding participants in the Kangaroo
math contest or the Pythagoras Olympiad were invited.
In total 132 students were invited, of which 119 participated in the second round on September
14th, 2007. The second or final round contains five problems for which the students have to
give extensive solutions and proofs. They have three hours for the paper (see page 5–7).
After the prices had been awarded in the beginning of November, the Dutch Mathematical
Olympiad concluded its 46th edition 2007. In 2011 we will have our 50th edition.
The 24 most outstanding candidates of the Dutch Mathematical Olympiad 2007 were invited
to an intensive seven-month training programme, consisting 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. At this time, only 14 candidates
were left. Out of those, the team was selected by a final selection test on June 7, 2008 (see
page 8–12). The team will have a training camp in Spain from July 7 until July 14, together
with the team from New Zealand.
In the meantime, a new edition of the Dutch Mathematical Olympiad has started at the
participating schools on January 25, 2008 (see page 13–16). We are happy to see that the
number of participating schools as well as the number of participants has increased to 201
respectively 3004. As a new initiative, we have made some training material and we have
organized training sessions at six universities in the country for the 144 students who have
been invited for the second round in September 2008.
i
The Dutch team for IMO 2008 Madrid consists of
We are grateful to Jinbi Jin for the translation into English of both the problems and the solutions.
ii
SE
W
AND
IS
K
U
ERL
N
D
E
NED Dutch Mathematical Olympiad
P IADE
OLYM Round 1
Friday, January 26, 2007
Please read the following before you start: Available time: 2 hours
• The A-problems are multiple choice questions. Only one of the five given options is correct. Please state
clearly behind which letter the correct solution is stated. You get 2 points for each correct answer.
• The B-problems are open questions, which have a number as answer. You get 5 points for each correct
answer. Please work accurately, since an error in your calculations
√ can cause your solution to be considered
wrong. Please give exact answers, for example 1181
or 2 + 3 or π + 1.
• This is a competition, not an exam. So it is to be expected that only few people will get all the answers
right, and you don’t need to worry if you have only solved a part of the problems.
• The use of formula sheets, calculators and similar tools is not allowed.
• The point is that you have fun solving unusual mathematical problems. Good luck!
A-problems
A1. The number M has 2007 digits, each of which is equal to 1, M = 111 . . . 111. What is the
sum of the digits of 2007 × M ?
A2. The next five sequences all contain the same numbers. Which sequence is ordered correctly?
117 513
A) 0.16 < <
101 < 33
7 97 <
17 1 5 13
B) 101 < 0.16 < 7 < 33 < 97
1 5 17 13
C) 7 < 0.16 < 33 < 101 < 97
5 1 13 17
D) 33 < 7 < 97 < 101 < 0.16
13 1 5 17
E) 97 < 7 < 33 < 0.16 < 101
A3. In the figure, nine grid points are drawn. In how many ways can we
draw a triangle, the vertices of which lie on three of these nine grid
points?
A) 76 B) 84 C) 92 D) 496 E) 504
1
A4. How many pairs of positive integers (a, b) with a + b < 100 satisfy the equation a + b =
13 × (b + a1 )?
A) 5 B) 7 C) 9 D) 13 E) 28
A) B) C) D) E)
1
A6. We construct a sequence of numbers as following:
The first number is equal to 2, and so is the second number. Every following number
is the product of the two previous ones. So the first few numbers of the sequence are:
2, 2, 4, 8, 32, . . .
What is the final digit of the 2007th number in this sequence?
A) 0 B) 2 C) 4 D) 6 E) 8
A7. Does the equation 9n + 9n + 9n = 32007 have any integral solutions? If so, which of the
following is a solution?
A8. Behind a long table, nine chairs are arranged in a row, on which six students and three
teachers are going to sit. The three teachers arrive first and they decide to pick their chairs in
such a way that every teacher sits between two students. In how many ways can the teachers
pick their chairs, given this constraint?
A) 12 B) 36 C) 60 D) 84 E) 630
B-problems
B1. Mark one card with a ‘1’, two cards with a ‘2’, . . ., fifty cards with a ‘50’. Put these 1 + 2 +
· · · + 50 = 1275 cards into a box and shuffle them. How many cards do you need to take from
the box to be certain that you will have taken at least 10 cards with the same mark?
D C
B2. We have a quadrilateral ABCD with side lengths |AB| = 16, |BC| =
21, |CD| = 2 and |DA| = 28. The sides AB and CD are parallel. Two
lines, both parallel to AB and CD, divide ABCD into three similar
quadrilaterals.
Calculate the perimeter of the smallest of these quadrilaterals.
A B
B3. For every pair of integers a, b, we define an operator a ⊗ b with the following three properties.
1. a ⊗ a = a + 2;
2. a ⊗ b = b ⊗ a;
a⊗(a+b) a+b
3. a⊗b = b .
Calculate 8 ⊗ 5.
B4. A flag with the shape of an equilateral triangle is hung up on the tops
of two vertical poles by two of its vertices. One of these poles has
length 4, while the other has length 3. Also, the third vertex of the
flag touches the ground. Calculate the length of the flag’s side.
2
SE
W
AND
IS
K
U
ERL
N
D
E
NED
P IADE Solutions
OLYM
of the problems of Round 1 of the
Dutch Mathematical Olympiad 2007
A-problems
A1. C) 18063 If M had 4 ones, then 2007 × M = 2229777, and the sum of the digits equals
3 × (2 + 7) + 9 = 4 × 9.
If M had 5 ones, then 2007×M = 22299777, and the sum of the digits equals 3×(2+7)+2×9 =
5 × 9.
Continuing like this, we see that if M has 2007 ones, then the sum of the digits equals
2007 × 9 = 18063.
A2. E) We don’t need to give all the fractions the same denominator; we can choose
16
the denominators in such a way that they are near 100. So 0, 16 = 100 , 71 = 14 5 15
98 , 33 = 99 .
Since, for k < n, nk < n+1
k+1 k+1
, (because of n+1 − nk = n(n+1)
n−k
> 0,) we now can see that
13 14 15 16 17
97 < 98 < 99 < 100 < 101 .
2, 2, 4, 8, 2, 6, 2, 2, 4, 8, 2, 6, 2, . . . ,
and that we have a repeating part of length 6, namely 2, 2, 4, 8, 2, 6. Since 2007 = 334 × 6 + 3,
the final digit of the 2007th number must be a 4.
3
A8. C) 60 Imagine that the six students sit on their chairs in a row. The teachers then can
pick two neighbouring students, between which they are going to sit. For the first teacher,
there are exactly five ways to pick a chair that is between the chairs of two students. For
the second teacher, there are four, and for the third, there are three. So the total number of
ways for the teachers to pick their chairs in such a way that every teacher sits between two
students, is 5 × 4 × 3 = 60.
B-problems
B1. 415 Considering the worst case scenario, we see that we can get all the cards marked
‘1’ to ‘9’, and 9 of each of the cards marked ‘10’ to ‘50’, without having 10 cards of one kind.
In that case, we have 45 + 9 × 41 = 414. Now picking one card more (so 415 in total) ensures
that we have 10 cards of one kind.
B2. 13 Let P, Q, R, S be the intersections of the two lines where
P and R are on DA, and Q and S are on CB, see the figure. Now the
quadrilaterals ABSR, RSQP and P QCD are similar. So
D 2 C
|DC| |P Q| |RS| 4 3
= = . P
4
Q
|P Q| |RS| |AB|
R S
8
Hence |P Q|2 = 2 × |RS|, and |RS|2 = |P Q| × 16, so |P Q|3 = 64,
and |P Q| = 4, |RS| = 8. So the sides of RSQP are twice as large
as the corresponding sides of P QCD, and the sides of ABSR are A B
16
twice as large as the corresponding sides of RSP Q. Thus |RP | =
2 × |P D|, |AR| = 2 × |RP | = 4 × |P D|, and |AD| = 7 × |P D|, so we
conclude that |P D| = 4. Similarly, we see that |QC| = 3.
Thus the perimeter of P QCD equals 13.
5 ⊗ (5 + 3) 8 8
8 ⊗ 5 = (5 + 3) ⊗ 5 = 8⊗5= × (5 ⊗ 3)
5⊗3 3 3
3 ⊗ (3 + 2) 5 5
5 ⊗ 3 = (3 + 2) ⊗ 3 = 5 ⊗ 3 = × (3 ⊗ 2)
3⊗2 2 2
2 ⊗ (2 + 1) 3
3 ⊗ 2 = (2 + 1) ⊗ 2 = 3 ⊗ 2 = 3 × (2 ⊗ 1)
2⊗1 1
1 ⊗ (1 + 1) 2
2 ⊗ 1 = (1 + 1) ⊗ 1 = 2 ⊗ 1 = 2 × (1 ⊗ 1)
1⊗1 1
Since 1 ⊗ 1 = 1 + 2 = 3, we have 2 ⊗ 1 = 2 × 3 = 6, then 3 ⊗ 2 = 3 × 6 = 18, and
5 ⊗ 3 = 52 × 18 = 45, and finally, 8 ⊗ 5 = 38 × 45 = 120.
2
√
B4. 3 39 Let x denote the side of the equilateral triangle. By re- C
2 2 x
peated application of Pythagoras’
√ Theorem, √ we get 2x = |BD|2 +
1
D B
2 2 2 2 2
√ + |AF
1 = (|EA| √ |) + 1 = ( x − 4 + x − 3 ) + 1 = 2x −
25√ + 2 x√− 4 x2 − 32 + 1. Rewriting this equation gives us
2 2 x
3 3
2 x2 − 42 x2 − 32 = −x2 + 24. Then we square both sides: 4(x4 − x
25x2 + 144) = x4 − 48xq 2 + 576. Thus 3x4 − 52x2 = 0. Now the only
√ E A F
positive solution is x = 523 = 2
3 39.
4
SE
W
AND
IS
K
U
ERL
N
D
E
NED Dutch Mathematical Olympiad
P IADE
OLYM Round 2
Friday, September 14, 2007
Available time: 3 hours
• Writing down just the answer itself is not sufficient; you also need to describe the way you solved the problem.
• Calculators and formula sheets are not allowed; you are only allowed to use a pen, a compass and a ruler or
set square. And of course your common sense.
• Please write the solutions of each problem on a different sheet of paper.
• Good luck!
2. Is it possible to partition the set A = {1, 2, 3, . . . , 32, 33} into eleven subsets that contain
three integers each, such that for every one of these eleven subsets, one of the integers
is equal to the sum of the other two? If so, give such a partition; if not, prove that such
a partition cannot exist.
3. Does there exist an integer having the form 444 · · · 4443 (all fours, and ending with a
three) that is divisible by 13? If so, give an integer having that form that is divisible
by 13; if not, prove that such an integer cannot exist.
5
Solutions of Round 2 of the Dutch Mathematical Olympiad 2007.
Problem 1:
A2 B1
Draw AD through M , and√draw M A1 . Since AD is a me-
dian, |M D| = 13 |AD| = 61 3 and |BD| = 21 . Now, apply-
C
ing Pythagoras’ Theorem
√ to 4M DA1 , we find |M A1√ |2 =
D
2 2
|M D| + |DA1 | = ( 16 3)2 + ( 32 )2 = 37 . So |M A1 | = 13 21.
M
(Similarly, one can compute
√ the distance from M to the A B
C1 C2
other points, yielding 13 21 every time.) So the radius of
the circle passing
√ through the points A1 , B2 , C1 , A2 , B1 , C2
is equal to 31 21.
B2 A1
Problem 2:
Suppose that we have such a partition into eleven subsets. Then for each of those eleven
subsets {a, b, c} we have that, say, a + b = c. So a + b + c = 2c, and the sum of the integers
in each subset is even. Hence also the sum of 1, 2, . . . , 32, 33 must be even.
But since 1 + 2 + 3 + · · · + 33 = 21 × 33 × (33 + 1) = 33 × 17 is an odd number, we have a
contradiction. So there doesn’t exist a partition with the desired properties.
Problem 3:
| · {z
Suppose that such an integer 444 · · 4443} exists and that it is divisible by 13, then substracting
k
13 from that integer gives another integer that is divisible by 13. So 444 · · · 4430 is divisible
by 13. Since 444 · · · 4430 = 444 · · · 443 × 10, and since 10 is not divisible by the prime 13,
| ·{z
444 · · 443}, the integer with one 4 less, must also be divisible by 13. We can repeat this
k−1
argument until we get the integer 43, from which we conclude that 43 must also be divisible
by 13. But 43 is not a multiple of 13, hence the integer we started with cannot be divisible
by 13 either. So there is no integer having that form that is divisible by 13.
(When the amount of 4’s equals 0, we have the number 3, which obviously isn’t a multiple of
13 either.)
6
Problem 4:
In the prime factorisation of a perfect square, every prime factor occurs an even number of
times. Also, if in the prime factorisation of an integer every prime factor occurs an even
number of times, then that integer must be a perfect square.
- Suppose a is even, and write a = 2c, then aa = (2c)2c = (2c cc )2 . So in this case, aa
always is a perfect square.
So we conclude that every even integer and every odd perfect square no larger than 100 satisfy
the condition, and there are 50 + 5 = 55 of those.
Problem 5:
C
we denote the area of a quadrilateral ABCD as [ABCD].
A B
In a similar way, we find for 4BCP and 4CAP respectively that [ESF P ] = 31 [BCP ] and
[F T DP ] = 13 [CAP ]. Hence
7
Team Selection Test
2. Julian and Johan are playing a game with an even number of cards, say 2n cards,
(n ∈ Z>0 ). Every card is marked with a positive integer. The cards are shuffled and
are arranged in a row, in such a way that the numbers are visible. The two players
take turns picking cards. During a turn, a player can pick either the rightmost or the
leftmost card.
Johan is the first player to pick a card (meaning Julian will have to take the last card).
Now, a player’s score is the sum of the numbers on the cards that player acquired during
the game.
Prove that Johan can always get a score that is at least as high as Julian’s.
√
4. Let n be positive√integer such that 1 + 12n2 is an integer.
Prove that 2 + 2 1 + 12n2 is the square of an integer.
5. Let 4ABC be a right triangle with ∠B = 90◦ and |AB| > |BC|; and let Γ be the
semicircle with diameter AB that lies on the same side as C. Let P be a point on Γ
such that |BP | = |BC| and let Q be on AB such that |AP | = |AQ|.
Prove that the midpoint of CQ lies on Γ.
8
Solutions of the Team Selection Test 2008
The above proves that the only function that can satisfy the functional equation is f (n) = n.
Checking this,
f (f (f (n))) + f (f (n)) + f (n) = n + n + n = 3n,
also proves that f also satisfies the functional equation.
Problem 2. Let 2n be the number of cards, and assume that the cards are marked with
(from left to right) a1 , a2 , . . . , a2n . We’ll prove by induction on n that Johan can always pick
either the odd cards a1 , a3 , . . . , a2n−1 , or the even cards a2 , a4 , . . . , a2n . This is clear for n = 1;
Johan picks the first card if he wants the odd cards, and he picks the second if he wants the
even cards.
So now suppose that the above statement is true for n = k, and consider the situation with
2k + 2 cards, a1 , a2 , . . . , a2k+2 . If Johan wants the odd cards, he first picks a1 , and Julian
will have to either pick a2 or a2k+2 in that case. In the first case, the sequence of cards after
2 turns becomes {bi } where bi = ai+2 , and according to the induction hypthesis, Johan can
get the odd bi , which in combination with a1 , will get him all the odd ai . In the second case,
the sequence of cards becomes {bi } where bi = ai+1 , and again, according to the induction
hypothesis, Johan can get the even bi , which in combination with a1 , will get him all odd ai .
So Johan can pick the odd cards. In a similar way we can prove that Johan can pick the even
cards. This completes the induction.
Now if the sum of the numbers on the odd cards is larger than or equal to the sum of the
numbers on the even numbers, Johan picks the odd cards. Otherwise he picks the even cards.
This ensures that Johan will get a score at least as high as Julian’s.
9
Problem 3.
aj ≥ i ⇐⇒
a1 , a2 , . . . , aj ≥ i ⇐⇒
at least j terms of a are greater than or equal to i ⇐⇒
bi ≥ j ⇐⇒
b1 , b 2 , . . . , b i ≥ j ⇐⇒
at least i terms of b are greater than or equal to j ⇐⇒
cj ≥ i.
Any term al with k > al is counted exactly al times in this sum (for every i ≤ al ), so
this sum must equal the sum of all such al . By solution 1 of a), we now have k ≤ al if
and only if l ≤ bk , so k > al if and only if l ≤ bk . Thus,
k
X X n
X
(bi − bk ) = al = al ,
i=1 l:k>al l=bk +1
10
Solution 2. By induction on k. If k = 1, then we need to prove
n
X
b1 = b1 + aj ,
j=b1 +1
which is obviously true since b1 = #{j : aj ≥ 1} = n, so the sum on the right hand side
is empty. So suppose that what we need to prove is true for k = s, 1 ≤ s ≤ m − 1. Then
s+1
X s
X
bi = bi + bs+1
i=1 i=1
n
IH
X
= s · bs + aj + bs+1
j=bs +1
n
X bs
X
= (s + 1) · bs+1 + aj + s(bs − bs+1 ) − aj
j=bs+1 +1 j=bs+1 +1
n
X bs
X
= (s + 1) · bs+1 + aj + s(bs − bs+1 ) − s
j=bs+1 +1 j=bs+1 +1
X n
= (s + 1) · bs+1 + aj .
j=bs+1 +1
Note that the left hand side is even, hence so is the right hand side, and a has to be odd.
Since the left hand side has an even number of factors 2, both a + 1, a − 1 have an odd number
of factors 2, since one of these two has exactly one factor 2. Also, (a + 1, a − 1) = 2, from
which follows that for any odd prime divisor p of the left hand side, p divides exactly one
of a + 1, a − 1. Thus, if p ≥ 5, the right hand side has an even number of factors p, and if
p = 3, the right hand side has an odd number of factors p. Hence we have the following two
possibilities:
a + 1 = 6b2 and a − 1 = 2c2
for certain integers b, c such that bc = n, or
11
Problem 5. Solution 1. Let S be the intersection of Γ en CQ. We need to prove that
|QS| = |SC|.
Note that BC is tangent to Γ, so ∠CBP = ∠BAP = ∠QAP . Since both 4CBP and 4QAP
are both isosceles, they’re similar. Let α = ∠BCP , then
So P , just like B, lies on the circle with diameter CQ. S lies on this diameter, and all that’s
left to show, is
∠BSP = 2α = 2∠BCP.
Considering the cyclic quadrilateral QBCP , we see that ∠CP B = ∠CQB, so
Considering the cyclic quadrilateral ABSP , with diameter AB, we see that
Solution 2. First of all, as in solution 1, the two isosceles triangles are similar. Since BC is
orthogonal to AB, one of the triangles is rotated by 90◦ with respect to the other. Now let
l1 be the angle bisector of ∠P BC and l2 be the angle bisector ∠P AQ, and note that these
angle bisectors are perpendicular. Let T be the intersection of l1 and l2 . Then ∠AT B = 90◦ ,
so T lies on Γ.
Note that P is the image of C under reflection in l1 , and Q is the image of P under reflection
in l2 . The composition of these two reflections is the rotation about T by 2 · 90 = 180 degrees.
This rotation sends C to Q, so CT Q is a straight line. Thus T = S. Also, |T C| = |T Q|, so
S is the midpoint of CQ.
12
SE
W
AND
IS
K
U
ERL
N
D
E
NED Dutch Mathematical Olympiad
P IADE
OLYM Round 1
Friday, January 25, 2008
Available time: 2 hours
Please read the following before you start:
• The A-problems are multiple choice questions. Only one of the five given options is correct. Please state
clearly behind which letter the correct solution is stated. You get 2 points for each correct answer.
• The B-problems are open questions, which have a number as answer. You get 5 points for each correct
answer. Please work accurately, since an error in your calculations
√ can cause your solution to be considered
wrong. Please give exact answers, for example 1181
or 2 + 3 or π + 1.
• This is a competition, not an exam. So it is to be expected that only few people will get all the answers
right, and you don’t need to worry if you have only solved a part of the problems.
• The use of formula sheets, calculators and similar tools is not allowed.
• The point is that you have fun solving unusual mathematical problems. Good luck!
A-problems
A1. Alex, Birgit, Cedric, Dion and Ersin all write their names on a sheet of paper, and they put
those five sheets into a large box. They each take one sheet out of the box at random. Now
it turns out that Birgit has Alex’ sheet, Cedric has Dion’s and Dion has Ersin’s. Also, Ersin
doesn’t have Cedric’s sheet. Whose sheet does Alex have?
A2. In a magic 3 × 3 square, the three row sums, the three column sums
and the two diagonal sums are all equal to each other. (A row sum 7
being the sum of the numbers on a certain row, etc.) In the magic
3 × 3 square shown here three numbers have already been filled in. ?
What number must be filled in instead of the question mark?
A) 2 B) 4 C) 6 D) 8 E) 9
10 3
A3. Calculating 6 × 5 × 4 × 3 × 2 × 1 yields 720. How many divisors does
720 have? (A divisor of an integer n is a positive integer by which n is
divisible. For example: the divisors of 6 are 1, 2, 3 and 6; the divisors
of 11 are 1 and 11.)
A) 6 B) 8 C) 20 D) 30 E) 36
D 5
A4. Of a quadrilateral ABCD, we know that |AB| = 3, |BC| = 4, |CD| = C
5, |DA| = 6 en ∠ABC = 90◦ . (|AB| stands for the length of segment
AB, etc.) What is the area of quadrilateral ABCD? 6
√ 4
A) 16 B) 18 C) 18 21 D) 20 E) 6 + 5 11
A 3 B
13
A5. How many five-digit numbers (like 12345 or 78000; the first digit must
be non-zero) are there that end on a 4 and that are divisible by 6?
D C
A6. We have a square ABCD with |AB| = 3. On AB, there is a point E
such that |AE| = 1 and |EB| = 2. AC and DE intersect in H. What
is the area of triangle CDH?
9 21 27
H
A) 8 B) 2 C) 8 D) 3 E) 8
A E B
A7. The seven blocks S E T T E E S are shuffled. For example, you can
get E E E S S T T or T E S E T E S .
How many different “words” of length 7 can we get this way? (Any
combination of the 7 letters counts as word.)
A) 4 B) 5 C) 6 D) 7 E) 8
B-problems
B1. We number both the rows and the columns of an 8×8 chessboard with
the numbers 1 to 8. A number of grains is placed onto each square, in
such a way that the number of grains on a certain square equals the
product of its row and column numbers. How many grains are there
on the entire chessboard?
B2. We take 50 distinct integers from the set {1, 2, 3, . . . , 100}, such that
their sum equals 2900. What is the minimal number of even integers
amongst these 50 numbers?
A B
14
SE
W
AND
IS
K
U
ERL
N
D
E
NED
P IADE Solutions
OLYM
of the problems of Round 1 of the
Dutch Mathematical Olympiad 2008
A-problems
A1. C) Cedric When we put the data into a table, we see that Birgit’s
and Cedric’s sheets haven’t been picked yet. Since Ersin didn’t pick A B C D E
Cedric’s sheet, he must have picked Birgit’s. So Alex must have picked ? A D E ?
Cedric’s sheet.
A5. C) 3000 If x is a positive multiple of 6 that ends with a 4, then the next multiples of
6 end with a 0 (x + 6), a 6 (x + 12), a 2 (x + 18), an 8 (x + 24), a 4 (x + 30), so the next
multiple of 6 that ends with a 4 is x+30. So any 30 consecutive positive integers must contain
exactly one integer with the desired properties. How many such integers lie between 10000
and 99999? Since we have 90000 consecutive positive integers, we find 90000 ÷ 30 = 3000
such integers amongst them.
D Q C
27
A6. E) 8 Draw a line through H parallel to AD, and let P and Q
be the intersections of this line with AB and CD, respectively. Now
we have |HP | : |HQ| = |AE| : |CD| = 1 : 3, so |HQ| = 34 × |P Q| =
3 9 1 9 27 H
4 × 3 = 4 . So the area of triangle CDH equals 2 × 3 × 4 = 8 .
A PE B
15
S S . . . . .
We have 6 + 5 + 4 + 3 + 2 + 1 = 21 (or 72 ) possibilities
A7. A) 210 S . S . . . .
to arrange the S-blocks onto the 7 places (see figure). For each choice, S . . S . . .
we have 4 + 3 + 2 + 1 = 10 (or 52 ) possibilities to arrange the T-blocks ..
.
on the remaining 5 places; after which the positions of the E-blocks . S S . . . .
are determined. So we have 21 × 10 = 210 possibilities.
. S . S . . .
Alternative solution: If all the blocks were different, we would have got
. S . . S . .
7! possibilities. But the three E-blocks aren’t different, so we end up ..
counting each word 3! times this way. Similarly for the two S-blocks .
7!
and the two T-blocks. So we find 3!×2!×2! = 7×6×5×4×6 3!
= 7×6×5 = . . S S . . .
63!×4
210 different words. . . S . S . .
..
.
A8. B) 5 This equation is equivalent to
(x2 − 2)2 − 5 = 1 or (x2 − 2)2 − 5 = −1.
√ √
The first is equivalent to x2 − √
2 = 6 of x2 − 2 = − 6, with 2 and 0
solutions respectively (since − 6 + 2 < 0).
The latter is equivalent to x2 − 2 = 2 of x2 − 2 = −2, with 2 and 1
solution(s) respectively. So we have 2 + 0 + 2 + 1 = 5 solutions in total.
B-problems
B1. 1296 In the first column, we have, successively, 1 × 1, 1 × 2, 1 × 3, . . . , 1 × 8 grains.
So, in the first column: 1 × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
In the second column: 2 × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
In the third column is: 3 × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
..
.
Finally, in the eighth column: 8 × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
So, in total: (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8) × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
Since 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 = 21 × 8 × (1 + 8) = 36, there are 362 = 1296 grains on the
board.
B2. 6 The 50 odd integers from the set {1, 2, 3, . . . , 100} sum up to 12 × 50 × (1 + 99) =
2500, which is still 400 short of 2900. Now exchange the smallest odd integers for the largest
even integers, in pairs, since 400 is even. First exchanging 1 and 3 for 100 and 98 makes the
sum equal to 2694. The next step gives us 2694 − 5 − 7 + 96 + 94 = 2872. Which is still less
than 2900, so we require another exchange. Now exchanging 9 and 11 for 20 and 28 works,
making the sum 2900 with 6 even integers, showing along the way that that is the minimal
number of even integers we need to do so.
1
B3. 110 We know that x must satisfy x + = 5, so x2 − 5x + 1 =
√x
5± 21
0, from which follows that x = x1,2 = 2 . Note that x1 x2 = 1.
√ 3 √ 3
Now we have x3 + x−3 = x31 + x32 = 1
8 5 + 21 + 5 − 21 =
1
√ √ 2 √ 3
√ √ 2 √ 3
8 53 + 3 · 52 · 21 + 3 · 5 · 21 + 21 + 53 − 3 · 52 · 21 + 3 · 5 · 21 − 21 =
2 3
8 5 + 3 · 5 · 21 = 110.
Alternative solution: From (x + x1 )3 = x3 + 3x2 ( x1 ) + 3x( x1 )2 + ( x1 )3 = x3 + 3x + x3 + x13 we
can deduce that x3 + x13 = (x + x1 )3 − 3(x + x1 ) = 53 − 3 × 5 = 110.
S
√ D C
B4. 2 3 Let Q, R, S, T be the orthogonal projections of P on
AB, BC, CD, DA, respectively. Then we have P
|AQ|2 + |QP |2 = 36 and |BQ|2 + |SP |2 = 25 (since |BQ| = |CS|), T R
2 2 2 2 2
so |AQ| + |QP | + |BQ| + |SP | = 61. We also have |BQ| + |QP | = 2
49. √
So |DS| 2 + |SP |2 = |AQ|2 + |SP |2 = 61 − 49 = 12 and |DP | = 12 =
√
2 3. A B
Q
16
The 47th
Dutch
Mathematical
Olympiad
2008
W
AND
IS
K
U
ERL
N
D
E
NED
PIADE
OLYM
Contents
Introduction i
The Dutch Mathematical Olympiad consists of two rounds. The first round is held on the
participating schools and consists of eight multiple choice questions and four open questions
(see page 1–4). Students get two hours to work on the paper. On January 25, 2008, in total
3004 students of 201 secondary schools participated in this first round.
The best students are invited for the second round. For the first time we set different thresh-
olds for students from different grades, in order to stimulate young students to participate.
Those students from grade 5 (4, ≤ 3) that scored 26 (22, 18) points or more on the first round
(out of a maximum of 36 points) were invited to the second round. Also some outstanding
participants in the Kangaroo math contest or the Pythagoras Olympiad were invited.
As a new initiative, we organized training sessions at six universities in the country for the
143 students who had been invited for the second round in September 2009. Former Dutch
IMO-participants are involved at each of the universities.
From those 143, in total 123 participated in the second round on September 12th, 2008 at
the Eindhoven University of Technology. This final round contains five problems for which
the students have to give extensive solutions and proofs. They have three hours for the paper
(see page 5–7). After the prices had been awarded in the beginning of November, the Dutch
Mathematical Olympiad concluded its 47th edition 2008. In 2011 we will have our 50th
edition.
The 31 most outstanding candidates of the Dutch Mathematical Olympiad 2008 were invited
to an intensive seven-month training programme, consisting 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.
In May, ten of them were invited to participate to the first Benelux Mathematical Olympiad,
held in Bergen op Zoom, the Netherlands (see page 8–12). We think this has been a very
stimulating experience for our students.
In June, 24 out of these 31 candidates were left. Out of those, the team was selected by a
final selection test on June 13, 2009 (see page 13–16). The team will have a training camp
from July 6 until July 13, together with the team from New Zealand.
In the meantime, a new edition of the Dutch Mathematical Olympiad had started. In October
2008, at the VU University Amsterdam we organized the first Junior Mathematical Olympiad
for the winners of the popular Kangaroo math contest (see page 17–20). The 30 best students
of grade 2, grade 3 and grade 4 were invited. We hope all of them will enjoy those problems
so much, that they will participate to the first round in the years that come.
The new first round took place at participating schools on January 30, 2009 (see page 21–24).
We are very happy to see that the number of schools as well as the number of participants
has increased again: 230 schools and 4379 participants, a new record since ages!
i
The Dutch team for IMO 2009 Bremen consists of
• Raymond van Bommel (17 y.o., participated in IMO 2007, 2008 as well)
We are grateful to Jinbi Jin for the translation into English of most of the problems and the solutions.
We also thank Lieneke Notenboom-Kronemeijer for her useful remarks concerning the formulation of
the First Round 2009 paper.
ii
First Round
SE
W
AND
IS
Dutch Mathematical Olympiad K
U
ERL
N
D
E
NED
Friday, January 25, 2008
PIADE
Problems OLYM
• Time available: 2 hours.
• The A-problems are multiple choice questions. Only one of the five options given is correct.
Please state clearly which letter precedes the correct solution. Each correct answer is worth 2
points.
• The B-problems are open questions; the answers to these questions are a number, or numbers.
Each correct answer is worth 5 points. Please work accurately, since an error in your calculations
may cause your solution to be considered incorrect and then you won’t√get points at all for
11
that question. Please give your answers exactly, for example 81 or 2 + 12 5 or 14 π + 1.
• You are not allowed to use calculators and formula sheets; you can only use a pen, a compass
and a ruler or set square. And your head, of course.
• This is a competition, not an exam. The main thing is that you have fun solving unusual
mathematical problems. Good luck!
A-problems
A1. Alex, Birgit, Cedric, Dion and Ersin all write their names on a sheet of paper, and they put
those five sheets into a large box. They each take one sheet out of the box at random. Now
it turns out that Birgit has Alex’ sheet, Cedric has Dion’s and Dion has Ersin’s. Also, Ersin
doesn’t have Cedric’s sheet. Whose sheet does Alex have?
A2. In a magic 3 × 3 square, the three row sums, the three column sums
and the two diagonal sums are all equal to each other. (A row sum 7
being the sum of the numbers on a certain row, etc.) In the magic
3 × 3 square shown here three numbers have already been filled in. ?
What number must be filled in instead of the question mark?
A) 2 B) 4 C) 6 D) 8 E) 9
10 3
A3. Calculating 6 × 5 × 4 × 3 × 2 × 1 yields 720. How many divisors does
720 have? (A divisor of an integer n is a positive integer by which n is
divisible. For example: the divisors of 6 are 1, 2, 3 and 6; the divisors
of 11 are 1 and 11.)
A) 6 B) 8 C) 20 D) 30 E) 36
D 5
A4. Of a quadrilateral ABCD, we know that |AB| = 3, |BC| = 4, |CD| = C
5, |DA| = 6 en ∠ABC = 90◦ . (|AB| stands for the length of segment
AB, etc.) What is the area of quadrilateral ABCD? 6
√ 4
A) 16 B) 18 C) 18 21 D) 20 E) 6 + 5 11
A 3 B
1
A5. How many five-digit numbers (like 12345 or 78000; the first digit must
be non-zero) are there that end on a 4 and that are divisible by 6?
D C
A6. We have a square ABCD with |AB| = 3. On AB, there is a point E
such that |AE| = 1 and |EB| = 2. AC and DE intersect in H. What
is the area of triangle CDH?
9 21 27
H
A) 8 B) 2 C) 8 D) 3 E) 8
A E B
A7. The seven blocks S E T T E E S are shuffled. For example, you can
get E E E S S T T or T E S E T E S .
How many different “words” of length 7 can we get this way? (Any
combination of the 7 letters counts as word.)
A) 4 B) 5 C) 6 D) 7 E) 8
B-problems
B1. We number both the rows and the columns of an 8×8 chessboard with
the numbers 1 to 8. A number of grains is placed onto each square, in
such a way that the number of grains on a certain square equals the
product of its row and column numbers. How many grains are there
on the entire chessboard?
B2. We take 50 distinct integers from the set {1, 2, 3, . . . , 100}, such that
their sum equals 2900. What is the minimal number of even integers
amongst these 50 numbers?
A B
2
First Round
SE
W
AND
IS
Dutch Mathematical Olympiad K
U
ERL
N
D
E
NED
Friday, January 25, 2008
PIADE
Solutions OLYM
A1. C) Cedric When we put the data into a table, we see that Birgit’s
and Cedric’s sheets haven’t been picked yet. Since Ersin didn’t pick A B C D E
Cedric’s sheet, he must have picked Birgit’s. So Alex must have picked ? A D E ?
Cedric’s sheet.
A5. C) 3000 If x is a positive multiple of 6 that ends with a 4, then the next multiples of
6 end with a 0 (x + 6), a 6 (x + 12), a 2 (x + 18), an 8 (x + 24), a 4 (x + 30), so the next
multiple of 6 that ends with a 4 is x+30. So any 30 consecutive positive integers must contain
exactly one integer with the desired properties. How many such integers lie between 10000
and 99999? Since we have 90000 consecutive positive integers, we find 90000 ÷ 30 = 3000
such integers amongst them.
D Q C
27
A6. E) 8 Draw a line through H parallel to AD, and let P and Q
be the intersections of this line with AB and CD, respectively. Now
we have |HP | : |HQ| = |AE| : |CD| = 1 : 3, so |HQ| = 34 × |P Q| =
3 9 1 9 27 H
4 × 3 = 4 . So the area of triangle CDH equals 2 × 3 × 4 = 8 .
A PE B
3
S S . . . . .
We have 6 + 5 + 4 + 3 + 2 + 1 = 21 (or 72 ) possibilities
A7. A) 210 S . S . . . .
to arrange the S-blocks onto the 7 places (see figure). For each choice, S . . S . . .
we have 4 + 3 + 2 + 1 = 10 (or 52 ) possibilities to arrange the T-blocks ..
.
on the remaining 5 places; after which the positions of the E-blocks . S S . . . .
are determined. So we have 21 × 10 = 210 possibilities.
. S . S . . .
Alternative solution: If all the blocks were different, we would have got
. S . . S . .
7! possibilities. But the three E-blocks aren’t different, so we end up ..
counting each word 3! times this way. Similarly for the two S-blocks .
7!
and the two T-blocks. So we find 3!×2!×2! = 7×6×5×4×6 3!
= 7×6×5 = . . S S . . .
63!×4
210 different words. . . S . S . .
..
.
A8. B) 5 This equation is equivalent to
(x2 − 2)2 − 5 = 1 or (x2 − 2)2 − 5 = −1.
√ √
The first is equivalent to x2 − √
2 = 6 of x2 − 2 = − 6, with 2 and 0
solutions respectively (since − 6 + 2 < 0).
The latter is equivalent to x2 − 2 = 2 of x2 − 2 = −2, with 2 and 1
solution(s) respectively. So we have 2 + 0 + 2 + 1 = 5 solutions in total.
B1. 1296 In the first column, we have, successively, 1 × 1, 1 × 2, 1 × 3, . . . , 1 × 8 grains.
So, in the first column: 1 × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
In the second column: 2 × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
In the third column is: 3 × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
..
.
Finally, in the eighth column: 8 × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
So, in total: (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8) × (1 + 2 + 3 + 4 + 5 + 6 + 7 + 8).
Since 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 = 21 × 8 × (1 + 8) = 36, there are 362 = 1296 grains on the
board.
B2. 6 The 50 odd integers from the set {1, 2, 3, . . . , 100} sum up to 12 × 50 × (1 + 99) =
2500, which is still 400 short of 2900. Now exchange the smallest odd integers for the largest
even integers, in pairs, since 400 is even. First exchanging 1 and 3 for 100 and 98 makes the
sum equal to 2694. The next step gives us 2694 − 5 − 7 + 96 + 94 = 2872. Which is still less
than 2900, so we require another exchange. Now exchanging 9 and 11 for 20 and 28 works,
making the sum 2900 with 6 even integers, showing along the way that that is the minimal
number of even integers we need to do so.
1
B3. 110 We know that x must satisfy x + = 5, so x2 − 5x + 1 =
√x
5± 21
0, from which follows that x = x1,2 = 2 . Note that x1 x2 = 1.
√ 3 √ 3
Now we have x3 + x−3 = x1 + x32
3 = 1
85 + 21 + 5 − 21 =
1
√ √ 2 √ 3 3 √ √ 2 √ 3
8 53 + 3 · 52 · 21 + 3 · 5 · 21 + 21 + 5 − 3 · 52 · 21 + 3 · 5 · 21 − 21 =
2 3
8 5 + 3 · 5 · 21 = 110.
Alternative solution: From (x + x1 )3 = x3 + 3x2 ( x1 ) + 3x( x1 )2 + ( x1 )3 = x3 + 3x + 3
x + 1
x3
we
can deduce that x3 + x13 = (x + x1 )3 − 3(x + x1 ) = 53 − 3 × 5 = 110.
S
√ D C
B4. 2 3 Let Q, R, S, T be the orthogonal projections of P on
AB, BC, CD, DA, respectively. Then we have P
|AQ|2 + |QP |2 = 36 and |BQ|2 + |SP |2 = 25 (since |BQ| = |CS|), T R
so |AQ|2 + |QP |2 + |BQ|2 + |SP |2 = 61. We also have |BQ|2 + |QP |2 =
49. √
2 2 2 2
So√ |DS| + |SP | = |AQ| + |SP | = 61 − 49 = 12 and |DP | = 12 =
2 3. A B
Q
4
Second Round
SE
W
AND
Dutch Mathematical Olympiad IS
K
U
ERL
N
D
Friday, September 12, 2008 E
NED
Eindhoven University of Technology
PIADE
OLYM
Problems
• Available time: 3 hours.
• Writing down just the answer itself is not sufficient; you also need to describe the way you
solved the problem.
• Usage of calculators and formula sheets are not allowed; you are only allowed to use a pen, a
compass and a ruler or set square. And your common sense of course.
• Please write the solutions of each problem on a different sheet of paper. Good luck!
1. Suppose we have a square ABCD and a point S in the interior of this square. A D
A D
Under homothety with centre S and ratio of magnification k > 1, this square
becomes another square A0 B 0 C 0 D0 . S
Prove that the sum of the areas of the two quadrilaterals A0 ABB 0 and
C 0 CDD0 are equal to the sum of the areas of the two quadrilaterals B 0 BCC 0 B C
and D0 DAA0 . B C
3 · 2n + 1 = m2 .
3. Suppose that we have a set S of 756 arbitrary integers between 1 and 2008
(1 and 2008 included).
Prove that there are two distinct integers a and b in S such that their sum
a + b is divisible by 8.
C2
4. Three circles C1 , C2 , C3 , with radii 1, 2, 3 respectively, are externally tangent. C1
In the area enclosed by these circles, there is a circle C4 which is externally
tangent to all three circles.
Find the radius of C4 .
C3
5
Second Round
SE
W
AND
Dutch Mathematical Olympiad IS
K
U
ERL
N
D
Friday, September 12, 2008 E
NED
Eindhoven University of Technology
PIADE
OLYM
Solutions
2. We can rewrite the equation as 3 · 2n = (m − 1)(m + 1). Since n > 0, we see that 3 · 2n is an even
number, so at least one of m − 1, m + 1 is even as well. Hence they’re both even. As the factors
m − 1 and m + 1 differ by 2, they can’t both contain multiple factors of 2. Hence one of these
factors contains exactly one factor 2. This factor either contains the factor 3 as well, or it doesn’t.
So it must be equal to either 2 or 6. The other factor then must differ from this one by exactly 2.
If this factor is equal to 2, then the other one has to be either 0 or 4. Since none of these contain
a factor 3, none of these solve the equation. If this factor is equal to 6, then the other one has to
be either 4 or 8, both of which yield a solution, as they are powers of 2. Hence there are exactly
two solutions, given by (m, n) = (5, 3) and (m, n) = (7, 4).
3. We divide the integers 1 up to 2008 amongst eight distinct subsets V1 to V8 , where Vi is the subset
consisting of the 251 integers of the form 8k + i with 0 ≤ k ≤ 250. So we have
6
4. Let A, B, C be the centres of the circles C1 , C2 , C3 respectively. The triangle
formed by these three centres, has sides |AB| = 3, |AC| = 4 and |BC| = 5,
hence it is a right-angled triangle. Now choose the x- and the y-axis in such C
a way that A = (0, 0), B = (3, 0) and C = (0, 4). Let r and M = (x, y) be
the radius and centre of C4 , respectively. Then we see that |AM | = r + 1,
y
|BM | = r + 2 and |CM | = r + 3. This implies the three following equations A x B
with three unknowns.
r2 + 2r + 1 = |AM |2 = x2 + y 2 (1)
2 2 2 2 2 2
r + 4r + 4 = |BM | = (3 − x) + y = x − 6x + 9 + y (2)
2 2 2 2 2 2
r + 6r + 9 = |CM | = x + (4 − y) = x + y − 8y + 16 (3)
3−r
Taking the difference of (1) and (2), we see that 6x − 9 = −2r − 3, so 6x = 6 − 2r, hence x = 3 .
2−r
Taking the difference of (1) and (3), we see that 8y − 16 = −4r − 8, so 8y = 8 − 4r, hence y = 2 .
2 2 2 2
Substituting this in (1) yields r2 + 2r + 1 = x2 + y 2 = (3−r)
9 + (2−r)
4 = 9−6r+r
9 + 4−4r+r
4 , thus
23 2 11 r
2 r
r 2
36 r + 3 r −1 = 23 6 +22 6 −1 = 0. Substituting p = 6 , we then see that 0 = 23p +22p−1 =
(23p − 1)(p + 1). This gives the two possible solutions 6r = p = −1 and 23 · 6r = 23p = 1. But
6
since r has to be positive, we see that r has to be equal to 23 . (We can deduce from this that
21 20
M = (x, y) = ( 23 , 23 ).)
5. Consider an arbitrary sequence n1 , n2 , . . . , n2008 and an arbitrary sequence of moves. The first
integer n1 is reduced by 1 every time we pick b = n2 , since n1 is one of the neighbours of n2 .
If we pick b = nk , where k > 2, then we see that n1 remains unchanged. Hence we can pick
b = n2 at most n1 times. So in our sequence of moves, either there is a last time that the
move b = n2 occurs, or that move doesn’t occur at all. In the former case, we consider the
sequence of moves following that move, and in the latter case, we simply consider the sequence
of all moves. In this new sequence of moves, we never pick b = n2 . So all we do in this new
sequence, is picking b = n3 up to b = n2007 . Let n2 now be the value of the second integer at
the beginning of this new sequence (this integer need not be the same as before, since every
time we’ve picked b = n2 until then, it is increased by 7, and every time we’ve picked b = n3
until then, it is reduced by 1). In the new sequence, n2 is reduced by 1 every time we pick
b = n3 , and remains unchanged if we pick any other integer (i.e. b = n4 up to b = n2007 ).
Hence we can conclude that in the new sequence, either there is a last time that the move
b = n3 occurs, or this move doesn’t occur at all. So from that point onward, we only pick the
integers b = n4 up to b = n2007 . We can repeat this argument to see that, from some point
onward, only b = n2007 is picked, and that there either is a last time that the move b = n2007 oc-
curs, or this move doesn’t occur at all, indicating that our sequence of moves ends at some point.
Variant Suppose that there exists an infinite sequence of moves. By the pigeonhole
principle, there is a move b = nk that occurs infinitely often. Now let k be the smallest
integer for which the move b = nk occurs infinitely often. Then we see that the integer
b = nk−1 is only picked finitely often, hence is increased only finitely often. But this integer
is decreased infinitely often; namely for every time we pick b = nk . This yields our contradiction.
7
1st BENELUX MATHEMATICAL OLYMPIAD
Bergen op Zoom (Netherlands)
May 9, 2009
Language: English
Problem 1. Find all functions f : Z>0 → Z>0 that satisfy the following two conditions:
Problem 2. Let n be a positive integer and let k be an odd positive integer. Moreover, let
a, b and c be integers (not necessarily positive) satisfying the equations
an + kb = bn + kc = cn + ka.
Prove that a = b = c.
Problem 3. Let n ≥ 1 be an integer. In town X there are n girls and n boys, and each girl
knows each boy. In town Y there are n girls, g1 , g2 , . . . , gn , and 2n − 1 boys, b1 , b2 , . . . ,
b2n−1 . For i = 1, 2, . . . , n, girl gi knows boys b1 , b2 , . . . , b2i−1 and no other boys. Let r be an
integer with 1 ≤ r ≤ n. In each of the towns a party will be held where r girls from that town
and r boys from the same town are supposed to dance with each other in r dancing pairs.
However, every girl only wants to dance with a boy she knows. Denote by X(r) the number
of ways in which we can choose r dancing pairs from town X, and by Y (r) the number of
ways in which we can choose r dancing pairs from town Y .
Prove that X(r) = Y (r) for r = 1, 2, . . . , n.
Problem 4. Given trapezoid ABCD with parallel sides AB and CD, let E be a point on
line BC outside segment BC, such that segment AE intersects segment CD. Assume that
there exists a point F inside segment AD such that ∠EAD = ∠CBF . Denote by I the point
of intersection of CD and EF , and by J the point of intersection of AB and EF . Let K be
the midpoint of segment EF , and assume that K is different from I and J.
Prove that K belongs to the circumcircle of 4ABI if and only if K belongs to the circumcircle
of 4CDJ.
We are grateful to the authors of these problems for making them available for the Benelux Olympiad. We have
agreed upon not publishing them on any forum or by any other means before August 1st, 2009. Participating
to the BxMO entails that you respect this agreement.
8
1st BENELUX MATHEMATICAL OLYMPIAD
Bergen op Zoom (Netherlands)
May 9, 2009
Language: English
Solutions
Problem 1.
Solution 1. Let a be a positive integer satisfying f (1) = a2 . We will prove that f (n) =
na2 + n(n − 1) for all n by induction on n. For n = 1 it follows from the definition of a. Now
suppose we have f (n) = na2 + n(n − 1) for a certain positive integer n. Then by the second
condition with m = 1, we have
Suppose a > 1 and let p be a prime divisor of a. We know that f (p) = pa2 + p(p − 1) is a
square, and as it is obviously divisible by p, it must be divisible by p2 . Hence a2 + p − 1 is
divisible by p. But this is a contradiction, as a and p are both divisible by p.
We conclude a = 1 and f (n) = n2 for all n ∈ Z>0 . This function indeed satisfies the condi-
tions.
Comment. Proving that a = 1 can be done in various ways. Here is another possibility. We
have f (n) = na2 + n(n − 1) = n(a2 − 1) + n2 for all n > 0. Suppose a > 1, then n = a2 − 1 > 0
and f (a2 − 1) = 2(a2 − 1)2 . As all prime factors in (a2 − 1)2 occur an even number of times,
the prime factor 2 occurs an odd number of times in f (a2 − 1) = 2(a2 − 1)2 . Hence this is
not a square, which contradicts the first condition.
Solution 2. Define g(n) = f (n)−n2 for all n ∈ Z>0 . Then we can rewrite the second condition
as
g(m + n) + (m + n)2 = g(m) + m2 + g(n) + n2 + 2mn,
hence g(n) satisfies the functional equation
Let b = g(1) = f (1) − 1 ≥ 0. By setting n = 1 in (4) we find g(m + 1) = g(m) + b, and hence
by induction we have g(n) = nb for all n. Therefore f (n) = nb + n2 for all n. Suppose b > 0,
then taking n = b yields f (b) = 2b2 , which is not a square, contradicting the first condition.
We conclude b = 0 and f (n) = n2 for all n ∈ Z>0 . This function indeed satisfies the condi-
tions.
9
Problem 2.
From an + kb = bn + kc we find an − bn = k(c − b). In a similar way we find two more such
equations, so we have
an − bn bn − cn cn − an an − bn bn − cn cn − an
· · = · · = −k 3 . (6)
a−b b−c c−a b−c c−a a−b
Note that the left-hand side is the product of three integers.
Consider a, b and c modulo 2. According to the box principle, at least two of those are
n −bn
congruent, say a ≡ b mod 2. The integer a a−b = an−1 + an−2 b + · · · + abn−2 + bn−1 is the
sum of n terms, and as a ≡ b mod 2, all of these terms are even, or all of these terms are
odd. In any case, the sum of the n terms is even. So the left-hand side of (6) is even, while
the right-hand side is odd. This is a contradiction, which finishes the proof.
First consider the case that n is odd. Then xn is a monotonically increasing function of x. So
if a < b, then an < bn , and so on. So suppose a > b, then we have an > bn , hence an − bn > 0.
Hence by (5) we have c − b > 0. So bn − cn < 0 and therefore a − c < 0. From this we have
cn − an > 0, so b − a > 0, which contradicts the assumption. Similarly we get a contradiction
if a < b.
Now assume n is even. Consider the equalities in (5) modulo 2. As k is odd, the first
n −an
equality gives a − b ≡ c − b mod 2, hence a ≡ c mod 2. Now the integer c c−a =
cn−1 + cn−2 a + · · · + can−2 + an−1 is the sum of n terms, and as a ≡ c mod 2, all of these
terms are even, or all of these terms are odd. In any case, the sum of the n terms is even. Let
i be the exponent of the prime factor 2 in c − a. Then the exponent of the prime factor 2 in
cn − an is at least i+ 1. From the third equality in (5) we derive that the exponent of 2 in b − a
is at least i + 1 as well. Similarly to the previous argument, the exponent of 2 in an − bn must
be at least i + 2, and the same holds for the exponent of 2 in c − b. Finally the exponent of
2 in bn −cn and hence also in a−c must be at least i+3. This contradicts the definition of i.
10
Problem 3.
As X(r) and Y (r) are dependent on n, we will from now on denote them by Xn (r) and Yn (r).
There are nr ways to pick r girls from town X, and nr ways to pick r boys from town X,
and r! ways to make r pairs of these boys and girls. As all girls in town X know all boys in
town X, each girl is then with a boy she knows. Hence
2
n (n!)2
Xn (r) = · r! = .
r r!((n − r)!)2
Let An (r) be the number of different ways in which r girls from town Y can dance with r
boys from town Y , forming r pairs, each girl with a boy she knows, such that gn is one of the
girls in the pairs. Let Bn (r) = Yn (r) − An (r).
If girl gn is not in one of the pairs, then boys b2n−2 and b2n−1 are not in one of the pairs
either. So for n ≥ 2 and r ≤ n − 1 we have Bn (r) = Yn−1 (r).
On the other hand, if girl gn is in one of the pairs, then we can delete that pair to find a way
in which r −1 girls from a town with n−1 girls can dance with r −1 boys from the same town.
Given such a set of r − 1 pairs, we can extend this set to a set of r pairs from a town with
n girls by adding the pair (gn , bi ) for some i. For bi we can choose from {b1 , b2 , . . . , b2n−1 }
except that the r − 1 boys that are already in one of the pairs are not allowed. So there are
(2n − 1) − (r − 1) = 2n − r possibilities for bi . We conclude that for n ≥ 2 and r ≥ 2 we have
An (r) = (2n − r)Yn−1 (r − 1).
(n!)2
Yn (r) = , (7)
r!((n − r)!)2
For n = 1 we just need to prove that Y1 (1) = 1. As there is only one girl and one boy, this is
trivial. Now let k ≥ 1 and suppose (7) is true for n = k. Then for r = 2, 3, . . . , k we have
Yk+1 (1) = Ak+1 (1) + Bk+1 (1) = (2k + 1) + Yk (1) = (2k + 1) + k 2 = (k + 1)2 .
11
Finally, for r = k + 1 we have
(k!)2 ((k + 1)!)2
Yk+1 (k + 1) = Ak+1 (k + 1) = (k + 1)Yk (k) = (k + 1) = .
k! (k + 1)!
This completes the induction.
Problem 4.
We use signed distances throughout the proof. Assume that B is inside segment CE; in the
other case a similar proof can be used. We have
∠F AE = ∠DAE = ∠CBF = ∠EBF.
As B and F are on different sides of the line AE, we conclude that ABEF is a cyclic
quadrilateral. Hence JA · JB = JE · JF . Furthermore, K belongs to the circumcircle of
ABI if and only if JA · JB = JK · JI. Therefore K belongs to the circumcircle of ABI
if and only if JE · JF = JK · JI. Expressing JI = JF + F I, JE = JF + F E and
JK = 12 (JE + JF ) = 12 F E + JF , we find that K belongs to the circumcircle of ABI if and
only if
(JF + F E)JF = ( 21 F E + JF )(JF + F I) ⇔
1
2 · F E · JF = ( 21 F E
+ JF )F I ⇔
FE · FI
JF = .
F E − 2F I
Since ABEF is cyclic and AB is parallel to CD, we have
∠F EC = ∠F EB = 180◦ − ∠F AB = ∠F DC.
Hence CEDF is cyclic as well, yielding IC · ID = IE · IF . Furthermore, K belongs to the
circumcircle of CDJ if and only if IC · ID = IK · IJ. Therefore K belongs to the circumcircle
of CDJ if and only if IE · IF = IK · IJ. Expressing IJ = IF + F J, IE = IF + F E and
IK = 12 (IE + IF ) = 12 F E + IF , we find that K belongs to the circumcircle of CDJ if and
only if
(IF + F E)IF = ( 12 F E + IF )(IF + F J) ⇔
1
2 · F E · IF = ( 21 F E
+ IF )F J ⇔
F E · IF
FJ = ⇔
F E + 2IF
FE · FI
JF = .
F E − 2F I
We conclude that K belongs to the circumcircle of ABI if and only if K belongs to the cir-
cumcircle of CDJ.
Comment 1. After deriving that what needs to be proved is: JE · JF = JK · JI if and only if
IE · IF = IK · IJ, the solution comes down to eliminating all but three well-chosen distances
and then manipulating the equalities until it is clear that they are equivalent. The above
solution is just one way of doing this.
12
Team Selection Test
Problem 1. Let n ≥ 10 be an integer, and consider n in base 10. Let S(n) be the sum of
the digits of n. A stump of n is a positive integer obtained by removing a number of digits
(at least one, but not all) from the right side of n. E.g.: 23 is a stump of 2351. Let T (n) be
the sum of all the stumps of n.
Prove that n = S(n) + 9 · T (n).
Problem 2. Let ABC be a triangle, P the midpoint of BC, and Q a point on segment CA
such that |CQ| = 2|QA|. Let S be the intersection of BQ and AP . Prove that |AS| = |SP |.
Problem 3. Let a, b and c be positive reals such that a + b + c ≥ abc. Prove that
√
a2 + b2 + c2 ≥ 3 abc.
for all m, n ∈ Z.
Problem 5. Suppose that we are given an n-gon of which all sides have the same length,
and of which all the vertices have rational coordinates. Prove that n is even.
13
Solutions of the Team Selection Test 2009
Problem 1.
n = a0 + 10a1 + · · · + 10k ak .
j
k X k k
X X X 10j − 1
10j−i aj = 1 + 10 + · · · + 10j−1 aj =
= aj ,
10 − 1
j=1 i=1 j=1 j=1
where we used the summation formula for the geometric series in the final step. So we get
k
X k
X
9 · T (n) = (10j − 1)aj = (10j − 1)aj .
j=1 j=0
Pk
Recall that S(n) = j=0 aj . Hence we have
k
X k
X
S(n) + 9 · T (n) = (10j − 1 + 1)aj = 10j aj = n.
j=0 j=0
Problem 2.
Solution 1. Let T be a point on BQ such that P T is parallel to AC. Then P T joins the
midpoints of BC and BQ, so |P T | = 21 |CQ| = |QA|. So we see that AT P Q is a quadrilateral
of which two sides are parallel and of the same length. This implies that it is a parallelogram.
Since any parallelogram has the property that its diagonals bisect each other, we see that
|AS| = |SP |.
14
Solution 2. We apply Menelaos’ Theorem to triangle P CA. Since the points B, S and Q are
collinear, we have
P B CQ AS −1 2 AS AS
−1 = · · = · · =− .
BC QA SP 2 1 SP SP
AS
Hence SP = 1, from which we can deduce that S is the midpoint of the segment AP .
Solution 4. Let R be the intersection of CS and AB. According to Ceva’s Theorem, we have
AR BP CQ
· · = 1,
RB P C QA
from which we deduce that 2|AR| = |RB|. Let a = [P SC], b = [QSC], c = [QSA], d = [RSA],
e = [RSB] and f = [P SB]. Then we see that
b + a + f = 2(c + d + e) en a + e + f = 2(b + c + d),
which implies that
b − e = 2e − 2b,
hence b = e. Furthermore, we have 2d = e and 2c = b, hence c = d and c + d = e. Now
a + e + f = 2(b + c + d) implies that
a + f = b + 2c + 2d = b + c + d + e,
so
2(a + f ) = a + b + c + d + e + f = [ABC].
Hence 2|P S| = |P A|, implying that |P S| = |AS|.
Problem 3.
First of all, note that a2 + b2 + c2 ≥ ab + bc + ca, and hence that 3(a2 + b2 + c2 ) ≥ (a + b + c)2 .
1
Applying AM-GM we see that a+b+c ≥ 3(abc) 3 . On the other hand, note that a+b+c ≥ abc.
Now we have the following two inequalities:
2
a2 + b2 + c2 ≥ 31 (a + b + c)2 ≥ 3(abc) 3 ,
a2 + b2 + c2 ≥ 13 (a + b + c)2 ≥ 13 (abc)2 .
Raising the former equation to the 34 -th power, and the latter one to the 1
4 -th (which is
allowed, since every expression is positive):
3 3 1
(a2 + b2 + c2 ) 4 ≥ 3 4 (abc) 2 ,
1 1 1
(a2 + b2 + c2 ) 4 ≥ 3− 4 (abc) 2 .
Taking the product of these two equations, we then see that
1
a2 + b2 + c2 ≥ 3 2 (abc),
which is as desired.
15
Problem 4.
Suppose that there exists a c ∈ Z such that f (n) = c for all n. Then 2c = c2 + 2, so we
get the equation c2 − 2c + 2 = 0, which has no solutions in Z. Hence f cannot be constant.
Now substitute m = 0. This yields the equation f (n) + f (−1) = f (n)f (0) + 2, from which
we deduce that f (n)(1 − f (0)) is constant. Since f (n) isn’t a constant function, we see that
f (0) = 1. Using the same equation, we then get f (−1) = 2. Now substitute m = n = −1.
This yields f (−2) + f (0) = f (−1)2 + 2, from which follows that f (−2) = 5. Substituting
m = 1 and n = −1 now yields f (0) + f (−2) = f (1)f (−1) + 2, which implies that f (1) = 2.
f (n + 1) = 2f (n) + 2 − f (n − 1).
By induction, it then follows that f (n) = n2 +1 for all non-negative n, and, using the equation
f (n − 1) = 2f (n) + 2 − f (n + 1)
also for all non-positive n. Hence f (n) = n2 + 1 for all n, and we can easily check that this
function satisfies the given equation.
Problem 5.
Let (x1 , y1 ), . . . , (xn , yn ) be the coordinates of the vertices of the given n-gon. Define ai =
xi+1 − xi , bi = yi+1 P − yi for i = P1, 2, . . . , n, where xn+1 = x1 and yn+1 = y1 . Then we note
that ai , bi ∈ Q and ni=1 ai = ni=1 bi = 0; and that the sum a2i + b2i does not depend on i.
By multiplying with a suitable factor we can get rid of denominators and common divisors of
the ai and bi , so we may assume that ai , bi ∈ Z and that gcd(a1 , . . . , an , b1 , . . . , bn ) = 1. Let
c be the integer such that a2i + b2i = c for all i.
Now suppose that c is even. Then, for all i, we have ai ≡ bi mod 2. If there exists an i
such that ai and bi are both odd, then c = a2i + b2i ≡ 1 + 1 ≡ 2 mod 4. If there exists an
i such that ai and bi are both even, then c = a2i + b2i ≡ 0 + 0 ≡ 0 mod 4. These two cases
cannot occur together. Hence either all of the ai and bi are odd, or all of them are even.
The latter statement contradicts our assumption about the greatest common divisor of the ai
and bi . The former statement yields 0 = ni=1 ai ≡ n mod 2, implying again that n is even.
P
16
Junior Mathematical Olympiad October 3, 2008
Vrije Universiteit Amsterdam
Problems part 1
• The problems in part 1 are five-choice questions. At each question exactly one of the given
five answers is correct. Indicate clearly on the answer sheet which letter corresponds to the
right answer.
• Each correctly given answer will get you 2 points. For wrong answers no points are sub-
tracted.
• You are allowed scrap paper, as well as a compass and a ruler or protractor. Calculators and
other electronic devices are not allowed.
• The time allowed for this part is 60 minutes.
• Good luck!
1. A rectangle has been divided into four smaller rectangles. The areas of
three of the small rectangles are 6, 8 and 9 (see the figure). Determine 6 9
the area of the fourth small rectangle.
? 8
A) 4 32 B) 5 C) 5 31 D) 5 23 E) 7
2. We multiply all odd numbers between 0 and 100. What is the last digit of the result?
A) 1 B) 3 C) 5 D) 7 E) 9
3. Six students are sitting next to each other on six chairs, numbered from 1 to 6. Now they all
get up at the same time, and then they sit down on a chair again, according to the table below.
A) 1 B) 2 C) 3 D) 4 E) 6
A) 0 B) 1 C) 2 D) 3 E) 4
D
5. In a quadrilateral ABCD the sides AB and CD have equal length.
A 65
Moreover, three angles are given: ∠A1 = 65◦ , ∠A2 = 80◦ and ∠B =
1
2
80 ?
50◦ . Determine ∠C2 . 2
1
C
A) 30◦ B) 40◦ C) 50◦ D) 60◦ E) 65◦
50
B
12
6. How many (positive or negative) integers n exist such that n+5 is an integer?
A) 2 B) 6 C) 8 D) 10 E) 12
17
7. What is
1 1 1 1 1 1
1− × 1− × 1− × 1− × 1− × 1− ?
4 9 16 25 36 49
1 2 3 4 5
A) 7 B) 7 C) 7 D) 7 E) 7
8. How many integers of the form bbcac exist, where a, b and c are digits (0, 1, . . . , 9) with c > a
and with b equal to the mean of a and c?
A) 20 B) 21 C) 22 D) 23 E) 24
9. Two identical rhombuses are lying on top of each other, one of them
rotated 90 degrees compared to the other. The area where the two
rhombuses overlap, happens to be a regular octagon: all eight sides
have equal length and all eight angles have equal sizes. Determine the α
10. How many of the integers 1, 2, 3, . . . , 2008 are not divisible by 2 and not divisible by 5 either?
11. A large pond is being emptied by means of three pumps. Using just the first pump, emptying
the pond would take four days. Using just the second pump, it would take three days, and
using just the third pump, it would take two days. How many days does it take to empty the
pond using all three pumps at the same time?
1 12 13
A) 9 day B) 13 day C) 1 day D) 12 day E) 3 days
B
12. Through the vertices A, B and C of a triangle ABC passes a circle 2
1
with midpoint M . Of the three angles at the midpoint, two are equal 139 A
123
to 123◦ and 139◦ (see figure). Determine ∠B12 . M
13. If we divide the number 22008 − 22007 + 22006 − 22005 + 22004 − 22003 + 22002 − 22001 by the
number 22000 , then the result is an integer (that is, there is no remainder). Determine this
integer.
A) 4 B) 36 C) 72 D) 170 E) 200
14. Consider an integer abcd consisting of four distinct digits a, b, c and d (where a is not allowed
to be 0). Of this integer you know that abcd × 11 = ac9bd. The digit b can be equal to only
one of the following five options. Which one?
A) 2 B) 3 C) 4 D) 5 E) 6
A) 5 B) 8 C) 10 D) 11 E) 12
18
Junior Mathematical Olympiad October 3, 2008
Vrije Universiteit Amsterdam
Problems part 2
• The problems in part 2 are open questions. At each question your answer should be a
number or an expression (for example 2 23 or a2 ). Write this answer on your answer sheet
at the indicated place.
• Each correctly given answer will get you 2 points. For wrong answers no points are sub-
tracted.
• You are allowed scrap paper, as well as a compass and a ruler or a protractor. Calculators
and other electronic devices are not allowed.
• Good luck!
1. What is the maximum number of points of intersection between a circle and a triangle?
2. During a tournament with six players, each player plays a match against each other player.
At each match there is a winner; ties do not occur. A journalist asks five of the six players
how many matches each of them has won. The answers given are 4, 3, 2, 2 and 2. How many
matches have been won by the sixth player?
4. I have written four numbers on a piece of paper. In six different ways I can choose two and
add them up. The resulting sums are 11, 15, 16, 16, 17 and 21. Now I multiply all of the four
numbers on my piece of paper. What is the result?
5. What is the sum of the positive integers smaller than 100,000 (a hundred thousand) in which
only the digits 0 and 1 occur?
19
6. Find the smallest positive integer with the property that if you multiply its digits, the result
is 1890.
1
2
3
4
5
7. In this stretched chess board the small rectangles have sides with 6
12 3 4 5 6 7 8
8. Write the following expression as one fraction in its simplest terms, given that for a, b and c
we have a × b × c = 1.
1 1 1
+ +
1 + a + (a × b) 1 + b + (b × c) 1 + c + (c × a)
9. The two smallest circles in the figure have radiuses 2 and 3. Find the
radius of the largest circle.
3
2008
10. A hexagon has six angles all equal to 120 degrees. The lengths of 2006 2009
four consecutive sides are 2000, 2006, 2008 and 2009. Determine the
perimeter of the hexagon.
2000
www.wiskundeolympiade.nl/junior
20
First Round
SE
W
AND
IS
Dutch Mathematical Olympiad K
U
ERL
N
D
E
NED
Friday, January 30, 2009
PIADE
Problems OLYM
• Time available: 2 hours.
• The A-problems are multiple choice questions. Only one of the five options given is correct.
Please state clearly which letter precedes the correct solution. Each correct answer is worth 2
points.
• The B-problems are open questions; the answers to these questions are a number, or numbers.
Each correct answer is worth 5 points. Please work accurately, since an error in your calculations
may cause your solution to be considered incorrect and then you won’t√get points at all for
11
that question. Please give your answers exactly, for example 81 or 2 + 12 5 or 14 π + 1.
• You are not allowed to use calculators and formula sheets; you can only use a pen, a compass
and a ruler or set square. And your head, of course.
• This is a competition, not an exam. The main thing is that you have fun solving unusual
mathematical problems. Good luck!
A-problems
A1. Ella has answered three sets of questions. Of the first set, consisting of 25
questions, she answered 60% correctly. Of the second set, consisting of 30
questions, she answered 70% correctly, and of the third set, consisting of 45
questions, she answered 80% correctly. Now, if we combine the three sets to
form one set of 100 questions, what percentage of these 100 questions did
Ella answer correctly?
(A) 68% (B) 70% (C) 72% (D) 74% (E) 76%
A2. How many of the integers from 10 to 99 (10 and 99 included) have the
property that the sum of their digits is equal to the square of an integer?
(An example: The sum of the digits of 27 is equal to 2 + 7 = 9 = 32 .)
(A) 13 (B) 14 (C) 15 (D) 16 (E) 17
A3. Ronald rolls three dice. These dice look like normal dice, but the numbers
on their sides are unusual.
On the sides of the first dice are the numbers 1, 1, 2, 2, 3, 3.
On the sides of the second dice are the numbers 2, 2, 4, 4, 6, 6.
On the sides of the third dice are the numbers 1, 1, 3, 3, 5, 5.
He then adds up the three numbers he gets from rolling the three dice.
What is the probability that the resulting number is odd?
1 1 1 2 3
(A) 4 (B) 3 (C) 2 (D) 3 (E) 4
A4. Three distinct numbers from the set {1, 2, 3, 4, 5, 6, 7, 8, 9} are placed in the
three squares at the top of the figure to the right, after which the numbers
are added as indicated in said figure. We call Max the highest number that + +
can appear in the bottom square, and Min the lowest number that can appear
there. What is the value of Max − Min? +
(A) 16 (B) 24 (C) 25 (D) 26 (E) 32
21
A5. The ratio between the lengths of the diagonals of a rhombus is 3 to 4. (A
rhombus is an equilateral quadrilateral.) The sum of the lengths of the
diagonals is 56. What is the perimeter of this rhombus?
(A) 80 (B) 96 (C) 100 (D) 108 (E) 160
A6. Wouter walks from his home to the fitness center. He could also have chosen
to go by bike, in which case he would have covered the distance between his
home and the fitness center 7 times as fast. However, he has left his bike at
home. After having walked 1 km he reaches a bridge. Continuing on foot
will take Wouter just as long as walking back home to get his bike and then
cycle to the fitness center. What is the distance between the bridge and the
fitness center in kilometers?
(A) 87 (B) 67 (C) 65 (D) 54 (E) 43
A7. On the sides of an equilateral triangle, we draw three squares. The sides of
these squares that are parallel to the sides of the triangle are extended until
they intersect. These three intersections form another equilateral triangle.
Suppose that the length of a side of the original triangle is equal to 1. What
is the length of a side of the large equilateral triangle? 1 1
√ √ √ √ √
(A) 1 + 2 2 (B) 5 − 12 3 (C) 3 2 (D) 1 + 2 3 (E) 2 6 1
A8. Consider all four-digit numbers in which each of the digits 3, 4, 6 and 7
occurs exactly once. How many of these numbers are divisible by 44?
(A) 2 (B) 4 (C) 6 (D) 8 (E) 12
B-problems
B1. A sheet of paper shows a grid of 101 by 101 white squares. A chain is formed
by coloring squares grey as shown in the figure to the right. The chain starts
in the upper left-hand corner and goes on until it cannot go on any further.
Only part of the grid is shown. In total, how many squares are colored grey
in the original grid of 101 by 101 squares?
B3. Using a wide brush, we paint the diagonals of a square tile, as in the figure. 1
Exactly half of the surface of this tile is covered with paint. Given that the
width of the brush is 1, as indicated in the figure, what is the length of the
side of the tile?
22
First Round
SE
W
AND
IS
Dutch Mathematical Olympiad K
U
ERL
N
D
E
NED
Friday, January 30, 2009
PIADE
Solutions OLYM
A1. (C) 72% 60% of 25 is 15; 70% of 30 is 21; and 80% of 45 is 36. So in total, Ella answered
15 + 21 + 36 = 72 of the 100 questions correctly.
A2. (E) 17 We check how many of these numbers have sum of digits equal to 1, 2,
etc. There is 1 number with sum 1 (being 10); there are 2 with sum 2 (being 20 and
11); etc.; 9 with sum 9 (being 90, 81, . . . , 18); also, 9 with sum 10 (being 91, 82, . . . , 19);
etc.; and finally, 1 with sum 18 (being 99); see the table below. Then the sum of dig-
its is a square of an integer (i.e. 1, 4, 9 or 16) in 1 + 4 + 9 + 3 = 17 of the 90 cases.
sum: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
number: 1 2 3 4 5 6 7 8 9 9 8 7 6 5 4 3 2 1
A3. (B) 13 Note that the second die only has even numbers on it, and that the third die
only has odd numbers on it. So essentially the question is to find the probability that rolling the
first die gives an even number. Since 2 of the 6 numbers on this die are even, this probability is
equal to 62 = 13 .
A4. (D) 26 Let’s say we put a, b, and c in the top three squares. Then a b c
the result in the bottom square is a + 2b + c. So we can maximize this result + +
by making first b, then a and c as large as possible. Taking b = 9, a = 8 and
a+b b+c
c = 7 then yields 33 as result. In the same way, we can minimize the result
by making first b, then a and c as small as possible. Taking b = 1, a = 2, +
c = 3 yields 7 as result. The difference between these numbers is 33 − 7 = 26. a+2b+c
3 4
A5. (A) 80 Note that the diagonals have lengths 7 · 56 = 24 and 7 · 56 = 5 5
32. So the halves of diagonals have lengths 12 = 3 · 4 and 16 = 4 · 4. So the 3
4 4
rhombus is 4 times larger than the rhombus in the figure, which consists of
5 3
four triangles with sides 3, 4 and 5. Hence the sides of the original rhombus
5
have length 4 · 5 = 20, and thus the perimeter has length 4 · 20 = 80.
A6. (E) 34 Let x be said distance and let us suppose that he has spent a quarter of an hour
walking by then. Then continuing walking will take him x quarters of an hour. On the other
hand, if he decides to walk back home to pick up his bike, he’ll first have to spend one quarter
of an hour to get back, and then 1+x 7 quarters of an hour by bike; since he travels 7 times faster
that way. Then we have x = 1 + 7 , so 7x = 7 + (1 + x), or 6x = 8. We deduce that x = 86 = 43 .
1+x
Taking for ‘quarter of an hour’ any other time unit will give us the same result.
√
A7. (D) 1 + 2 3 In 4ABC, ∠A is half of 60◦ , so 30◦ . Also, ∠C is a right
angle, so 4ABC is a 30◦ -60◦ -90◦ -triangle, where |BC| = 1. So it’s half of B
an equilateral triangle with sides 2: |AB|
√ = 2. Now
√ we calculate |AC| with 2
1
the√Theorem√of Pythagoras: |AC| = 22 − 12 = 3. So the required length A 3 C
is 3 + 1 + 3.
A8. (A) 2 Suppose that n, having digits a, b, c and d (so n = 1000a + 100b + 10c + d) is
divisible by 44. Then it is also divisible by 11. Since the number m = 1001a + 99b + 11c is also
divisible by 11, so is m − n. Hence m − n = a − b + c − d is a multiple of 11. But this number
is at most the sum of the two highest digits, minus the sum of the lowest two, so 13 − 7 = 6,
and in the same way, we see that this number is at least −6. So it has to be equal to 0. Thus
a + c = b + d, and since the sum of the digits is 20, we have a + c = b + d = 10. First suppose that
d = 4. Then b = 6 so we get two possibilities for n, namely 3674 and 7634. But neither of these
is divisible by 4, let alone by 44. Now suppose that d = 6, then we have b = 4, and in this case
we get 3476 and 7436, both of which are divisible by 44. Finally, note that since n is divisible by
4, d must be even. We deduce that we have only 2 such numbers that are divisible by 44.
23
B1. 5201 We can subdivide this grid of 1012 squares as follows. In
the upper left corner, we have one (grey) square, then two L-shaped pieces,
one having 3 squares (one of which grey), the other having 5 squares (all
of which grey). Then we have two more L-shaped pieces, one having 7
squares (one of which grey), the other having 9 squares (all of which grey),
etc. Of the last two L-shaped pieces, the first one has 199 squares (one of
which grey), and the second one has 201 squares (all of which grey). We
have 50 pairs of L-shapes in total, so the total number of grey squares is
1 + (1 + 5) + (1 + 9) + (1 + 13) + . . . + (1 + 201) = 1 + (6 + 10 + 14 + . . . + 202) =
1 + 12 · 50 · (6 + 202) = 5201.
Alternative solution In each pair of these L-shapes, there are 4 more
grey squares than white squares. So there are 50 · 4 = 200 more grey squares
than white squares in the 1012 − 1 = 10200 squares contained in the 50 pairs
of L-shapes, so we have 5000 white squares and 5200 grey ones. Since the
upper left square is grey, in total, we have 5201 grey squares.
B2. 4017 93 = 729; 993 = 970299; 9993 = 997002999. It seems to
be the case that in general, the third power of a number n consisting of k
consecutive nines takes the following form: first k − 1 nines; then a 7; then
k−1 zeroes; then a 2; and finally k nines. To prove this, we write n = 10k −1.
Indeed: (10k − 1)3 = 103k − 3 · 102k + 3 · 10k − 1 = 102k (10k − 3) + (3 · 10k − 1). |99999 .{z
. . 99999}
The number 10k −3 can be written as 999 . . . 997 with k −1 nines. Multiplied 2009×
with 102k this gives a number that ends in 2k zeroes. Adding 3 · 10k − 1 to
this number, the last k + 1 zeroes are replaced with 2999 . . . 999 with k nines.
So in total, we have (k − 1) + k nines; in our case, k = 2009, so we have 4017
nines.
√
B3. 2 + 2 2 We only need to look at a quarter of the tile: 4ABC. The
B Q
area of 4P QR is half of the area 1
2 2
of 4ABC.
√ The triangles are√similar, so corresponding sides have a ratio of Q
1 : 2, so |QR| : |BC| = 1 : 2. x P
Now we calculate |BQ| using the Theorem q of Pythagoras in 4BQQ0 : A
√ R
2|BQ|2 = |BQ0 |2 + |BQ|2 = 12 , so |BQ| = 12 = 12 2. Now let us write x 1
2
√ √ √ √ 2
C
for |QR|. Then √
we√
find x + 2 = 2 · x, so x( 2 − 1) = 2 or equivalently,
√
2 2( 2+1) √ √ √
x = √2−1 = 2−1 = 2 + 2. Hence |BC| = x + 2 = 2 + 2 2 (or
√ √ √ √
|BC| = 2 · x = 2 · (2 + 2) = 2 2 + 2).
B4. (a, b, c) = (−12, 6, 24) of (a, b, c) = (−12, 24, 6) (one answer is enough)
We calculate (b + c)2 in two different ways. (b + c)2 = (18 − a)2 = 324 − 36a + a2 and (b + c)2 =
b2 + 2bc + c2 = (756 − a2 ) + 2a2 . So a2 − 36a + 324 = a2 + 756, or −36a = 756 − 324 = 432,
so a = −12. Substituting this in the first and in the last equation, we obtain the equations
b + c = 30 and bc = 144. Trying some divisors of 144 = 122 , we then should be able to find a
solution. Or we can just substitute c = 30−b in the last equation, yielding the quadratic equation
b(30 − b) = 144, or equivalently b2 − 30b + 144 = 0. We can factorize this as (b − 6)(b − 24) = 0
(or we can use the abc-formula) to see that we have two solutions b = 6 (and c = 24) or b = 24
(and c = 6).
Alternative solution Just as above, we see that a = −12. Then by substituting this in all three
equations, we see that b + c = 30, b2 + c2 = 756 − 144 = 612 and bc = 144. Combining√the last two
equations yields (b−c)2 = b2 +c2 −2bc = 612−2·144 = 324, or equivalently b−c = ± 324 = ±18.
Adding the equations b + c = 30 and b − c = −18, we get 2b = (b + c) + (b − c) = 12 so b = 6
(and c = 24). Adding the equations b + c = 30 and b − c = 18, we get 2b = (b + c) + (b − c) = 48
so b = 24 (and c = 6).
24
48th Dutch Mathematical Olympiad 2009
and the team selection for IMO 2010 Astana
MY
CY
We eat problems
CMY
for breakfast.
Preferably unsolved ones...
Olympiad 2009
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
WISKUNDE
Hoofdsponsors van de OLYMPIADE
We thank our sponsors
NEDERLANDSE
WISKUNDE 2010
OLYMPIADE
Overige sponsors
en donateurs
Contents
1 Introduction
3 First Round, January 2009
9 Final Round, September 2009
14 BxMO Team Selection Test, March 2010
18 Benelux Mathematical Olympiad, April 2010
24 IMO Team Selection Test 1, June 2010
29 IMO Team Selection Test 2, June 2010 Centraal Bureau voor de Statistiek
The best students were invited for the final round. In order to stimulate
young students to participate, we set different thresholds for students from
different grades. Those students from grade 5 (4, ≤ 3) that scored 26 (23,
20) points or more on the first round (out of a maximum of 36 points)
were invited to the final round. Also some outstanding participants in the
Kangaroo math contest or the Pythagoras Olympiad were invited.
For the second consecutive year we organised training sessions at six uni-
versities in the country for the 155 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 155, in total 131 participated in the final round on 18 Septem-
ber 2009 at Eindhoven University of Technology. This final round contained
five problems for which the students had to give extensive solutions and
proofs. They had three hours for the paper. After the prizes had been
awarded in the beginning of November, the Dutch Mathematical Olympiad
concluded its 48th edition 2009. In 2011 we will have our 50th edition.
On 5 March 2010 the first selection test was held. The best ten students
participated in the second Benelux Mathematical Olympiad (BxMO), held
in Amsterdam, the Netherlands.
In June, out of those 10 students and 3 reserve candidates, the team for
the International Mathematical Olympiad 2010 was selected by two team
selection tests on 9 and 12 June 2010. A seventh, young, promising student
was selected to accompany the team to the IMO. The team had a training
camp in Astana from 28 June until 5 July, together with the team from
1
New Zealand.
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 2009
Problems
A-problems
A1. Ella does three tests. During the first one, she answered 60% of
the 25 questions correctly, during the second one, she answered 70%
of the 30 questions correctly, and during the last one, she answered
80% of the 45 questions correctly. Now if we merge these three tests
together to form one of 100 questions, what is the percentage of these
100 questions that Ella answered correctly?
A) 68% B) 70% C) 72% D) 74% E) 76%
A2. How many of the integers from 10 to 99 (10 and 99 included) have the
property that the sum of its digits is equal to the square of an integer?
(An example: The sum of the digits of 27 is equal to 2 + 7 = 9 = 32 .)
A) 13 B) 14 C) 15 D) 16 E) 17
A3. Ronald throws three dice. These dice look just like ordinary dice, but
their faces are numbered differently. The first die has the numbers 1,
1, 2, 2, 3 and 3 on it. The second die has the numbers 2, 2, 4, 4, 6
and 6 on it. And the third die has the numbers 1, 1, 3, 3, 5 and 5 on
it. He then adds up the three numbers he gets from rolling the three
dice. What is the probability that the resulting number is odd?
1 1 1 2 3
A) 4 B) 3 C) 2 D) 3 E) 4
3
A5. The lengths of the diagonals of a rhombus have a ratio of 3 : 4. (A
rhombus is an equilateral quadrilateral.) The sum of the lengths of
the diagonals is 56. What is the diameter of this rhombus?
A) 80 B) 96 C) 100 D) 108 E) 160
A6. Wouter is traveling by foot from his home to the fitness center. He
also could have chosen to travel by bike, in which case he would travel
7 times as fast. But he left his bike at home. After walking for 1 km,
continuing to walk would take just as long as walking back to get his
bike, and then travel further by bike. By then, what is the distance
in km to the fitness center?
8 7 6 5 4
A) 7 B) 6 C) 5 D) 4 E) 3
A8. Consider all four-digit numbers where each of the digit 3, 4, 6 and 7
occurs exactly once. How many of these numbers are divisible by 44?
A) 2 B) 4 C) 6 D) 8 E) 12
B-problems
The answer to each B-problem is a number.
4
B2. The integer N consists of 2009 consecutive nines.
A computer calculates N 3 = (99999 . . . 99999)3 . 99999
| .{z
. . 99999}
3 2009×
How many nines does the number N contain in
total?
a2 = bc
Solutions
A-problems
A3. B) 13 Note that the second die only has even numbers on it,
and that the third die only has odd numbers on it. So essentially
the question is to find the probability that rolling the first die gives
an even number. Since 2 of the 6 numbers on this die are even, this
probability is equal to 26 = 13 .
5
A4. D) 26 Let’s say we put a, b, and c in the
top three squares. Then the result in the bottom a b c
square is a+2b+c. So we can maximize this result
+ +
by making first b, then a and c as large as possible. b+c
a+b
Taking b = 9, a = 8 and c = 7 then yields 33 as
result. In the same way, we can minimize the result +
a+2b+c
by making first b, then a and c as small as possible.
Taking b = 1, a = 2, c = 3 yields 7 as result. The
difference between these numbers is 33 − 7 = 26.
√
A7. D) 1 + 2 3 In 4ABC, ∠A is half of 60◦ , so
30◦ . Also, ∠C is a right angle, so 4ABC is
a 30◦ -60◦ -90◦ -triangle, where |BC| = 1. So
it’s half of an equilateral triangle with sides
2: |AB| = 2. Now we calculate |AC| √ with the B
|AC| 2 12 =
Theorem
√ of Pythagoras: √ 2 −√
= 2
1
3. So the required length is 3 + 1 + 3.
A 3 C
6
A8. A) 2 Suppose that n, having digits a, b, c and d (so n =
1000a + 100b + 10c + d) is divisible by 44. Then it is also divisible by
11. Since the number m = 1001a + 99b + 11c is also divisible by 11,
so is m − n. Hence m − n = a − b + c − d is a multiple of 11. But this
number is at most the sum of the two highest digits, minus the sum
of the lowest two, so 13 − 7 = 6, and in the same way, we see that this
number is at least −6. So it has to be equal to 0. Thus a + c = b + d,
and since the sum of the digits is 20, we have a + c = b + d = 10.
First suppose that d = 4. Then b = 6 so we get two possibilities for
n, namely 3674 and 7634. But neither of them is divisible by 4, let
alone by 44. Now suppose that d = 6, then we have b = 4, and in this
case we get 3476 and 7436, both of which are divisible by 44. Finally,
note that since n is divisible by 4, d must be even. We deduce that
we have only 2 such numbers that are divisible by 44.
Alternative solution Check all 24 possibilities, or just the 12 even
possibilities, of even only the 6 multiples of 4.
B-problems
7
B2. 4017 93 = 729; 993 = 970299; 9993 = 997002999. It seems to
be the case that in general, the third power of a number n consisting
of k consecutive nines takes the following form: first k − 1 nines; then
a 7; then k − 1 zeroes; then a 2; and finally k nines. To prove this, we
write n = 10k − 1. Indeed: (10k − 1)3 = 103k − 3 · 102k + 3 · 10k − 1 =
102k (10k − 3) + (3 · 10k − 1). The number 10k − 3 can be written as
999 . . . 997 with k − 1 nines. Multiplied with 102k this gives a number
that ends in 2k zeroes. Adding 3 · 10k − 1 to this number, the last
k + 1 zeroes are replaced with 2999 . . . 999 with k nines. So in total,
we have (k − 1) + k nines; in our case, k = 2009, so we have 4017
nines.
√
B3. 2 + 2 2 We only need to look at a
quarter of the tile: 4ABC. The area of
4P QR is half of the area of 4ABC. The B Q
1
triangles are similar,
√ so corresponding sides 2 2
have
√ a ratio of 1 : 2, so |QR| : |BC| = 1 : Q
2. Now we calculate |BQ| using the The-
orem of Pythagoras in 4BQQ0 : 2|BQ| q =
2 x P
A
0 2 2 2 1
|BQ | + |BQ| = 1 , so |BQ| = 2 = R
1
√
2 2. Now √ let us write x for |QR|. Then 1
2
√ √ √ 2
we find x + 2 = 2 · x, so x( √ 2 −√1) = 2 C
√
or equivalently, x = √2−12
= 2(2−1 2+1)
=
√ √ √
2+ 2. Hence
√ |BC|
√ = x+ √ 2 = 2+2
√ 2 (or
|BC| = 2 · x = 2 · (2 + 2) = 2 2 + 2).
B4. (a, b, c) = (−12, 6, 24) of (a, b, c) = (−12, 24, 6) (one answer is enough)
We calculate (b + c)2 in two different ways. (b + c)2 = (18 − a)2 =
324 − 36a + a2 and (b + c)2 = b2 + 2bc + c2 = (756 − a2 ) + 2a2 .
So a2 − 36a + 324 = a2 + 756, or −36a = 756 − 324 = 432, so
a = −12. Substituting this in the first and in the last equation, we
obtain the equations b + c = 30 and bc = 144. Trying some divisors
of 144 = 122 , we then should be able to find a solution. Or we can
just substitute c = 30 − b in the last equation, yielding the quadratic
equation b(30 − b) = 144, or equivalently b2 − 30b + 144 = 0. We can
factorize this as (b − 6)(b − 24) = 0 (or we can use the abc-formula)
to see that we have two solutions b = 6 (and c = 24) or b = 24 (and
c = 6).
8
Final Round, September 2009
Problems
For these problems not only the answer is important; you also have to describe the way you
solved the problem.
9
4. Let ABC be an arbitrary triangle. On the perpendicular bisector of
AB, there is a point P inside of triangle ABC. On the sides BC and
CA, triangles BQC and CRA are placed externally. These triangles
satisfy 4BP A ∼ 4BQC ∼ 4CRA. (So Q and A lie on opposite
sides of BC, and R and B lie on opposite sides of AC.)
Show that the points P , Q, C and R form a parallelogram.
Solutions
1. Note that the two integers mentioned in the second condition cannot
end in a 0; since otherwise the product would also have to end in
a 0. Thus, the two numbers have an equal number of digits. But
a product of two integers of four or more digits consists of at least
seven digits, hence is too large, and a product of two integers of at
most two digits consists of at most four digits, hence is too small.
So to satisfy the second condition, we have to consider products of
integers consisting of three digits, say abc and cba.
If we work this out, we obtain: abc · cba = (100a + 10b + c)(100c +
10b + a) = 10000ac + 1000b(a + c) + 100(a2 + b2 + c2 ) + 10b(a + c) + ac.
If ac > 9, then this integer consists of more than 5 digits. Hence
ac 6 9. Now if b(a + c) > 9, then the first digit will be greater
than ac, hence it will no longer be equal to the last digit. Hence we
also have b(a + c) 6 9. We can use a similar argument to show that
a2 +b2 +c2 6 9. We deduce from this that 1+b2 +1 6 a2 +b2 +c2 6 9,
and hence that b is equal to one of 0,1,2.
10
If b = 1, the conditions become a + c 6 9 and a2 + c2 6 8, and we
must also have 1 6 ac 6 9. From the second condition, we obtain
1 6 a 6 2 and 1 6 c 6 2.
If b = 2, the conditions become a + c 6 4 and a2 + c2 6 5. Hence we
have three possibilities for (a, b).
In the above table, we have worked out all the possibilities. It turns
out that there are eight possible palindromic products: 10201, 12321,
14641, 20502, 23632, 26962, 40804 and 44944.
11
Alternative solution 1: Pick a participant A who has won the least
number of matches, and a participant B who lost to A. Consider
the set X of participants who lost to B. Then the set X does not
contain A. If there is no triple of participants as desired, then A
must have won against all participants in X. But A also won against
B. But then A has won at least one match more than B, yielding a
contradiction.
4. Note that |AP | = |P B|, and hence that |BQ| = |QC| and |CR| =
|RA|. Since 4BP A ∼ 4BQC, it follows that |P B| |QB|
|AB| = |CB| , and hence
|P B| |AB|
that |QB| = |CB| . Moreover, since Q lies outside triangle ABC and
P lies inside of it, we have
12
5. We first consider the effect of two consecutive steps. If before step
2k −1, the stack of cards, from top to bottom, consists of a1 , . . . , a100 ,
the first 2k − 1 cards are reversed in order, so we obtain
a2k−1 , a2k−2 . . . , a2 , a1 , a2k , a2k+1 . . . , a100 . Then, at the 2k th step,
the top 2k cards are reversed in order, yielding
a2k , a1 , a2 , . . . , a2k−2 , a2k−1 , a2k+1 , . . . , a100 . Hence the effect of these
two steps is that the 2k th card is moved up to the top, and the rest
is shifted down. So if we consider the first 100 steps, then first 2 is
put on top, then 4 is put on top of that, then 6, and so on, so we ob-
tain 100, 98, 96, . . . , 4, 2, 1, 3, 5, . . . , 97, 99. We call this a megastep.
In general, a megastep changes the order a1 , a2 , . . . , a99 , a100 into
a100 , a98 , a96 , . . . , a4 , a2 , a1 , a3 , a5 , . . . , a97 , a99 . This megastep is in-
vertible; the order after a megastep can only come from one unique
order. Now we there are only finitely many possible ways (namely
100!) to order the 100 cards, so after applying megasteps for a while,
we’ll obtain an order that has occurred before. Consider the first
megastep that yields an order that has occurred before. If this is not
the starting position, then one megastep before, the position should
have occurred earlier as well; a contradiction. Hence we return to the
starting position after a finite number of (mega)steps.
13
BxMO Team Selection Test, March 2010
Problems
f (x)f (y) = f (x + y) + xy
for all x, y ∈ R.
Is N even or odd?
14
Solutions
2 First, note that the function given by f (x) = 0 for all x ∈ R, does
not satisfy the functional equation. Hence there exists an x0 ∈ R for
which f (x0 ) 6= 0. Substituting x = x0 and y = 0 gives f (x0 )f (0) =
f (x0 ). Since f (x0 ) 6= 0, we obtain f (0) = 1. Now substituting x = 1
and y = −1 gives f (1)f (−1) = f (0) − 1 = 1 − 1 = 0. Hence f (1) = 0
or f (−1) = 0.
We now distinguish the two cases. First, suppose that f (1) = 0, and
substitute x = 1. We then get 0 = f (1)f (y) = f (1 + y) + y for all
y ∈ R, so f (1 + y) = −y for all y ∈ R. Substituting y = t − 1 now
gives f (t) = −t + 1 for all t ∈ R. This is our first candidate solution.
Now suppose that f (−1) = 0 and substitute x = −1. We obtain
0 = f (−1)f (y) = f (−1 + y) − y for all y ∈ R, hence f (−1 + y) = y
for all y ∈ R. Substituting y = t + 1, we obtain f (t) = t + 1 for all
t ∈ R. This is our second candidate solution.
Since we have covered all the cases, these are the only two possible
solutions. Let’s check them.
If f (t) = −t + 1 for all t ∈ R, then for all x, y ∈ R:
15
3 Solution I. Consider an unordered 5-tuple that satisfies the given
equation, and suppose that it consists of the distinct integers b1 , . . . ,
bk , where bi occurs exactly ti times. Hence t1 +· · ·+tk = 5. Note that
5!
this unordered 5-tuple can be ordered in t1 !···t k!
ways. This is odd
if and only if three factors 2 occur in the denominator, which holds
if and only if 4! or 5! occurs in the denominator. Hence the only
unordered 5-tuples that can be ordered in a odd number of ways, are
those that consist of at least 4 numbers that are the same.
Such a 5-tuple has the form (a, a, a, a, b), where b can be equal to a,
and where a, b are positive integers. We obtain the equation a4 + 1b = 1,
or, equivalently, 4b + a = ab. We deduce that b has to be a divisor
of a and that a is a divisor of 4b. Hence there are three possibilities:
a = b, a = 2b and a = 4b. In the first case, we obtain 5b = b2 ,
hence b = 5, yielding the solution (5, 5, 5, 5, 5). In the second case,
we obtain 6b = 2b2 , hence b = 3, yielding the solution (6, 6, 6, 6, 3).
In the third and final case, we obtain 8b = 4b2 , so b = 2, yielding the
solution (8, 8, 8, 8, 2). Hence there are three unordered 5-tuples which
can be ordered in an odd number of ways. Hence N is odd.
Solution II. Let (a1 , a2 , a3 , a4 , a5 ) be an ordered 5-tuple that satisfies
the given equation. Then also the 5-tuple (a2 , a1 , a3 , a4 , a5 ). If a1 6=
a2 , this is a different ordered 5-tuple. Hence there is an even number
of solutions (a1 , a2 , a3 , a4 , a5 ) with a1 6= a2 . So we may only consider
the solutions with a1 = a2 . In the same way, we show that there is
an even number of solutions (a1 , a1 , a3 , a4 , a5 ) with a3 6= a4 . Hence
we may only consider the solutions with a1 = a2 , a3 = a4 . For
each solution (a1 , a1 , a3 , a3 , a5 ) with a1 6= a3 , there is yet another
solution (a3 , a3 , a1 , a1 , a5 ), hence there is an even number of solutions
(a1 , a1 , a3 , a3 , a5 ) with a1 6= a3 . Thus we may only consider the
solution of the form (a, a, a, a, b), where b can be equal to a.
We obtain the equation a4 + 1b = 1, or equivalently, 4b + a = ab. We
can rewrite this last equation as (a − 4)(b − 1) = 4. Since b is a
positive integer, we have b − 1 ≥ 0. Hence b − 1 is a positive divisor
of 4, namely 1, 2 or 4. This yields the three solutions (a, b) = (8, 2),
(a, b) = (6, 3) and (a, b) = (5, 5). Since the number of solutions that
we covered previously, is even, we deduce that N is odd.
16
4 We express all relevant angles in terms of α = ∠BAP and β =
∠P BA. The midpoint of AB is, by definition, also the midpoint of
P M , hence the diagonals of quadrilateral divide each other into equal
segments. Hence AP BM is a parallelogram and we have ∠AM B =
∠AP B. Note that (using the sum of the angles of a triangle) ∠AP B+
α + β = 180◦ . Hence 180◦ − ∠AM B = α + β.
By the inscribed angle theorem, applied on the tangent AB to Γ1 ,
we have α = ∠ADP . Then, by the inscribed angle theorem, applied
on the chord AP of Γ1 , it follows that ∠ADP = ∠AQP . Hence
α = ∠AQP . In a similar way, we deduce that β = ∠P CB = ∠P QB.
Hence ∠AQB = ∠AQP + ∠P QB = α + β = 180◦ − ∠AM B, so
AQBM is a cyclic quadrilateral.
Now let S be the intersection of DP and AB. Using the inscribed an-
gle theorem, we deduce that ∠SP B = ∠P CB = ∠P BS = ∠P BA =
β. Hence ∠DP F = ∠SP B = β, by opposite angles. Now we apply
the exterior angle theorem to 4DF P , to obtain ∠AF B = ∠AF P =
∠F DP + ∠DP F = α + β = ∠AQB. Hence AF QB is a cyclic quadri-
lateral. Hence F lies on the circumcircle of the cyclic quadrilateral
AQBM . Similarly, we see that E also lies on that circle. Hence
AM BEQF is a cyclic hexagon.
17
Benelux Mathematical Olympiad, April 2010
Problems
1. A finite set of integers is called bad if its elements add up to 2010. A fi-
nite set of integers is a Benelux-set if none of its subsets is bad. Deter-
mine the smallest integer n such that the set {502, 503, 504, . . . , 2009}
can be partitioned into n Benelux-sets.
(A partition of a set S into n subsets is a collection of n pairwise
disjoint subsets of S, the union of which equals S.)
p(a+b−2c)+p(b+c−2a)+p(c+a−2b) = 3p(a−b)+3p(b−c)+3p(c−a)
for all a, b, c ∈ R.
18
Solutions
19
2. For a = b = c, we have 3p(0) = 9p(0), hence p(0) = 0. Now set
b = c = 0, then we have
(a+b−2c)+(b+c−2a)+(c+a−2b) = 0 = 3(a−b)+3(b−c)+3(c−a)
for all a, b, c ∈ R.
Now note that if p(x) is a solution to our problem, then so is λp(x)
for all λ ∈ R. Also, if p(x) and q(x) are both solutions, then so is
p(x) + q(x). We conclude that for all real numbers a2 and a1 the
polynomial a2 x2 + a1 x is a solution. Since we have already shown
that there can be no other solutions, these are the only solutions.
20
3. Solution 1.
(a) Since P , R and Q are collinear, we have 4P AQ ∼ 4P BR,
hence
|AQ| |AP |
= .
|BR| |BP |
Conversely, P , T and S are collinear if it holds that
|AS| |AP |
= .
|BT | |BP |
So it suffices to prove
|BT | |AS|
= .
|BR| |AQ|
A1 A : AS = B1 B : BT,
A1 A : AS = B1 R : RB.
21
This gives
BB1 RB1 RB + BB1 BB1 BB1
= = =1+ =1− ,
BT RB RB RB BR
so
1
BB1 = 1 1 .
BT + BR
B2 B : BR = A2 A : AQ = B2 T : T B.
This gives
BB2 T B2 T B + BB2 BB2 BB2
= = =1+ =1− ,
BR TB TB TB BT
so
1
BB2 = 1 1 .
BR + BT
We conclude that B1 = B2 , which implies that P , K and L are
collinear.
22
4. Let (a, b, p, n) be a solution. Note that we can write the given equation
as
(a + b)(a2 − ab + b2 ) = pn .
As a and b are positive integers, we have a + b ≥ 2, so p | a + b.
Furthermore, a2 − ab + b2 = (a − b)2 + ab, so either a = b = 1 or
a2 −ab+b2 ≥ 2. Assume that the latter is the case. Then p is a divisor
of both a+b and a2 −ab+b2 , hence also of (a+b)2 −(a2 −ab+b2 ) = 3ab.
This means that p either is equal to 3 or is a divisor of ab. Since p
is a divisor of a + b, we have p | a ⇔ p | b, hence either p = 3, or
p | a and p | b. If p | a and p | b, then we can write a = pa0 , b = pb0
with a0 and b0 positive integers, and we have (a0 )3 + (b0 )3 = pn−3 , so
(a0 , b0 , p, n − 3) then is another solution (note that (a0 )3 + (b0 )3 is a
positive integer greater than 1, so n − 3 is positive).
Now assume that (a0 , b0 , p0 , n0 ) is a solution such that p - a. From the
reasoning above it follows that either a0 = b0 = 1, or p0 = 3. After
all, if we do not have a0 = b0 = 1 and we have p0 6= 3, then p | a.
Also, given an arbitrary solution (a, b, p, n), we can divide everything
by p repeatedly until there are no factors p left in a.
Suppose a0 = b0 = 1. Then the solution is (1, 1, 2, 1).
Suppose p0 = 3. Assume that 32 | (a20 − a0 b0 + b20 ). As 32 | (a0 + b0 )2 ,
we then have 32 | (a0 + b0 )2 − (a20 − a0 b0 + b20 ) = 3a0 b0 , so 3 | a0 b0 .
But 3 - a0 by assumption, and 3 | a0 + b0 , so 3 - b0 , which contradicts
3 | a0 b0 . We conclude that 32 - (a20 − a0 b0 + b20 ). As both a0 + b0 and
a20 − a0 b0 + b20 must be powers of 3, we have a20 − a0 b0 + b20 = 3. Hence
(a0 − b0 )2 + a0 b0 = 3. We must have (a0 − b0 )2 = 0 or (a0 − b0 )2 = 1.
The former does not give a solution; the latter gives a0 = 2 and b0 = 1
or a0 = 1 and b0 = 2.
So all solutions with p - a are (1, 1, 2, 1), (2, 1, 3, 2) and (1, 2, 3, 2).
From the above it follows that all other solutions are of the form
(pk0 a0 , pk0 b0 , p0 , n0 + 3k), where (a0 , b0 , p0 , n0 ) is one of these three
solutions. Hence we find three families of solutions:
• (2k , 2k , 2, 3k + 1) with k ∈ Z≥0 ,
• (2 · 3k , 3k , 3, 3k + 2) with k ∈ Z≥0 ,
• (3k , 2 · 3k , 3, 3k + 2) with k ∈ Z≥0 .
23
IMO Team Selection Test 1, June 2010
Problems
5 Find all triples (x, y, z) of real (but not necessarily positive) numbers
satisfying
3(x2 + y 2 + z 2 ) = 1,
x2 y 2 + y 2 z 2 + z 2 x2 = xyz(x + y + z)3 .
Solutions
24
hence AP ⊥ BC.
Conversely, suppose that AP ⊥ BC, or equivalently, ∠CEP = 90◦ .
Then we have
√
2 If M ≤ A, then automatically, M < A + B, so we’re done in this
case.
Hence suppose that M > A. Let k be such that ak = M 2 . So
Ak + B = M 2 . Since 0 < M − A < M , the square (M − A)2 is smaller
than M 2 . We have (M − A)2 = M 2 − 2M A + A2 = M 2 − A(2M − A).
So if k − (2M − A) ≥ 0, then
kn + 1 | k(n2 + n + 1) − (n + 1)(kn + 1) = k − n − 1.
25
left hand side is non-negative, and the right hand side is negative,
yielding a contradiction.
Case 3: k = n + 1. Now we have p = (n + 1)n + 1 = n2 + n + 1.
Hence 4p − 3 = 4n2 + 4n + 1 = (2n + 1)2 is a square.
We deduce that in the only possible case, 4p − 3 is a square.
5 We first show that for reals x, y and z satisfying the first condition,
we have
x2 y 2 + y 2 z 2 + z 2 x2 ≥ xyz(x + y + z)3 . (2)
So the solutions that we are looking for, are the cases in which equality
holds in 2.
2 2
Let a, b and c be reals. Then (a−b)2 ≥ 0, so a +b
2 ≥ ab with equality
if and only if a = b. We repeat this for the pairs (b, c) and (c, a), and
after summing, it follows that
a2 + b2 + c2 ≥ ab + bc + ca
26
with equality if and only if a = b = c.
We apply this inequality to the triple (xy, yz, zx), and obtain
x2 y 2 + y 2 z 2 + z 2 x2 ≥ xy 2 z + yz 2 x + zx2 y = xyz(x + y + z)
x2 y 2 + y 2 z 2 + z 2 x2 ≥ xyz(x + y + z) (3)
and
1 ≥ (x + y + z)2 (4)
2 2
To this end, we first multiply (4) by the non-negative factor x y +
y 2 z 2 + z 2 x2 ; we obtain
x = y = z or x2 y 2 + y 2 z 2 + z 2 x2 = 0
and (xy = yz = zx or x + y + z = 0) .
From the second condition, it follows that indeed, equality must hold.
Hence (x, y, z) must satisfy the conditions above. Furthermore, we
27
still have the first condition, which says that 3(x2 + y 2 + z 2 ) = 1:
3(x2 + y 2 + z 2 ) = 1,
x=y=z or x2 y 2 + y 2 z 2 + z 2 x2 = 0,
xy = yz = zx or x + y + z = 0.
3(x2 + y 2 + z 2 ) = 1,
xy = yz = zx,
x=y=z or xyz(x + y + z) = 0.
28
IMO Team Selection Test 2, June 2010
Problems
for all x ∈ R.
(In general the expression a = max g(s) means: a ≥ g(s) for all s ∈ S
s∈S
and furthermore there is an s ∈ S for which a = g(s).)
1
b be positive integers such that M (a, b) = a −
3. (a) Let a and b +
b b + a3 is an integer. Proof that M (a, b) is a square.
(b) Find nonzero integers a and b such that M (a, b) is a positive
integer, but not a square.
29
Solutions
an ≥ 2n−1 − 1 + a1 ≥ 2n−1
We conclude f (x) = x2 .
30
We check if this function satisfies. Let x ∈ R be given. Because
(x − y)2 ≥ 0, it is true that x2 ≥ 2xy − y 2 for all y ∈ R with equality
iff x = y, therefore
x2 = max (2xy − y 2 ).
y∈R
31
that M , N and P are collinear and that QN is parallel to DB. This
yields because of F-angles that ∠P SB = ∠P QN = ∠N P Q, because
|N P | = |N Q|. By Z-angles we see that ∠P M B = ∠M N Q and be-
cause of the exterior angle theorem inside triangle P QN that is equal
to ∠P QN + ∠N P Q = 2∠P SB. So ∠P M B = ∠P SB, together with
the inscribed angle theorem this yields that S lies on Γ1 . So S = D,
from which follows that P , Q and D are collinear.
Because ∠DP B = ∠DM Q = 90◦ and ∠P DB = ∠QDM , it is true
|DP |
that 4DP B ∼ 4DM Q. This yields |DB| = |DM |
|DQ| . Since |DB| =
2|DM |, this is equivalent to |DP ||DQ| = 2|DM |2 .√The power theo-
rem on Γ2 yields |DP ||DQ| = |DR|2 , so |DR| = 2|DM | = |DA|.
5. Let p be a prime and let k be such that A(k) and A(k + 1) are both
divisible by p. The difference between A(k) and A(k + 1) is also
divisible by p and this is equal to
The right hand side is independent of k. We now see that for each
prime p the number −a2 + 4b + 1 has to be divisible by p. This is
only possible if −a2 + 4b + 1 = 0. So we have a2 = 4b + 1. We now
see that a has to be odd and we write a = 2c + 1 with c ∈ Z. Then
it is true that 4b + 1 = a2 = 4c2 + 4c + 1, so b = c2 + c. Therefore,
the polynomial can be written as A(x) = x2 + (2c + 1)x + (c2 + c)
for an integer c. We can factor this as A(x) = (x + c)(x + c + 1),
consequently x = −c and x = −c − 1 are zeroes of the polynomial.
These are both integer points. We conclude that m = −c − 1 satisfies
A(m) = A(m + 1) = 0.
32
Junior Mathematical Olympiad, October 2009
Problems
Part 1
4. Bram has got coins of 5 cent, 10 cent, 20 cent and 50 cent, of each
type at least one. In total he has got 9 coins and together they are
worth 2,10 euro. How many coins of 20 cent does Bram have?
A) 1 B) 2 C) 3 D) 4 E) 5
33
6. Alice, Bas, Chris, Daan and Eva know each other very well. Each
of them always tells the truth or always lies. Chris says: “Alice is
honest”, on which Eve replies: “Chris lies! ” Chris says: “Bas is a
liar.” Eva claims: “Daan is honest.” Which two persons could both
be telling the truth?
A) Alice and Bas B) Bas and Chris C) Chris and Daan
D) Daan and Eva E) Eva and Alice
7. Jan is looking for a triangle which has got circumference 21 and in-
teger side lengths. Moreover, for each pair of sides the length of one
of them has to divide the length of the other. How many different
triangle with these properties exist?
A) 0 B) 1 C) 3 D) 5 E) 6
10. A large bag contains a number of balls: more then 100, but less than
150. If you divide the balls (evenly) over 7 children, then 2 balls will
remain. If you divide the balls over 6 children, 3 balls will remain.
How many balls will remain if you divide the balls over 5 children?
A) 0 B) 1 C) 2 D) 3 E) 4
34
12. Mister Jansen departs every morning at 8:00 to his work place by
car. If he (constantly) drives 40 km/h, he would arrive three minutes
late at his work place. If he drives 60 km/h, he would arrive three
minutes too early. How fast does he need to drive to be exactly on
time?
A) 48 B) 49 C) 50 D) 51 E) 52
14. Sir Tuinder has got a ‘half-open’ fence around his garden made of
shelves. The shelves are alternating at the front and the rear side,
exactly in front of the middle of the gap between the two shelves on
the other side. Here you can see the fence from above. The width of
the shelves is 21 cm and the thickness is 3 cm. Between the front- and
the rear side is a distance of 2 cm. How many cm has the distance
between two shelves next to each other be such that people you can’t
look into the garden from the outside?
A) 7 B) 8 C) 9 D) 10 E) 11
21 ?
3
2
35
Part 2
The answer to each problem is a number.
4. Jaap has got eight integers. Of each integer the first and last dig-
its are 1 and the other digits are 0. The eight integers have got
2, 3, 5, 9, 17, 33, 65 and 129 digits. If Jaap multiplies these eight digits
and adds up the digit of the result, which number will he get?
36
6. A square with area 54 is divided into four equal
squares. The upper left square is colored grey;
the lower right square is again divides into four
equal squares, and so on. The pattern continues
infinitely long. What is the area of the grey region?
7. Start with a number of two digits (the first digit is nonzero) and do
the following: multiply the two digits and add the same two digits
to the product. In this way 27 will yield 14 + 2 + 7 = 23. For how
many two digit numbers the result is equal the the number, you began
with?
9. Anneke can drive from A to B on the high way or via a shorter road.
On the high way Anneke can drive 120 km/h and on the shorter road
she can drive 60 km/h. The high way is 8 km longer, but the trip
takes 8 minutes less. How many kilometers long is the high way from
A to B?
32
10. In the figure you see a square. The square is di-
vided into four rectangles. Of three rectangles the
circumference is given and written inside the rect- 36 42
angle. What is the area of the fourth rectangle?
37
Solutions
Part 1
1
1. C) 8 6. D) Daan and Eva 11. D) hexagon
2. A) 40 7. C) 3 12. A) 48
3. E) 12 8. D) 35 13. A) 9
4. A) 1 9. B) 23 14. C) 9
5. E) 1 34 10. A) 0 15. C) 5
16
Part 2
1. 81 5. 90 8. 8
2. 597 6. 18 9. 32
3. 4 12 7. 9 10. 42
4. 111
| .{z
. . 111}
256 ones
38
WISKUNDE
Hoofdsponsors van de OLYMPIADE
We thank our sponsors
NEDERLANDSE
WISKUNDE 2010
OLYMPIADE
Overige sponsors
en donateurs
Contents
1 Introduction
3 First Round, January 2009
9 Final Round, September 2009
14 BxMO Team Selection Test, March 2010
18 Benelux Mathematical Olympiad, April 2010
24 IMO Team Selection Test 1, June 2010
29 IMO Team Selection Test 2, June 2010 Centraal Bureau voor de Statistiek
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.
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.
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
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
√ 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?
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
5
A
A3. B) 2
3 Connect the centre M to A, C and E. B F
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.
√ √
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
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.
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?
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.
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 ).
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.
10
B4. It is easy to calculate some more numbers of the sequence:
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:
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.
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.
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
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.
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.
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.
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.
17
Solutions
18
The inscribed angle theorem gives ∠BAK = ∠EAK + ∠BAE = ∠EAK +
∠BCE, therefore we find
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.
or, equivalently,
1
2 (x − y)2 + 12 (y − z)2 + 12 (z − x)2 + 1 = |(x − y) + (z − y)|. (1)
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:
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:
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
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:
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.
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
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,
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
(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.
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
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.
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
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
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
If e = 0, then
0 = P (0) = −n − 2,
so n = −2. So we can indeed factor P (x):
28
If e = 2, then
If e = 6, then
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|
Thus 1
2 |AC|·(|AB|−|BC|) 1
|DE| |AB|+|BC| 2 (|AB|− |BC|)
= |BC|·|AC|
= .
|CE| |BC|
|AB|+|BC|
|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
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
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
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
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
(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 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 .
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.
37c 6 30b + a.
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
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
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
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
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.
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?
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?
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?
41
Solutions
Part 1
1. E) 9 6. B) 5 11. D) 2684
2. E) 6 7. B) 45 12. C) 10
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
50th Dutch Mathematical Olympiad 2011
and the team selection for IMO 2012 Argentina
MY
We eat problems
CMY
Olympiad 2011
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, February 2011
10 Second Round, March 2011
15 Final Round, September 2011
19 BxMO/EGMO Team Selection Test, March 2012
23 IMO Team Selection Test 1, June 2012
28 IMO Team Selection Test 2, June 2012
33 Junior Mathematical Olympiad, October 2011
In the meantime the entire selection process for IMO 2012 started with the
first round on 4 February 2011, held 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 5258 students of 245 secondary schools
participated in this first round.
Those 799 students from grade 5 (4, 6 3) that scored 15 (13, 11) points or
more on the first round (out of a maximum of 36 points) were invited to the
second round, which was held in March 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 32 (28, 22) 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. In total 149 students
were invited. They also received an invitation to some training sessions at
the ten universities, in order to prepare them for their participation in the
final round.
Out of those 149, in total 142 participated in the final round on 16 Septem-
ber 2011 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 50th edition 2011.
1
six-day training camp in the beginning of June.
Among the participants of the training programme, there were some extra
girls, as this year we would participate in the first European Girls’ Mathe-
matical Olympiad (EGMO). In total there were eight girls competing to be
in the EGMO team. The team of four girls was selected by a selection test,
held on 16 March 2012. They attended the EGMO in Cambridge from 10
until 16 April, and the team returned with two honourable mentions and
a bronze medal.
The same selection test was used to determine the ten students partici-
pating in the Benelux Mathematical Olympiad (BxMO), held in Namur,
Belgium, from 20 until 22 April. The Dutch team managed to come first
in the country ranking, and received two honourable mentions, two bronze
medals, three silver medals and two gold medals.
In June the team for the International Mathematical Olympiad 2012 was
selected by two team selection tests on 6 and 9 June 2012. A seventh,
young, promising student was selected to accompany the team to the IMO
as an observer C. The team had a training camp in Buenos Aires, from 29
June until 8 July, together with the team from New Zealand.
For younger students the Junior Mathematical Olympiad was held in Oc-
tober 2011 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 consisted
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.
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
Dutch delegation
The Dutch team for IMO 2012 in Argentina consists of
3
First Round, February 2011
Problems
A-problems
B A
A3. A heptagon ABCDEF G is given, all sides of which
have length 2. Moreover: ∠E = 120◦ , ∠C = C G
∠G = 90◦ and ∠A = ∠B = ∠D = ∠F .
What is the area of the heptagon?
√ √ D F
A) 10 + 2 √2 B) 8 + 3√ 3 C) 14
D) 10 + 2 6 E) 8 + 3 6 E
A4. Alice, Brian and Carl have participated in a math contest consisting of 12
problems. Before the contest they were pessimistic and made the following
statements.
Alice: “Brian will answer at least two more problems correctly than I will.”
Brian: “I will not answer more than five problems correctly.”
Carl: “I will at the most answer correctly as many problems as Alice.”
Their teacher tried to encourage them by saying: “Together, you will an-
swer more than 18 problems correctly.” Afterwards, it transpired that all
three students and their teacher had made a wrong prediction.
4
Who has/have answered the least number of problems correctly?
A) only Alice D) both Alice and Brian
B) only Brian E) you cannot be sure of that
C) only Carl
A5. Jack wants to write down some of the numbers from 1 to 100 on a piece of
paper. He wants to do it in such a way, that no two numbers on the piece
of paper will add up to 125.
How many numbers, at most, can Jack write down on his piece op paper?
A) 50 B) 61 C) 62 D) 63 E) 64
B-problems
The answer to each B-problem is a number.
1 1
B1. A number x satisfies: x = . Determine x − . Simplify your answer
1+x x
as much as possible.
5
B2. An escalator goes up from the first to the second floor of a department
store. Dion, while going up the escalator, also walks at a constant pace.
Raymond, going in the opposite direction, tries to walk downwards, from
the second to the first floor, on the same escalator. He walks at the same
pace as Dion. They both take one step of the escalator at a time. Dion
arrives at the second floor after exactly 12 steps; Raymond arrives at the
first floor after exactly 60 steps.
How many steps would it take Dion to get upstairs if the escalator would
stand still?
B3. Six scouts are on a scouting expedition. They are going to the woods on
Saturday, and into the mountains on Sunday. On both days, they have
to go in pairs. The scoutmaster wants to group them into pairs for both
expeditions in such a way that nobody has the same partner on the second
day as on the first day.
In how many ways can he do that?
C
B4. In the figure you see a pointed arch ABC and its
inscribed circle. The pointed arch consists of a line
segment AB of length 1, a circular arc BC with cen-
ter A and a circular arc AC with center B.
What is the radius of the inscribed circle of this
pointed arch?
A B
Solutions
A-problems
2
A1. D) 5 0
Notice that all squares in the second
1
row and second column must be white. We consider 1
two cases, depending on the color of the upper left 2 0 1 1
square. If this square is white, then the last two
squares in the first row and column must be black.
This determines the coloring. See the top figure.
If this square is black, both the first row and first
column require exactly one more black square. For
each of the resulting 2×2 = 4 choices, there is exactly
one solution. Indeed, one row and one column are
left that need an additional black square.
6
Therefore the square at the intersection of this row
and column must be colored black, and the remain-
ing squares must be colored white, see bottom four figures.
A2. C) June The year of the date we are looking for, starts with a digit
2 or higher. We will look for the first date of which the year starts with
digit 2, and all eight digits are different. If such a date exists, we are done.
For the month, both 11 (two equal digits) and 12 (digit 2 is already used)
can be rejected. Therefore, the month (01 to 10) contains digit 0. This
implies that the day starts with digit 1 or 3. In the second case, it’s the
31st, since digit 0 is already taken. In both cases, the day contains digit
1. Both digit 0 and 1 being taken, the smallest possible year is 2345. The
smallest number we can use for the month is then 06, that is, June. Finally,
the day will be the 17th. Observe that the constructed date 17-06-2345
consists of eight different digits, as required.
√
A3. B) 8 + 3 3 The heptagon can be partitioned into B A
A4. A) only Alice As Brian’s prediction was wrong, he has at least six correct
answers. Alice’s prediction was wrong as well, which implies that Brian
answered at the most one more question correctly than Alice. Hence, Alice
has at least five correct answers. Since Carl made a wrong prediction, he
answered more questions correctly than Alice, hence at least six. Alice
cannot have more than five correct answers. Indeed, then Carl would have
seven correct answers, leading to a total of at least 6 + 6 + 7 = 19 correct
answers, as the teacher (incorrectly) predicted. We can conclude that Alice
answered five questions correctly. Since the others have at least six correct
answers, Alice is the only one with the smallest number of correct answers.
7
A5. C) 62 Jack can certainly write down 62 numbers, for example: the
numbers from 1 to 62 (since 62 + 61 < 125). More than 62 numbers will
not be possible. Indeed, the numbers from 25 to 100 can be partitioned
into pairs of sum 125: 25 + 100 = 125, 26 + 99 = 125, and so on up to
62 + 63 = 125. Jack must skip at least one number from each of the 38
pairs. In total, therefore, he can write down no more than 100 − 38 = 62
of the numbers.
A7. C) 35 After 140 seconds, Ann has made 7 rounds and Bob only 5.
At that moment, Ann leads by two full rounds. Hence after only 1404 = 35
seconds, Anne leads by half a round. That is exactly the first moment she
and Bob are at maximal distance from each other.
8
B-problems
1
B1. −1 We are given that x = 1+x . Clearly, x 6= 0, since 0 6= 11 .
Hence, on both sides of the equation, we may flip the numerator and
denominator of the fraction. This results in: x1 = 1 + x. Combining both
formulas, we obtain x − x1 = x − (1 + x) = −1.
B2. 20 Consider three escalators in a row: the first one going up,
the second standing still, and the third going down. If Dion walks up
the first escalator, he arrives at the top after exactly 12 steps. Raymond,
walking up the third escalator, takes 60 steps to reach the top and reaches
only 15 of the escalator after 12 steps. A third person, say Julian, takes
the second escalator and walks at the same pace as Dion and Raymond.
After 12 steps, he will be positioned exactly in between Dion and Julian,
at ( 55 + 51 )/2 = 35 of the escalator. Therefore, he will need 53 · 12 = 20 steps
to reach the top.
B3. 120 Let’s call the eldest scout A. There are 5 possibilities for
finding him a partner B for the first day. Then, there are 4 possible
partners C for B on the second day, because he cannot be paired with A
twice. Now for C, there are 3 possible partners D on the first day, since he
cannot go with B again, and A is already paired. For D, there are now 2
possible partners E on the second day, since B and C are already paired,
and A cannot be his partner because that would leave two scouts that are
forced to form a pair on both days. Finally, there is one scout left. He
has no choice but to team op with E on the first day, and with A on the
second. In total there are 5 × 4 × 3 × 2 = 120 possibilities.
3
B4. 8 Consider the inscribed circle. We de-
C
note its center by O and its radius by r. The points
where the circle is tangent to AB and BC are de- R
noted by M and R, respectively. O
Since A, O and R are on a line, we have: |AO| =
|AR| − |OR| = 1 − r. We also know that |OM | = r
and |AM | = 12 |AB| = 21 . Using the Pythagorean
A M B
theorem, we find |AM |2 +|OM |2 = |AO|2 , and hence
1 2 2 2
4 + r = (1 − r) = r − 2r + 1. This implies that
2r = 4 and therefore r = 38 .
3
9
Second Round, March 2011
Problems
B-problems
The answer to each B-problem is a number.
B1. At a gala, a number of pairs (consisting of one man and one woman) are
dancing, in such a way that 23 of the women present is dancing with 35 of
the men present.
What part of those present at the gala is dancing?
B3. In a classroom, there are 23 students who chose to learn a foreign language,
namely German and French. Of those 23 students, 10 are girls, and 11
of those 23 students have chosen French as their foreign language. The
number of girls that have chosen French, plus the number of boys that
have chosen German, is equal to 16.
What is the number of girls that have chosen French?
B4. We have a deck of 10.000 cards, numbered from 1 to 10.000. A step consists
of removing all the cards that has a square on it, and then renumbering
the remaining cards, starting from 1, in a successive way.
What is the number of steps needed to remove all but one card?
10
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 triples (a, b, c) of positive integers with the following prop-
erties:
• we have a < b < c, and a, b and c are three successive odd integers;
• the number a2 + b2 + c2 consists of four equal digits.
Solutions
B-problems
12
B1. 19 Let w be the number of women present, and let m be the
number of men present. The problem tells us that 32 w = 35 m and hence
9
that w = 10 m. The number of people dancing, is exactly twice the number
of men dancing, namely 65 m.
9
The number of people present is of course m + w = m + 10 m = 1910 m. So
it follows that the part of those present that is dancing, is equal to
6
5m 6 10 12
19 = 5 · 19 = 19 .
10 m
11
B2. 10 We split the black part into four tri-
angles, of which we coloured two of them gray. The
two gray triangles both have base 2, and their com-
bined height is 7 − 2 = 5, namely the height of the
larger square minus the height of the smaller square.
Hence the area of the two grey triangles together is
equal to 21 · 2 · 5 = 5. The same holds for the two
black triangles. It follows that the combined area is 5 + 5 = 10.
12
√
B5. π 2 cm Imagine the pole as a paper cylin-
der. Cut it open along its length, then un-
roll it, to get a rectangular strip of paper. So
points A and D correspond to the same point A D
on the cylinder, just like points B and C. The
width of the strip is equal to the perimeter of 45o
the cylinder, so |AD| = |BC| = 2π · 2 cm =
4π cm.
Note that the grey ribbon forms a 45◦ angle
with the cutting line, ABCD is a square. √The
length of the diagonal BD is equal to 2 · B C
4π cm and also equal to four times the width
of the grey ribbon, since the white and grey√stripes have the same width.
It follows that the grey ribbon has width π 2 cm.
C-problems
C1. Since a, b and c are three successive positive odd integers, we can write:
a = 2n − 1, b = 2n + 1 and c = 2n + 3, with n a positive integer.
A calculation then gives:
This needs to be equal to an integer that consists of four digits p. Hence the
integer 12n2 + 12n consists of four digits, of which the first two are equal
to p, and the last two are equal to p − 1. Since 12n2 + 12n is divisible by 2,
p − 1 has to be even. So we have the following possibilities for 12n2 + 12n;
1100, 3322, 5544, 7766 and 9988. This integer also must be divisible by
3, so the only integer remaining is 5544, so n2 + n = 5544 12 = 462. We
can rewrite this as n2 + n − 462 = 0. Factorizing this quadratic equation
then gives: (n − 21)(n + 22) = 0. Since n is a positive integer, the only
solution is n = 21. So the only triple satisfying the given properties is
(a, b, c) = (41, 43, 45).
13
C2. Note that the possible scores are multiples of 5. The lowest score a student
can get is 0, and the highest score is 16 · 10 = 160. Now suppose that there
are no two students with the same score. Then the combined score of the
students is at most 160 + 155 + 150 + · · · + 15 = 12 · 175 · 30 = 2625. We’ll
derive a contradiction from this.
Let A be the combined number of correct answers that were given within
one minute, B be the combined number of correct answers that were not
given within a minute, and C be the combined number of incorrect answers.
The students answered 16·30 = 480 questions together, so A+B+C = 480.
More than half of the questions was answered correctly within one minute,
so A > 240. Also note that B = C, so B = C = 480−A 2 . We now can
express the combined score in A. This is equal to:
480 − A 15
10 · A + 5 · B + 0 · C = 10 · A + 5 · = 2 A + 1200.
2
Since A > 240, the combined scores of the students is greater than 15
2 ·
240 + 1200 = 3000. But from the assumption that no two students have
the same score, we deduced that the combined score was at most 2625.
This is a contradiction. We deduce that this assumption was wrong, so
that there are two students with the same score.
14
Final Round, September 2011
Problems
For these problems not only the answer is important; you also have to describe the way you
solved the problem.
1. Determine all triples of positive integers (a, b, n) that satisfy the following
equation:
a! + b! = 2n .
Notation: k! = 1×2×· · ·×k, for example: 1! = 1, and 4! = 1×2×3×4 = 24.
2. Let ABC be a triangle. Points P and Q lie on side BC and satisfy |BP | =
|P Q| = |QC| = 31 |BC|. Points R and S lie on side CA and satisfy |CR| =
|RS| = |SA| = 13 |CA|. Finally, points T and U lie on side AB and satisfy
|AT | = |T U | = |U B| = 13 |AB|. Points P, Q, R, S, T and U turn out to lie
on a common circle.
Prove that ABC is an equilateral triangle.
3. In a tournament among six teams, every team plays against each other
team exactly once. When a team wins, it receives 3 points and the losing
team receives 0 points. If the game is a draw, the two teams receive 1
point each.
Can the final scores of the six teams be six consecutive numbers a, a +
1, . . . , a + 5? If so, determine all values of a for which this is possible.
4. Determine all pairs of positive real numbers (a, b) with a > b that satisfy
the following equations:
√ √ √ √
a a + b b = 134 and a b + b a = 126.
5. The number devil has coloured the integer numbers: every integer is
coloured either black or white. The number 1 is coloured white. For
every two white numbers a and b (a and b are allowed to be equal) the
numbers a − b and a + b have different colours.
Prove that 2011 is coloured white.
15
Solutions
In all, there are four solutions: (1, 1, 1), (2, 2, 2), (2, 3, 3) and (3, 2, 3).
Triangles ASL and ACK are similar (aa), because ∠SAL = ∠CAK and
∠LSA = ∠T SA = ∠BCA = ∠KCA. Hence |CK| |CA|
|SL| = |SA| = 3. This
|SL| 3·|SL| |CK|
implies that L is the midpoint of ST , because |ST | = 3·|ST | = |CB| = 12 .
Consider the center M of the circle through P , Q, R, S, T and U . It
is incident to both the perpendicular bisector of P Q, and that of ST .
16
However, since P Q and ST are parallel, the two perpendicular bisectors
must coincide: they are the same line. This line is incident to L and K,
and is therefore equal to line AK, which shows that AK ⊥ BC.
It follows that |AC| = |AB|, because AK is the perpendicular bisector of
BC.
In a similar fashion, one can show that |AC| = |BC|, concluding the proof
that triangle ABC is equilateral.
3. In all, 15 matches are played. In each match, the two teams together earn
2 or 3 points. The sum of the final scores is therefore an integer between
15 · 2 = 30 (all matches end in a draw) and 15 · 3 = 45 (no match is a
draw).
On the other hand, the sum of the six scores equals a+(a+1)+· · ·+(a+5) =
15 + 6a. Hence 30 6 15 + 6a 6 45, which shows that 3 6 a 6 5. We will
prove that a = 4 is the only possibility.
First consider the case a = 5. The sum of the scores equals 15 + 30 = 45,
so no match ends in a draw. Because in every match the teams earn either
0 or 3 points, every team’s score is divisible by 3. Therefore, the scores
cannot be six consecutive numbers.
Next, consider the case a = 3. The scores sum up to 3+4+5+6+7+8 = 33.
The two teams scoring 6 and 7 points must both have won at least one out
of the five matches they played.
The team scoring 8 points must have won at least two matches, because
3 + 1 + 1 + 1 + 1 = 7 < 8. Hence at least 4 matches did not end in a draw,
which implies
A B C D E F
that the sum of the scores is at least 4 · 3 +
A - 3 1 0 0 0 4
11 · 2 = 34. But we have already see that B 0 - 1 0 3 1 5
this sum equals 33, a contradiction. C 1 1 - 3 0 1 6
Finally, we will show that a = 4 is possi- D 3 3 0 - 1 0 7
E 3 0 3 1 - 1 8
ble. The table depicts a possible outcome
F 3 1 1 3 1 - 9
in which teams A to F have scores 4 to
9. The rightmost column shows the total
scores of the six teams.
17
√ √
4. For convenience, write y = a and z = b. The equations transform to
y 3 + z 3 = 134 and y 2 z + yz 2 = 126.
Combining these two equations in a handy way, we find
(y + z)3 = (y 3 + z 3 ) + 3(y 2 z + yz 2 ) = 134 + 3 · 126 = 512 = 83 .
This immediately implies that y + z = 8.
Rewrite the first equation as follows: (y + z)yz = y 2 z + yz 2 = 126. Since
y + z = 8, we see that yz = 126 63
8 = 4 .
From y + z = 8 and yz = 63 4 , we can determine y and z by solving a
2
√ the roots of the equation x −
quadratic equation: y and z are precisely
63 8± 64−4· 63 9
8x + 4 = 0. The two solutions are 2
4
, that is 2 and 72 .
9
Since a > b, also y > z holds. Hence y = 2 and z = 72 . We therefore find
that (a, b) = ( 81 49
4 , 4 ).
Because (a, b) = ( 81 49
4 , 4 ) satisfies the given equations, as required, we
conclude that this is the (only) solution.
We have just shown the base case k = 0. Assume that the claim holds true
for k = `.
Since 1 is white, and 3` + 2 is white by the induction hypothesis, the
numbers (3` + 2) − 1 = 3` + 1 and (3` + 2) + 1 = 3(` + 1) have different
colours. As 3` + 1 is white by the induction hypothesis, 3(` + 1) must be
black.
Since 2 and 3` + 2 are both white, the numbers (3` + 2) + 2 = 3(` + 1) + 1
and (3` + 2) − 2 = 3` must have different colours. As 3` is black by the
induction hypothesis, 3(` + 1) + 1 must be white.
Since 3(` + 1) + 1 and 1 are both white, the numbers 3(` + 1) + 1 + 1 =
3(` + 1) + 2 and 3(` + 1) have different colours. We already know that
3(` + 1) is black, so 3(` + 1) + 2 must be white.
This proves the claim for k = ` + 1.
Because 2011 = 3 · 670 + 1, this shows that 2011 is white.
18
BxMO/EGMO Team Selection Test, March 2012
Problems
1. Do there exist quadratic polynomials P (x) and Q(x) with real coefficients
such that the polynomial P (Q(x)) has precisely the zeros x = 2, x = 3, x =
5 and x = 7?
2. Let 4ABC be a triangle and let X be a point in the interior of the triangle.
The second intersection points of the lines XA, XB and XC with the
circumcircle of 4ABC are P, Q and R. Let U be a point on the ray XP
(these are the points on the line XP such that P and U lie on the same
side of X). The line through U parallel to AB intersects BQ in V . The
line through U parallel to AC intersects CR in W .
Prove that Q, R, V , and W lie on a circle.
x3 + y 3 = 4(x2 y + xy 2 − 5).
4. Let ABCD a convex quadrilateral (this means that all interior angles are
smaller than 180◦ ), such that there exist a point M on line segment AB
and a point N on line segment BC having the property that AN cuts the
quadrilateral in two parts of equal area, and such that the same property
holds for CM .
Prove that M N cuts the diagonal BD in two segments of equal length.
5. Let A be a set of positive integers having the following property: for each
positive integer n exactly one of the three numbers n, 2n and 3n is an
element of A. Furthermore, it is given that 2 ∈ A. Prove that 13824 ∈ / A.
19
Solutions
20
3. It holds that (x + y)3 = x3 + 3x2 y + 3xy 2 + y 3 , hence the given equality
yields
For each of the possible values for x + y we check whether its third power
is congruent to 1 modulo 7. It holds that 53 ≡ −1 mod 7, hence x + y = 5
is impossible. It holds that 103 ≡ −1 mod 7, hence x + y = 10 is also
impossible. It holds that 203 ≡ −1 mod 7, hence x + y = 20 is also
impossible. We are left with the possibilities x + y = 2 and x + y = 4.
If x + y = 2, then x = y = 1 and then the left hand site is positive and the
right hand side is negative, hence also x + y = 2 is impossible. If x + y = 4,
then (x, y) is equal to (1, 3), (2, 2) or (3, 1). By substituting this into the
equality we find that only (1, 3) and (3, 1) are solutions.
Triangles AT C and AN C have the same base AC, hence their altitudes
have equal lengths. This means that the line N T is parallel to the base
AC.
Analogously, we show that M T is parallel to AC. Hence, N T k M T and
this yields that M , N and T lie on a line. Hence, the midpoint of BD lies
on the line M N .
21
5. We will prove the following two assertions:
(i) If m ∈ A and 2 | m, then 6m ∈ A.
(ii) If m ∈ A and 3 | m, then 43 m ∈ A.
As 29 · 32 ∈ A, it holds that 29 · 33 ∈
/ A. Since 13824 = 29 · 33 , this is what
we had to prove.
22
IMO Team Selection Test 1, June 2012
Problems
a a+1
3. Determine all positive integers that cannot be written as b + b+1 where a
and b are positive integers.
5. Let Γ be the circumcircle of the acute triangle ABC. The angle bisector
of angle ABC intersects AC in the point B1 and the short arc AC of Γ in
the point P . The line through B1 perpendicular to BC intersects the short
arc BC of Γ in K. The line through B perpendicular to AK intersects AC
in L.
Prove that K, L and P lie on a line.
23
Solutions
where the second equality holds because the segments on the tangent lines
to the incircle have equal lengths.
Furthermore, we have ∠IKE = ∠IKC = ∠BN I = ∠DN I and ∠KEI =
90◦ = ∠IDN , hence 4IKE ∼ 4IN D (AA). Since line segments EI and
DI both are radii of the incircle, they have equal lengths, hence we even
have 4IKE ∼ = 4IN D. This yields that |EK| = |N D|.
Analogously, we find that |EL| = |M F |. Altogether, we have
2. It holds that
a−b a−b+b+c a+c
= −1= − 1.
b+c b+c b+c
By applying the equality also to the other three fractions and adding 4 to
both sides, we find that it is sufficient to prove that:
a+c b+d c+a d+b
LHS := + + + > 4. (1)
b+c c+d d+a a+b
Next, we apply the inequality of the harmonic and geometric mean to the
two positive numbers b + c and d + a:
2 (b + c) + (d + a)
1 1 6 ,
b+c + d+a
2
hence,
1 1 4
+ > .
b+c d+a a+b+c+d
24
Analogously, it holds that
1 1 4
+ > .
c+d a+b a+b+c+d
Using this, we find the following inequality for the left-hand side of (1):
1 1 1 1
LHS = (a + c) + + (b + d) +
b+c d+a c+d a+b
4 4
> (a + c) · + (b + d) ·
a+b+c+d a+b+c+d
(a + c) + (b + d)
=4·
a+b+c+d
= 4.
3. It holds that
a a+1 2ab + a + b
+ = .
b b+1 b(b + 1)
Next, suppose that this is equal to an integer n. Then we have b | 2ab+a+b
and b + 1 | 2ab + a + b. The former yields that b | a and hence b | a − b.
The latter yields b + 1 | (2ab + a + b) − (b + 1) · 2a = −a + b and hence also
b + 1 | a − b. Since the gcd of b and b + 1 is equal to 1, we may conclude
that b(b + 1) | a − b. Hence, we can write a as a = b(b + 1) · k + b. Since a
is positive, k must be a non-negative integer. Substitution yields
2ab + a + b
n=
b(b + 1)
2 · b(b + 1) · k + b · b + b(b + 1) · k + b + b
=
b(b + 1)
b(b + 1) · (2kb + k) + 2b2 + 2b
=
b(b + 1)
= (2b + 1)k + 2.
Hence, n is of the form n = (2b + 1)k + 2. This yields that n > 2 (because
k > 0) and that n − 2 is divisible by an odd integer greater than 1 (namely
2b + 1 > 3).
For the converse, suppose that for a number n we have that n > 2 and that
n − 2 is divisible by an odd integer greater than 1, say 2b + 1 where b > 1
is an integer. Then there is a k > 0 for which we have n = (2b + 1)k + 2.
25
Now choose a = b(b + 1) · k + b, then a is a positive integer and it holds
that
a a+1
+ = (b + 1)k + 1 + bk + 1 = (2b + 1)k + 2 = n.
b b+1
We conclude that the numbers that can be written as ab + a+1 b+1 , where a
and b are positive integers, are exactly the numbers n > 2 for which n − 2
is divisible by an odd integer greater than 1. Hence, the numbers that
cannot be written like this, are exactly 1 and the numbers n > 2 for which
n − 2 has no odd divisor greater than 1, otherwise stated for which n − 2
is a power of two, say 2m where m > 0.
We conclude that the numbers that cannot be written as ab + a+1 b+1 , are
exactly 1 and the numbers of the form 2m + 2 for m > 0.
at = u,
au = n + 1 − t,
an+1−t = n + 1 − u,
an+1−u = t.
Since u 6= t and u 6= n + 1 − t, the four numbers on the right hand side are
pairwise distinct. Furthermore, we can divide the four numbers into two
pairs of the form (v, n + 1 − v). We have found four numbers for which
it holds that on the same positions in the permutation are the same four
numbers, but in a different order. Now we may choose a t0 not equal to
one of the four mentioned numbers and a u0 with at0 = u0 and in the same
way we find a quadruple containing t0 . Note that n + 1 − t0 and n + 1 − u0
cannot be a number of the first quadruple, since then also u0 and t0 would
be a number of the first quadruple. We can continue this way until all n
numbers have been divided into quadruples.
We can generate all permutations by using the following procedure:
26
• Choose the smallest number k for which ak is not yet determined.
Choose ak = u for a certain u for which au is not yet determined and
for which u 6= k, u 6= n + 1 − k. This also determines the values of
au , an+1−u and an+1−k .
• Repeat the last step until all values of ak are determined.
For the first k we have n − 2 possible choices of u. During the next step
we have n − 6 possible choices left. One step later, we have n − 10, and so
on. Hence, the total number of permutations having the property, is equal
to
2 · 6 · 10 · . . . · (n − 10) · (n − 6) · (n − 2).
Write n = 4m, then we can write this as
(2m)!
2m · 1 · 3 · 5 · . . . ·(2m − 5) · (2m − 3) · (2m − 1) = 2m ·
2 · 4 · . . . · (2m)
(2m)! (2m)! ( 1 n)!
= = = 21 .
1 · 2 · 3 · ... · m m! ( 4 n)!
5. The fact that the angle bisector of angle ABC intersects the short arc AC
in P , implies that P lies exactly on the middle of this arc AC. We have
to prove that KL also goes through P , hence that KL intersects the arc
AC in the middle. Because K lies on Γ, this means that we have to prove
that KL is the angle bisector of ∠AKC.
Let S be the intersection point of B1 K and BC and let T be the intersection
point of BL and AK. Then we have ∠BSK = 90◦ and ∠BT K = 90◦ ,
hence BT SK is a cyclic quadrilateral. This yields that
∠CBL = ∠SBT = ∠SKT = ∠B1 KA. (2)
Because ABKC is a cyclic quadrilateral, we have ∠B1 AK = ∠CAK =
∠CBK. By the exterior angle theorem, we have ∠LB1 K = ∠B1 AK +
∠B1 KA, hence because of (2) we get
∠LB1 K = ∠B1 AK + ∠B1 KA = ∠CBK + ∠CBL = ∠LBK.
This yields that LKBB1 is a cyclic quadrilateral, which means that ∠LBB1 =
∠LKB1 . If we add (2) to this, we get
∠CBB1 = ∠CBL + ∠LBB1 = ∠B1 KA + ∠LKB1 = ∠LKA.
Hence,
∠LKA = ∠CBB1 = 12 ∠CBA = 21 ∠CKA,
where we used that ABKC is a cyclic quadrilateral. This yields that KL
is the angle bisector of ∠AKC, as desired.
27
IMO Team Selection Test 2, June 2012
Problems
a−b
1. For all positive integers a and b, we define a b = .
gcd(a, b)
Show that for every integer n > 1, the following holds: n is a prime power
if and only if for all positive integers m such that m < n, it holds that
gcd(n, n m) = 1.
2. There are two boxes containing balls. One of them contains m balls, and
the other contains n balls, where m, n > 0. Two actions are permitted:
x + y + 1 | 2xy and x + y − 1 | x2 + y 2 − 1.
for all x, y ∈ R.
28
Solutions
1. First suppose that n = pk with p prime and k > 0. Then we need to show
that for all positive integers m with m < n, we have gcd(n, n m) = 1.
So let m be a positive integer such that m < n. Write m = pl s with p - s
and 0 6 l < k. Then gcd(n, m) = gcd(pk , pl s) = pl , so
pk − pl s
n m= = pk−l − s.
pl
n − (n − q t+1 ) q t+1
n m= = = q.
qt qt
As q | n, it follows that gcd(n, n m) = q > 1.
2. First suppose that k = 2. Then we can remove all of the balls from both
boxes as follows.
If m = n, we remove m balls from both boxes, and we are done. If m 6= n,
we may assume without loss of generality that m < n. If furthermore,
2m < n holds, then we double the number of balls in the first box, until
there are m0 balls in the first box, where m0 satisfies both m0 < n and
2m0 > n. Hence we may assume without loss of generality that 2m > n.
Now remove 2n − m > 0 balls from both boxes. In the first box, m − (2m −
n) = n−m > 0 balls remain; in the second box, n−(2m−n) = 2(n−m) > 0
balls remain. Doubling the number of balls in the first box, then removing
2(n − m) balls from both boxes now makes both boxes empty.
Now consider the case k = 3. Let S denote the number of balls in the
first box, minus the number of balls in the second one. The first action
does not change S. Applying the second action to some box with t balls,
does not change the parity of S either, since we have 3t ≡ t (mod 2).
Hence the parity of S does not change. We conclude that the parity of
S is an invariant under both actions. Now suppose that we start with
29
(m, n) = (1, 2), then S = −1 ≡ 1 (mod 2). Then it is impossible to reach
the situation in which both boxes are empty, since then one would have
S ≡ 0 (mod 2). Hence we cannot empty both boxes if k = 3.
3. Note that
Note that the last inequality holds, as for all real numbers x and y, we
have x2 + y 2 > 2xy.
If k > 2, then 4xy > 2·(4xy−1), so 4xy 6 2. This contradicts x and y being
positive integers. Hence k = 1, so we deduce that 4xy = x2 + y 2 + 2xy − 1.
This implies that x2 + y 2 − 1 − 2xy = 0, or equivalently, (x − y)2 = 1.
Hence either x = y − 1, or x = y + 1.
Hence the only pairs that can satisfy the conditions are (x, x + 1) with
x > 1, and (x, x − 1) with x > 2. We check that they indeed satisfy the
conditions. Indeed, for the first family of pairs, we note that 2x + 2 divides
2x(x + 1), and that 2x divides x2 + (x + 1)2 − 1 = 2x2 + 2x, and similarly
for the second family of pairs. We conclude that the solutions are precisely
the pairs (x, x + 1) with x > 1 and the pairs (x, x − 1) with x > 2.
30
As M , N and T are collinear, Menelaus’ Theorem gives
BT CM AN
· · = −1.
TC MA NB
From these two identities, it follows that
BL BT
=− .
LC TC
5. Substituting y = −1 gives
c = c + 12 c + 12 ,
f f (− 12 ) = f (0) + 12 f (− 12 ) + 12 ,
hence
1
· 21 ,
f (0) = f (0) + 2
so f (0) = 12 .
Now suppose that there exists some a 6= −1 such that f (a) = − 21 . Substi-
tuting y = a now gives
f x(1 + a) − 12 = 0.
31
As 1 + a 6= 0, the function mapping x to x(1 + a) − 12 attains all values
in R. Hence f must be identically zero, hence constant, contradicting our
earlier findings. So the unique x ∈ R such that f (x) = − 12 is x = −1.
Now let b be such that f (b) = 0. Substitute x = b − 21 , and y = 0:
f b − 12 + 21 = f b − 12 + 12 12 + 12 ,
or equivalently,
1 1
f (b) = f b − 2 + 2 .
1
Since f (b) = 0, we have f (b − 2) = − 21 . Hence it follows by an earlier
result that b − 12 = −1, hence b = − 12 . Hence the unique x ∈ R with
f (x) = 0 is x = − 21 .
Now substitute x = −1. This gives
f −1 − y + f (y) = 0.
32
Junior Mathematical Olympiad, October 2011
Problems
Part 1
4. In a cube with edges of length 3 cm, some holes are made. In the center of
each face, a square hole with sides of length 2 cm is made, such that the
sides of this square are parallel to the edges of the cube. This hole runs
through the entire cube.
What is (in cm3 ) the volume of the object that remains after making these
three holes?
A) 6 B) 7 C) 8 D) 9 E) 10
33
5. In a classroom, one or more pupils always speak the truth. The other
pupils sometimes do speak the truth, and sometimes do not. The pupils
were asked how many of them always speak the truth. The answers were:
5, 6, 2, 3, 4, 6, 3, 6, 3, 4, 6, 5, 4, 3, and 6.
How many pupils do always speak the truth?
A) 2 B) 3 C) 4 D) 5 E) 6
9. Adrian puts stones in the nine squares of a 3×3-board. Into each of the
squares, he is allowed to put any number of stones. He is also allowed to
leave squares empty. When Adrian is finished, he counts the numbers of
stones lying in each of the columns and rows. He wants these six numbers
to be pairwise distinct.
At least how many stones does Adrian need to achieve this?
A) 7 B) 8 C) 9 D) 10 E) 11
34
10. Peter received 100 euro for his birthday. He uses all of it to buy exactly
100 objects. What he buys, consists of liquorice wheels of 10 cents a piece,
bouncing balls of 2 euro a piece, and decks of playing cards of 5 euro a
piece.
How many decks of playing cards does Peter buy?
A) 10 B) 11 C) 15 D) 16 E) 19
11. How many integers have the following properties; the digits are pairwise
distinct, they are non-zero, and the sum of the digits is 10?
A) 32 B) 48 C) 56 D) 511 E) 512
12. Eight children together enumerate all of the numbers from 1 up to 2011,
in the following manner.
• Angie enumerates all of the numbers from 1 up to 2011 in groups
of three, skipping the middle number of each triple. So she says:
“1, 3, 4, 6, 7, 9, . . . , 2005, 2007, 2008, 2010, 2011”.
• Ben enumerates all of the numbers Angie skipped, in groups of three,
skipping the middle number of each triple.
• Catherine enumerates all of the numbers that were skipped by both
Angie and Ben, again in groups of three, and skipping the middle
number of each triple.
• Dorothy, Eve, Francis, and Gerald continue in the same way.
• In the end, Henry enumerates the only remaining number.
Which number is enumerated by Henry?
A) 712 B) 1094 C) 1123 D) 1265 E) 1387
13. In this magic square, the three rows, the three columns,
2
and the two diagonals all have the same sum.
Which number is represented by the question mark?
?
A) 5 B) 6 C) 7 D) 10 E) 16 9
35
14. Ten ants walk upwards on a long and thin blade of grass. When an ant
reaches the tip, it turns around. The ants cannot pass each other, so when
they ‘collide’, they both turn around as well. In the end, all ten ants reach
the bottom of the blade of grass safely.
What is the total number of times that an ant turned around?
A) 45 B) 50 C) 90 D) 95 E) 100
Part 2
The answer to each problem is a number.
1. Use the digits 4, 5, 6, and 7 to make two numbers (every digit may only be
used once in total), such that the first number is a multiple of the second
one.
What is the largest possible ratio between these two numbers?
3. A group of 10 boys and 9 girls goes to watch a movie. They all buy a ticket
for 6 euro. Fourteen of these 19 children buy a bag of popcorn. After the
movie finished, it turns out that the combined amount of money that the
boys spent (on the tickets and bags of popcorn), is equal to the combined
amount of money that the girls spent.
Determine all of the possible prices of a bag of popcorn.
36
4. Quintijn writes down four pairwise distinct numbers that sum up to 44.
Of every pair of numbers, he calculates their difference. These are: 1, 3,
4, 5, 6 and 9.
What possibilities are there for the value of the largest number of the four
numbers that were written down? Write down all possibilities.
5. Julian has a big jar of liquorice wheels. Every day, he eats exactly a quarter
of the liquorice wheels that are in the jar at the beginning of the day. At
the end of the third day, less than 100 liquorice wheels remain in the jar.
At most how many liquorice wheels were in the jar at the beginning of the
first day?
6. The number 1782379 has seven digits. Subtract this number from 9999999,
and you get another seven digit number. Put this number at the end of
1782379, giving a fourteen digit number. Divide this number by 9999999.
What is the result?
7. At the beginning of class, all boys are present, but only two girls are
present. After a while, three more girls enter class. This doubles the
percentage of girls in class.
How many boys are in class?
8. We denote by max(a, b) the largest of the numbers a and b. E.g. max(4, −7) =
4.
What is the smallest possible value of max 5 − max(a, 3), a + 3 ?
10. Given a three digit number, we do the following. We multiply all of its
digits, and subtract the three digits one by one from the result.
Find the smallest three digit number such that this procedure gives the
number 221.
37
Solutions
Part 1
1. E) 15 6. C) 99 11. C) 56
1
2. C) 5 7. A) 4 12. B) 1094
3. E) 25 8. C) 5 13. E) 16
4. B) 7 9. B) 8 14. E) 100
Part 2
1. 189 6. 1782380
2. 11 7. 10
4. 15 and 16 9. 5
38
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, February 2011
10 Second Round, March 2011
15 Final Round, September 2011
19 BxMO/EGMO Team Selection Test, March 2012
23 IMO Team Selection Test 1, June 2012
28 IMO Team Selection Test 2, June 2012
33 Junior Mathematical Olympiad, October 2011
MY
We eat problems
CMY
Olympiad 2012
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
sponsored by: Centrum Wiskunde & Informatica sterdam 2011
OmslagNWO2013-v1.indd 1 14-6-2013 14:56:52
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
5 First Round, January 2012
13 Second Round, March 2012
19 Final Round, September 2012
25 BxMO/EGMO Team Selection Test, March 2013
29 IMO Team Selection Test 1, June 2013
35 IMO Team Selection Test 2, June 2013
40 Junior Mathematical Olympiad, October 2012
Those 811 students from grade 5 (4, ≤ 3) that scored 20 (16, 12) points
or more on the first round (out of a maximum of 36 points) were invited
to the second round, which was held in March at twelve 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, ≤ 3) that scored 32 (27, 22) 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. In total 150 students
were invited. They also received an invitation to some training sessions at
the universities, in order to prepare them for their participation in the final
round.
Out of those 150, in total 143 participated in the final round on 14 Septem-
ber 2012 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 51st edition 2012.
1
Among the participants of the training programme, there were some extra
girls, as this year we participated for the second time in the European Girls’
Mathematical Olympiad (EGMO). In total there were nine girls competing
to be in the EGMO team. The team of four girls was selected by a selection
test, held on 8 March 2013. They attended the EGMO in Luxembourg from
8 until 14 April, and the team returned with an honourable mention and a
silver medal. For more information about the EGMO (including the 2013
paper), see www.egmo.org.
The same selection test was used to determine the ten students particip-
ating in the Benelux Mathematical Olympiad (BxMO), held in Dordrecht,
the Netherlands, from 26 until 28 April. The Dutch team managed to
come first in the country ranking, and received five bronze medals, two
silver medals and two gold medals. For more information about the BxMO
(including the 2013 paper), see www.bxmo.org.
In June the team for the International Mathematical Olympiad 2013 was
selected by two team selection tests on 5 and 8 June 2013. A seventh,
young, promising student was selected to accompany the team to the IMO
as an observer C. The team had a training camp in Santa Marta, from 14
July until 21 July.
For younger students the Junior Mathematical Olympiad was held in Oc-
tober 2012 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 consisted
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.
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
Dutch delegation
The Dutch team for IMO 2013 in Colombia consists of
3
Love, fire and mathematics
All four fireflies are equally in love. Therefore each will fly at the same speed. As time passes the flies
will eventually meet in the centre of the tetrahedron. An outburst of passion, lovesickness and
heartbreaks follows.
Love is for fools. If they were smart, they all would directly fly to the centre of the figure. But they are
not. They each long for their own centre of the world: their loved one.
Question:
What is the distance each firefly will travel before he or she reaches the centre of our geometric
form?
Please give the answer in units of the length of an edge of the tetrahedron.
If you have found the solution to this puzzle, please send your answer to wiskunde@transtrend.com.
First Round, January 2012
Problems
A-problems
A2. A palindromic number is a number that does not change when the order of
its digits is reversed, like 707 and 154451. Leon lists all 5-digit palindromic
numbers (numbers do not start with the digit 0) in increasing order.
Which number is 12th on Leon’s list?
A) 11111 B) 11211 C) 12221 D) 12321 E) 12421
A4.
F
G H I J
K L M
5
A4. On the previous page is a paper model of a dipiramid. The unfolded version
is depicted on the left, and the resulting folded dipiramid is depicted on
the right.
Which three points of the unfolded model correspond to vertices of the
grey triangle?
A) F , G and L B) H, I and M C) G, H and I
D) H, K and L E) F , I and K
A5. Frank has a drawer containing single socks. There are 10 red socks, the
other socks are blue. Without looking, he picks a number of socks from
the drawer. The number of socks he needs to pick to be sure of getting
at least two red socks, is twice the number he needs to pick for getting at
least two blue socks.
How many socks does the drawer contain?
A) 14 B) 18 C) 26 D) 32 E) 40
D E C
A6. Point E lies on side CD of a square ABCD. Line
segment AE is divided into four equal parts by points T U
P , R, and T . Line segment BE is divided into four
R S
equal parts by points Q, S, and U . The length of
P Q equals 3. P Q
What is the area of quadrilateral P QU T ? 3
15 17 9
A) 4 B) 4 C) 4 D) 2 E) 5
A B
A7. Carry has six cards. On each card a positive integer number is written.
Carry chooses three cards and calculates the sum of the three numbers on
these cards. Doing this for each of the 20 possible selections of three cards,
she finds the following sums: 16 (ten times) and 18 (also ten times).
What is the smallest number occurring on the cards?
A) 1 B) 2 C) 3 D) 4 E) 5
A8. A sequence of numbers starts as 27, 1, 2012, . . . The sequence has the fol-
lowing property: the numbers in positions 1, 2, and 3 add up to 2040. The
numbers in positions 2, 3, and 4 add up to 2039, the numbers in positions
3, 4, and 5 add up to 2038, and so on. More generally: the numbers in
positions k, k + 1, and k + 2 add up to 2041 − k.
What number is in position 2013?
A) −670 B) −669 C) 670 D) 1341 E) 1342
6
B-problems
The answer to each B-problem is a number.
B1. Consider all 5-digit numbers. Of these numbers, a numbers have the prop-
erty that the product of their digits is 25, and b numbers have the property
that the product of their digits is 15.
Determine ab .
C
B4. For all positive numbers x and y, the operation ¯ obeys the following
three rules:
1
Rule 1: (2x) ¯ y = 2 + (x ¯ y).
2 2
Rule 2: y ¯ x = x ¯ y.
Rule 3: 2 ¯ 2 = 32 .
Compute 32 ¯ 8.
7
Solutions
A-problems
× ∗∗ 7
A1. D) 12 Consider the top figure. The cell 24 56
marked by the star must contain a number that di- 36 8
vides both 27 and 6. Hence this number is either 1 ∗ 27 6
or 3. The first option implies that the cell marked 6 18 42
by the double star contains the number 27. But 36
is not divisible by 27, so this option is off. × 3 9 2 7
Using the second option, we can complete the full 8 24 72 16 56
table and arrive at the bottom figure. Clearly, 12 is 4 12 36 8 28
the largest number occurring more than once. 3 9 27 6 21
6 18 54 12 42
A4. B) H, I and M Different letters may represent the same vertex of the
dipiramid. The five vertices are G = F = J, K = M , H, I and L. Since
H and I both have four neighbours, they represent vertices of the grey
triangle. Also K = M has four neighbours: I, H, L and G = F = J, so that
must be the third vertex of the grey triangle.
8
A5. D) 32 Denote the number of blue socks in the drawer by b. To be
sure of getting two blue socks, Frank must take at least 12 socks. Indeed,
in the worst case he will start by picking the ten red socks. To be sure of
getting at least two red socks, he must take at least b + 2 socks, because
in the worst case he will first pick all b blue socks.
Since we know that this second number is twice the first number, we find
that b + 2 = 2 × 12 = 24. This implies that b = 22. The total number of
socks is therefore 22 + 10 = 32.
9
B-problems
1
B1. 2 The 5-digit numbers with product of the digits equal to 25,
consist of two fives and three ones. There are a of these numbers, where
a is the number of ways to place the three ones (because the positions of
the fives is then fixed as well).
The 5-digit numbers with product of the digits equal to 15, consist of one
three, one five, and three ones. There are b of those numbers, where b
equals the number of ways to first place the three ones, and then place the
digits 3 and 5 in one of two possible ways.
We see that b = 2a, which implies that ab = 21 .
B3. 48 and 60 Let a and b denote the number of cells in the length and
width of the rectangle. We may assume that a ≥ b. Therefore, the total
number of cells in the rectangle is ab and the number of cells at the edges
equals 2a + 2b − 4. Given the fact that half of the cells are at the edge of
the rectangle, we know that ab = 2(2a + 2b − 4). Rearranging terms gives
ab−4a−4b+16 = 8. Factoring the left-hand side, we get (a−4)(b−4) = 8.
Since a and b are positive integers and a ≥ b, the only possibilities are
a − 4 = 8, b − 4 = 1 and a − 4 = 4, b − 4 = 2 (note that a − 4 and b − 4
cannot be negative because in that case b − 4 is at most −4, contradicting
the fact that b is positive). We find a = 12, b = 5 or a = 8, b = 6. This gives
two possible rectangles with 12 × 5 = 60 and 8 × 6 = 48 cells respectively.
10
11 1 2
B4. 2 Using Rule 1 we find: 32 ¯ 8 = 2 + 16 ¯ 8 = 2 + 8¯8 =
3
2 + 4 ¯ 8.
From Rule 2 with y = 2 and x = 8 it follows that : 4 ¯ 8 = 64 ¯ 2.
Repeated application of Rule 1 gives: 64 ¯ 2 = 12 +32 ¯ 2 = 22 +16 ¯ 2 =
. . . = 25 + 2 ¯ 2.
Collecting these results and using Rule 3, we obtain:
3 3 8 11
32 ¯ 8 = 2 + 4¯8 = 2 + 64 ¯ 2 = 2 + 2¯2 = 2 .
11
Faculty of Science
B2. All 2012 camels in the Netherlands are to be distributed among 40 pastures.
No two pastures are allowed to get the same number of camels. The pasture
in the city centre of Amsterdam has to get the largest number of camels.
At least how many camels have to be placed in that pasture?
B3. One of the four dwarfs Anne, Bert, Chris and Dirk stole the king’s gold.
The dwarfs, who know each other very well, each make two statements. If
a dwarf is a liar, then at least one of these two statements is a lie. If the
dwarf is not a liar, then both statements are true.
Anne says: “Bert is a liar.” and “Chris or Dirk stole the gold.”
Bert says: “Chris is a liar.” and “Dirk or Anne stole the gold.”
Chris says: “Dirk is a liar.” and “Anne or Bert stole the gold.”
Dirk says: “Anne is a liar.” and “Bert or Chris stole the gold.”
13
B5. A square ABCD with side length 8 is fol- D A C
C-problems
For the C-problems not only the answer is important; you also
have to describe the way you solved the problem. B
C1. You have one card with the number 12 on it. You are allowed to add new
cards to your collection according to the following rules.
• If you have a card with the number a on it, then you are allowed to
make a new card with the number 2a + 1 on it.
• If you have a card with the number b on it and b is divisible by 3,
then you are allowed to make a new card with the number 3b on it.
(a) Show that you can make a card with the number 29 on it.
(b) Show that you can make make a card with the number 22012 − 1 on
it.
(c) Show that you can never make a card with the number 100 on it.
D
2
E 1 S 2
1 M T
B C
14
Solutions
B-problems
B2. 70 We may put 23 camels in one pasture, and put 32, 33, 34, . . . ,
70 camels in the remaining 39 pastures, respectively. The last pasture is
the one in the centre of Amsterdam. The total number of assigned camels
indeed equals 23+(32+33+· · ·+70) = 23+39· 32+70
2 = 2012. Hence, a valid
distribution having 70 camels in the pasture in the centre of Amsterdam
exists.
It is not possible to do it with less than 70. Indeed, suppose that at most
69 camels are put in the pasture in the centre of Amsterdam. Because the
number of camels is different for each pasture, the second most populated
pasture has at most 68 camels, the third most populated pasture has at
most 67 camels, and so on. In total the pastures have at most 69 + 68 +
. . . + 30 = 40 · 30+69
2 = 1980 camels, leaving at least 32 unassigned camels.
Therefore, no solution exists with fewer than 70 camels in the pasture in
the centre of Amsterdam. The minimum is therefore 70.
B3. 5 First consider the case that Anne has stolen the king’s gold.
The last statement of both Bert and Chris is true in this case, and the last
statement of both Anne and Dirk is false. Hence, Anne and Dirk are liars.
Since both of Chris’ statements are true, Chris is no liar. This implies
that Bert is a liar, because his first statement was a lie. Now that we know
who is a liar and who is not, it is easy to see that exactly five of the eight
15
statements are true.
In the cases where Bert, Chris or Dirk is the thief, a similar reasoning
holds. By the symmetry of the problem (cyclically permuting the names
‘Anne’, ‘Bert’, ‘Chris’, and ‘Dirk’ does not change the problem), exactly
five of the statements are true in each case.
B4. −178 Colour the 10,000 squares in a chess board pattern: the upper
left square and bottom right square will be white and the squares in the
upper right and bottom left will be black. Consider the 99 · 99 = 9801 2×2-
blocks. The 4901 blocks having a white square in the upper left corner are
called the white blocks and the 4900 blocks having a black square in the
upper left corner are called the black blocks. For each white block, add
the four numbers it contains. Let W be the result of adding these 4901
outcomes (some squares are counted twice!). Let Z be the result when
applying the same procedure to the 4900 black blocks. Since the number
of white blocks is one more than the number of black blocks, we have
W − Z = 20.
Now consider for each square how many times it is counted in total. Each
of the four corner squares is counted only in one white block. Each of the
other squares at the edge of the board are counted in exactly one white
block and one black block. Each of the remaining squares is counted in
exactly two white blocks and two black blocks. Considering the difference
W − Z, only the four corner squares are counted, each exactly once. For
the number x in the bottom right corner, we find that x + 0 + 99 + 99 =
W − Z = 20, hence x = −178.
B5. 2
3 Introduce points E, F , and D A C
G as in the figure. Suppose that |DE| = x.
Then |AE| = 8−x, because the square has
sides of length 8. Using the Pythagorean
theorem, we get (8−x)2 = |AE|2 = |DE|2 + E
|AD|2 = x2 + 16. Solving this quadratic
equation gives x = 3. Observe that we
F
have ∠CAF = 180◦ − ∠DAE − ∠EAB =
90◦ − ∠DAE = ∠DEA. Also, ∠ADE = B
90◦ = ∠F CA. Therefore, triangles DEA G
|AF | |EA|
and CAF are similar (AA). This implies that 14 |AF | = 5
|AC| = |DE| = 3 ,
|CF | |AD|
and hence |AF | = 20 1
3 . We also see that 4 |CF | =
4
|AC| = |DE| = 3 and so
|CF | = 3 . It follows that |BF | = 8 − |AF | = 34 . Since ∠CF A = ∠BF G
16
16
are similar (AA). This implies that 34 |BG| = |BG| |AC| 4
|BF | = |CF | = 16 = 43 and
3
therefore |BG| = 1. It follows that the area of the grey triangle equals
1 1 4 2
2 · |BG| · |BF | = 2 · 1 · 3 = 3 .
C-problems
C1. (a) Applying the two rules, we can make the following sequence of cards:
12 → 4 → 9 → 3 → 7 → 15 → 31 → 63 → 21 → 43 → 87 → 29.
(b) We have seen in part (a) how to make a card with the number 3 =
22 − 1. Iterating the first rule, we can sequentially construct cards
with numbers 2 · (22 − 1) + 1 = 23 − 1, 2 · (23 − 1) + 1 = 24 − 1, and
so on. In particular, we can make the number 22012 − 1.
(c) Applying rule 1 to any card produces a card with an odd number.
Applying rule 2 to a card with an odd number (assuming it is a
multiple of three), produces a card with an odd number. As soon as
we apply rule one, the resulting card will only give rise to cards with
an odd number. To get even numbers, we must therefore restrict to
using rule 2 only (repeatedly). The only even numbers that can be
made are therefore 12 and 4.
C2. Observe that ∠ESB = ∠DSC (vertical angles). Since triangles BES
and SDC are isosceles, ∠EBS = ∠ESB = ∠DSC = ∠DCS. Hence
|BS|
triangles BES and SDC are similar (AA). In particular, |BS| = |BE| =
|SC|
|SD|= 12 |SC| = |SM |. Therefore, triangle BSM is isosceles and ∠SBM =
∠SM B = ∠T M C (vertical angles). Using that the angles of a triangle sum
to 180◦ , we find that ∠T M C = 180◦ − ∠M T C − ∠T CM and hence also
∠AT B = 180◦ − ∠M T C = ∠T M C + ∠T CM . We had already shown
that ∠T M C = ∠SBM and ∠T CM = ∠DCS = ∠ABS. We therefore see
that ∠AT B = ∠SBM + ∠ABS = ∠ABT . It follows that triangle BAT is
isosceles with top A, and therefore |AB| = |AT | holds.
17
// I DREAM OF
SOFTWARE EVEN
MY GRANDAD
UNDERSTANDS//
www.tue.nl
Final Round, September 2012
Problems
For these problems not only the answer is important; you also have to describe the way you
solved the problem.
1 2 3
3 1 2
1 2 3
2 3 1
(a) Suppose that n = 5. Can the numbers be placed in such a way that
each row contains the same number of grey cells?
(b) Suppose that n = 10. Can the numbers be placed in such a way that
each row contains the same number of grey cells?
p3 + m(p + 2) = m2 + p + 1
holds.
19
4. We are given an acute triangle ABC and points D on BC and E on AC
such that AD is perpendicular to BC and BE is perpendicular to AC.
The intersection of AD and BE is called H. A line through H intersects
line segment BC in P , and intersects line segment AC in Q. Furthermore,
K is a point on BE such that P K is perpendicular to BE, and L is a
point on AD such that QL is perpendicular to AD.
E
D
Q H
P
L K
A B
20
Solutions
21
3. All terms containing a factor p are brought to the left-hand side of the
equation. In this way we obtain p3 + mp − p = m2 − 2m + 1, or
p(p2 + m − 1) = (m − 1)2 .
E
D
Q H
P
L K
A B
22
5. A sequence satisfying the condition is called ‘good’. We will determine the
number of good sequences in an indirect way, by first counting something
else.
First consider the number of ways in which we can colour the numbers
1 to 12, each number being coloured either red or blue. As there are
two options (red or blue) for each of the 12 numbers independently, there
are 212 possible colourings. We call a colouring ‘good’ if there is at least
one number of each colour, and the largest red number is larger than the
smallest blue number. There are exactly 13 colourings that are not good:
the colouring having only blue numbers, and the twelve colourings in which
the red numbers are precisely the numbers 1 to k, for some k = 1, 2, . . . , 12
(in the case k = 12, there are no blue numbers.) In total, there are
212 − 13 = 4083 good colourings.
We will now show that the number of good sequences is equal to the number
of good colourings, by perfectly matching colourings and sequences. Take a
good sequence and suppose that a is the number smaller than its immediate
predecessor in the sequence. To this sequence we associate the following
colouring: the numbers preceding a in the sequence are coloured red, and
the other numbers are coloured blue. This will be a good colouring.
As an example, consider the good sequence 1, 3, 4, 5, 8, 9, 2, 6, 7, 10, 11, 12.
Hence, a = 2. The numbers 1, 3, 4, 5, 8, and 9 will be coloured red, the
numbers 2, 6, 7, 10, 11, and 12 will be blue. This is a good colouring,
because 9 > 2.
In this way, each good colouring is obtained precisely once from a good
sequence. Indeed, for a given good colouring, we can find the unique
corresponding good sequence as follows: write down the red numbers in
increasing order, followed by the blue numbers in increasing order.
We conclude that there are 4083 sequences satisfying the condition.
23
TTENT
O RK ! ATTENT ION ! MATH A
IO ! M
NA! TMHATH AT W
ATTENTNIO AT WO
RK! AT
TENTI
ON! M
ATH A
T WO
WWW.IMC.NL
BxMO/EGMO Team Selection Test, March 2013
Problems
3. Find all triples (x, n, p) of positive integers x and n and primes p for which
the following holds:
x3 + 3x + 14 = 2 · pn .
for all x, y ∈ R.
5. Let ABCD be a cyclic quadrilateral for which |AD| = |BD|. Let M be the
intersection of AC and BD. Let I be the incentre (centre of the inscribed
circle) of 4BCM . Let N be the second point of intersection of AC and
the circumscribed circle of 4BM I. Prove that |AN | · |N C| = |CD| · |BN |.
25
Solutions
a1 = b + c − a = 671 − 2u.
The difference with 671 is multiplied by -2 after one step. The same holds
for the second and third number of the triple. By using induction we get
3. The left hand side can be factorised as (x + 2)(x2 − 2x + 7). The equation
becomes
(x + 2)(x2 − 2x + 7) = 2pn .
26
First consider the case in which x is even. Then x + 2 = 2pa for a certain
integer a ≥ 0 and x2 − 2x + 7 = pn−a . (Notice that this also holds in the
case in which p = 2.) Because x is a positive integer, we have x+2 ≥ 3 and
we can rule out a = 0. Furthermore we have x2 − 3x + 5 > (x − 23 )2 ≥ 0
for all x, hence x2 − 2x + 7 > x + 2. This yields that n − a > a. We now
substitute x = 2pa − 2 in x2 − 2x + 7 = pn−a :
4. Substituting x = y = 0 gives
f (0) = f (0) − 0 + f f (0) ,
hence f f (0) = 0. Substituting x = y = 1 gives
f 1 + f (1) = f f (1) − 1 + f 1 + f (1) ,
hence f f (1) = 1. Now substitute x = 1 and y = 0 to find
f (1) = f f (0) − 1 + f f (1) .
27
We know that f f (0) = 0 and f f (1) = 1, hence we now find f (1) = 0.
Because f f (1) = 1, this also yields that f (0) = 1. Now substitute y = 0
to get
f (x) = f (x) − x + f f (x) for all x ∈ R,
hence f f (x) = x for all x ∈ R. Substitute x = 1 to get
f (1) = f f (y) − 1 + f (y) for all x ∈ R.
Together with f (1) = 0 and f f (y) = y we now find 0 = y − 1 + f (y),
hence f (y) = 1 − y for all y ∈ R. Now we know for sure that f (x) = 1 − x
for all x ∈ R is the only possible solution to the equation. If we substitute
this in the original equation both the left and right hand side are equal to
1 − x − y + xy.
Hence this function is a solution and it is the only solution.
28
IMO Team Selection Test 1, June 2013
Problems
ab + c + d = 3,
bc + d + a = 5,
cd + a + b = 2,
da + b + c = 6.
4n−2
2. Determine all integers n for which n+5 is the square of a rational number.
29
Solutions
4n−2
2. Suppose that n+5 is the square of a rational number. Then we can
2
write this number as pq2 , where p, q are non-negative integers, q 6= 0, and
gcd(p, q) = 1. As gcd(p, q) = 1, we also have gcd(p2 , q 2 ) = 1, so there
exists a non-zero integer c 6= 0 such that 4n − 2 = cp2 and n + 5 = cq 2 .
This implies that
22 = 4(n + 5) − (4n − 2) = 4cq 2 − cp2 = c (2q)2 − p2 = c(2q − p)(2q + p).
Hence c is a divisor of 22. Note that 2 divides 22, so the right hand side
is divisible by 2. Also note that 2q − p contains a factor 2 if and only if
2q + p does; their difference 2p is even. As 22 contains exactly one factor
2, if follows that 2q − p and 2q + p cannot both be even. Hence they must
be odd, and c must be even.
As p ≥ 0 and q ≥ 1, we have 2q + p ≥ 2, which implies that the factor
11 of 22 must be factor of 2q + p. We deduce that there are only two
possibilities.
30
• c = 2, 2q + p = 11, 2q − p = 1;
• c = −2, 2q + p = 11, 2q − p = −1.
In the first case, we find 2p = 11 − 1 = 10, so p = 5 and q = 3. Hence
52
4n−2 = 50, so n = 13. Indeed, we see that 4n−2 50 25
n+5 = 18 = 9 = 32 , so n = 13
has the desired property. In the second case, we find 2p = 11 − (−1) = 12,
so p = 6, but then q is not an integer. Hence this case cannot occur. We
conclude that n = 13 is the unique integer having the desired property.
31
Hence we can cover a n × n-board by 2n − 1 rectangles. We now show that
any covering with N rectangles must have N ≥ 2n − 1. So fix a covering
of the n × n-board satisfying the conditions, and let N be the number of
rectangles used. Let k be the number of type I rectangles that are not of
type II. Then the number of type II rectangles that are not of type I is also
k. Furthermore, let l be the number of 1 × 1-rectangles. Then N = 2k + l.
If k ≥ n, then N ≥ 2n, in which case there is nothing to prove. Hence
assume that k < n. It then suffices to show l ≥ 2n − 2k − 1.
Every rectangle of type I can only cover squares in a single row. Hence
there are n − k rows in which no square is covered by a rectangle that is
of type I, but not of type II. Similarly, every rectangle of type II can only
cover squares in a single column. Hence there are n − k columns in which
no square is covered by a rectangle that is of type II, but not of type I.
Consider the (n − k)2 squares that lie on such a row and such a column.
They must be covered by 1 × 1-rectangles, hence l ≥ (n − k)2 .
Since (k − n + 1)2 ≥ 0, we have k 2 + n2 + 1 − 2kn + 2k − 2n ≥ 0, so
n2 − 2kn + k 2 ≥ 2n − 2k − 1. Hence l ≥ (n − k)2 ≥ 2n − 2k − 1, as desired.
We conclude that the minimal value of N is 2n − 1.
5. Note that a, b, and c are positive, so the conditions of AM-GM are satisfied
in each of the cases below.
First, apply AM-GM to a2 and 1 to get
√
a2 + 1 ≥ 2 a2 = 2a.
We repeat this for b and c, and add the resulting inequalities. This yields
a2 + b2 + c2 + 3 ≥ 2a + 2b + 2c. (2)
Now we add 2 times (2), 4 times (3), and 1 time (4) to deduce that
32
Note that the left hand side is equal to 3(a2 + b2 + c2 + 2bc + 2ca + 2ab) =
3(a + b + c)2 . For the right hand side, note that 9 = 8 + abc. Hence the
right hand side is equal to (a + 2)(b + 2)(c + 2). Both sides are positive, so
dividing by 3 and taking square roots gives us
q
a + b + c ≥ 13 (a + 2)(b + 2)(c + 2).
33
Optimize your career
ORTEC addresses complex optimization issues in various sectors worldwide. Our employees
assist customers in making well-founded decisions based on mathematical models and the use of
simulation and optimization techniques.
ORTEC is a young, professional organization with great career opportunities. You can work
for ORTEC during or after your studies. You will be involved in projects immediately and given
considerable personal responsibility. We provide a working environment with plenty of room to
develop your talents within your own line of interest, both in the Netherlands and internationally.
Do you like the sound of this? Are you currently pursuing, or have you previously completed a
degree in Econometrics, Operational Research, Computer Science or Mathematics? Are you good
with statistical models and do you feel at home in the world of logistics or finance? Then ORTEC
is the place to be!
ORTEC Logistics
Houtsingel 5
2719 EA Zoetermeer Tell us how your talents will improve our products and services, and how you can help ORTEC
Tel.: 088-678 32 65 continue its international growth. Please visit our website www.ortec.com/carriere for more
088-ORTECNL
recruitment@ortec.com information about a career at ORTEC and an up-to-date overview of our vacancies, internships and
graduation projects. If you do not find your ideal job or subject, please send an open application.
www.ortec.com
IMO Team Selection Test 2, June 2013
Problems
1. Show that
2013
X 4026!
2
n=0 n!(2013 − n)!
is the square of an integer.
i + j ≡ ni + nj
(mod 2)
35
Solutions
We have
(2m)! (m!)2 (2m)! m 2
2m
2 = 2 · = n · m .
n!(m − n)! n!(m − n)! (m!)2
2m
Hence this sum is equal to n . This proves (5).
36
|AB| |BC|
hence |BC| = |CD| . As 4XBY ∼ 4Y CZ, we also have ∠ABC =
∠XBY = ∠Y CZ = ∠BCD. Hence 4ABC ∼ 4BCD (SAS). We de-
duce that ∠CAB = ∠DBC. From this, it follows that
or equivalently,
(t + 3)aq = (q + 3)at .
In particular, we have 5a3 = 6a2 and 5a7 = 10a2 . Substituting k = 3 and
p = 2 now gives
6
a7 = 2a3 − 3a2 + 13 = 2 · a2 − 3a2 + 13.
5
10
As a7 = 5 a2 , this implies that
13
a2 = 13,
5
(p+3)a2
hence a2 = 5. We deduce that for all primes p, we have ap = 5 =
p + 3.
Substituting k = 4 and p = 3 gives
37
Finally, substitute k = 4 and p = 503 to get
as desired.
4. We first show that n satisfies the condition if and only if ni ≡ i+1 mod 2
for all i such that0 ≤ i ≤ n. Suppose that ni ≡ i+1 mod 2 for all i, then
we have ni + nj ≡ i + 1 + j + 1 ≡ i + j mod 2 for all i and j. Hence n
satisfies
the condition. Conversely,
suppose that n satisfies the condition.
As n0 = 1, we have i ≡ 1 + ni mod 2 for all i, so ni ≡ i − 1 ≡ i + 1
mod 2 for all i.
Write n as n = 2k + m with 0 ≤ m < 2k . Since n ≥ 2, we may assume
that k ≥ 1. Consider
k
(2k + m)(2k + m − 1) · · · (m + 4)(m + 3)
n 2 +m
k
= k
= .
2 −2 2 −2 (2k − 2)(2k − 3) · · · 2 · 1
k k
The product in the denominator has b 2 2−2 c factors divisible by 2, b 2 4−2 c
k
−2
factors divisible by 4, . . . , b 22k−1 c factors divisible by 2k−1 and no factors
divisible by 2 . The product in the numerator consists of 2k −2 consecutive
k
k k
factors, hence has at least b 2 2−2 c factors divisible by 2, at least b 2 4−2 c
k
−2
factors divisible by 4, . . . , at least b 22k−1 c factors divisible by 2k−1 . We
deduce that the number of factors 2 of the product in the numerator is
at least that of the one in the denominator. If 2k occurs as factor of the
numerator, then the number of factors 2 of the product in the numerator
is greater than that of the one in the denominator. This is the case if
m + 3 ≤ 2k , i.e. if m ≤ 2k − 3. So if m ≤ 2k − 3, then 2kn−2 is even,
whereas 2k − 2 is even as well. Hence n = 2k + m does not satisfy the
condition.
We deduce that n can only satisfy the condition if there exists a k ≥ 2
such that 2k − 2 ≤ n ≤ 2k − 1. If n is odd, then n0 + n1 = 1 + n ≡ 0
mod 2, so n does not satisfy the condition. This implies that n can only
38
satisfy the condition if n is of the form n = 2k − 2 with k ≥ 2 is. Suppose
that n is of that form. We show that n satisfies the condition. We have
the following.
k
(2k − 2)(2k − 3) · · · (2k − c)(2k − c − 1)
2 −2
= .
c c · (c − 1) · · · 2 · 1
39
Junior Mathematical Olympiad, October 2012
Problems
Part 1
1. A bag contains red and blue marbles. Of these marbles 53 is blue, the rest
is red.
If we double the number of red marbles in the bag, what portion of the
marbles will be blue?
1 3 4 3 4
A) 5 B) 7 C) 7 D) 5 E) 5
2. Alyssa sums odd numbers. First she takes 1, then she calculates 1 + 3,
next she calculates 1 + 3 + 5, then 1 + 3 + 5 + 7, etc. In this way she obtains
a very long list of numbers. The largest number that she calculates is
1 + 3 + 5 + 7 + · · · + 197 + 199.
How many of the numbers on Alyssa’s list end with a 4?
A) 9 B) 10 C) 19 D) 20 E) 50
3. We have got a dice and we will colour its pips. For each of the six faces we
choose one of the colours red, white and blue and colour all pips on that
face in that colour. There is one rule: in each vertex we must choose three
different colours for the three adjacent faces.
In how many ways can we colour the dice?
A) 3 B) 6 C) 12 D) 27 E) 36
40
5. Ans, Ben, Carla and Dirk participate to a lottery. In a hat there are eight
cards with the numbers 1 to 8. One by one each person takes two cards
out of the hat such that Dirk takes the last two cards. To make it even
more exciting, everybody adds his or her two numbers and tells the result
to the others. For Ans this is 10, for Ben 14 and for Carla 5.
What are the numbers that are on Dirk’s cards?
A) 1 and 6 B) 1 and 7 C) 2 and 5 D) 2 and 6 E) 3 and 4
7. How many zeros does the number that is the result of the division
1010101010101010101 : 101
contain?
A) 8 B) 9 C) 10 D) 11 E) 12
8. Jan and Katrijn live on a straight dike, exactly 11,7 kilometres from each
other. Jan cycles to Katrijn. On each place on the road he can see how
far he is from his own house and how far is he from Katrijn’s house. Both
distances we round to whole kilometres. On some part of the dike this
rounded distances are equal.
How many kilometres long is this part?
A) 0,1 B) 0,3 C) 0,5 D) 0,7 E) 0,9
41
9. Anne draws a triangle of which the sides have have different lengths and
the shortest side is 6 cm long. Bert redraws Anne’s triangle, but he draws it
twice as big (all sides twice as long). Christa also redraws Anne’s triangle,
but she draws it three times as big. It appears that Anne’s triangle has
a side that has the same length as one of the sides of Bert’s triangle and
that Bert’s triangle has a side that has the same length as one of the sides
of Christa’s triangle.
How many centimetres long is the longest side of Anne’s triangle?
A) 12 B) 18 C) 24 D) 30 E) 36
a b c
11. Of a 3×3-board with squares a1 till c3 three squares
are painted black. This is done in such a way that 1
no two black fields are neighbouring fields (i.e. shar-
2
ing an edge). In the figure there is one way to do
this.
3
In how many ways can this be done?
A) 14 B) 18 C) 20 D) 22 E) 30
12. While Snow White is cooking, the seven dwarfs play chess. Doc plays
against all other six dwarfs. Happy plays against five others, Grumpy
against four others, Sneezy against three others, Bashful again two others
and Sleepy plays against one other dwarf, after which he falls asleep.
Against how many dwarfs does Dopey, the seventh dwarf, play?
A) 0 B) 1 C) 3 D) 5 E) 6
42
13. In a regular hexagon ABCDEF the diagonals are C B
drawn as in the figure. The area of the hexagon
ABCDEF is 1.
What is the area of the grey hexagon?
D A
1 1 1 3 5
A) 2 B) 3 C) 4 D) 8 E) 16
E F
14. Peter has some numbers of which the average is 20. If he leaves out the
smallest number, the average of the other numbers will be 22. If he leaves
out the biggest number, then the average of the others is 13. If he leaves
out both the smallest and the biggest number, then the average becomes
14.
How many numbers does Peter have?
A) 4 B) 5 C) 6 D) 7 E) 8
15. Ann, Bo, Cas, Dex and Eva cycle in the same direction on a long road,
each with a constant speed. Their speeds are different and for each pair
there is a moment when they cycle exactly next to each other and the one
takes over the other. After the trip Ann tells: “First I took over Bob, then
I got taken over by Cas. Then I took over Dex and finally I got taken over
by Eva.”
Bo tells: “I got taken over by Eva. That was after the moment when I
cycled next to Ann, but before the moment I cycled next to Cas or Dex.”
Put the five in order from fastest to slowest.
A) Eva, Cas, Ann, Bo, Dex D) Cas, Eva, Ann, Dex, Bo
B) Cas, Eva, Ann, Bo, Dex E) This cannot be determined
C) Eva, Cas, Ann, Dex, Bo
43
Part 2
The answer to each problem is a number.
2. Huey, Dewey and Louie are going to run. They start at the same time
and each of them walks at constant speed. When Huey finished three laps,
Dewey is exactly halfway his third lap. When Dewey finished his third
lap, Louie is exactly halfway his third lap. After a while Huey, Dewey and
Louie are together at the start at the same moment.
How many laps did Louie finish at this moment?
3. What is the largest number of 2×2×1-bricks that one can put in a box of
7×7×6? The bricks must be aligned parallel to the sides of the box.
4. Anne is completely broke, but today (day 1) she got a new job. She works
every day (even during the weekends) and after every shift, she is always
paid 13 euro. After every shift, she spends the money on groceries. On
the first day Anne spends 1 euro on groceries, on the second day 2 euro,
on the third day 3 euro, etc.
What is the first day on which Anne cannot afford her groceries anymore?
5. Ronald has bought a large barrel of water. The first day of the year, 1
January 2012, he uses half of it. The second day he uses a third of the
remaining water, and on the third day a fourth of the then remaining
water, and so on until the 365th day of the year. On the last day of the
year (2012 has 366 days) he has only one litre of water left and he drinks
it.
How many litres of water did the barrel contain when Ronald bought it?
44
6. To a three digit number you add its three digits. For example, the number
216 becomes 216 + 2 + 1 + 6 = 255.
What is the largest three digit number that you cannot make in this way?
For example with 41 you get: 41×41 = 1681, 1681 − 6×41 = 1681 − 246 =
1435, 1435 + 9 = 1444. If the result is equal to the initial number, but
then with the two digits interchanged, followed by two zeros, then we call
the initial number great. For example, if the result for 41 would have been
equal to 1400, then 41 would have been great. There is exactly one great
two digit number.
What is that number?
45
Solutions
Part 1
3
1. B) 7 6. E) 1, 2, 3 and 4 11. D) 22
2. D) 20 7. E) 12 12. C) 3
1
3. B) 6 8. D) 0,7 13. B) 3
48
4. D) 5 9. A) 12 14. D) 7
17
5. C) 2 and 5 10. D) 8 15. B) C, E, A, B, D
Part 2
1. 200 6. 995
2. 25 7. DBAC
3. 72 8. 63
4. 26 9. 10
5. 366 10. 81
46
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
5 First Round, January 2012
13 Second Round, March 2012
19 Final Round, September 2012
25 BxMO/EGMO Team Selection Test, March 2013
29 IMO Team Selection Test 1, June 2013
35 IMO Team Selection Test 2, June 2013
40 Junior Mathematical Olympiad, October 2012
MY
We eat problems
CMY
Olympiad 2013
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2013
10 Second Round, March 2013
16 Final Round, September 2013
21 BxMO/EGMO Team Selection Test, March 2014
25 IMO Team Selection Test 1, June 2014
30 IMO Team Selection Test 2, June 2014
36 Junior Mathematical Olympiad, October 2013
Those 800 students from grade 5 (4, ≤ 3) that scored 24 (21, 18) points or
more on the first round (out of a maximum of 36 points) were invited to
the second round, which was held in March at twelve universities in the
country. This round contained five open 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, ≤ 3) that scored 26 (24, 21) 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. In total 149 students
were invited. They also received an invitation to some training sessions at
the universities, in order to prepare them for their participation in the final
round.
Out of those 149, in total 143 participated in the final round on 13 Septem-
ber 2013 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 52nd edition 2013.
1
Among the participants of the training programme, there were some extra
girls, as this year we participated for the third time in the European Girls’
Mathematical Olympiad (EGMO). In total there were eight girls competing
to be in the EGMO team. The team of four girls was selected by a selection
test, held on 21 March 2014. They attended the EGMO in Antalya, Turkey
from 10 until 16 April, and the team returned with a gold and a bronze
medal. For more information about the EGMO (including the 2014 paper),
see www.egmo.org.
The same selection test was used to determine the ten students particip-
ating in the Benelux Mathematical Olympiad (BxMO), held in Brugge,
Belgium, from 2 until 4 May. The Dutch team managed to come first in
the country ranking, and received two bronze medals, two silver medals
and two gold medals. For more information about the BxMO (including
the 2014 paper), see www.bxmo.org.
In June the team for the International Mathematical Olympiad 2014 was
selected by two team selection tests on 6 and 7 June 2014. A seventh,
young, promising student was selected to accompany the team to the IMO
as an observer C. The team had a training camp in Cape Town, from 28
June until 6 July.
For younger students the Junior Mathematical Olympiad was held in Oc-
tober 2013 at the VU University Amsterdam. The students invited to
participate in this event were the 70 best students of grade 1, grade 2 and
grade 3 of the popular Kangaroo math contest. The competition consisted
of two one-hour parts, one with eight multiple choice questions and one with
eight open 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.
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
Dutch delegation
The Dutch team for IMO 2014 in South Africa consists of
3
First Round, January 2013
Problems
A-problems
A1. A traffic light is alternately green and red. The periods green and red
are equally long and always of the same length: either 1, 2, or 3 minutes.
There are four combinations for the colour of the light at 12:08 pm and at
12:09 pm: red–red, red–green, green–red, and green–green.
How many of these four combinations are possible, given that the light is
red at 12:05 pm and also red at 12:12 pm?
A) 1 B) 2 C) 3
D) 4 E) The light cannot be red at both times.
D C
A2. The rectangle ABCD is divided into five equal rect-
angles. The perimeter of each of these small rect-
angles is 20.
What is the area of rectangle ABCD?
A) 72 B) 112 C) 120 D) 140 E) 150
A B
a + b + 1 = b + c − 2 = c + d + 3 = d + e − 4 = e + a + 5.
4
A5. Out of a shipment of boxes, one fourth is empty. We open one fourth of
all boxes and notice that one fifth of them is non-empty.
Which part of the unopened boxes is empty?
4 1 1 1 1
A) 15
B) 4
C) 15
D) 16
E) 20
A6. A regular hexagon and an equilateral triangle have the same perimeter.
What is the ratio area hexagon : area triangle?
A) 2 ∶ 3 B) 1 ∶ 1 C) 4 ∶ 3 D) 3 ∶ 2 E) 2 ∶ 1
A8. Twenty students did a test. No two students answered the same number
of questions correctly. Each question was answered correctly by at most
three students.
What is the smallest number of questions that the test could have had?
A) 63 B) 64 C) 67 D) 70 E) 71
5
B-problems
The answer to each B-problem is a number.
B1. What is the smallest positive integer consisting of the digits 2, 4 and 8,
such that each digit occurs at least twice and the number is not divisible
by 4?
D C
B2. A rectangle ABCD has sides of length a and b, where 4
a < b. The lines through A and C perpendicular
5
to the diagonal BD divide the diagonal into three a
segments of lengths 4, 5, and 4. 4
Calculate ab .
A b B
B3. A bus calls at three stops. The middle bus stop is equally far from the
first stop as from the last stop. Fred, standing at the middle bus stop, has
to wait for 15 minutes for the bus to arrive. If he cycles to the first stop,
he will arrive there at the same time as the bus. If instead he runs to the
last stop, he will also arrive there at the same moment as the bus.
How long would it take Fred to cycle to the last stop and then run back
to the middle stop?
B4. We write down the numbers from 1 to 30 000 one after the other to form
a long string of digits:
123456789101112 . . . 30000.
6
Solutions
A-problems
A1. B) 2 It is given that the light is red at 12:05 pm. The colours at
times 12:05 pm to 12:12 pm are now fixed for a traffic light of period 1:
they are alternately red and green. When the period is 2, there are two
possibilities and for period 3 there are three possibilities:
period 12:05 12:06 12:07 12:08 12:09 12:10 12:11 12:12
1 min red green red green red green red green
2 min red red green green red red green green
2 min red green green red red green green red
3 min red red red green green green red red
3 min red red green green green red red red
3 min red green green green red red red green
In three of the six cases, the light is red at 12:12 pm, as required. This
gives us two colour combinations for the light at 12:08 pm and at 12:09
pm: red–red and green-green.
A2. C) 120 Twice the length of a small rectangle equals three times its
width. Therefore, the ratio between length and width equals 3 ∶ 2. As the
perimeter is 20, the length must be 6 and the width must be 4. We see
that the area of each small rectangle equals 6 × 4 = 24, hence the area of
rectangle ABCD is 5 × 24 = 120.
Every sum of four of the numbers can be obtained by adding two of the
pairs. For example, a+b+c+e is equal to (e+a)+(b+c). Of all four-tuples,
a, c, d, e has the smallest sum, because a + c + d + e = (e + a) + (c + d) is the
sum of the two smallest pairs. The remaining number b must be the largest
among the five numbers. Indeed, the largest number is the one for which
the remaining four numbers have the smallest sum.
A4. A) 3 The order in which the moves are made is irrelevant for the
final result. By pressing the three bulbs in the top row, all bulbs will change
7
from being on to being off. Indeed, the bulbs in the top row change their
state three times, and the other bulbs change their state exactly once.
It is not possible to turn off all light bulbs by pressing two or fewer bulbs,
because some bulb will not be in the same row or column as the chosen
bulbs, and hence remain on.
1
A5. C) 15 Without changing the problem, we may assume that there
are 20 boxes. Hence, a total of 5 boxes is empty. Out of the 5 boxes
that are opened, one fifth turns out to be non-empty, exactly 1 box. As
4 of the opened boxes are empty, there is exactly 1 empty box among the
remaining 15 unopened boxes.
51 = 0005 55 = 3125
52 = 0025 56 = 1 5625
53 = 0125 57 = 7 8125
54 = 0625 58 = 39 0625
The last four digits of a power of 5 are already determined by the last
four digits of the previous power of 5. For example: the last four digits of
5 × 390625 and 5 × 0625 are both 3125. Because the last four digits of 54
and 58 are the same, the last four digits of powers of 5 will repeat every
four steps. The last four digits of 52013 will be the same as those of 52009
and those of 52005 , continuing all the way down to the last four digits of
55 = 3125. We conclude that 3125 are the last four digits of 52013 .
8
total. Because every question was answered correctly at most three times,
there must be at least 190
3
= 63 31 questions. That is, there were at least 64
questions.
B-problems
3
B2. 2
Triangles ADF , BDA en BAF are
similar (AA). Hence
D C
b ∣BA∣ ∣AF ∣ ∣BF ∣
= = = .
a ∣DA∣ ∣DF ∣ ∣AF ∣ F
a
Therefore
b 2 ∣AF ∣ ∣BF ∣ ∣BF ∣ 9
( ) = ⋅ = = .
a ∣DF ∣ ∣AF ∣ ∣DF ∣ 4 A b B
We conclude that b
a
= 32 .
B3. 30 minutes The time needed by Fred to bike from the middle stop to the
last stop and run back to the middle stop, is equal to the time he needs to
bike from the middle stop to the first stop plus the time he needs to run
from the middle stop to the last stop. This is because the distance to both
stops is the same.
It is given that this amount of time equals the time the bus needs to get to
the first stop plus the time it needs to get to the last stop. This is exactly
twice the time the bus needs to get to the middle stop: 2 × 15 = 30 minutes.
9
Second Round, March 2013
Problems
B-problems
The answer to each B-problem is a number.
B1. A number of students took a test for which the maximum possible score
was 100 points. Everyone had a score of at least 60 points. Exactly five
students scored the maximum of 100 points. The average score among the
students was 76 points.
What is the minimum number of students that could have taken the test?
D C
A B
B3. Consider two clocks, like the ones in the figure
12 12
on the right, whose hands move at a constant 11 1 11 1
10 2 10 2
speed. Both clocks are defective; the hands
9 3 9 3
of the first clock turn at a pace that is 1%
8 4 8 4
faster than it should be, while the hands of 7 5 7 5
6 6
the second clock turn at a pace that is 5% too
fast. At a certain moment, both clocks show a time of exactly 2 o’clock.
Some time passes until both clocks again show exactly the same time.
At that moment, what time do the clocks show?
10
B5. A regular hexagon is divided into seven parts by lines
parallel to its sides, see the figure. Four of those 11
C-problems
For the C-problems not only the answer is important; you also 5
2 ⋅ 2 ≥ 3 + 1, 2 ⋅ 5 ≥ 1 + 2, and 2 ⋅ 4 ≥ 2 + 5.
(a) Give a 4-digit number that is above average and has ‘4’ as its first
digit.
(b) Show that no 4-digit number that is above average has ‘4’ as its second
digit.
(c) For 7-digit numbers that are above average, determine all possible
positions of the digit ‘7’.
C2. We will call a triple (x, y, z) good if x, y, and z are positive integers such
that y ≥ 2 and the equation x2 − 3y 2 = z 2 − 3 holds.
An example of a good triple is (19, 6, 16), because 6 ≥ 2 and 192 − 3 ⋅ 62 =
162 − 3.
(a) Show that for every odd number x ≥ 5 there are at least two good
triples (x, y, z).
(b) Find a good triple (x, y, z) with x being even.
11
Solutions
B-problems
ACD which is 21 ⋅ 4 ⋅ 4 = 8.
A B
B3. 5 o’clock In 12 hours, the amounts by which the hour hands of the
1 5
first and the second clock are ahead increase by 100 and 100 of a turn
respectively. Therefore, in 12 hours, the second clock increases its lead
compared to the first clock by 5−1
100
= 25
1
of a full turn. Hence after 25 ⋅ 12
hours, the hour hand of the second clock has made exactly one extra full
turn compared to that of the first clock. This is the first time the two
clocks again display the same time.
During those 25⋅12 hours, the hour hand of the first clock has made exactly
101
100
⋅ 25 = 25 41 full turns. Both clocks will then display a time of 2 + 3 = 5
o’clock.
12
1 1
B4. 4
The product of the eight numbers in 2 32 8 1
the second and fourth row equals the product of the 4 2 8 2
eights numbers in the first and second column. Writ- 4 1 8 4
ing this out, we get:
16 2 14 16
C ⋅ 2 ⋅ 8 ⋅ 2 ⋅ F ⋅ G ⋅ H ⋅ 16 = 1
2
⋅ C ⋅ 4 ⋅ F ⋅ 32 ⋅ 2 ⋅ 1 ⋅ G.
Since C, F , and G are nonzero, we may divide both sides of the equation
by C, F , and G. The resulting equation is 512 ⋅ H = 128, which implies
that H = 41 . The figure shows one solution.
13
C-problems
C1. (a) The number 4132 starts with a ‘4’ and is above average because 2 ⋅ 3 ≥
4 + 1 and 2 ⋅ 2 ≥ 1 + 3.
(b) Suppose that a4bc is a 4-digit number that is above average, where a,
b, and c are the digits ‘1’, ‘2’, and ‘3’ (possibly in a different order).
Then 2 ⋅ b ≥ a + 4 ≥ 5. So b ≥ 3.
Similarly, we find that 2 ⋅ c ≥ 4 + b ≥ 7, hence c ≥ 4. However, this is
impossible because c was at most 3.
(c) The numbers 1243756, 1234576, and 1234567 are above average and
have digit ‘7’ in the fifth, sixth, and seventh position, respectively.
Digit ‘7’ cannot be in the first position. Indeed, suppose that 7abcdef
would be above average. Then 2 ⋅ b ≥ 7 + a ≥ 8, hence b ≥ 4. Then we
must have 2 ⋅ c ≥ a + b ≥ 5, hence c ≥ 3. Now we find (in turn) that
also d, e, f ≥ 4. It follows that both digit ‘1’ and digit ‘2’ must be in
the position of a, which is impossible.
Digit ‘7’ cannot be in the second or third position. Indeed, otherwise
the digit following ‘7’ must be at least 4, which implies that also the
digits following it must be at least 4. Digits ‘1’, ‘2’, and ‘3’ must
therefore all be in the first two positions, which is impossible.
Finally, digit ‘7’ cannot be in the fourth position. Digit ‘1’ cannot be
in the third position since 2 ⋅ 1 < 2 + 3. Because the digit in the third
position must be at least 2, the digit in the fifth position must be at
least 5. The next digit must therefore be at least 6, as must be the
digit following it. The digits ‘1’ to ‘4’ must therefore all be in the first
three positions, which is impossible.
14
C2. (a) Since x ≥ 5 is odd, we can write x = 2n + 1 for an integer n ≥ 2. Now
are two different good triples. In both cases it is clear that y and
z are indeed positive integers and that y ≥ 2. Substitution into the
equation shows that they are indeed solutions:
15
Final Round, September 2013
Problems
For these problems not only the answer is important; you also have to describe the way you
solved the problem.
4. For a positive integer n the number P (n) is the product of the positive
divisors of n. For example, P (20) = 8000, as the positive divisors of 20 are
1, 2, 4, 5, 10 and 20, whose product is 1 ⋅ 2 ⋅ 4 ⋅ 5 ⋅ 10 ⋅ 20 = 8000.
(a) Find all positive integers n satisfying P (n) = 15n.
(b) Show that there exists no positive integer n such that P (n) = 15n2 .
16
Solutions
17
2. The first equation yields z = x+y +1. Substitution into the second equation
gives x2 − y 2 + (x + y + 1)2 = 1. Expanding gives 2x2 + 2xy + 2x + 2y = 0, or
2(x + y)(x + 1) = 0. We deduce that x + y = 0 or x + 1 = 0. We consider the
two cases.
3. First, we prove that some triangles in the figure are isosceles (the top angle
coincides with the middle letter).
(1) Triangle ADC is isosceles, because ∠DAC = ∠ACB = ∠ACD. The
first equality holds because AD and BC are parallel and the second
equality follows from the fact that AC is the interior angle bisector
of angle BCD.
(2) Triangle DAO is isosceles, because ∣AD∣ = ∣CD∣ = ∣AO∣. The first
equality follows from (1) and the second equality is given in the prob-
lem statement.
(3) Triangle BCO is isosceles, because it is similar to triangle DAO (hour-
glass shape).
(4) Triangle COD is isosceles, because ∣DO∣ = ∣BC∣ = ∣CO∣. Here the first
equality is given in the problem statement and the second one follows
from (3).
(5) Triangle BDC is isosceles, because it is similar to triangle BCO as
two pairs of corresponding angles are equal: ∠DBC = ∠OBC and
∠BDC = ∠DCO = ∠OCB. Here the last equality follows from (4).
(6) Triangle ADB is isosceles, because ∣AD∣ = ∣CD∣ = ∣BD∣ because of (1)
and (5).
Denote ∠ACB = α and ∠CBD = β (in degrees). From (5) it follows that
2α = β. From (3) it follows that 180○ = 2β + α = 5α, hence α = 180 = 36○
○
5
and β = 72 . In the isosceles triangle ADB ○the ○ top angle is equal to
○
∠ADB = β, hence its equal base angles are 180 2−72 = 54○ . The requested
angle therefore equals ∠ABC = ∠ABD + ∠DBC = 54○ + 72○ = 126○ .
18
A D
α β
α
β
O
β
α
β α
B C
◻
19
5. To illustrate the idea, we first calculate the number of fives in the result s
of the sum 1 + 10 + 19 + ⋯ + 100000. First notice that each term in the sum
is a multiple of nine plus 1:
s = 1 + (1 + 9) + (1 + 2 ⋅ 9) + ⋯ + (1 + 11111 ⋅ 9).
The number of terms in the sum is 11111 + 1 = 11112 and the average value
of a term is 1+100000
2
. It follows that s = 11112 ⋅ 100001
2
= 11112
2
⋅ 100001.
Because 2 = 5⋅2 = 10 = 5556, we find that s = 5556 + 555600000 =
11112 5⋅11112 55560
S = 1 + (1 + 9) + (1 + 2 ⋅ 9) + ⋯ + (1 + n ⋅ 9)
1+101023
has exactly n + 1 terms, with an average value of 2
. Hence S =
n+1
2
⋅ (1 + 102013 ).
n+1
Calculating the fraction 2
gives:
n + 1 5n + 5 555 . . . 560
= = = 555 . . . 56,
2 10 10
a number with 2011 fives followed by a 6. Because the last 2013 digits of
the number 102013 ⋅ n+1
2
are all zeroes, there is no ‘overlap’ between the
non-zero digits of 2 and 102013 ⋅ n+1
n+1
2
. We deduce that
S = n+1
2
⋅ (1 + 102013 )
= n+1
2
+ 102013 ⋅ n+1
2
= 55 . . . 56055 . . . 56.
20
BxMO/EGMO Team Selection Test, March 2014
Problems
1. Find all non-negative integers n for which there exist integers a and b such
that n2 = a + b and n3 = a2 + b2 .
4. Let m ≥ 3 and n be positive integers such that n > m(m − 2). Find the
largest positive integer d such that d ∣ n! and k ∤ d for all k ∈ {m, m +
1, . . . , n}.
5. Let n be a positive integer. Daniël and Merlijn are playing a game. Daniël
has k sheets of paper lying next to each other on a table, where k is a
positive integer. On each of the sheets, he writes some of the numbers
from 1 up to n (he is allowed to write no number at all, or all numbers).
On the back of each of the sheets, he writes down the remaining numbers.
Once Daniël is finished, Merlijn can flip some of the sheets of paper (he is
allowed to flip no sheet at all, or all sheets). If Merlijn succeeds in making
all of the numbers from 1 up to n visible at least once, then he wins.
Determine the smallest k for which Merlijn can always win, regardless of
Daniël’s actions.
21
Solutions
22
4. We prove that d = m − 1 is the largest integer satisfying the conditions.
First note that m − 1 ∣ n! and that for k ≥ m we have k ∤ m − 1, so d = m − 1
indeed satisfies the conditions.
Now suppose that for some d we have d ∣ n! and k ∤ d for all k ∈ {m, m +
1, . . . , n}. We prove that d ≤ m − 1. Write d = p1 p2 ⋯pt , where the pi
are prime for all i (but not necessarily pairwise distinct). If t = 0, then
d = 1 ≤ m − 1, so we may assume that t ≥ 1. From the first condition on d,
it follows that pi ≤ n for all i. From the second condition on d, it follows
that pi ∈/ {m, m + 1, . . . , n} for all i. Hence pi ≤ m − 1 for all i. Now consider
the integers p1 , p1 p2 , . . . , p1 p2 ⋯pt . These are divisors of d and hence all are
not in {m, m + 1, . . . , n}. Moreover, we know that p1 ≤ m − 1. Consider the
largest j ≤ t such that p1 p2 ⋯pj ≤ m − 1. If j < t, then
23
Now suppose that 2k ≤ n. Then Daniël first fills the boxes with sets of
chips, in such a way that every possible set of colours occurs in some box.
This is possible, since there are 2k possible sets of colours, and we have at
least 2k boxes. Now Daniël writes on the front of each sheet of coloured
paper precisely those numbers whose boxes contain the colour of that sheet,
and on the back the remainder of the numbers. In this way, the chips in
the boxes correspond to the numbers on the fronts of the sheets. Now
Merlijn cannot choose a set of colours not occurring in any of the boxes;
he cannot win. ◻
24
IMO Team Selection Test 1, June 2014
Problems
a2 + b ∣ a2 b + a and b2 − a ∣ ab2 + b.
3. Let a, b and c be rational numbers for which a + bc, b + ac and a + b are all
non-zero and for which we have
1 1 1
+ = .
a + bc b + ac a + b
√
Prove that (c − 3)(c + 1) is rational.
4. Let △ABC be a triangle with ∣AC∣ = 2∣AB∣ and let O be its circumcentre.
Let D be the intersection of the angle bisector of ∠A and BC. Let E be
the orthogonal projection of O on AD and let F ≠ D be a point on AD
satisfying ∣CD∣ = ∣CF ∣. Prove that ∠EBF = ∠ECF .
25
Solutions
a2 + b ∣ (a2 b + a) − b(a2 + b) = a − b2 .
b2 − a ∣ (ab2 + b) − a(b2 − a) = b + a2 .
26
3. We have
1 1 (b + ac) + (a + bc) a + b + ac + bc
+ = = .
a + bc b + ac (a + bc)(b + ac) ab + a2 c + b2 c + abc2
Hence, the problem statement’s equality yields
or equivalently,
or equivalently,
a2 + 2abc + ab − abc2 + b2 = 0.
We can consider this as a quadratic equation in a, of which we know that
it has a rational solution. Hence, the discriminant of this equation must
be the square of a rational number. This discriminant equals
27
on a chord of this circle is the midpoint of that chord. Let M be the
midpoint of AC. Because ∣AB∣ = ∣AC∣
2
= ∣AM ∣ and because AD is the angle
bisector of ∠BAM , we have that M is the image of B under reflection
in AD. Now we have ∠DGC = ∠AGC = ∠ABC = ∠ABD = ∠DM A =
180○ − ∠DM C, hence DM CG is a cyclic quadrilateral. Then we have
AM 2 = AM2⋅AC = AD⋅AG2
= AD ⋅ AE, hence △AM E ∼ △ADM (SAS).
Now we have 180○ − ∠EM C = ∠EM A = ∠M DA = ∠BDA = ∠CDF =
∠DF C = ∠EF C, hence EM CF is a cyclic quadrilateral. Hence, ∠EBF =
∠EM F = ∠ECF , which is exactly what we needed to prove. ◻
5. Number the rows from 1 up to 2014 and also number the columns. Consider
the following beginning situation: in row i the light bulbs in columns i,
i + 1, . . . , i + 1005 are on and the rest are off, in which we take the column
numbers modulo 2014. In each row and in each column there are exactly
1006 light bulbs that are on. Hence, there is no move possible. It is not
always possible to get to a situation in which less than 2014 ⋅ 1006 light
bulbs are on.
Now we shall show that it is always possible to get to a situation in which
at most 2014 ⋅ 1006 are on. Suppose, from the contrary, that in a certain
situation at least 2014 ⋅ 1006 + 1 light bulbs are on and it is not possible to
get to a situation in which less bulbs are on. If there is a row or column in
which at least 1008 bulbs are on, then we can change the states of these
bulbs in this row or column and there will be less bulbs that are on. This
is a contradiction, hence in each row and column at most 1007 bulbs are
on.
Let I be the set of rows in which exactly 1007 bulbs are on and let J1
be the set of columns in which exactly 1007 bulbs are in. We will try to
change the state of the bulbs in all rows in I. This we call the big plan.
If after executing the big plan there is a column in which at least 1008
bulbs are on, then we get a contradiction. Hence we shall assume that this
does not happen. Let J2 be the set of columns that, after executing the
big plan, contain exactly 1007 bulbs that are on. If there exists a square
(i, j) with i ∈ I and j ∈ J1 containing a bulb that is off, we can switch row
i and then column j gets more than 1007 bulbs that are on, which is a
contradiction. Hence every bulb on (i, j) with i ∈ I and j ∈ J1 is on. If
there exists a square (i, j) with i ∈ I and j ∈ J2 containing a bulb that is
off, after executing the big plan, then we get a contradiction in the same
way. Hence every bulb on (i, j) with i ∈ I and j ∈ J2 is off after executing
the big plan. Because the columns in J2 contain exactly 1007 bulbs that
are on after executing the big plan, there are 1007 − ∣I∣ bulbs that are on
before executing the big plan. (For a set X, we denote by ∣X∣ its number
28
of elements.) In the columns of J1 there are exactly 1007 that are on, and
this also means that J1 and J2 are disjoint. In the other columns at most
1006 bulbs are on. The total number of bulbs that are on before the big
plan is at most
This has to be at least 1006 ⋅ 2014 + 1, hence ∣J1 ∣ + ∣J2 ∣ − ∣I∣ ⋅ ∣J2 ∣ ≥ 1. This
yields ∣J1 ∣ > (∣I∣ − 1)∣J2 ∣. If ∣I∣ ≥ 2, then ∣J1 ∣ > ∣J2 ∣. By executing the big
plan, the number of columns in which 1007 bulbs are on decreases. But
after that, we again have a situation with ∣I∣ rows containing 1007 bulbs
that are on, and in which J1 and J2 have been interchanged. Then we can
again apply the big plan, to decrease the number of columns that contain
1007 bulbs that are on again. This is a contradiction, because now we are
back in the beginning situation. We conclude that we must have ∣I∣ = 1.
We have a situation in which there is exactly one row with exactly 1007
bulbs that are on. Analogously, we can show that there must be exactly one
column in which exactly 1007 bulbs are on. Because there are 1006⋅2014+1
bulbs that are on, each other row and column must contain exactly 1006
bulbs that are on. Now change the state of the bulbs in the row with 1007
bulbs that are on. In 1007 columns there will be 1006 + 1 = 1007 bulbs that
are on. We have already seen that in this situation we can decrease the
number of bulbs that are on.
We conclude that if there are more than 1006 ⋅ 2014 bulbs that are on,
it is always possible to decrease this number. Hence, the smallest k is
1006 ⋅ 2014. ◻
29
IMO Team Selection Test 2, June 2014
Problems
1. Let f ∶ Z>0 → R be a function such that for all n > 1 there is a prime divisor
p of n such that
n
f (n) = f ( ) − f (p).
p
Furthermore, it is given that f (22014 ) + f (32015 ) + f (52016 ) = 2013.
Determine f (20142 ) + f (20153 ) + f (20165 ).
2. The sets A and B are subsets of the positive integers. The sum of any two
distinct elements of A is an element of B. The quotient of any two distinct
elements of B (where we divide the largest by the smallest of the two) is
an element of A. Determine the maximum number of elements in A ∪ B.
4. Determine all pairs (p, q) of prime numbers for which pq+1 +q p+1 is a square.
30
Solutions
31
2. Suppose that A contains at least three elements, say a < b < c. Then B
contains the three distinct elements a + b < a + c < b + c. Hence, A certainly
b+c
contains the element a+c . Apparently this fraction is an integer, hence
a + c ∣ b + c. But then it follows that a + c ∣ (b + c) − (a + c) = b − a. We know
that b > a, hence b − a is positive, hence we must have a + c ≤ b − a. This
yields c ≤ b − 2a < b, which is in contradiction with c > b. Hence, A contains
at most two elements.
Suppose that B contains at least four elements, say a < b < c < d. Then
A contains the three distinct elements ad , db and dc . But A cannot contain
three distinct elements, contradiction. Hence, B contains at most three
elements.
In total A ∪ B contains at most 5 elements. This number can be attained,
take for example A = {2, 4} and B = {3, 6, 12}. Now 2 + 4 = 6 ∈ B and
12
6
= 63 = 2 ∈ A and 12
3
= 4 ∈ A, hence this pair of sets satisfies the conditions.
We conclude that the maximum number of elements of A ∪ B is 5. ◻
3. The triangle is acute, hence H lies inside the triangle. This means that E
and F lie on the short arcs AC and BC. Suppose that E lies on the short
arc AC and that F lies on the short arc BC.
If we reflect H in AC, then the reflection H ′ lies on the circumcircle of
△ABC. (This is a well-known fact, which can be proved by angle chasing
to prove that ∠AHC = 180○ − ∠ABC.) On the other hand, this reflection
also lies on the circle with centre C through H, because ∣CH ′ ∣ = ∣CH∣.
Hence, H ′ is the intersection point of the two circles and that is E. We
conclude that E is the image of H under the reflection in AC.
This means that EH is perpendicular to AC and hence it is the same
line as BH. Because AD is also perpendicular to AC, the lines BE are
AD parallel. Furthermore, D lies on the circumcircle of △ABC because
∠CAD + ∠CBD = 90○ + 90○ = 180○ . We have already seen that E lies on
the short arc AC, hence EADB is a cyclic quadrilateral (in this order).
Now we have ∠BEA + ∠EAD = 180○ since BE and AD are parallel, but
also ∠EBD + ∠EAD = 180○ because of the cyclic quadrilateral. Hence,
∠BEA = ∠EBD, hence the corresponding chords BA and ED have the
same length.
Analogously we can prove that ∣AB∣ = ∣DF ∣, which solves the problem. ◻
32
4. First suppose that both p and q are odd. Then the exponents in the
sum pq+1 + q p+1 are both even, from which it follows that both terms are
congruent to 1 mod 4. Hence, the sum is congruent to 2 mod 4, but this
is never a square.
Now suppose that both p and q are even. Then, they are both equal to
2. That yields pq+1 + q p+1 = 23 + 23 = 16 = 42 , hence this pair satisfies the
conditions.
Finally, suppose that one of both, say p, is even and the other one is odd.
We then have p = 2 and 2q+1 +q 3 = a2 for a certain positive integer a. Write
q + 1 = 2b with b a positive integer, then the equality becomes 22b + q 3 = a2 ,
or equivalently
q 3 = a2 − 22b = (a − 2b )(a + 2b ).
Both factors on the right hand side now have to be a power of q, say
a − 2b = q k and a + 2b = q l with l > k ≥ 0. Both factors are divisible by
q k , hence also the difference is divisible by it. Hence, q k ∣ 2 ⋅ 2b = 2b+1 .
However, q is an odd prime, hence the only power of q that is a divisor of
a power of two, is 1. Hence, k = 0. Now we get q 3 = a + 2b and a − 2b = 1,
hence q 3 = (2b + 1) + 2b = 2b+1 + 1. This yields
33
5. For i = 1, 2, . . . , 10 let ci be an integer such that P (ci ) = i. For i ∈
{1, 2, . . . , 9} we have that
hence ci+1 −ci = ±1 for all i ∈ {1, 2, . . . , 9}. Furthermore, it holds that ci ≠ cj
for i ≠ j, because P (ci ) = i ≠ j = P (cj ). We conclude that c1 , c2 , . . . , c10 are
ten consecutive integers, either in ascending or descending order. Hence,
we will consider the following two cases:
or equivalently,
10
P (x) = 1 + x − c1 + R(x) ⋅ ∏(x − ci ).
i=1
10 10
= 10 + a ⋅ (∏(10 − ci ) − ∏(0 − ci )) .
i=1 i=1
34
The numbers 10 − c1 , 10 − c2 , . . . , 10 − c10 are ten consecutive numbers and
the numbers 0−c1 , 0−c2 , . . . , 0−c10 are the next ten consecutive numbers.
Hence, there is an N such that
10 10
∏(10−ci )−∏(0−ci ) = (N +20)(N +19)⋯(N +11)−(N +10)(N +9)⋯(N +1).
i=1 i=1
We will find a bound for this quantity. First we suppose that N + 1 > 0.
Then we have
P (x) = 1 − x + c1 .
And now it follows again that for any integer k there is an integer m with
P (m) = k. ◻
35
Junior Mathematical Olympiad, October 2013
Problems
Part 1
3. A football tournament with five teams is held, in which every pair of teams
plays one match against each other. Two points are awarded for winning
a match, one point is awarded for a draw, and zero points are awarded for
losing a match. After the tournament, every team has a number of points
different from all of the other teams. What are the possibilities for the
number of points of the winning team?
A) 8 B) 8, 7 C) 8, 6 D) 8, 7, 6 E) 8, 7, 6, 5
4
4. A small 4 by 4 square lies partially on a larger 5 by
5 square, in such a way that one of the vertices of
the larger square lies directly beneath the centre of
the small square. What part of the large square is
covered by the small square? 5
36
5. A large sheet of paper can be divided into at most 6 1 6
pieces with two horizontal lines and one other line,
see figure. With five horizontal lines and five other
lines, what is the maximal number of pieces that you
2 5
can obtain?
3 4
A) 43 B) 44 C) 45 D) 46 E) 47
6. Five children, Ahmed, Bob, Celine, Dan and Eve, are standing in a queue
to buy ice cream at an ice cream van. They all have an integer number of
euros. The one or more children in front of Ahmed in the queue, together
have 4 euros. The children between Celine and Dan together have 7 euros.
The children in front of Eve together have 6 euros, and Eve has 2 euros
herself. How many euros does Ahmed have?
A) 1 B) 2 C) 3 D) 4 E) 5
8. John has six squares of equal size: two red ones, two grey ones, and two
blue ones. He makes a cube out of them by gluing them together. How
many different cubes can John make? Two cubes are different if they
cannot be transformed into one another by a rotation of the cube.
A) 3 B) 4 C) 5 D) 6 E) 8
37
Part 2
1. What is the smallest number that can be obtained by adding three consec-
utive positive even integers, but also by adding four consecutive positive
even integers? (A number is positive if it is larger than 0.)
2. Nick has red and blue marbles. All red marbles have the same weight, and
so do all blue marbles. Nick weighs a red marble and a blue one: together
they weigh 30 grams. Then he weighs a number of red marbles and a blue
one: they weigh 180 grams. Exactly the same total number of marbles,
but now with one red instead of one blue marble, together weigh 60 grams.
What is the weight of a red marble in grams?
3. You have square tiles with edges of length 1 metre, 2 metres, 3 metres, and
so on. You have more than enough tiles of each kind. What is the length
of the side of the smallest square you can make using precisely 11 of these
tiles?
4. How many numbers between 1 and 1000 are not divisible by 2, divisible
by 3, not divisible by 5, and divisible by 7?
38
C
A D B
39
Solutions
Part 1
1. D) 8 5. D) 46
2. E) 62, 5% 6. B) 2
3. D) 8, 7, 6 7. D) 4
4. B) 16% 8. D) 6
Part 2
1. 36 5. 6
2. 25 grams 6. 5367
5
3. 5 metres 7. 12
4. 19 8. 6 and 18
40
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2013
10 Second Round, March 2013
16 Final Round, September 2013
21 BxMO/EGMO Team Selection Test, March 2014
25 IMO Team Selection Test 1, June 2014
30 IMO Team Selection Test 2, June 2014
36 Junior Mathematical Olympiad, October 2013
MY
We eat problems
CMY
Olympiad 2014
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2014
12 Second Round, March 2014
18 Final Round, September 2014
24 BxMO/EGMO Team Selection Test, March 2015
28 IMO Team Selection Test 1, June 2015
33 IMO Team Selection Test 2, June 2015
39 Junior Mathematical Olympiad, October 2014
The 1000 best students were invited to the second round, which was held in
March at twelve universities in the country. This round contained five open
questions, and two problems for which the students had to give extensive
solutions and proofs. The contest lasted 2.5 hours.
The 130 best students were invited to the final round. Also some outstand-
ing participants in the Kangaroo math contest or the Pythagoras Olympiad
were invited. In total about 150 students were invited. They also received
an invitation to some training sessions at the universities, in order to pre-
pare them for their participation in the final round.
The final round in September contained five problems for which the stu-
dents 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 53rd edition
2014.
1
Among the participants of the training programme, there were some extra
girls, as this year we participated again in the European Girls’ Mathemat-
ical Olympiad (EGMO). In total there were eight girls competing to be in
the EGMO team. The team of four girls was selected by a selection test,
held on 6 March 2015. They attended the EGMO in Minsk, Belarus from
14 until 20 April, and the team returned with two honourable mentions.
For more information about the EGMO (including the 2015 paper), see
www.egmo.org.
The same selection test was used to determine the ten students partici-
pating in the Benelux Mathematical Olympiad (BxMO), held in Mersch,
Luxembourg, from 8 until 10 May. The Dutch team received four bronze
medals and three silver medals. For more information about the BxMO
(including the 2015 paper), see www.bxmo.org.
In June the team for the International Mathematical Olympiad 2015 was
selected by two team selection tests on 5 and 6 June 2015. A seventh,
young, promising student was selected to accompany the team to the IMO
as an observer C. The team had a training camp in Chiang Mai, from 30
June until 8 July.
For younger students the Junior Mathematical Olympiad was held in Oc-
tober 2014 at the VU University Amsterdam. The students invited to
participate in this event were the 100 best students of grade 2 and grade 3
of the popular Kangaroo math contest. The competition consisted of two
one-hour parts, one with eight multiple choice questions and one with eight
open 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.
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
Dutch delegation
The Dutch team for IMO 2015 in Thailand consists of
3
First Round, January 2014
Problems
A-problems
2. A pond contains both red and yellow carp. Two fifths of the carp are
yellow, the rest of the carp are red. Three quarters of the yellow carp are
female. In total, there are an equal number of male and female carp.
Which fraction of the total carp population are red males?
1 1 3 2 1
A) 5 B) 4 C) 10 D) 5 E) 2
3. Seven lily pads are numbered 1 through 7 from left to right. A frog jumps
along these pads. It can jump to the left and to the right, but only by
leaps of three or five pads at once. For example, starting from pad 2, it
can only leap to pads 5 and 7. The frog wants to make a journey in which
he visits each pad exactly once (so the first and last pad on his journey
will be different).
1 2 3 4 5 6 7
4
4. A square paper ring has height 1, and the sides have length 4. The ring is
depicted in the left hand figure. By folding it flat on the tabletop, we get
the right hand figure, where ABCD is a square.
1 A D
A D
4 B C
B C
What is the length of side AB?
5 7 9
A) 2 B) 3 C) 2 D) 4 E) 2
5. Tom and Jerry were running a race. The number of runners finishing before
Tom was equal to the number of runners finishing after him. The number
of runners finishing before Jerry was three times the number of runners
finishing after him. In the final ranking, there are precisely 10 runners in
between Tom and Jerry. All runners finished the race, and no two runners
finished at the same time.
How many runners participated in the race?
A) 22 B) 23 C) 41 D) 43 E) 45
5
7. In the figure, a quadrilateral ABCD is drawn. The midpoint of side AB
is called M . The four line segments AM , BM , BC, and AD each have
length 8, and the line segments DM and CM both have length 5.
What is the length of line segment CD?
D C
?
8 8
5 5
A 8 M 8 B
6
B-problems C
The answer to each B-problem is a number.
A B
2. Carl has a large number of apples and pears. He wants to pick ten pieces
of fruit and place them in a row. He wants to do it in such a way that there
is no pear anywhere between two apples. For example, the fruit sequences
AAAAAAAAAA and AAPPPPPPPP are allowed, but AAPPPPPPPA
and APPPPPPPAA are not.
How many sequences can Carl make?
999
| {z. . . 99} × 444
| {z. . . 44}
2014 nines 2014 fours
and then add up all digits of the resulting number, what number would
the final outcome be?
7
Solutions
A-problems
2 2 3 3
A2. D) 5 From the given data, we deduce that 5 · 4 = 10 of the carp
3
are yellow females. Since half the carp are female, we find that 12 − 10 = 15
of the carp are red females. Finally, using the fact that three fifths of the
carp are red, we see that 53 − 15 = 52 of the carp are red males.
A3. C) 3 and 5 The journey that visits the pads in the order 3, 6, 1, 4, 7, 2,
and 5 (or in the opposite order), shows that pads 3 and 5 are the starting
pad of a possible journey. We will see that these are the only possible
starting points.
We say that two pads are neighbours if the frog can jump from one pad
to the other (and hence also the other way around). Every intermediate
pad in the frog’s journey must have at least two neighbours: the pad the
frog came from and the pad it will go next. Since pads 3 and 5 have only
one neighbour (pad 6 and pad 2 respectively), these must be the starting
point and end point of the frog’s journey. The other pads must therefore
be intermediate pads.
8
A4. B) 3 Consider the top rim of the paper ring (indicated in bold).
The rim has length 4 × 4 = 16. In the folded state, the rim becomes
rectangle EF GH. Since |AE| = |BF | = |CG| = |DH| = 1, we find that
|AB| + |F G| + |CD| + |EH| = 16 − 4 = 12. These four lengths equal the
length of the sides of square ABCD. It follows that the square has sides
of length 12
4 = 3.
E H
A D
1
B C
4
F G
9
A7. C) 25
8 Observe that AM B is a straight angle. This implies that
∠AM D + ∠DM C + ∠CM B = 180◦ . Since triangles AM D and BM C are
equal (three equal sides), we see that ∠CM B = ∠M DA. Hence ∠DM C =
180◦ − ∠AM D − ∠M DA = ∠DAM , because the angles of triangle AM D
sum to 180 degrees. It follows that DM C and DAM are isosceles triangles
with equal apex angles. Hence these two triangles are equal up to scaling.
|CD| |DM | 5
This means that |DM | = |AD| . Therefore, the length of CD equals 8 · 5 =
25
8 .
B-problems
C
B1. 24
5 The six rectangles have equal areas.
f d c a
Rectangles c and d are twice as tall as rectangle
a and therefore also twice as thin. Hence they 5
B
have width 25 . Rectangle e thus has a width of
5 5 b
2 + 2 +5 = 10 and must be half as tall as rectangle
a. This means that rectangle f is precisely 25
times as tall as rectangle a and therefore has a e
5
width of 5/2 = 2. It follows that the square has
sides of length 5 + 52 + 25 + 2 = 12. Because
the square has a height of 25 times the height of
12 24
rectangle a, the height of rectangle a equals |BC| = 5/2 = 5 .
10
containing 1 apple. In total there are 1 + 2 + 3 + · · · + 10 = 55 sequences
containing at least one apple. The total number of sequences is therefore
55 + 1 = 56.
9×4 = 40 − 4 = 36,
99 × 44 = 4400 − 44 = 4356,
999 × 444 = 444000 − 444 = 443556,
9999 × 4444 = 44440000 − 4444 = 44435556.
The pattern should be clear. To solve the problem, observe that 999 . . . 99 =
1000 . . . 00 − 1, where the first number has 2014 zeroes. The product is
therefore equal to
444
| {z . . . 00} − |444 {z
. . . 44} |000 {z . . . 44} = 444
| {z. . . 44} 3 |555 {z
. . . 55} 6.
2014 fours 2014 zeroes 2014 fours 2013 fours 2013 fives
11
Second Round, March 2014
Problems
B-problems
The answer to each B-problem is a number.
B1. Brenda is filling pouches from an unlimited supply of red and blue marbles.
In each pouch she puts more red than blue marbles, and each pouch can
contain at most 25 marbles. For example, she can make a pouch containing
6 red marbles and 2 blue marbles, or a pouch containing 1 red marble and
0 blue marbles.
How many differently filled pouches can she make in total?
B3. For how many of the integers n from 1 up to and including 100 is the
number 8n + 1 a perfect square?
B4. Evan and nine other people are standing in a circle. All ten of them think
of an integer (that may be negative) and whisper their number to both
of their neighbours. Afterwards, they all state the average of the two
numbers that were whispered in their ear. Evan states the number 10,
his right neighbour states the number 9, the next person along the circle
states the number 8, and so on, finishing with Evan’s left neighbour who
states the number 1.
What number did Evan have in mind?
12
B5. The numbers of dots on two opposite faces of a die
always sum to 7. Nine identical dice are glued in
a 3×3-array. This is done in such a way that when
?
two faces are glued together, they must contain
the same number of dots. In the figure you can
see the top view of the array. For five of the dice
the number of dots is not shown.
What number of dots must be on the place of the
question mark?
C-problems For the C-problems not only the answer is important; you also have to
describe the way you solved the problem.
C2. A positive integer n is called a jackpot number if it has the following prop-
erty: there exists a positive integer k consisting of two or more digits, all of
which are equal (such as 11111 or 888), for which the product n · k is again
a number consisting of equal digits. For example, 3 is a jackpot number
because 3 · 222 = 666.
(a) Find a jackpot number consisting of 10 digits and prove that it is a
jackpot number.
(b) Prove that 11 is not a jackpot number.
(c) Determine whether 143 is a jackpot number and prove that your an-
swer is correct.
13
Solutions
B-problems
B1. 169 Brenda can make 25 differently filled pouches without blue
marbles as she can put 1 to 25 red marbles in a pouch. There are 23
differently filled pouches possible containing 1 blue marble because 2 to 24
red marbles may be added. Using 2 blue marbles there are 21 possibilities,
namely by adding 3 to 23 red marbles. In total, there are 25+23+. . .+1 =
169 differently filled pouches that Brenda can make.
14
B4. 5 The number that Evan came up with is denoted by c10 , the
number of his right neighbour is denoted by c9 , continuing in this way until
the left neighbour of Evan, whose number is denoted by c1 . From the data
we deduce that
c10 + c8 = 2 · 9 = 18,
c8 + c6 = 2 · 7 = 14,
c6 + c4 = 2 · 5 = 10,
c4 + c2 = 2 · 3 = 6,
c2 + c10 = 2 · 1 = 2.
Adding up these equations yields 2(c2 + c4 + c6 + c8 + c10 ) = 50, hence
c2 + c4 + c6 + c8 + c10 = 25.
Finally, we find c10 = (c2 + c4 + c6 + c8 + c10 ) − (c2 + c4 ) − (c6 + c8 ) =
25 − 6 − 14 = 5.
?
g h
a 7_ a
7_ b e f
7_ a b b c d
15
We will now apply this to the six dice in the first two columns of the
3 × 3 array, see the figure on the right. The two dice in the third row
show complementary numbers of dots on their top faces, namely 2 and 5.
Therefore, the numbers of dots c and d, on the faces where the dice are
glued to dice in the second row, must be equal. This in turn implies that
the numbers of dots e and f on the top faces must be complementary.
Therefore, the numbers g and h are equal. Finally, the top faces of the
dice in the first row must show complementary numbers of dots. Hence,
there must be 3 dots on the place of the question mark.
C-problems
C1. (a) First, we notice that a + d is the area of triangle AEF and p + q is the
area of triangle BEF . The base AE of triangle AEF has the same
length as the base BE of triangle BEF . Because the two triangles
have the same height, they must also have equal areas.
(b) Here we use that triangles DEF and CEF have bases of the same
length (|DF | = |CF |), and equal corresponding heights. Hence, they
have equal areas. This implies that c + d = q + r. Subtracting the
equation of part (a) yields c − a = r − p, or a + r = c + p, as required.
(c) The heights of triangles AED, BEC and ABF with respect to the
bases AE, BE, and AB are denoted by x, y, and z. Because F is the
midpoint of CD, the height z is the average of the heights x and y,
in formulas: x+y 1
2 = z. The area of triangle AED is 2 · x · |AE|, the
1
area of triangle BEC is 2 · y · |BE|, and the area of triangle ABF is
1
2 · z · |AB|. Because E is the midpoint of AB we have |AE| = |BE|
and |AB| = 2 · |AE|. The sum of the areas of triangles AED and
BEC is thus equal to 12 · (x + y) · |AE| = z · |AE|, while the area of
triangle ABF is equal to 21 · z · 2|AE| = z · |AE|. Hence, these areas
are equal and we find a + b + p + s = a + d + p + q. By subtracting
a + p on both sides of the equation, we find b + s = d + q, as required.
16
C2. (a) Consider the ten digit number 1001001001. If we multiply it with 111,
we get the number 111111111111 which consists of ones only. Hence,
the number 1001001001 is a jackpot number.
(b) Let k be a number of at least two digits, all of which are equal, say
equal to a. Remark that a 6= 0. We have to prove that the digits of
the number
11k = k + 10k = a · · · a + a · · · a0
17
Final Round, September 2014
Problems
1. Determine all triples (a, b, c), where a, b, and c are positive integers that
satisfy a ≤ b ≤ c and abc = 2(a + b + c).
D C
C
2. Version senior students
W
On the sides of triangle ABC, isosce-
V
les right-angled triangles AU B, CV B,
U
and AW C are placed. These three tri-
angles have their right angles at ver-
tices U , V , and W , respectively. Tri-
angle AU B lies completely inside tri-
angle ABC and triangles CV B and A B
AW C lie completely outside ABC. See
the figure.
Prove that quadrilateral U V CW is a parallelogram.
18
4. A quadruple (p, a, b, c) of positive integers is called a Leiden quadruple if
• p is an odd prime number,
• a, b, and c are distinct and
• ab + 1, bc + 1 and ca + 1 are divisible by p.
19
Solutions
D C
2. Version for junior students
We know that ∠ABH = ∠CBG, because
these are opposite angles. Because trian-
gles ABH and CBG are isosceles, we have
G
∠AHB = ∠ABH and ∠CBG = ∠CGB.
A B
Triangles ABH and CBG are similar (AA)
and hence we have ∠BAH = ∠BCG. Be-
cause ABCD is a parallelogram, we have
H
∠DAB = ∠DCB and hence ∠DAH =
∠DAB + ∠BAH = ∠DCB + ∠BCG = ∠DCG holds. Because ABCD is
a parallelogram, we have |CD| = |AB| = |AH| and |AD| = |BC| = |CG|.
Therefore, triangles DAH and GCD are congruent (SAS) and we have
|DH| = |DG|. In other words, triangle DGH is isosceles.
20
2. Version for senior students C
Because triangle AU B is isosceles with top
W
angle ∠AU B = 90◦ , we have ∠U AB =
45◦ . In the same way, we have ∠CAW = V
U
45◦ . Combining these two equalities, we
find ∠W AU = 45◦ + ∠CAU = ∠CAB.
By the Pythagorean theorem, we find 2|AW |2 =
2 2 2
|AW
√ | +|CW | = |AC| and hence |AW | =
1
2 √2·|AC|. In the same way we find |AU | = A B
1
2 2 · |AB|. Hence, √
triangles W AU and CAB are similar (SAS) with scale
√
factor |AW | 1 |AU | 1
|AC| = 2 2 = |AB| . In particular, we find |W U | = 2 2 · |BC| =
|CV |.
In the same√ way, we see that triangles V BU and CBA are similar and that
|V U | = 21 2 · |AC| = |CW |. It follows that in quadrilateral U V CW the
opposite sides have equal lengths, hence U V CW is a parallelogram.
21
b) In the table below there is a possible outcome for 7 teams called A
through G. In the row corresponding to a team, crosses indicate wins
against other teams. Row 2, for example, indicates that team B won
against teams C and D and obtained a total score of 2 points. Each
team (except A) has indeed lost exactly one match against a team
with a lower score. These matches are indicated in bold.
A B C D E F G Score
A - X 1
B - X X 2
C X - X X 3
D X X - X 3
E X X X - X 4
F X X X - X 4
G X X X X - 4
4. a) Without loss of generality, we may assume that a < b < c. The inte-
gers a and c are not divisible by p because that would imply that ac+1
is a multiple of p plus 1, hence not divisible by p. Since bc+1 and ac+1
are both divisible by p, their difference (bc+1)−(ac+1) = (b−a)c is di-
visible by p as well. Hence, since c is not divisible by p, it must be the
case that b − a is divisible by p. Similarly, (ac + 1) − (ab + 1) = a(c − b)
is divisible by p and since a is not divisible by p, this implies that c−b
is divisible by p.
Thus, we find that b = a + (b − a) ≥ a + p and c = b + (c − b) ≥ a + 2p.
a+b+c a + (a + p) + (a + 2p)
≥ = p + a ≥ p + 2.
3 3
22
b) Again, we may assume that a < b < c. In part a) we have seen
that a+b+c
3 ≥ a+(a+p)+(a+2p)
3 = p + a ≥ p + 2. We can only have
a+b+c
3 = p + 2 if b = a + p, c = a + 2p, and a = 2. Since ab +
1 = 2(2 + p) + 1 = 2p + 5 must be divisible by p, it follows that
5 is divisible by p. We conclude that p = 5, b = 7, and c = 12.
The quadruple (p, a, b, c) = (5, 2, 7, 12) is indeed a Leiden quadruple,
because ab + 1 = 15, ac + 1 = 25, and bc + 1 = 85 are all divisible by
p.
We conclude that p = 5 is the only number for which there is a Leiden
quadruple (p, a, b, c) that satisfies a+b+c
3 = p + 2.
1−x
x A
1
2
3
.
B . D
.
24
25
26
C
23
BxMO/EGMO Team Selection Test, March 2015
Problems
24
Solutions
1. There is a positive integer k with (5m + n)k = 5n + m. Hence, 5km − m =
5n − kn, or (5k − 1)m = (5 − k)n. The left hand side is positive, hence also
the right hand side is positive, which yields k < 5. If k = 1, then 4m = 4n,
hence m = n, hence m | n. If k = 2, then 9m = 3n, hence 3m = n, hence
m | n. If k = 3, then 14m = 2n, hence 7m = n, hence m | n. If k = 4,
then 19m = n, hence m | n. We conclude that in all cases we have that
m | n.
2. We will prove this by contradiction. Suppose that such a t does not exist.
If ak = 1, then akk = k would hold, contradicting our assumption. Hence,
ak ≥ 2. We will now prove by induction to i that aik ≥ i + 1. We just
proved the base case. Now suppose that for certain i ≥ 1 we have that
aik ≥ i + 1. Then we also have that a(i+1)k ≥ i + 1. If a(i+1)k = i + 1, then
(i+1)k
a(i+1)k = k, which is a contradiction. Hence, a(i+1)k ≥ i + 2. This finishes
the induction. Now take i = ar , then we have aar k ≥ ar + 1. Moreover,
because r = ar (k + 1) we have ar = aar (k+1) ≥ aar k ≥ ar + 1, which is a
contradiction.
2
3. We will prove that k = b n4 c is the largest possible integer.
2
Suppose that n is even. Then it is possible to cover the board with n2
2
dominoes (without considering the colours). Because there are n2 domi-
2 2
noes, each of which is either colourful or even, there are at least b n4 c = n4
2 2
colourful or at least b n4 c = n4 even dominoes.
2
When n is odd, we can cover the board with n 2−1 dominoes. (Notice
that this number is an even integer.) Of these dominoes either at least
n2 −1 2 2 2
25
2
all the other squares blue. Then there are b n4 c red squares, hence we can
2
put at most b n4 c non-overlapping colourful dominoes on the board as each
of these dominoes covers one red square. An even domino cannot cover
two red squares, because there are no pairs of adjacent squares coloured
red. Hence, it must cover two blue squares. One of these blue squares must
have been black, hence the number of even dominoes is at most the number
2
of black-blue squares and that is b n4 c. Hence of both the colourful as the
2
even dominoes we can put at most b n4 c simultaneously on the board.
2
We conclude that the maximum k is indeed k = b n4 c.
We therefore see (AA) that 4ABE ∼ 4ACF . From this, it follows that
|AB| |BE|
= .
|AC| |CF |
26
5. Substituting x = y = 0 gives 0 = f (0)3 , so f (0) = 0. We consider two
more cases; either f has another zero, or f has no other zeroes. In the first
case there is some a 6= 0 such that f (a) = 0. Substituting x = a then gives
(a2 + y 2 )f (ay) = 0. As a2 + y 2 > 0 (since a 6= 0) we find that f (ay) = 0
for all y. As ay can attain all values in R, it follows that f (x) = 0 for all x;
i.e. f is the constant function with value 0. Note that this function indeed
satisfies the given equation.
So now consider the second case, in which f (x) 6= 0 for all x 6= 0. Substi-
tuting x 6= 0 and y = 1 then gives (x2 + 1)f (x) = f (x)f (1)f (x2 + 1) and
f (x) 6= 0, so we can divide this equation by f (x). Therefore (x2 + 1) =
2
f (1)f (x2 + 1). Let c = f (1). Note that c 6= 0, so f (x2 + 1) = x c+1 . Since
x
x2 + 1 attains all reals that are at least 1, we find f (x) = c for all x > 1.
Substituting x = y = 2 now gives (4 + 4)f (4) = f (2)f (2)f (4 + 4). Since
we know the values f (x) takes for x > 1, we know the values of f (2), f (4),
and f (8). Therefore 8 · 4c = 2c · 2c · 8c , or equivalently, 1c = c13 . We deduce
that c2 = 1, so c = 1 or c = −1. As c2 = 1, it follows that f (x) = cx for
all x > 1.
Let x > 1, and substitute y = x1 . This gives (x2 + x12 )f (1) = f (x)f ( x1 )f (x2 +
1 2 1 2 1
x2 ). Since x > 1, we have x + x2 > 1, so we deduce that (x + x2 )c = xc ·
1 1 1 1 1
2
f ( x )·c(x + x2 ), or equivalently, 1 = xc·f ( x ). Therefore f ( x ) = xc = c· x1 .
We deduce that f (x) = cx for all x > 0 with x 6= 1. Since we also have
f (1) = c, it follows that f (x) = cx for all x > 0.
Substituting x = y = −1 gives 2f (1) = f (−1)2 f (2). Since f (1) = c
and f (2) = 2c, it follows that 2c = f (−1)2 · 2c, so since c 6= 0 we have
f (−1)2 = 1. Therefore either f (−1) = 1 or f (−1) = −1.
Let x > 0. Substituting y = −1 then gives (x2 +1)f (−x) = f (x)f (−1)f (x2 +
1), so (x2 + 1)f (−x) = cx · f (−1) · c(x2 + 1), or equivalently, f (−x) =
c2 xf (−1) = xf (−1). Let d = f (−1). Then f (x) = −dx for all x < 0,
where d2 = 1.
So aside from f (x) = 0 there are four more possible solutions, namely
f (x) = x, f (x) = −x, f (x) = |x|, and f (x) = −|x|. We first check
f (x) = tx with t = ±1. The left hand side then reads (x2 + y 2 ) · txy, and
the right hand side then reads tx · ty · t(x2 + y 2 ). As t2 = 1, the left hand
side and the right hand side are equal, so these two functions are indeed
solutions of the functional equation.
Next, we check the two functions f (x) = t|x|, with t = ±1. Now the left
hand side reads (x2 + y 2 ) · t|xy|, and the right hand side reads t|x| · t|y| ·
t|x2 + y 2 |. Since x2 + y 2 = |x2 + y 2 |, |xy| = |x||y|, and t2 = 1, the left hand
side and the right hand side are equal. So these two functions are indeed
solutions of the functional equations. Therefore there are five solutions;
f (x) = 0, f (x) = x, f (x) = −x, f (x) = |x| and f (x) = −|x|.
27
IMO Team Selection Test 1, June 2015
Problems
2. Find all polynomials P (x) with real coefficients such that the polynomial
is constant.
Show that c1 + c2 + · · · + cn = n2 − 1.
1. Show that for all positive integers n, the number Dn is of the form
3k with k ≥ 0 an integer.
2. Show that for all integers k ≥ 0 there exists a positive integer n such
that Dn = 3k .
28
Solutions
29
respectively 0, a2 − a1 + a0 , 0, and a0 . These sum up to a2 − a1 + 2a0 . On
the other hand, we can compute Q(1);
This is indeed constant, so all such P (x) satisfy the condition. In fact,
every constant polynomial is also of this form, with b = 0. Therefore
the polynomials satisfying the condition are precisely those of the form
P (x) = bx2 + bx + a with a, b ∈ R.
30
4. We consider only the configuration in which A, B, E, and D lie in that
order on a circle, in which O1 , A, and C lie in that order on a line, and O2 ,
A, and D lie in that order on a line; the proof is analogous for the other
configurations.
As ABED is a cyclic quadrilateral, we have ∠BED = 180◦ − ∠DAB.
Moreover, using the parallel lines, we see that ∠BED = ∠DAO1 and as
|O1 A| = |O1 D|, we have ∠DAO1 = ∠ADO1 . We deduce that 180◦ −
∠DAB = ∠ADO1 . Therefore DO1 and AB are parallel.
We already know that ∠ADO1 = ∠DAO1 . Since |O2 A| = |O2 C|, it follows
that ∠DAO1 = ∠O2 AC = ∠O2 CA. So ∠O2 DO1 = ∠ADO1 = ∠O2 CA =
∠O2 CO1 , so O1 DCO2 is a cyclic quadrilateral.
The line O1 O2 is the perpendicular bisector AB, therefore is also perpen-
dicular to DO1 , as this line is parallel to AB. Therefore ∠O2 O1 D = 90◦ .
As O1 DCO2 is a cyclic quadrilateral, we now also have ∠O2 CD = 90◦ .
31
Modulo 3m+2 , this is congruent to 3m+1 c, and since 3 - c, it follows
that this is divisible by 3m+1 , but not by 3m+2 , as desired. This
completes our inductive argument.
Next we show that for n = 3k−1 , we have that (a+3)n −an is divisible
by 3k for all positive integers a. Again, we prove this by induction on
k. For k = 1 we have n = 1, so indeed we see that (a + 3) − a = 3
is divisible by 3 for all positive integers a. Now suppose that m ≥ 1,
and suppose that we proved our claim for k = m. Let n = 3m−1 .
Then (a + 3)n − an is divisible by 3m for all positive integers a, so
we can write (a + 3)n = an + 3m c for some integer c. Taking third
powers of both sides then yields
so
(a + 3)3n − a3n = a2n · 3m+1 c + an · 32m+1 c2 + 33m c3 ,
which is divisible by 3m+1 . This completes our inductive argument.
We have now shown for n = 3k−1 that 3k | 1n + 2n + 3n and 3k |
(a + 3)n − an for all positive integers a, from which we, by induction
on a, immediately deduce that 3k | an + (a + 1)n + (a + 2)n for all
positive integers a. Therefore 3k | Dn . As 3k+1 - 1n + 2n + 3n , we
also have 3k+1 - Dn . Therefore Dn = 3k , as desired.
32
IMO Team Selection Test 2, June 2015
Problems
2. Determine all positive integers n for which there exist positive integers a1 ,
a2 , . . . , an with
a1 + 2a2 + 3a3 + . . . + nan = 6n
and
1 2 3 n 1
+ + + ... + =2+ .
a1 a2 a3 an n
5. Find all functions f : Z>0 → Z>0 such that f (1) = 2 and such that for
all m, n ∈ Z>0 we have that min(2m + 2n, f (m + n) + 1) is divisible by
max(f (m) + f (n), m + n).
33
Solutions
1. We will prove that the grid point (x, y) is reachable if and only if x + y ≡ 0
mod 2.
If we move the pawn from (x, y) to (x ± a, y ± a), then the sum of the new
coordinates equals x + y + 2a, x + y or x + y − 2a, hence it is congruent to
the sum of the old coordinates modulo 2. The same holds when making a
step of type B. Because the pawn starts at (0, 0), after executing any finite
number of steps, we will have that the sum of the coordinates of the grid
point where the pawn is standing, is even. Hence, the points (x, y) with
x + y ≡ 1 mod 2 are not reachable.
Now we will show that all other points are reachable. Because gcd(a, b) =
1, there exist integers m, n with ma+nb = 1. Then we have 2ma+2nb = 2.
Of the numbers m and n one must be positive and the other negative. We
assume that m is positive. The other case is treated analogously. We will
now first make 2m steps of type A and 2m steps of type B. For the steps
of type A we choose to do m times the step (x, y) 7→ (x + a, y + a) and m
times (x, y) 7→ (x + a, y − a). For the steps of type B we choose m times
(x, y) 7→ (x + b, y + b) and m times (x, y) 7→ (x − b, y − b). The effect of
all these steps together is that the x-coordinate increased by 2ma and the
y-coordinate did not change (in face, the B-steps cancel each other). After
this we do 2|n| steps of type A and 2|n| steps of type B. For the steps of
type A we choose to do |n| times the step (x, y) 7→ (x + a, y + a) and |n|
times (x, y) 7→ (x − a, y − a). For the steps of type B we choose to do |n|
times (x, y) 7→ (x − b, y + b) and |n| times (x, y) 7→ (x − b, y − b). The
effect of all these steps together is that the x-coordinate decreased by 2|n|
and the y-coordinate did not change (in fact, now the A-steps cancel each
other). Altogether, after these 4m + 4|n| steps, we moved the pawn from
a starting point (x, y) to the point (x + 2ma − 2|n|b, y) = (x + 2, y). We
can construct an analogous series of steps that move the pawn from (x, y)
to (x, y + 2), and also analogous series of steps that move the pawn from
(x, y) to (x − 2, y) and (x, y − 2). Each of these series starts with a step of
type A and ends with a step of type B.
By combining series like these, we can move the pawn from (0, 0) to any
point (x, y) with x ≡ y ≡ 0 mod 2. Now consider a point (x, y) with
x ≡ y ≡ 1 mod 2. Because gcd(a, b) = 1, at least one of a and b is odd.
Suppose that a is odd. Then the point (x − a, y − a) is a point with two
even coordinates, to which we can construct a series of steps, ending with
a step of type B. After that, we can make a step of type A that moves the
pawn from (x − a, y − a) to (x, y). Next, suppose that a is even. Then b is
odd. The point (x − b, y − b) now has two even coordinates, hence we can
34
reach this point with a series of steps ending with a step of type B. Now
we execute three more steps after this:
A B A
(x − b, y − b) 7−→ (x − b − a, y − b − a) 7−→ (x − a, y − a) 7−→ (x, y).
We have
6n 12 12
1 = <
2 n(n + 1) n+1 n
and
1 1 2 1 2
2 n(n + 1) 2 n (n+ 1) 2 n (n+ 1)
1 = > = 41 n2 .
2+ n
2n + 1 2n + 2
12
Altogether we find that n > 14 n2 , or 48 > n3 , which yields n ≤ 3.
1
For n = 1 we get a1 = 6 and a1 = 3, which is a contradiction. Hence,
n = 1 is not possible.
For n = 2 we get a1 + 2a2 = 12 and a11 + a22 = 2 + 12 . If a2 ≥ 2 we have
1 2
a1 + a2 ≤ 1 + 1 and this is too small. Hence, we have a2 = 1, but then
we find that a1 = 12 − 2 = 10 and hence a11 + a22 = 10 1
+ 2, which is a
contradiction. Hence, n = 2 is impossible.
For n = 3 there is a solution, namely a1 = 6, a2 = 3 and a3 = 2. Hence,
n = 3 is possible and we conlude that n = 3 is the only solution.
35
As BD is parallel to AC, we have ∠CBD = ∠ACB = 60◦ = ∠CBA.
Hence, the point E is the intersection point of the perpendicular bisector
of CD and the exterior angle bisector of ∠CBD. This means that E lies
on the circumcircle of triangle CDB. (This is a known fact, but it is
also possible to prove it as follows. Let E 0 be the intersection point of the
exterior angle bisector of ∠CBD and the circumcircle of 4CBD. Because
BE 0 is the exterior angle bisector, we have ∠CBE 0 = 180◦ − ∠DBE 0 .
Hence, chords CE 0 and DE 0 have the same length, which means that E 0
lies on the perpendicular bisector of CD.) We conclude that CEBD is a
cyclic quadrilateral. Hence, ∠CED = ∠CBD = 60◦ .
36
We conclude that at least two numbers of 3 up to and including 2n − 1
are not expressible as the sum of a red and a blue number. We shall now
show that we can colour the numbers in such a way that all number from
4 up to and inlcuding 2n − 2 are expressible as the sum of a red and a blue
number, implying that the maximum k equals 2n − 5.
To obtain this, colour all the even numbers, except n, and also the number
1 blue. All odd numbers, except 1, and also the number n we colour red.
By adding 1 to an odd number (unequal to 1) we can obtain all even
numbers from 4 up to and inlcuding n as the sum of a red and a blue
number. By adding 2 to an odd number (unequal to 1), we can obtain all
odd numbers from 5 up to and including n + 1 as the sum of a red and a
blue number. By adding n − 1 to an even number (unequal to n), we can
obtain all odd numbers from n + 1 up to and including 2n − 3 as the sum
of a red and a blue number. By adding n to an even number (unequal to
n), we can obtain all even numbers from n + 2 up to and including 2n − 2
as the sum of a red and a blue number. Altogether, we can express all
numbers from 4 up to and including 2n − 2 as the sum of a red and a blue
number.
We conclude that the maximum k equals 2n − 5 = 4023.
37
Now suppose that f (2) = 4. We will prove by induction to n that f (n) =
2n for all n. Namely, suppose that this holds for n = r − 1 for some
r ≥ 4. We will prove that also f (r) = 2r holds. First substitute m =
1 and n = r − 1. Then we find that min(2r, f (r) + 1) is divisible by
max(f (1) + f (r − 1), 1 + r − 1) = max(2r, r) = 2r. Hence, min(2r, f (r) + 1)
cannot be smaller than 2r, hence f (r) ≥ 2r − 1. However, we already knew
that f (r) ≤ 2r, hence f (r) ∈ {2r−1, 2r}. Suppose that f (r) = 2r−1. Then
substitute m = 1 and n = r. Then we find that min(2(r + 1), f (r + 1) + 1)
is divisible by max(f (1) + f (r), 1 + r) = max(2 + 2r − 1, r + 1) = 2r + 1.
Because 2r + 1 - 2(r + 1), the minimum does not equal 2(r + 1), hence we
have f (r +1)+1 < 2r +2 and, moreover, it must also be divisible by 2r +1,
hence f (r + 1) = 2r. Now also substitute m = 2 and n = r − 1. Then we
find that min(2(r+1), f (r+1)+1) = min(2r+2, 2r+1) = 2r+1 is divisible
by max(f (2) + f (r − 1), 1 + r) = max(4 + 2r − 2, r + 1) = 2r + 2, which is a
contradiction,. We conclude that f (r) = 2r, which finishes the induction.
Hence, we find the candidate function f (n) = 2n for all n. We will check
whether this function satisfies the conditions. In this case, for all m, n we
have that min(2m+2n, f (m+n)+1) = min(2m+2n, 2m+2n+1) = 2m+2n
and max(f (m) + f (n), m + n) = max(2m + 2n, m + n) = 2m + 2n. The
former is divisible by the latter (because it equals the latter), hence this
function satisfies the conditions.
We conclude that there are exactly two solutions: the function given by
f (n) = n + 1 for all n and the function given by f (n) = 2n for all n.
38
Junior Mathematical Olympiad, October 2014
Problems
Part 1
39
5. In a triangle ABC, we have ∠A = 84◦ . Moreover, D is a point on the line
segment AB such that ∠D1 = 3 · ∠C2 and such that the line segments DC
and DB have equal lengths.
What is ∠C1 ?
A) 27◦ B) 28◦ C) 30◦ D) 32◦ E) 36◦
1 D
2
1
2 84◦
C A
6. A piece of apple pie had been stolen, and five children are being questioned
on this. They all know who the culprit is, but not all of them are speaking
the truth. Whenever one of the children lies, the next one will feel so
guilty about this that he or she will tell the truth. The children make the
following claims in the order shown:
• Asim: “Coen and I both didn’t do it.”
• Bob: “Either Coen or Dilan is the culprit.”
• Coen: “Eva and I both didn’t do it.”
• Dilan: “Asim is the culprit.”
• Eva: “At least two of Asim, Bob, Coen, and Dilan lied.”
Who stole the apple pie?
A) Asim B) Bob C) Coen D) Dilan E) Eva
40
8. Max has a lot of white and red paint. He starts with a 2-litre bucket in
which one litre of red paint and one litre of white paint. Max now repeats
the following step a number of times.
Step. Max pours precisely one litre out of the bucket, into a large con-
tainer. Next, he fills the bucket back up to 2 litres of paint, using
either the white paint, or the red paint. After this, he mixes the
paint in the bucket.
After a number of steps, the percentage of red paint in the bucket must
be between 83 and 84 percent. What is the smallest number of steps Max
needs to attain this?
A) 5 B) 6 C) 7
D) 8 E) Max cannot obtain such a percentage.
Part 2
2. There are 36 balls, numbered from 1 up to 36. We want to put these into
boxes in such a way that the following two conditions are satisfied:
(1) Every box contains at least 2 balls.
(2) Whenever you pick up two balls from a box, the sum of the two
numbers of these balls is always a multiple of 3.
What is the smallest number of boxes for which this is possible?
41
3. We are given a square ABCD. A line is drawn through A that intersects
the segment BC in E, and the line through C and D in F . The ratio of
the lengths of the segments BE and EC is 1 : 2. The area of the grey area
is 60.
What is the area of the square?
D C F
A B
4. We want to exchange a 200-euro bill for bills of 5, 10, and 20 euros. One
possibility is to exchange it for 5 bills of 20 euros, 6 bills of 10 euros, and
8 bills of 5 euros. Another possibility is to exchange it for 20 bills of 10
euros.
How many possibilities are there to exchange a 200-euro bill for bills of 5,
10, and 20 euros?
5. By stacking small cubes (all of the same size) neatly, a larger cube is
formed. Two small cubes with faces placed against one another are called
neighbours. So a cube can have at most six neighbours. The number of
cubes having precisely four neighbours is 96.
How many small cubes are there having precisely five neighbours?
6. Michelle colours each of the numbers from 1 up to 2014. The first two
numbers (1 and 2) are coloured red, the next two (3 and 4) are coloured
white, the following two (5 and 6) are coloured blue, the two following
those are coloured red, the two after those are coloured white, the two
after that are coloured blue, and so on and so forth. Michelle then sums
all blue numbers, and subtracts from that the sum of all red numbers.
What is the result?
42
7. The figure below represents a puzzle. The goal is to fill each of the 16 cells
with a number from 1 up to 4. This has to be done in such a way that in
each column and in each row, the four numbers are distinct. Moreover, in
each of the four 2×2-squares, the four numbers have to be distinct as well.
Finally, the four numbers in the grey squares also need to be distinct.
How many solutions does this puzzle have?
4
3
2
1
A B C D
8. Mies has drawn a regular hexagon with area 1. She notices that the mid-
points of the six sides also form a regular hexagon.
What is the area of this small hexagon?
43
Solutions
Part 1
1. C) 4 5. A) 27◦
2. D) 9 6. B) Bob
4. E) 328 8. B) 6
Part 2
1. 320 5. 384
2. 13 6. −1343
3. 36 7. 168
3
4. 121 8. 4
44
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2014
12 Second Round, March 2014
18 Final Round, September 2014
24 BxMO/EGMO Team Selection Test, March 2015
28 IMO Team Selection Test 1, June 2015
33 IMO Team Selection Test 2, June 2015
39 Junior Mathematical Olympiad, October 2014
MY
We eat problems
CMY
Olympiad 2015
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
sponsored by: Centrum Wiskunde & Informatica sterdam 2011
OmslagNWO2016-v2wit.indd 1 06-06-16 20:49
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2015
9 Second Round, March 2015
14 Final Round, September 2015
19 BxMO Team Selection Test, March 2016
23 IMO Team Selection Test 1, June 2016
28 IMO Team Selection Test 2, June 2016
31 IMO Team Selection Test 3, June 2016
35 Junior Mathematical Olympiad, October 2015
The 1000 best students were invited to the second round, which was held in
March at twelve universities in the country. This round contained five open
questions, and two problems for which the students had to give extensive
solutions and proofs. The contest lasted 2.5 hours.
The 130 best students were invited to the final round. Also some outstand-
ing participants in the Kangaroo math contest or the Pythagoras Olympiad
were invited. In total about 150 students were invited. They also received
an invitation to some training sessions at the universities, in order to pre-
pare them for their participation in the final round.
The final round in September contained five problems for which the stu-
dents 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 54rd edition
2015.
In February a team of four girls was chosen from the training group to
represent the Netherlands at the EGMO in Buşteni, Romania, from 10
until 16 April. The team brought home a silver medal and two bronze
medals, a very nice achievement. For more information about the EGMO
(including the 2016 paper), see www.egmo.org.
In March a selection test of three and a half hours was held to determ-
ine the ten students participating in the Benelux Mathematical Olympiad
(BxMO), held in Soest, the Netherlands, from 29 April until 1 May. The
Dutch team received six bronze medals and three silver medals, and man-
aged to get the highest total score. For more information about the BxMO
(including the 2016 paper), see www.bxmo.org.
1
In June the team for the International Mathematical Olympiad 2016 was
selected by three team selection tests on 2, 3 and 4 June 2016, each lasting
four hours. A seventh, young, promising student was selected to accompany
the team to the IMO as an observer C. The team had a training camp in
Hong Kong, from 30 June until 9 July.
For younger students the Junior Mathematical Olympiad was held in Oc-
tober 2015 at the VU University Amsterdam. The students invited to
participate in this event were the 100 best students of grade 2 and grade 3
of the popular Kangaroo math contest. The competition consisted of two
one-hour parts, one with eight multiple choice questions and one with eight
open 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.
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
Dutch delegation
The Dutch team for IMO 2016 in Hong Kong consists of
3
First Round, January 2015
Problems
A-problems
2. Five suspects are being questioned about the order of arrival at a crime
scene. They make the following statements.
4. The difference of two integers is 10. If you multiply the two integers, you
will get one of the following five numbers.
Which number do you get?
A) 22398 B) 22420 C) 22442 D) 22453 E) 22475
4
5. Jan has got a wooden cube. He divides each of the
faces into 2×2 squares that he subsequently paints in a
black-white pattern: two diagonally opposite squares
are painted black, the other two are painted white. In
each vertex of the cube three squares meet. If two or
three of these squares are black, we call the vertex a
dark vertex.
What is the smallest number of dark vertices that the cube can have?
A) 0 B) 1 C) 2 D) 3 E) 4
6. In how many ways can you get the number 100 by adding some consecutive
integers between 1 and 99 inclusive?
A) 1 B) 2 C) 3 D) 4 E) 5
8. A tree grows in the following manner. On day 1, one branch grows out of
the ground. On day 2, a leaf grows on the branch and the branch tip splits
up into two new branches. On each subsequent day, a new leaf grows on
every existing branch and each branch tip splits up into two new branches.
See the figure below.
How many leaves does the tree have at the end of the tenth day?
A) 172 B) 503 C) 920 D) 1013 E) 2047
5
B-problems
The answer to each B-problem is a number.
1. Julia constructs a sequence of numbers. She starts with two integers she
chooses herself. Then, she calculates the next numbers in the sequence as
follows: if the last number she wrote down is b and the number before that
is a, then the next number will be 2b − a. The second number in Julia’s
sequence is 55 and the hundredth number is 2015.
What is the first number in Julia’s sequence?
2. Two points A and B and two circles are given, one having A as centre and
going through B and the other one having B as centre and going through
A. Point C lies on the second circle and on line AB. Point D also lies on
the second circle. Point E lies on the first circle and on line BD. See the
figure below. Moreover, ∠D = 57◦ .
What is the value of ∠A in degrees?
57◦
A C
? B
6
4. On a school trip, twenty students will be abseiling. In each round, one stu-
dent will descend the mountain. Hence, after twenty rounds, all students
will have gone down the mountain safely. In the first round, cards bearing
the numbers 1 to 20 are distributed among the students. The student get-
ting number 1 will go down first. In round 2, cards bearing the numbers 1
to 19 are distributed among the remaining students. The student receiving
the number 1 is next to descend. They continue in this way, until there is
only one student left in round 20, who automatically gets a card bearing
the number 1. By an amazing coincidence, no student gets the same num-
ber twice. In the first round, Sara gets a card with number 11.
What is the sum of the numbers on the cards received by Sara?
7
Solutions
A-problems
1. A) 5 5. C) 2
2. E) Erik 6. B) 2
3. A) 12081 7. C) 18
4. E) 22475 8. D) 1013
B-problems
1. 35
2. 48◦
3. 70
4. 66
8
Second Round, March 2015
Problems
B-problems
The answer to each B-problem is a number.
3. Berry has picked 756 raspberries. He divides the raspberries among himself
and his friends in such a way that everyone gets the same number of
raspberries. However, three of his friends are not feeling hungry and they
each return a number of raspberries: exactly one quarter of their share.
Berry has a healthy appetite and eats not only his own share, but the
returned raspberries as well. Berry has lost count, but does know for a
fact that he has eaten more than 150 raspberries.
How many raspberries did Berry eat?
4. Four line segments divide a rectangle into eight pieces as indicated in the
figure. For three of the pieces, the area is indicated as well: 3, 5, and 8.
8
5
9
5. In the cells of a 5×5-table, the numbers 1 to 5 are placed in such a way
that in every row and in every column, each of the five numbers occurs
exactly once. A number in a given row and column is well-placed if the
following conditions are met.
• In that row, all smaller numbers are to the left of the number and all
larger numbers are to the right of it, or conversely.
• In that column, all smaller numbers are below the number and all
larger numbers are above it, or conversely.
C-problems For the C-problems not only the answer is important; you also have to
describe the way you solved the problem.
1. A set of different numbers are evenly spread if after sorting them from
small to large, all pairs of consecutive numbers have the same difference.
For example: 3, 11 and 7 are evenly spread, because after sorting them,
both differences are 4.
a) Kees starts out with three different numbers. He adds each pair of
these numbers to obtain three outcomes. According to Jan, these
three outcomes can be evenly spread only if the three starting numbers
were evenly spread.
Is Jan right? If so, prove this; if not, use an example to prove that
Jan is wrong.
b) Jan starts out with four different numbers. He also adds each pair of
them to obtain six outcomes. He wants to choose his four numbers in
such a way that the six resulting numbers are evenly spread.
Is this possible? If so, give an example; if not, prove that it is im-
possible.
10
2. We consider rectangular boards consisting of m × n cells that are arranged
in m (horizontal) rows and n (vertical) columns. We want to colour each
cell of the board black or white in such a way that the following rules are
obeyed.
• For every row, the number of white cells equals the number of black
cells.
• If a row and a column meet in a black cell, the row and column contain
equal numbers of black cells.
• If a row and a column meet in a white cell, the row and column
contain equal numbers of white cells.
Determine all pairs (m, n) for which such a colouring is possible.
11
Solutions
B-problems
1. 6
√
2. 6± 11
3. 189
4. 16
5. 5
C-problems
1. a) Kees starts with three numbers a < b < c. The three sums are then
ordered as follows: a + b < a + c < b + c. If these are evenly spread,
then the difference (b+c)−(a+c) = b−a equals (a+c)−(a+b) = c−b.
This is exactly the condition for the three original numbers a, b and
c to be evenly spread. Hence, Jan was right.
b) Jan can accomplish this by taking the four numbers 0, 1, 2 and 4 to
start with. The six results then are 0 + 1 = 1, 0 + 2 = 2, 1 + 2 = 3,
0 + 4 = 4, 1 + 4 = 5, and 2 + 4 = 6. These six numbers are evenly
spread.
12
Now we shall show that these are the only possible board sizes. Consider a
coloured board that meets all requirements. Because each row has equally
many black and white squares, the total number of squares in a row must
be divisible by 2. Write n = 2k. Each row has exactly k white and k black
squares. Now consider the left column. If all its squares are white, then
the column has k white squares because of the second requirement. Hence,
we have m = k in this case. The same happens when all squares in the
left column are black. If there are both black and white squares in the left
column, then there must be exacly k white and k black squares because of
the second requirement. Hence, we find m = 2k = n in this case.
We conclude that for a pair (m, n) there exists a colouring if and only if
n = 2m or if m = n and n is even.
13
Final Round, September 2015
Problems
14
3. Version for senior students E
Points A, B, and C are on a line in this order.
D
Points D and E lie on the same side of this line, S
in such a way that triangles ABD and BCE are
equilateral. The segments AE and CD intersect
in point S. A B C
Prove that ∠ASD = 60◦ .
7pq 2 + p = q 3 + 43p3 + 1.
5. Given are (not necessarily positive) real numbers a, b, and c for which
15
Solutions
1. (a) It is possible to make 2015 groups. For example, take the 2015 groups
{−4, −3, −2, −1, i}, where i runs from 1 to 2015. Each group consists
of five distinct numbers, as required, and any two groups have exactly
four numbers in common: −4, −3, −2, and −1.
(b) Using six available numbers, there are only six possible groups of
five numbers (each obtained by leaving out one of the six numbers).
Those six groups do satisfy the requirement that any two of them
have exactly four numbers in common. We conclude that six is the
greatest number of groups we can make in this case.
(c) A way to make three groups is to take {1, 2, 3, 4, 5}, {1, 2, 3, 4, 6}, and
{1, 2, 3, 4, 7}.
More than three groups is not possible. Indeed, suppose we have
four or more groups. The first two groups are A = {a, b, c, d, e} and
B = {a, b, c, d, f }, where a, b, c, d, e, and f are distinct numbers.
Then there must be a third group C containing a seventh number g.
The remaining four numbers in C must be in both A and B, hence
C = {a, b, c, d, g}.
Now consider a hypothetical fourth group D. This group cannot
contain the number g since otherwise, using a similar reasoning as
for C, we would have D = {a, b, c, d, g}. Because D does not contain
the number g, it must contain the remaining four numbers a, b, c,
and d from C. Comparison with groups A and B then shows that D
can contain neither e nor f . It follows that besides a, b, c, and d, D
cannot contain a fifth number, contradicting the requirements.
We conclude that the greatest number of groups we can make is three.
16
To complete the proof, we need to show that it is not possible to place more
than 250,000 dominoes on the board. Partition the board into 500×500
patches consisting of 2×2 squares each. Of each patch, at most two out of
the four squares can be covered by dominoes since otherwise two dominoes
would be adjacent. Hence, no more than 2 · 500 · 500 = 500,000 squares can
be covered by dominoes. This shows that no more than 250,000 dominoes
can fit on the board.
17
4. We start by observing that in the equation 7pq 2 + p = q 3 + 43p3 + 1 the
numbers p and q cannot both be odd. Otherwise, 7pq 2 + p would be even,
while q 3 + 43p3 + 1 would be odd. Since 2 is the only even prime number,
we conclude that p = 2 or q = 2.
In the case p = 2, we obtain the equation 14q 2 +2 = q 3 +344+1, which can
be rewritten as q 3 − 14q 2 = −343. This shows that q must be a divisor of
343 = 73 , hence q = 7. Substitution confirms that (p, q) = (2, 7) is indeed
a solution since 14q 2 + 2 = 2 · 7 · 72 + 2 and q 3 + 344 + 1 = 73 + (73 + 1) + 1 =
2 · 73 + 2 are equal.
Next, we consider the case that q = 2 and p is odd. This results in the
equation 28p + p = 8 + 43p3 + 1. Since p is odd, we see that 28p + p is odd,
while 8 + 43p3 + 1 is even. Hence, no solutions exist in this case.
We conclude that (p, q) = (2, 7) is the only solution.
Since the first and the last term are equal, we can conclude that all of the
above inequalities must be equalities. In particular, a − b = |a − b| = |c|.
Since c ≤ 0, this implies that a + c = b. This shows that one of the three
numbers equals the sum of the other two.
18
BxMO Team Selection Test, March 2016
Problems
1. For a positive integer n that is not a power of two, we define t(n) as the
greatest odd divisor of n and r(n) as the smallest positive odd divisor of
n unequal to 1. Determine all positive integers n that are not a power of
two and for which we have
n = 3t(n) + 5r(n).
2. Determine all triples (x, y, z) of non-positive real numbers that satisfy the
following system of equations
x2 − y = (z − 1)2 ,
y 2 − z = (x − 1)2 ,
z 2 − x = (y − 1)2 .
4. The Facebook group Olympiad training has at least five members. There
is a certain integer k with following property: for each k-tuple of members
there is at least one member of this k-tuple friends with each of the other
k − 1. (Friendship is mutual: if A is friends with B, then also B is friends
with A.)
(a) Suppose k = 4. Can you say with certainty that the Facebook group
has a member that is friends with each of the other members?
(b) Suppose k = 5. Can you say with certainty that the Facebook group
has a member that is friends with each of the other members?
19
Solutions
1. Let p be the smallest odd prime divisor of n. Then r(n) = p holds. Now
we can write n = 2t mp with m odd and t ≥ 0. Then we have t(n) = pm,
hence the given equality becomes 2t mp = 3pm + 5p, or (2t − 3)mp = 5p,
hence (2t − 3)m = 5. We see that m must be a divisor of 5, hence m = 1
or m = 5. If m = 1 holds, then 2t = 8, hence t = 3. We get that n = 8p
with p an odd prime number. This is indeed a solution for all odd prime
numbers p. If m = 5 holds, then 2t = 4, hence t = 2. We get that n = 4·5·p
with p an odd prime. This only gives a solution if p is the smallest odd
prime divisor, which is if p = 3 or p = 5. In this way, we find two more
solutions: n = 60 and n = 100.
2. Expanding the right hand sides and adding up all equations gives
x2 + y 2 + z 2 − (x + y + z) = x2 + y 2 + z 2 − 2(x + y + z) + 3,
3. (a) The midpoint E of the arc AB not containing C, lies on the angu-
lar bisector CI. In the same way, F lies on BI. We have ∠IF C =
∠BF C = ∠BAC = 90◦ , because ABCF is a cyclic quadrilateral
and ∠IDC = 90◦ because D is the point of tangency of the in-
circle to BC. Hence, ∠IF C + ∠IDC = 180◦ , which yields that
F IDC is a cyclic quadrilateral. Now we have ∠DF I = ∠DCI =
1 1
2 ∠ACB, while also ∠IF E = ∠BF E = ∠BCE = 2 ∠ACB. Hence,
1 1
∠DF E = 2 ∠ACB + 2 ∠ACB = ∠ACB. Analogously, we have
∠DEF = ∠ABC. Altogether, this yields 4DEF ∼ 4ABC.
20
(b) Let S be the intersection of EF and AB. In the previous part we
already saw that BF is the angular bisector of ∠DF E = ∠DF S.
Hence, BF is also the angular bisector of ∠ABC = ∠SBD. There-
fore, 4BDF ∼ = 4BSF because of (ASA). This means that |BD| =
|BS|. On the other hand, the distances of B to the points of tangency
of the incircle to BC and BA are equal and one of these points of
tangency is D, hence the other point of tangency must be S. Hence,
EF goes through the point of tangency of the incircle to AB. Ana-
logously, it also goes through the point of tangency of the incircle to
AC.
4. (a) Yes, you can. If everybody is friends with everyone else, then we are
done. Hence, suppose that there are two members, say A and B, who
are not friends with each other. If we consider a group of four with A,
B, and two other members, then one of the other two must be friends
with the other and with A and B. In particular, the two others are
friends with each other. This holds for any two members (unequal to
A and B) that we choose, hence each pair not containing A and B is
friends with each other. Now take A, B, C, and D and suppose that
C is friends with A, B, and D. Moreover, he is also friends with all
other members of the group, hence C is someone who is friends with
everybody else.
(b) No, this is not possible. We give a counterexample. Suppose that the
Facebook group has six members, called A, B, C, D, E, and F . They
are all friends with each other, except for the pair (A, B), the pair
(C, D) and the pair (E, F ). This means that no member is friends
with every other member. If we take a group of five then without loss
of generality this is A, B, C, D, and E. Here we can find someone
who is friends with the other four, namely E. Hence, the condition is
met.
21
5. Suppose that the quotient of 2n(3m2 + n2 ) + 8 and (m + n)3 is unequal to
1. Then it is at least 2, hence we have
(m + n)3 ≤ n(3m2 + n2 ) + 4,
or
m3 + 3m2 n + 3mn2 + n3 ≤ 3m2 n + n3 + 4,
or
m3 + 3mn2 ≤ 4.
This yields m < 2, and therefore m = 1. Then we have 1 + 3n2 ≤ 4, and
hence also n = 1. The pair (m, n) = (1, 1) is indeed a solution, because
23 | 2 · 4 + 8 holds.
The other possibility is that the quotient does equal 1. Then we have
(m + n)3 = 2n(3m2 + n2 ) + 8,
or
m3 + 3m2 n + 3mn2 + n3 = 6m2 n + 2n3 + 8,
or
m3 − 3m2 n + 3mn2 − n3 = 8.
The left hand side we can factor as (m − n)3 . Hence, we have m − n = 2,
or m = n + 2. From the previous calculations it also follows that (m, n) =
(n + 2, n) is indeed a solution for all positive integers n.
We conclude that the solutions are: (m, n) = (1, 1) and (m, n) = (n + 2, n)
for n ≥ 1.
22
IMO Team Selection Test 1, June 2016
Problems
for all x, y ∈ R.
23
Solutions
2. First note that a rectangle with integer dimensions has a power of two as
area if and only if the dimensions are powers of two.
Consider a covering in which no two rectangles have the same dimensions.
We first show that then there are no rectangles of width 1. Suppose to
the contrary that such a rectangle occurs in the covering. Colour every
square covered by a rectangle of width 1 with the colour blue. Let M be
the number of blue squares. Then M is the sum of all rectangles of width
1, so it is the sum of (at least one) distinct powers of 2. Let 2k be the
largest one. As 2k − 1 = 2k−1 + 2k−2 + · · · + 2 + 1, there are fewer than
2k blue squares not covered by this rectangle of width 1 and height 2k , so
there is at least one row which contains a blue square of this rectangle of
height 2k and no other blue square. But the remaining squares in this row,
of which there are an odd number, must then be covered by rectangles of
even width, which is a contradiction. So there are no rectangles of width
1. Analogously, there are no rectangles of height 1, either. Therefore all
rectangles have an even width and an even height.
Now consider the smallest n for which a covering as in the problem exists
in which no two rectangles have the same dimension. As all rectangles
have an even width and an even height, we can divide all dimensions (of
both the square and the rectangles) by 2 to obtain a square with smaller
dimensions covered by rectangles with integer dimensions, each having a
power of two as area. Since no two rectangles had the same dimensions,
this contradicts the minimality of n.
24
3. Let d = gcd(m, n) and write m = da and n = db. Then lcm(m, n) ·
gcd(m, n) = mn, so we can rewrite the given equation as
da·db
d − d = k(da − db),
or equivalently,
ab − 1 = k(a − b). (1)
So from now on we consider the following equivalent problem: find all
positive integers k for which (1) has no solutions (a, b) in positive integers
with a 6= b and gcd(a, b) = 1. Note that if a pair (a, b) satisfies this
equation, then it automatically follows that gcd(a, b) = 1; suppose that
t | a and t | b, then we have t | ab and t | a − b, so t | 1, and we deduce that
a and b have no common divisor greater than 1.
First suppose that k ≥ 3. We claim that (a, b) = (k 2 − k − 1, k − 1) is a
solution. Indeed, we have
ab − 1 = 2(a − b).
25
4. If f were constant, then the left hand side is constant, whereas the right
hand side is not, this is a contradiction. Therefore f is not constant.
Substitute x = 0. This gives f (−1) + f (0)f (y) = −1, so f (0)f (y) is a
constant function in y. As f is not constant, it follows that f (0) = 0, and
therefore it also follows that f (−1) = −1. Substituting x = y = 1 gives:
f (0) + f (1)2 = 1, so we have either f (1) = 1 or f (1) = −1.
Now substituting y = 1 + x1 with x 6= 0 gives f (x + 1 − 1) = f (x)f (1 + x1 ) =
2x + 2 − 1, so we have
f (x)f 1 + x1 = 2x + 1 − f (x)
for all x 6= 0. (2)
1
Substituting y = x with x 6= 0 gives f (1 − 1) + f (x)f ( x1 ) = 2 − 1, so we
have
1
f (x)f x =1 for all x 6= 0. (3)
Substituting y = 1, x = z + 1 gives f (z + 1 − 1) + f (z + 1)f (1) = 2z + 2 − 1,
so we have
f (z) + f (z + 1)f (1) = 2z + 1 for all z.
Using (3) and (2), we can rewrite the first and second term respectively as
follows:
If the second factor of the left hand side is non-zero, then we can divide
by this, so we obtain
Recall that we had two possible values for f (1). First suppose that f (1) =
1. Then we have
2 + 2x 2
f (x) = 2 =x if x 6= 0 and x + 2 6= 0.
x +2
26
Note that x2 + 2 = 0 only if x = −1. As we already know that f (−1) = −1
and f (0) = 0, it follows that f (x) = x for all x. We check that this f
satisfies the original equation: the left hand side then is xy − 1 + xy =
2xy − 1, so this f is indeed a solution.
Suppose on the other hand that f (1) = −1. Then we have
−2x
f (x) = 2 = −x2 if x 6= 0 and 2
x 6= 0.
x
27
IMO Team Selection Test 2, June 2016
Problems
2. Determine all pairs (a, b) of integers having the following property: there
is an integer d ≥ 2 such that an + bn + 1 is divisible by d for all positive
integers n.
S = {ai + aj | 1 ≤ i, j ≤ 1000 en i + j ∈ A}
is a subset of A.
28
Solutions
√
1. We can write 3 abc as 3 a4 · b · 4c. Applying the inequality of the arithmetic
p
and geometric mean on the positive reals a4 , b and 4c yields
√ a
r
3 a + b + 4c a b 4c
abc = 3 · b · 4c ≤ 4 = + + .
4 3 12 3 3
Next we apply AM-GM on a2 and 2b:
√ a
r
a 2 + 2b a
ab = · 2b ≤ = + b.
2 2 4
We add up these two inequalities and we also add a to both sides to obtain:
√ √
3 a a b 4c 4
a+ ab + abc ≤ a + +b+ + + = (a + b + c).
4 12 3 3 3
2. Consider a pair (a, b) that has the property, with the corresponding d. Let
p be a prime divisor of d (which exists as d ≥ 2). Because d | an + bn + 1
for all n, we also have p | an + bn + 1 for all n. Consider n = p − 1. Then
an ≡ 0 mod p holds if p | a and an ≡ 1 mod p holds if p - a, because of
the Fermat’s little theorem. Similarly, this holds for b. Hence, an + bn + 1
can attain the values 1, 2 and 3 modulo p. On the other hand, it must be
congruent to 0 modulo p, hence p = 2 or p = 3. We consider the two cases.
Suppose that p = 3. Then we must have 3 - a, 3 - b. The case n = 1
moreover yields that 3 | a + b + 1, hence a + b ≡ 2 mod 3. These two
requirements together are equivalent to a ≡ b ≡ 1 mod 3. In this case, it
is true that for all positive integers n we have an + bn + 1 ≡ 1 + 1 + 1 ≡ 0
mod 3, hence such a pair has the property, with d = 3.
Now suppose that p = 2. Then 2 must be a divisor of exactly one of a and b.
In this case we have for all positive integers n that an +bn +1 ≡ 0+1+1 ≡ 0
mod 2, hence each such pair has the property, with d = 2.
We conclude that these are the pairs that have the property: (a, b) with
a ≡ b ≡ 1 mod 3, (a, b) with a ≡ 1, b ≡ 0 mod 2, and (a, b) with a ≡ 0,
b ≡ 1 mod 2.
3. From the data and the angle bisector theorem it follows that
|BF | |BE| |CE|
= = .
|BD| |BD| |CD|
29
As 4ABC is isosceles, we have ∠F BD = ∠ECD, which yields together
with the first equality that 4BF D ∼ 4CED. This yields ∠BF D =
∠CED = ∠BED, hence BDEF is a cyclic quadrilateral. It is well-known
that a cyclic quadrilateral BDEF is a trapezoid with DE k BF if and
only if |BD| = |EF |. We have |AF | = |EC| if and only if |BF | = |AE| (as
|AB| = |AC|), that is to say, if and only if |EA| = |EB|, which holds if and
only if ∠BAE = ∠ABE. Because we have 2∠BED = ∠BEC = ∠BAE +
∠ABE by the exterior angle theorem, ∠BAE = ∠ABE is equivalent to
∠BED = ∠ABE which is again equivalent to DE k BF . To summarise,
|BD| = |EF | holds if and only if DE k BF which holds if and only if
|AF | = |EC|.
4. We will prove that there are 214 such sets. In particular, we will prove that
the sets A that satisfy these conditions are of the form B ∪ C, with C a
subset of {2001, . . . , 2014} and B = {1, 2, . . . , 1000 − |C|}. The sets of this
form will be called “nice”. As there are 214 subsets of {2001, . . . , 2014},
there are 214 nice sets.
First we will show that each nice set A satisfies the conditions. Suppose
that i, j ≤ 1000 with i+j ∈ A. Then we have i+j ≤ 2000, hence i+j ∈ B.
Therefore, there exists a k with k ≤ 1000 − |C| such that i + j = ak (= k).
Because ak ≤ 1000 − |C| holds, we also have i, j ≤ 1000 − |C|, hence we
have ai = i and aj = j. That means that ai + aj = i + j = ak is an element
of A. Hence each element of S is an element of A, from which we deduce
that A satisfies the conditions.
We will now show that each set A satisfying the conditions is nice. First
suppose that there exists an integer k with 1 ≤ k ≤ 1000 satisfying ak ∈
{1001, . . . , 2000}. Then we have ak = 1000+i for a certain i with i ≤ 1000,
hence a1000 + ai is a an element of S and hence it must also be an element
of A. However, we have a1000 + ai > a1000 , which yields a contradiction.
Hence such an ak cannot occur. This means that A can be written as a
disjoint union B ∪ C, with C ⊆ {2001, . . . , 2014} and B ⊆ {1, 2, . . . , 1000}.
Let b be the number of elements of B. Then b ≥ 986 holds, because
C has at most 14 elements. In order to prove that A is nice, we must
prove that B = {1, 2, . . . , b}. To prove this it is sufficient to prove that
ab , the maximum of B, equals b. Therefore, suppose the contrary, i.e.
that ab > b. For i with i = ab − b we then have b + i = ab ≤ 1000,
hence i ≤ 1000 − b ≤ 14 < b. Therefore we have ai ≤ 1000 and hence
ab + ai ≤ 2000. Because i + b = ab ∈ A, we have ab + ai ∈ S ⊂ A, but then
ab +ai is an element of B greater than ab , which was the maximum. This is
a contradiction. Hence, ab = b, from which it follows that B = {1, 2, . . . , b}.
Therefore, A is nice.
30
IMO Team Selection Test 3, June 2016
Problems
3. Let k be a positive integer, and let s(n) denote the sum of the digits
of n. Show that among the positive integers with k digits, there are as
many numbers n satisfying s(n) < s(2n) as there are numbers n satisfying
s(n) > s(2n).
31
Solutions
1. For each girl, call the boy with whom she wants to dance the most her
favourite.
We solve the problem by induction on n. If n = 1, the only girl will form a
couple with the only boy, who is therefore her favourite. So suppose that
k ≥ 1, and assume that the problem has been solved for n = k.
Consider the case n = k + 1. We distinguish two cases. First suppose
that every boy occurs exactly once as a favourite. In this case we can just
couple every girl to her favourite, and form n dancing couples that way.
In the remaining case, not every boy occurs exactly once as a favourite.
Since there are n favourites and n boys, once of the boys, say X, is not the
favourite of any girl (and someone else is the favourite of more than one
girl). Choose a pairing as in the problem; this exists by assumption. Let
Y be the girl coupled with boy X, and remove X and Y from the village.
There are k boys and k girls left. Note that the pairing chosen still has the
property that every girl is paired with a boy on her list. Moreover, every
girls still has a favourite among the k remaining boys, as boy X is not the
favourite of any girl. Therefore, by the induction hypothesis, we can form
k dancing couples, in such a way that every girl is paired with a boy on
her list, and at least one of the girls is paired with her favourite. Adding
the couple X − Y back in completes the induction.
2. For n = 3 we consider (a1 , a2 , a3 ) = (1, 2, 3). The sums of pairs are in this
case 3, 4, and 5, and these form an arithmetic progression. For n = 4 we
consider (a1 , a2 , a3 , a4 ) = (1, 3, 4, 5). The sums of pairs are in this case 4,
5, 6, 7, 8, and 9, and these form an arithmetic progression.
Now suppose that n ≥ 5, and suppose that a1 , a2 , . . . , an satisfies the
condition. Without loss of generality, we assume that a1 < a2 < · · · < an .
Let d be the difference between two consecutive terms of the corresponding
arithmetic progression. Note that the smallest sum is a1 + a2 , and the
second smallest is a1 + a3 . As the difference between these sums is d, we
have a3 − a2 = d. The largest sum is an + an−1 and the second largest is
an + an−1 , therefore we have an−1 − an−2 = d as well. Hence
If n ≥ 6, then the left hand side and the right hand side are sums of distinct
pairs, but the difference between such sums must be at least d. This is a
contradiction. Therefore there are no solutions for n ≥ 6.
32
For n = 5, we have a3 − a2 = d and a4 − a3 = d. Therefore the third
smallest sum must be a1 + a4 (as this one is d larger than a1 + a3 ), and
the third largest sum must be a5 + a2 . Between these, we have a2 + a3 <
a2 + a4 < a3 + a4 , and the difference between these consecutive sums is d,
and we also have a1 + a5 . Therefore a1 + a5 is either the fourth smallest
sum or the fourth largest sum.
Without loss of generality, assume that a1 + a5 is the fourth smallest sum.
Then we have
3. We show that among the positive integers with at most k digits there are
as many numbers n satisfying s(n) < s(2n) as there are numbers satisfying
s(n) > s(2n). The required result then follows by combining this result for
k and for k − 1.
We pair each number n with at most k digits to another number m with
at most k digits. Let m = 999 · · · 999 − n, where the first number consists
of k nines, so that m has at most k digits. We show that s(m) − s(2m) =
s(2n) − s(n).
To calculate m, we subtract n from 999 · · · 999, in which every digit of n
is of course at most 9. Therefore every digit of m is equal to 9 minus
the corresponding digit of n. Here we consider both m and n as numbers
having precisely k digits by adding zeroes to the left if necessary. Hence
s(m) + s(n) = s(999 · · · 999) = 9k.
Next, consider 2m and 2n. We have 2m = 1999 · · · 998 − 2n. Consider 2m
and 2n as numbers having exactly k + 1 digits, where the first digit of 2n
is either a 0 or a 1. Subtracting that from 1999 · · · 998, we find that the
first digit of 2m is either 1 − 1 = 0 or 1 − 0 = 1. The last digit of 2m is
8 minus the last digit of 2n, which cannot be a 9 as 2n is even. All other
digits of 2m are equal to 9 minus the corresponding digit of 2n. Therefore,
we have s(2m) + s(2n) = s(1999 · · · 998) = 1 + 9(k − 1) + 8 = 9k. Hence
s(m) + s(n) = s(2m) + s(2n), so s(m) − s(2m) = s(2n) − s(n).
33
Now we see that s(m) > s(2m) if and only if s(n) < s(2n). Moreover, no
number is paired to itself as 999 · · · 999 is odd. Hence there are as many
numbers with s(n) < s(2n) as there are numbers with s(n) > s(2n).
∠P SM = ∠P AM = ∠P AB = 90◦ − 21 ∠ABP
= 90◦ − (180◦ − ∠AQP ) = 90◦ − ∠AQS,
where in the second to last step, we applied the inscribed angle theorem.
Moreover, using the sum of angles in 4AQS, we find that 90◦ − ∠AQS =
∠QAS. As ASQT is a cyclic quadrilateral with |QT | = |QS| (since QS
and QT are both tangent to Γ1 ), we have ∠QAS = ∠QT S = ∠QST . To
summarise, we have ∠P SM = 90◦ − ∠AQS = ∠QAS = ∠QST = ∠P ST .
Hence S, T , and M are collinear.
From this, it follows that N is the intersection of ST and AQ. In the
cyclic quadrilateral ASQT we find, using the power of a point theorem,
that N T · N S = N A · N Q. Note that the left hand side of N is the power
of N with respect to Γ1 and that the right hand side is the power of N
with respect to Γ2 . Therefore N lies on the radical line of Γ1 and Γ2 .
34
Junior Mathematical Olympiad, October 2015
Problems
Part 1
3. Quintijn has three equally big and equally filled bottles of wine. Bottles
1 and 3 contain the same kind of white wine, while bottle 2 contains red
wine. Quintijn now pours a small amount of wine from bottle 1 into bottle
2. Next, after mixing the content of bottle 2 really well, he pours the same
amount from bottle 2 into bottle 3. In the same way, he pours the same
amount of wine from bottle 3 into bottle 1. Now all bottles contain the
same amount of wine as they did at the start. However, the content of
each bottle is polluted with wine of the other type.
Which bottle is polluted most?
A) bottle 1 B) bottle 2 C) bottle 3
D) all three equally E) you cannot determine this
35
4. Aad, Bep, Cor, Dirk, Eva, and Fenna are sitting in this order in a circle
around the campfire. Aad has a torch. He gives it to Bep who is sitting
one place to his right. She gives the torch to Dirk, who is sitting two places
to her right. He gives the torch to Aad, who is sitting three places to his
right, et cetera. It happens that someone must give the torch to the person
sitting, for instance, six or twelve positions to the right. Then this person
gives the torch to him- or herself.
When Dirk is given the torch for the hundredth time, who does he pass
the torch to?
A) Aad B) Bep C) Cor D) Dirk E) Eva
5. When a cube is cut with a plane, a cross section is created. This figure is
formed by the lines where the plane cuts the facets of the cube. In the left
figure, you can see an example in which the cross section is a triangle.
In the right figure, a flattened cube is drawn with the cut lines drawn on
the facets.
What is the cross section of the cube corresponding to this figure?
A) a triangle B) a square C) a rectangle, but not a square
D) a hexagon E) a parallelogram, but not a rectangle
6. On a machine there are three buttons. The first button can be used to
add 20 marbles to a tray in the machine. The second button can be used
to increase the number of marbles in the tray by 20%, after which 15
additional marbles are added. The third button can be used to increase
the number of marbles in the tray by 50%. If pressing a button would
cause the number of marbles to be non-integral, the pressing of the button
is not allowed. In the beginning the tray is empty. After some button
presses, there are 91 marbles in the tray.
How often has the first button been pressed?
A) 0 B) 1 C) 2 D) 3 E) 4
36
7. Ria, Sophie, and Tine are sitting around a round table in clockwise order
and are playing a game with chips. Ria starts with 3 chips, Sophie with 4
chips, and Tine with 5 chips. In each round they simultaneously give chips
to one of their neighbours. Each player can choose to give 2 chips to her
right neighbour or 1 chip to her left neighbour. If someone has no more
chips, the game ends.
Is it possible for the players to have the same number of chips after some
number of rounds, and if this is the case, how many rounds have to be
played at least to accomplish this?
A) no, it is impossible. B) yes, 3 rounds. C) yes, 6 rounds.
D) yes, 7 rounds. E) yes, 8 rounds.
8. Six people are sitting around a round table. Each of them is either a
knight or a knave. Knights always speak the truth, while knaves always
lie. Each of them has a card containing a number. All numbers are different
and everyone knows the numbers of their two neighbours. When asked:
“Is your number greater than the numbers of both your neighbours?”,
everyone answers with “Yes”. When asked: “Is your number smaller than
the numbers of both your neighbours?”, at least one person answers “Yes”
and at least one answers “No”.
What are the possible numbers of people answering “Yes” to this second
question?
A) 1 or 2 B) 1 or 3 C) 2 or 3 D) 2 or 4 E) 2, 3, or 4
37
Part 2
2. A palindromic number is a number that is the same when read from left
to right as when read from right to left. A number does not start with the
digit 0. To a six-digit palindromic number the palindromic number 21312
is added. The result is a seven-digit palindromic number.
What is this resulting number?
3. Using exactly six zeros and six ones we create two or more numbers which
we then multiply. For instance, we could get 10 × 10 × 10 × 10 × 10 × 10 =
1,000,000, or 10,011 × 100 × 1,100 = 1,101,210,000.
What is the largest possible result we could get in this way?
4. Pieter is staying at a hotel. The hotel has a ground floor (numbered 0) and
seven additional floors numbered 1 to 7. Pieter wants to make a trip by
elevator. He starts on one of the floors 1 to 7 and ends at the ground floor.
In between, he travels from floor to floor, never stopping at a previously
visited floor and never stopping at the ground floor (except for the last
stop).
The distance between any two consecutive floors is 3 metres. If, for ex-
ample, Pieter starts at floor 3, then goes to floor 6, then to floor 4 and
finally to the ground floor, he travels a total of (3 + 2 + 4) × 3 = 27 metres.
What is the maximum length of his trip in metres?
5. How many 3-digit numbers (the first digit cannot be 0) have the property
that adding all the digits gives a strictly greater result than multiplying
all the digits?
38
6. The mayor of a town wants to build a network of express trams. She wants
it to meet the following conditions:
• There are at least two distinct tram lines.
• Each tram line serves exactly three stops (also counting the start and
terminus).
• For each two tram stops in the town there is exactly one tram line
that serves both stops.
What is the minimum number of stops that the mayor’s tram network can
have?
E
7. A rectangle ABCD and a point E are given. D C
Line segments BE and BA have the same
length. Line segments CE and CB also have
the same length. Moreover, the area of rect-
angle ABCD is four times as large as the area
A B
of triangle BCE. Side AD has length 10.
What is the length of diagonal AC?
39
Solutions
Part 1
2. D) 2 6. D) 3
4. B) Bep 8. D) 2 or 4
Part 2
11
1. 12 5. 199
2. 1008001 6. 7
3. 12,321,000,000 7. 20
40
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2015
9 Second Round, March 2015
14 Final Round, September 2015
19 BxMO Team Selection Test, March 2016
23 IMO Team Selection Test 1, June 2016
28 IMO Team Selection Test 2, June 2016
31 IMO Team Selection Test 3, June 2016
35 Junior Mathematical Olympiad, October 2015
MY
We eat problems
CMY
Olympiad 2016
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
sponsored by: Centrum Wiskunde & Informatica sterdam 2011
OmslagNWO2017-v3.indd 1 10-06-17 17:15
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2016
8 Second Round, March 2016
13 Final Round, September 2016
19 BxMO Team Selection Test, March 2017
23 IMO Team Selection Test 1, June 2017
27 IMO Team Selection Test 2, June 2017
31 IMO Team Selection Test 3, June 2017
35 Junior Mathematical Olympiad, October 2016
The 1017 best students were invited to the second round, which was held in
March at twelve universities in the country. This round contained five open
questions, and two problems for which the students had to give extensive
solutions and proofs. The contest lasted 2.5 hours.
The 122 best students were invited to the final round. Also some outstanding
participants in the Kangaroo math contest or the Pythagoras Olympiad
were invited. In total about 150 students were invited. They also received
an invitation to some training sessions at the universities, in order to prepare
them for their participation in the final round.
The final round in September 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 55th edition 2016.
In February a team of four girls was chosen from the training group to
represent the Netherlands at the EGMO in Zürich, Switzerland, from 6
until 12 April. The team brought home a silver medal, a bronze medal,
and a honourable mention; a very nice achievement. For more information
about the EGMO (including the 2017 paper), see www.egmo.org.
In March a selection test of three and a half hours was held to determine the
ten students participating in the Benelux Mathematical Olympiad (BxMO),
held in Namur, Belgium, from 5 until 7 May. The Dutch team received
two gold medals, three silver medals and four bronze medals, and managed
to get the highest total score. For more information about the BxMO
(including the 2017 paper), see www.bxmo.org.
1
In June the team for the International Mathematical Olympiad 2017 was
selected by three team selection tests on 1, 2 and 3 June 2017, each lasting
four hours. A seventh, young, promising student was selected to accompany
the team to the IMO as an observer C. The team had a training camp in
Rio de Janeiro, from 8 until 15 July.
For younger students the Junior Mathematical Olympiad was held in Oc-
tober 2016 at the VU University Amsterdam. The students invited to
participate in this event were the 100 best students of grade 2 and grade 3
of the popular Kangaroo math contest. The competition consisted of two
one-hour parts, one with eight multiple choice questions and one with eight
open 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.
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
Dutch delegation
The Dutch team for IMO 2017 in Brazil consists of
3
First Round, January 2016
Problems
A-problems
1. Frank has two integers that add up to 26. Kees adds two more integers to
it and gets 41. Pieter adds another two integers and gets 58.
At least how many of the six integers that were added up are even?
A) 0 B) 1 C) 2 D) 3 E) 4
5. Five integers are written around a circle. Two neighbouring numbers never
add up to a multiple of three. Also, a number and its two neighbours never
add up to a multiple of three.
How many of the five integers are multiples of three?
A) 1 B) 2 C) 3 D) 4 E) 2 and 3 are both possible
4
6. In the figure you see a wire-frame model of a 2×2×2-cube
consisting of 8 small cubes with side length 1 dm. This
figure uses 54 dm of wire.
How many dm of wire are needed for a wire-frame model
of a 10×10×10-cube consisting of one thousand small
cubes with side length 1 dm?
A) 121 B) 1000 C) 1210 D) 3000 E) 3630
7. A square board is divided into 4×4 squares. At the start, all squares are
white. Now, we want to colour some of the squares blue, in such a way
that each blue square will be adjacent to exactly one white square (two
squares are called adjacent if they have a side in common).
What is the maximum number of squares that we can colour blue?
A) 6 B) 8 C) 10 D) 12 E) 14
B-problems
The answer to each B-problem is a number.
1. We construct a list of all positive integers that divide 707070. The numbers
are listed in decreasing order. The first number in the list is therefore
707070 and the last one is 1.
What is the seventh number in the list?
2. In the AO-language all words consist of only A’s and O’s and every possible
sequence of A’s and O’s is a word. There are, for example, eight three letter
words: ‘OOO’, ‘OOA’, ‘OAO’,. . . , ‘AAO’, and ‘AAA’. Words that contain
the letter combinations ‘AO’ and ‘OA’ equally often are called special. For
example, ‘AOAAOOOAA’ is special, because the word contains both letter
combinations ‘AO’ and ‘OA’ twice.
Find a special word consisting of four A’s and four O’s with the additional
property that after removing any of its letters, the resulting seven letter
word is again special.
5
D C
3. In the square ABCD lies a point U such that BU U
and AB have the same length. Point V is the
intersection of BU and the diagonal AC. The size V
?
of angle DAU is 28 degrees.
What is the size of the angle at V in triangle BV C?
28◦
A B
4. Seven people are suspects of a theft:
- Alex, a brown-haired man with blue eyes;
- Boris, a blond man with green eyes;
- Chris, a blond man with brown eyes;
- Denise, a blond woman with brown eyes;
- Eva, a brown-haired woman with blue eyes;
- Felix, a brown-haired man with brown eyes;
- Gaby, a blond woman with blue eyes.
Detectives Helga, Ingrid, and Julius know that one of the suspects is the
thief. After conducting some investigations they share their information.
Helga: “I know the eye and hair colour of the thief, but I do not know
who the thief is.”
Ingrid did not hear Helga and says:
“I know the hair colour and the gender, but I do not know who
the thief is.”
At last, Julius says:
“First I knew only the gender of the thief, but after hearing you
I know who the thief is.”
6
Solutions
A-problems
1. C) 2 5. B) 2
2. D) 36 6. E) 3630
3. B) 5 7. D) 12
4. E) 165 8. D) b + c − a < 3
B-problems
1. 70707
3. 101◦
4. Denise
7
Second Round, March 2016
Problems
B-problems
The answer to each B-problem is a number.
B1. How many of the integers from 10 to 99 have the property that the number
equals four times the sum of its two digits?
B2. In a box there are 100 cards that are numbered from 1 to 100. The numbers
are written on the cards. While being blindfolded, Lisa is going to draw
one or more cards from the box. After that, she will multiply together the
numbers on these cards.
Lisa wants the outcome of the multiplication to be divisible by 6. How
many cards does she need to draw to make sure that this will happen?
D C
P
B4. At championships of ‘The Settlers of Catan’, three participants play against
each other in each game. At a certain championship, three of the parti-
cipants were girls and they played against each other in the first game.
Each pair of participants met each other in exactly one game and in each
game at least one girl was playing.
What is the maximum number of participants that could have competed in
the championship?
8
B5. Triangle ABC has a right angle at B. Moreover,
the side length of AB is 1 and the side length of
A
BC is 2. On the side BC there are two points
D and E such that E lies between C and D and
1 G F
DEF G is a square, where F lies on AC and G
lies on the circle through B with centre A.
D E C
Determine the length of DE. B
2
C-problems For the C-problems not only the answer is important; you also have to
describe the way you solved the problem.
C1. A positive integer is called 2016-invariant if the sum of its digits does
not change when you add 2016 to the integer. For example, the integer
8312 is 2016-invariant: the sum of the digits of 8312 is 8 + 3 + 1 + 2 = 14,
and this equals the sum of the digits of 8312 + 2016 = 10328, which is
1 + 0 + 3 + 2 + 8 = 14.
(a) Determine the largest four-digit number that is 2016-invariant.
(b) There are 9999 positive integers having at most four digits.
Determine how many of these are 2016-invariant.
C2. For the upcoming exam, the desks in a hall are arranged in n rows containing
m desks each. We know that m ≥ 3 and n ≥ 3. Each of the desks is occupied
by a student. Students who are seated directly next to each other, in front
of each other, or diagonally from each other, are called neighbours. Thus,
students in the middle of the hall have 8 neighbours. Before the start of
the exam, each student shakes hands once with each of their neighbours.
In total, there are 1020 handshakes.
Determine the number of students.
9
Solutions
B-problems
1. 1
2. 68
3. 5
4. 7
2
5. 5
C-problems
C1. When adding two integers, we write one on top of the other and then add
the digits from right to left. If at a certain position the sum of the two
digits is greater than 9, a carry occurs at this position: we must carry a 1
to the addition of the two digits in the next position (the position to the
left).
Let n be a number whose sum of digits equals s. Now consider the
number n + 2016. We can derive the sum of its digits from the addition
procedure. If no carry occurs, then the sum of the digits of n + 2016 equals
s + 2 + 1 + 6 = s + 9. If carries do occur, then with each carry the sum
of the digits decreases by 9. After all, if two digits x and y add up to a
number greater than 9, the digit below these two will be x + y − 10 instead
of x + y, while we carry a 1 causing the digit to the left to increase by 1.
In total, the sum of the digits thus decreases by 9 for each carry.
For example, taking n = 1015 (and s = 1 + 0 + 1 + 5 = 7), then there will
be one carry in the addition n + 2016 = 3031. For the sum of the digits we
have 3 + 0 + 3 + 1 = (s + 9) − 1 · 9.
As another example, take n = 3084 (where s = 15). Then, there will be two
carries in the addition n + 2016 = 5100 and indeed we have 5 + 1 + 0 + 0 =
(s + 9) − 2 · 9.
Hence, a number is 2016-invariant if and only if exactly one carry occurs
when adding 2016 to it.
10
(a) After all observations in the previous paragraph, we conclude that
we are looking for the greatest four-digit number having the property
that you need to carry exactly one 1 when adding 2016 to it. This
number is 9983. In the addition 9983 + 2016 exactly one carry occurs
and if you raise one of the digits 8 or 3, a second carry occurs.
(b) To count how many of the numbers n from 1 to 9999 give exactly
one carry when adding 2016 to it, we consider four cases according to
the position in which the carry occurs. In each case we consider the
possible values of the digits of n.
• First, consider the case that the only carry occurs in the thousands
place. The thousands digit must then be either 8 or 9. The unit
digit can only be 0, 1, 2, or 3. The tens digit can be any digit
except for 9 (because 1 + 9 would give a carry in the tens place).
The hundreds digit can be any value. Hence, in total there are
2 · 10 · 9 · 4 = 720 numbers in this case.
• There are no numbers for which a carry occurs only in the hundreds
place (a carry there can only occur if we also carry a 1 from the
tens place).
• Next, consider the case that the only carry occurs in the tens
place. For the units digit, the only possible values are 0, 1, 2,
and 3. The tens digit must be a 9. Because you now need to
carry a 1 from the tens place to the hundreds place, the digit in
the hundreds place cannot be a 9 anymore since this would cause
another carry (for example, in the addition 990 + 2016). The digit
in the thousands place can take all values except 8 and 9. Hence,
in total there are 8 · 9 · 1 · 4 = 288 numbers in this case.
• Finally, consider the case that the only carry occurs in the units
place. For this digit, the possible values are 4, 5, 6, 7, 8, and 9.
The digit in the tens place cannot be 9 or 8, because then another
carry would occur (for example in 84 + 2016). The hundreds digit
can be any value and the thousands digit can be any value except
8 and 9. Hence, in total we find 8 · 10 · 8 · 6 = 3840 numbers in
this case.
Altogether, there are 720 + 288 + 3840 = 4848 numbers from 1 to 9999
that are 2016-invariant.
11
C2. There are four students seated in the corners, since n, m ≥ 2. They each
shook the hand of three other people. Not counting the students in the
corners, there are m − 2 students who are seated completely in the front,
m − 2 who are seated in the back, n − 2 students who are seated in the
leftmost row and n−2 in the rightmost row. These 2n+2m−8 students each
shook the hand of five other people. The remaining mn−4−(2n+2m−8) =
mn − 2n − 2m + 4 students each shook the hand of eight other people.
Adding the numbers of handshakes for each of the students gives a total of
12
Final Round, September 2016
Problems
13
3. Find all possible triples (a, b, c) of positive integers with the following
properties:
• gcd(a, b) = gcd(a, c) = gcd(b, c) = 1;
• a is a divisor of a + b + c;
• b is a divisor of a + b + c;
• c is a divisor of a + b + c.
5. Bas has coloured each of the positive integers. He had several colours at
his disposal. His colouring satisfies the following requirements:
• each odd integer is coloured blue;
• each integer n has the same colour as 4n;
• each integer n has the same colour as at least one of the integers n + 2
and n + 4.
Prove that Bas has coloured all integers blue.
14
Solutions
1. (a) For brevity, we will say that a number encircled in red is a red number,
and similarly for blue. So some numbers could be both red and blue.
Since there are 999 numbers written on the pavement, of which 500
are red and 500 are blue, we have at least one bicoloured number
by the pigeonhole principle. Consider such a bicoloured number and
suppose it is the number k. From left to right, the red numbers form
the sequence 1, 2, . . . , 500. Hence, to the left of the bicoloured number
we have the red numbers 1 to k − 1, and to the right we have the red
numbers k + 1 to 500. The blue numbers are written in the opposite
order: from left to right they form the sequence 500, 499, . . . , 1. To
the left of the bicoloured number we therefore have the blue numbers
k + 1 to 500, and to the right we have the blue numbers 1 to k − 1.
We count how many numbers there are on each side of the bicoloured
number. On the left we have the red numbers 1 to k − 1 and the blue
numbers k + 1 to 500. Hence, there are at least 499 distinct numbers
on the left. On the right we have the red numbers k + 1 to 500 and
the blue numbers 1 to k − 1. Again at least 499 distinct numbers.
Since there are only 999 = 499 + 1 + 499 numbers on the pavement, we
have already considered all numbers on the pavement. We conclude
that there are precisely 499 numbers on each side of the bicoloured
number. The bicoloured number is therefore precisely in the middle
of the sequence.
(b) Just as in part (a), the pigeonhole principle yields that at least one
number is bicoloured, i.e. both red and green. If more than one number
is bicoloured, one of them is not exactly in the middle of the sequence
and we are done. Therefore, it suffices to examine the case where
there is exactly one bicoloured number. Let this number be k.
Again, we count how many numbers there are on each side of the
bicoloured number. On the left we have the red numbers 1 to k − 1
and the green numbers 1 to k − 1. Since none of these numbers is
bicoloured, there are at least 2 · (k − 1) distinct numbers on the left.
On the right we have the red numbers k + 1 to 500 and the green
numbers k + 1 to 500. Since none of these numbers is bicoloured,
we have at least 2 · (500 − k) distinct numbers on the right. Since
2 · (k − 1) + 1 + 2 · (500 − k) = 999, we have already counted all numbers.
On the left of the bicoloured number we therefore have exactly 2·(k−1)
numbers, and on the right we have exactly 2 · (500 − k) numbers. Since
2 · (k − 1) is an even number, it is unequal to 499. We conclude that
the bicoloured number is not exactly in the middle of the sequence.
15
2. Suppose that our sequence has x ones, y minus ones (and hence 2n − x −
y zeroes). We calculate the sum product value of the sequence (as an
expression in x and y).
In the sum product value, six different types of terms occur: 1 · 1, 1 · −1,
−1 · −1, 1 · 0, −1 · 0, and 0 · 0. Only the first three types contribute since
the other types are equal to 0.
The number of terms of the type 1 · 1 = 1 equals the number of ways to
select two out of x ones. This can be done in x(x−1) 2 ways: there are x
options for the first 1, and then x − 1 options for the second 1. Since the
order in which we select the two ones does not matter, we effectively count
each possible pair twice.
y(y−1)
Similarly, the number of terms of the type −1 · −1 = 1 is equal to 2 .
The number of terms of the type 1 · −1 = −1 is equal to xy, since there
are x options for choosing a 1 and, independently, there are y options for
choosing a −1.
In total, we obtain a sum product value of
16
If a + b + c = 2c, then c = a + b. We know that b must be a divisor of
a + b + c = 2a + 2b. Since a > 0, we have 2a + 2b > 2b. Since b ≥ a, we
also have 2a + 2b ≤ 4b. Therefore, since 2a + 2b must be a multiple of b,
there are only two possibilities: 2a + 2b = 3b or 2a + 2b = 4b. Again, we
consider these cases separately.
If 2a + 2b = 3b, then b = 2a. Similarly to the first case, we find that
a = gcd(a, 2a) = gcd(a, b) = 1. Therefore, b = 2 and c = a + b = 3. The
resulting triple (a, b, c) = (1, 2, 3) is indeed a solution, since gcd(1, 2) =
gcd(1, 3) = gcd(2, 3) = 1 and 1 + 2 + 3 = 6 is divisible by 1, 2, and 3.
If 2a + 2b = 4b, then a = b. Again we see that a = gcd(a, a) = gcd(a, b) = 1.
From b = a = 1 it follows that c = a + b = 2 and hence (a, b, c) = (1, 1, 2).
This is indeed a solution since gcd(1, 1) = gcd(1, 2) = 1 and 1 + 1 + 2 = 4
is divisible by 1 and 2.
We conclude that the solutions for which a ≤ b ≤ c holds are: (a, b, c) =
(1, 1, 1), (a, b, c) = (1, 1, 2), and (a, b, c) = (1, 2, 3). Permuting the values of
a, b and c, we obtain a total of ten solutions (a, b, c):
(1, 2, 3), (1, 3, 2), (2, 1, 3), (2, 3, 1), (3, 1, 2), (3, 2, 1).
D
4. Version for junior students
Let K be the intersection of AX and BD, C
L
and let L be the intersection of CY and BD. Y
Consider the triangles P LY and P KX.
The angles ∠P LY and ∠P KX are both P
right angles. The angles ∠Y P L and ∠XP K
are opposite angles and therefore equal. X K
A
Since |P X| = |P Y |, we see that the tri-
angles P KX and P LY are congruent (SAA).
Hence, we have |P K| = |P L|. B
Now consider triangles P AK and P CL. Angles ∠AKP and ∠CLP are both
right angles. Angles ∠KP A and ∠LP C are opposite angles, hence equal.
We have already shown that |P K| = |P L|. Therefore, triangles P AK and
P CL are congruent (ASA). From this, we conclude that |AP | = |P C|.
In a similar fashion, we may deduce that |BP | = |DP |. The two diagonals
of ABCD bisect each other, hence ABCD is a parallelogram.
17
4. Version for senior students
We start by observing that we have that ∠CM Y = 12 ∠CM A = 12 (180◦ −
∠AM B) = 90◦ − ∠XM B. Since ∠M XB = 90◦ , and the angles of triangle
BM X sum to 180◦ , we see that ∠CM Y = 90◦ − ∠XM B = ∠M BX.
Looking at triangles CM Y and M BX, we observe that ∠CM Y = ∠M BX,
∠M Y C = 90◦ = ∠BXM , and |CM | = |M B|. The two triangles are
therefore congruent (SAA). In particular, we obtain the equalities |M X| =
|CY | and |M Y | = |BX|.
Now consider triangle XY M . We already know that |M Y | = |BX| and
that ∠Y M X = ∠Y M A + ∠AM X = 12 ∠CM A + 12 ∠AM B = 12 · 180◦ =
90◦ . Since triangles XY M and M BX also share the side M X, they are
congruent (SAS).
In particular, we see that ∠M XY = ∠XM B. Since M X is the angle
bisector of ∠AM B, we have ∠XM B = ∠AM X. This implies that triangle
M XZ has two equal angles and is therefore an isosceles triangle with vertex
angle Z. We conclude that |M Z| = |XZ|.
In a similar manner, we see that triangles XY M and CM Y are congruent,
and find that ∠XY M = ∠CM Y = ∠Y M A. Triangle M Y Z is therefore
isosceles with vertex angle Z. This implies that |Y Z| = |M Z|. Together
with |M Z| = |XZ|, this concludes the proof.
5. Suppose that not all numbers are coloured blue. Then, there must be a
number k that is not blue. We will use this to derive a contradiction.
Without loss of generality, we may assume that k is coloured red. Since all
odd numbers are blue, k must be even, say k = 2m for some integer m ≥ 1.
From the second requirement, it follows that 8m is red as well. From the
third requirement, it now follows that at least one of the numbers 8m + 2
and 8m + 4 is red. However, 2m + 1 is odd and therefore blue, which by
the second requirement implies that 8m + 4 = 4 · (2m + 1) is blue as well.
So 8m + 2 must be red.
By the third requirement, 8m − 2 must be the same colour as 8m or 8m + 2.
Since both 8m and 8m + 2 are red, this implies that 8m − 2 must be red
as well. Since 8m and 8m − 2 are red, this implies (again by the third
requirement) that 8m − 4 is also red. The second requirement now implies
that (8m − 4)/4 = 2m − 1 is also red. But that is impossible since 2m − 1
is odd, and therefore blue.
We conclude that the assumption that not all numbers are blue leads to a
contradiction. Therefore, all numbers must be blue.
18
BxMO Team Selection Test, March 2017
Problems
5. Determine all pairs of prime numbers (p, q) such that p2 + 5pq + 4q 2 is the
square of an integer.
19
Solutions
1. We will prove that the minimum number of integers in a complete sequence
is 2. First consider the case n = 2. Let a1 and a2 be the numbers in
the sequence. Then either a1 or a2 is integral. Without loss of generality
assume a1 is integral. Moreover, a1 + a2 is integral, hence also a2 is integral.
Therefore, the sequence contains at least two integers.
Now consider the case n > 2. Write n = 2k (because n is even) with
k ≥ 2. Then either a1 + a2 + . . . + ak or ak+1 + ak+2 + . . . + a2k is integral.
But, as the sum of both expressions is also integral, they are both integral.
Moreover, either a1 + a2 + . . . + ak−1 or ak+2 + ak+3 + . . . + a2k is integral.
This yields that either ak or ak+1 is integral. Moreover, we know that a1
or a2k is integral, and these do not coincide with ak or ak+1 because k ≥ 2.
Hence, at least two different numbers are integers.
Finally, we will show for each even integer n that it is possible to write a
complete sequence with exactly two integers. Again write n = 2k. If k is
odd, we take a1 = ak+1 = 1 and all other terms equal to 12 . The sum of
all numbers in the sequence is integral, hence it is sufficient to show that
the sum of the first or last m terms is integral for 1 ≤ m ≤ k; the cases in
which m > k follow directly. For odd m ≤ k, the sum of the first m terms
is integral, for even m < k, the sum of the last m terms is integral.
If k is even, we take a1 = ak = 1 and all other terms equal to 12 . For
odd m < k, the sum of the first m terms is integral, and for even m ≤ k,
the sum of the last m terms is integral. Moreover, the sum of all terms is
integral, hence the requirement is also met for m > k.
We conclude that the minimum number of integers in a complete sequence
of n numbers is 2.
2. We will first prove that for prime numbers p and positive integers k we
have f (pk ) = kpk−1 . We will prove this using mathematical induction to k.
For k = 1, the statement becomes f (p) = 1, which is known to hold. Now,
let l ≥ 1 and suppose that we proved the statement for k = l. Consider
k = l + 1. Then we apply the second property with x = p and y = pl :
This finished the induction. Now we will prove that for distinct prime
numbers p1 , p2 , . . . , pt and positive integers a1 , a2 , . . . , at we have
f (pa1 1 pa2 2 · · · pat t ) = pa1 1 pa2 2 · · · pat t · ap11 + ap22 + · · · + aptt .
20
We will prove this by induction to t. For t = 1, this is exactly the formula
that we have just proven. Now, let r ≥ 1 and suppose that the statement
is proved for t = r. Then we apply the second property again:
ar+1
f pa1 1 pa2 2 · · · par r · pr+1
ar+1 ar+1
= pa1 1 pa2 2 · · · par r · f pr+1 + pr+1 · f (pa1 1 pa2 2 · · · par r )
ar+1 −1 ar+1
= pa1 1 p2a2 · · · par r · ar+1 pr+1 + pr+1 · pa1 1 pa2 2 · · · par r · ap11 + ap22 + · · · + aprr
ar+1 ar+1 ar+1
= pa1 1 pa2 2 · · · par r pr+1 · pr+1 + pa1 1 pa2 2 · · · par r pr+1 · ap11 + ap22 + · · · + aprr
ar+1
= pa1 1 pa2 2 · · · par r pr+1 · ap11 + ap22 + · · · + aprr + apr+1 r+1
.
or
a1 a2 at
p1 + p2 + ··· + pt = 1.
Multiplying this with p1 p2 · · · pt yields
a1 p2 p3 · · · pt + a2 p1 p3 · · · pt + . . . + at p1 p2 · · · pt−1 = p1 p2 · · · pt .
21
Moreover, EF is a mid-parallel in triangle ADC, hence EF k BC. There-
fore, we have ∠CBF = ∠BF E. Using the inscribed angle theorem and
the sum of the angles in the isosceles triangle EBM , we obtain ∠BF E =
1 ◦ ◦
2 ∠BM E = 90 − ∠EBM . Therefore, ∠ABE = ∠CBF = 90 − ∠EBM .
◦
We conclude that ∠ABM = ∠ABE + ∠EBM = 90 , and AC k BM .
22
IMO Team Selection Test 1, June 2017
Problems
4. Let ABC be a triangle, let M be the midpoint of AB, and let N be the
midpoint of CM . Let X be a point satisfying both ∠XM C = ∠M BC and
∠XCM = ∠M CB such that X and B lie on opposite sides of CM . Let Ω
be the circumcircle of triangle AM X.
(a) Show that CM is tangent to Ω.
(b) Show that the lines N X and AC intersect on Ω.
23
Solutions
2. Number the vertices of the 2n-gon from 1 up to 2n, clockwise, and let ai be
the value of vertex i. Since the number of vertices of the polygon is even,
we can pair each vertex to the one directly opposite to it. Sum the values
of each pair of opposite vertices to get the numbers si = ai + ai+n .
If four vertices A, B, C, and D lie in that order on a rectangle, then
∠ABC = 90◦ and ∠ADC = 90◦ , so B and D lie on the circle with
diameter AC by Thales’s theorem. As all vertices of the polygon lie on its
circumcircle, AC is the diameter of this circumcircle. Therefore A and C
are opposite vertices, and the same holds for B and D. Conversely, if A and
C are opposite vertices, and so are B and D, then ABCD is a rectangle,
again by Thales’s theorem. In short, four vertices form a rectangle if and
only if they form two pairs of opposite vertices.
So the condition in the problem is equivalent to si + sj being prime for
all 1 ≤ i < j ≤ n. Suppose that at least three of the si have the same
24
parity, say sj , sk , and sl . Then the sum of each pair of these three numbers
is even, but prime as well, so this sum must be 2 in all cases. Hence
sj + sk = sk + sl = sj + sl , from which follows that sj = sk = sl . Since
sj + sk = 2, we have sj = sk = sk = 1. Therefore at most two of the si are
even, and if three or more of the si are odd, then the odd si all equal 1.
The sum of all si is equal to the sum of the values of the vertices, which is
On the one hand, this sum is congruent to n modulo 2. On the other hand,
modulo 2 this sum must be congruent tot the number of odd si . Hence
the number of odd si has the same parity as n. As there are n of the si , it
follows that the number of even si is even. Since we have already seen that
the number of even si is at most 2, it follows this number is either 0 or 2.
Suppose that n = 4 and and that there are two even si . Then there are
also two odd si . The sum of the even si , as well as that of the odd si , is
equal to 2, so the sum of the si is equal to 4 = n. Now suppose that n ≥ 5
and that there are two even si . Then n − 2 ≥ 3 of the si are odd, so all odd
si are equal to 1. The sum of the two even si is 2, so the sum of the si is
equal to (n − 2) + 2 = n. Now suppose that there are no even si . Then all
si are odd, and therefore equal to 1, so again, the sum of the si equals n.
So in all of the cases, the sum of the si is n. We have also seen that this
sum is equal to 2mn + n(2n + 1). Therefore 2mn + n(2n + 1) = n, so
dividing by the non-zero integer n gives the equality 2m + 2n + 1 = 1.
Hence m = −n.
Finally, we show that if m = −n, there indeed exists a solution. Let
ai = i and an+i = 1 − i for 1 ≤ i ≤ n. Then the integers 1, 2, . . . , n and
the integers 0, −1, . . . , −n + 1 are values of vertices; these are precisely
the integers m + 1, m + 2, . . . , m + 2n for m = −n. Moreover, we have
si = i + (1 − i) = 1 for all i. Hence all rectangular sums are equal to 2,
which is a prime number.
Therefore m = −n is the only value of m for which the there exists a
solution.
25
Using the identity a3 − b3 = (a − b)(a2 + ab + b2 ) we can also write this as
(x + y)3 − x3 y 3 = (x + y − xy) (x + y)2 + xy(x + y) + (xy)2
= (x + y − xy)(x2 + y 2 + 2xy + x2 y + xy 2 + x2 y 2 ).
Therefore the two right hand sides are equal, so we see that either x+y−xy =
0 or x2 +y 2 +2xy +x2 y +xy 2 +x2 y 2 = 3xy. The latter case can be rewritten
as
0 = x2 + y 2 − xy + x2 y + xy 2 + x2 y 2
= 12 (x − y)2 + 12 x2 (y 2 + 2y + 1) + 12 y 2 (x2 + 2x + 1)
= 21 (x − y)2 + 12 x2 (y + 1)2 + 12 y 2 (x + 1)2 .
So since the sum of three squares must be 0, each of the squares must be
0 as well. Hence x = y = −1. This indeed gives a solution of the given
equation, and we have x1 + y1 = 2.
In the remaining case we have x + y − xy = 0, so y1 + x1 − 1 = 0, from which
we deduce that x1 + y1 = 1. This is attained for example for x = y = 2.
Therefore the possible values are −2 and 1.
C
X
S
4. We consider the configuration in the figure; N
the proof is similar for the other configur-
ations. A B
M
(a) First, we have 4XM C ∼ 4M BC.
We have ∠AM X = 180◦ − ∠XM C − ∠BM C = 180◦ − ∠XM C −
|AM | |BM | |M C|
∠M XC = ∠M CX, and |M X| = |M X| = |CX| , so we have 4AM X ∼
4M CX. Hence ∠XAM = ∠XM C, so using the converse of the
tangent chord angle theorem we see that CM is tangent to Ω.
(b) Let S be the second intersection point of AC and Ω (or the point at
which AC and Ω are tangent; which then would be A). We need to
show that S lies on N X. By the tangent chord angle theorem we have
∠SM C = ∠SAM = ∠CAM . (If S = A, then ∠SM C = ∠CAM
follows from CA and CM both being tangents, as then 4CAM is
isosceles.) Therefore 4CSM ∼ 4CM A (AA), so |CM | |SM |
|CA| = |M A| .
1 |M N |
Since |CM | = 2|M N | and |M A| = 2 |AB|, it follows that |CA| =
|SM |
|AB| .
Togther with ∠SM N = ∠SM C = ∠CAM = ∠CAB, we get
4SN M ∼ 4BCA. Hence ∠M SN = ∠ABC. Moreover, we have
∠XSM = 180◦ − ∠XAM . In (a), we have seen that ∠XAM =
∠XM C = ∠M BC = ∠ABC, so ∠XSM = 180◦ − ∠ABC = 180◦ −
∠M SN . Therefore S lies on XN .
26
IMO Team Selection Test 2, June 2017
Problems
for all x, y ∈ R.
27
Solutions
1. (a) Suppose for a contradiction that two of a2 , b2 , and c2 have the same
remainder after division by p, say a2 ≡ b2 mod p. Then p | a2 − b2 =
(a − b)(a + b), so p | a − b or p | a + b. In the latter case, we have
p ≤ a + b ≤ c(a + b) < ab + bc + ca = p, which is a contradiction. In
the former case, we see that by a 6= b, we have p ≤ |a − b| ≤ a + b < p,
which is a contradiction.
(b) Suppose for a contradiction that two of a3 , b3 , and c3 have the same
remainder after division by p, say a3 ≡ b3 mod p. Then p | a3 − b3 =
(a−b)(a2 +ab+b2 ), so p | a−b or p | a2 +ab+b2 . The former case leads
to a contradiction, as we have seen in (a). Consider the remaining case:
p | a2 +ab+b2 . Then p | a2 +ab+b2 +(ab+bc+ca) = (a+b)(a+b+c), so
p | a+b or p | a+b+c. However, note that a+b < a+b+c < ab+bc+ca,
since a, b, and c cannot all be equal to 1. Therefore p is neither a
divisor of a + b, nor a + b + c, which is a contradiction.
28
Since HLIK is also a cyclic quadrilateral (as it has two opposite right
angles) we have
We deduce that ∠DF E + ∠DSE = 180◦ , and it follows that S lies on the
circumcircle of 4DEF , which is the incircle of triangle 4ABC.
as desired.
29
4. Substituting x = 0 gives (y + 1)f (0) + f (f (y)) = y, so f (f (y)) = y · (1 −
f (0)) − f (0). If f (0) 6= 1, the right hand side is a bijective function of y,
hence so is the left hand side. So in this case, f is bijective.
We will next show that in case f (0) = 1, the function f is bijective as well.
So suppose that f (0) = 1. Then f (f (y)) = −1 for all y ∈ R. Substituting
y = 0 gives f (x) + f (x + f (x)) = 0, so f (x + f (x)) = −f (x). Substituting
x = f (z) and y = z and replacing each occurrence of f (f (z)) by −1, we
see that
(z + 1) · −1 + f f (z)f (z) + f (z + f (z)) = z,
so using that f (z + f (z)) = −f (z), we see that
f f (z)2 − f (z) = 2z + 1.
30
IMO Team Selection Test 3, June 2017
Problems
3. Compute the product of all positive integers n for which 3(n!+1) is divisible
by 2n − 5.
31
Solutions
∠M V L = ∠P OL = ∠QAL = ∠M AL,
2. We have bn = 0. Suppose that there exists an i ≤ n−1 such that bi > 0 and
ai+1 ≥ 0. Then from bi (aj+1 − ai+1 ) ≥ 0 it follows that aj+1 ≥ ai+1 ≥ 0,
for all i ≤ j ≤ n − 1. So bn = bi + ai+1 + ai+2 + · · · + an ≥ bi > 0, which
is a contradiction. Therefore bi > 0 implies that ai+1 < 0. Analogously, if
bi < 0 then ai+1 > 0.
Let k now be such that |bk | = max1≤m≤n |bm |. We may assume without
loss of generality that bk > 0 (by multiplying all ai by −1 if necessary).
If k = 1, then bk = a1 , so |bk | = |a1 | ≤ max1≤`≤n |a` |, in which case we
are done. Now suppose k > 1. If bk−1 > 0, then by the above, we have
ak < 0. On the other hand, ak = bk − bk−1 ≥ 0 since bk was maximal,
which is a contradiction. Hence ak = bk − bk−1 = bk + |bk−1 | ≥ bk , so
|bk | ≤ |ak | ≤ max1≤`≤n |a` |, which is what we needed to prove.
32
n > 4 we have 3 - n! + 1, so 3(n! + 1) is divisible by at most one factor 3.
So in this case, n cannot satisfy the given condition either.
So for n > 4 satisfying the given condition, we must have that 2n − 5
is a prime number greater than 3. Write q = 2n − 5. Then q | n! + 1,
or equivalently, n! ≡ −1 mod q. Moreover, Wilson’s theorem states that
(q − 1)! ≡ −1 mod q. Therefore
11! = 1 · (2 · 9) · (3 · 6) · (5 · 7) · 4 · 8 · 10 · 11
≡ 4 · 8 · 10 · 11 = 88 · 40 ≡ 3 · 6 ≡ 1 mod 17,
n n+1
4. We prove the smallest such m equals 2 if n is even and 2 if n is odd.
Choose the n points in such a way that they all lie on a circle, and denote
them by P1 , P2 , . . . , Pn , in the order in which they lie on the circle. The
n segments P1 P2 , P2 P3 , . . . , Pn P1 must all be intersected by at least one
33
line. As every line intersects the circle at most twice, every line intersects
at most two of these segments. This shows that the number of lines is at
least n2 if n is even and at least n+1
2 if n is odd.
We now show this number of lines suffices. First we show that given four
distinct points P1 , P2 , and Q1 , Q2 , there always exists a line with respect
to which P1 and P2 lie on opposite sides, as well as Q1 and Q2 . Take the
line through the midpoints of the segments P1 P2 and Q1 Q2 . Suppose this
line passes through P1 . Then this line passes through P2 as well, and as no
three of the given points lie on a line, this line does not pass through Q1 ,
nor through Q2 . Hence one can rotate this line a bit around the midpoint of
P1 P2 in such a way that it still intersects the segment Q1 Q2 , and such that
neither P1 nor P2 lie on it. The three remaining cases are done analogously.
Now if this line still passes through one of the other given points, one can
move this line around a tiny bit for it to no longer pass through any of the
given points; this is always possible as there are only finitely many points.
Now P1 and P2 , as well as Q1 and Q2 , lie on opposite sides of this line.
Assume that n is even; if n is odd, we add an arbitrary point that is not on
any line through any pair of given points. The lines that we are about to
construct, still give a correct example if we remove the extra point. So we
need to construct n2 lines. Fix an arbitrary line such that no line through
two of the given points is parallel to it; this is possible as there are only
finitely many pairs of points, and infinitely many directions to choose from.
Translate this line across the plane. This line will meet the given points
one by one. So at some point, there are no points on the line, and on both
sides of the line there are n2 points. This will be our first line.
The plane is now split into two regions, say the left region, and the right
region. We now add lines in such a way that every line creates a new region
in the left region, as well as the right one. (We will only consider regions
containing at least one of the given points.) To this end, we pick two points
in the left region that aren’t separated yet by a line, and two points in
the right region that aren’t separated yet by a line, and pick a line that
separates the two points in the left region, as well as the points in the right
region; we have seen before that this is possible. This creates a new region
in both the left region and the right region, as the line separates pairs of
points in both regions that weren’t separated before. If at some point,
either the left or the right region no longer contains pairs of points that
are not separated, then we will only use the points in the remaining region.
After adding n2 − 1 lines, we will have, in both the left and the right region,
at least n2 regions, each containing at least one given point; so each region
has to contain exactly one given point. In total we have used n2 lines. Hence,
it is always possible to satisfy the given condition with n2 lines.
34
Junior Mathematical Olympiad, October 2016
Problems
Part 1
1. The sum of the digits of a number is obtained by adding the digits of this
number. For example, the sum of the digits of 76 equals 7 + 6 = 13. The
sum of the digits of the double of 76 is 1 + 5 + 2 = 8.
How many numbers consisting of two digits are there for which the sum of
the digits equals the sum of the digits of the double of the number?
Attention: the first digit cannot be a zero. Thus, the number 09, for example,
is ruled out.
A) 0 B) 8 C) 9 D) 10 E) 11
A) Only Dion and Thijs D) Only Dion, Huub, Peter, and Thijs
B) Only Dion, Jaap, and Thijs E) Everybody except Birgit
C) Only Huub, Jaap, and Peter
35
4. A baker’s helper is filling cream puffs for one and a half hours. He is not
in a hurry and he is filling two cream puffs every minute. At some point,
the baker comes in and supervises the helper for a while. This motivates
the helper to work a bit faster: during this time he is filling three cream
puffs per minute. As soon as the baker has left, the helper reverts to filling
the cream puffs at the initial, slow, pace. Afterwards, it turns out that
during any continuous one hour period within these one and a half hours,
the helper has filled exactly 140 cream puffs.
How many cream puffs are filled by the helper during the full one and a
half hours?
A) 180 B) 200 C) 210 D) 230 E) 270
5. Start with the number 60. Then, keep repeating the following two steps:
(1) Throw a die and look at the number that comes up.
(2) If your number is divisible by the number on the die, then you divide
your number by the number on the die. If not, then you multiply your
number by the number on the die.
In this way, you obtain a sequence of numbers. If your first three rolls are
5, 6, and 3, consecutively, then the first four numbers in your sequence are
60, 60/5 = 12, 12/6 = 2, and 2 × 3 = 6.
What is the greatest number you can obtain in this way?
A) 60 B) 120 C) 240
D) 360 E) You can obtain arbitrarily large numbers
6. Harry and Hermione are trapped in a room in which 6 bottles are put next
to each other. From left to right, the bottles are numbered 1 to 6. One
of the bottles contains a potion that helps them to escape. On a piece of
paper there are four clues to help them:
• 3 of the bottles contain poison, 2 of the bottles contain sleeping
draught; the remaining bottle contains the potion to escape.
• Immediately left to the sleeping draught there is poison.
• The smallest bottle contains poison.
• The second bottle from the left and the second bottle from the right
have the same content.
36
Hermione now has enough information to identify the bottle which contains
the potion to escape. Which bottle is the smallest bottle?
A) Bottle 1 B) Bottle 2 or 5 C) Bottle 3
D) Botttle 4 E) Bottle 6
8. Between the digits of the number 2016, we put one or more symbols from
×, +, and − (you are allowed to use a symbol multiple times). In this
way, we can create different numbers, such as 20 + 1 × 6, which is 26, and
201 × 6, which is 1206.
How many of the numbers from 1 to 10 are, just like 26 and 1026, the result
of such a calculation? (Attention: you cannot put a − before the 2!)
A) 6 B) 7 C) 8 D) 9 E) 10
37
Part 2
2. Alice, Bob, Carla, Daan, and Eva are standing in this order along a circle
(Bob is standing to the left of Alice). Each of them has a number of sweets,
they have 100 sweets in total. All at the same time, they give part of their
sweets to their left neighbour: Alice gives away 13 of her sweets, Bob 14 ,
Carla 15 , Daan 16 , and Eva 17 . After this, everybody has the same number
of sweets as before.
How many sweets does Eva have? E
D C
3. In the figure on the right, rectangles ABCD and
F
BDEF are shown. The length of AB is 8 and the
length of BC is 5.
What is the area of pentagon ABF ED?
A B
5. How many times a day (which is 24 hours) are the small hand and the big
hand of the clock perpendicular?
38
6. Janneke, Karin, Lies, Marieke, and Nadine participated in a running race.
They all finished at distinct times except for two of them; they finished at
the same time. Moreover, we know that:
7. For all positive integers a and b we make the number a ♥ b. The following
rules hold:
• rule 1: 1 ♥ 1 = 1;
• rule 2: a ♥ b = b ♥ a;
• rule 3: a ♥(b + c) = a + (a ♥ b) + (a ♥ c).
From these rules it follows, for example, that
2 ♥ 1 = 1 ♥ 2 = 1 ♥(1 + 1) = 1 + 1 ♥ 1 + 1 ♥ 1 = 1 + 1 + 1 = 3.
Calculate 20 ♥ 16.
39
Solutions
Part 1
1. D) 10
3. B) 2
4. B) 200
6. E) Bottle 6
7. C) 5
8. B) 7
Part 2
1. 90 5. 44
3. 60 7. 924
4. 8 8. 17
40
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2016
8 Second Round, March 2016
13 Final Round, September 2016
19 BxMO Team Selection Test, March 2017
23 IMO Team Selection Test 1, June 2017
27 IMO Team Selection Test 2, June 2017
31 IMO Team Selection Test 3, June 2017
35 Junior Mathematical Olympiad, October 2016
MY
We eat problems
CMY
Olympiad 2017
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
sponsored by: Centrum Wiskunde & Informatica sterdam 2011
OmslagNWO2018-v2.indd 1 17-06-18 13:42
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
3 First Round, January 2017
7 Second Round, March 2017
13 Final Round, September 2017
19 BxMO Team Selection Test, March 2018
23 IMO Team Selection Test 1, June 2018
27 IMO Team Selection Test 2, June 2018
31 IMO Team Selection Test 3, June 2018
36 Junior Mathematical Olympiad, September 2017
The 987 best students were invited to the second round, which was held in
March at twelve universities in the country. This round contained five open
questions, and two problems for which the students had to give extensive
solutions and proofs. The contest lasted 2.5 hours.
The 124 best students were invited to the final round. Also some outstanding
participants in the Kangaroo math contest or the Pythagoras Olympiad
were invited. In total about 160 students were invited. They also received
an invitation to some training sessions at the universities, in order to prepare
them for their participation in the final round.
The final round in September 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 56th edition 2017.
In February a team of four girls was chosen from the training group to
represent the Netherlands at the EGMO in Florence, Italy, from 9 until 15
April. The team brought home a silver medal, two bronze medals, and a
honourable mention; a very nice achievement. For more information about
the EGMO (including the 2018 paper), see www.egmo.org.
In March a selection test of three and a half hours was held to determine the
ten students participating in the Benelux Mathematical Olympiad (BxMO),
held in Mersch, Luxemburg, from 27 until 29 April. The Dutch team
received a gold medal, four silver medals and three bronze medals, and
managed to get the highest total score of the Benelux countries, beaten only
by guest country France. For more information about the BxMO (including
the 2018 paper), see www.bxmo.org.
1
In June the team for the International Mathematical Olympiad 2018 was
selected by three team selection tests on 7, 8 and 9 June 2018, each lasting
four hours. A seventh, young, promising student was selected to accompany
the team to the IMO as an observer C. The team had a training camp in
Cluj-Napoca, from 30 June until 6 July.
For younger students the Junior Mathematical Olympiad was held in Oc-
tober 2017 at the VU University Amsterdam. The students invited to
participate in this event were the 100 best students of grade 2 and grade 3
of the popular Kangaroo math contest. The competition consisted of two
one-hour parts, one with eight multiple choice questions and one with eight
open 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.
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.
Dutch delegation
The Dutch team for IMO 2018 in Romania consists of
• Nils van de Berg (18 years old) • Jovan Gerbscheid (15 years old)
– bronze medal at BxMO 2017 – silver medal at BxMO 2018
– gold medal at BxMO 2018 • Jippe Hoogeveen (15 years old)
– hon. mention at IMO 2017
• Matthijs van der Poel (17 years old)
• Szabi Buzogany (18 years old)
– bronze medal at BxMO 2016
– silver medal at BxMO 2018 – bronze medal at BxMO 2017
• Thomas Chen (17 years old) – observer C at IMO 2016
– gold medal at BxMO 2017 – silver medal at IMO 2017
– silver medal at BxMO 2018
2
First Round, January 2017
Problems
A-problems
3
5. In the expression below, the ten dots are replaced by ten distinct digits
(0 to 9) in such a way that none of the resulting two-digit numbers starts
with 0:
·· + ·· + ··
·· − ··
What is the largest possible outcome we can obtain?
255 219
A) 7 B) 2 C) 116 D) 222 E) 255
D C
A B
4
17 18 17 27 18
16 18 20 17 18 18 18 20 20 26 16 16 12 20 24 18 18 18
19 16 10 19 9 18 18
B-problems
The answer to each B-problem is a number.
1. Isaac writes down a three digit number. None of its digits is a zero. Isaac
gives his sheet with the number to Dilara, and below Isaac’s number she
writes down all three digit numbers that one can obtain by putting the
digits of Isaac’s number in a different order. Then she adds up all numbers
on the sheet. The outcome is 1221.
What is the greatest number that Isaac could have written down?
2. There are two triples (a, b, c) of positive integers that satisfy the equations
ab + c = 34,
a + bc = 29.
Which two triples are these?
B
5
Solutions
A-problems
1. C) Thursday 5. D) 222
2. D) 169 6. B) 3
29
3. C) 3 7. D) 2
1
4. E) 2 8. E) the fifth
B-problems
1. 911
162
3. 5
4. 27
6
Second Round, March 2017
Problems
B-problems
The answer to each B-problem is a number.
7
4. A parallelogram ABCD is intersected by a line D C
m. From each of the four vertices A, B, C, and ?
D we draw a perpendicular to m. The four 25 Q
feet are P , Q, R, and S, respectively. Point S R
m
A 7
is also the intersection of line m and AB . The
lengths of line segments AP , BQ, and DS are 6 S B
P
6, 7, and 25, respectively.
What is the length of CR?
Be careful: the figure is not drawn to scale.
5. Simon has 2017 blue blocks that are numbered from 1 up to and including
2017. He also has 2017 yellow blocks that are numbered from 1 up to and
including 2017. Simon wants to arrange his 4034 blocks in a row, in such a
way that, for every k = 1, 2, . . . , 2017, the following conditions are met:
• to the left of blue block number k there are k or more yellow blocks;
• to the right of yellow block number k there are k or fewer blue blocks.
Determine all possible numbers for the 1000th block from the left in the
row.
C-problems For the C-problems not only the answer is important; you also have to
describe the way you solved the problem.
1 0 1 1 1 0 1 1 0 0 0 1 0
A B C D E
0 1 0 1 1 1 0 1 0 1 0 0 1
You want to form a row of tiles such that the same sequence of zeroes and
ones is formed on the top and the bottom. We will call this a matching
combination. Consider, for example, the row ‘DDC’ consisting of three tiles
of types D, D, and C, in that order. The top sequence is 1001001, while
the bottom sequence is 010010101. Since the two sequences are not the
same, the row of tiles is not a matching combination.
(a) Construct a matching combination using only tiles of type A, B, and
C.
8
(b) Show that no matching combination using only tiles of types B, C,
and D exists.
(c) Does a matching combination using only tiles of types B, C, D, and E
exist?
If so, give an example. If not, prove that such a combination does not
exist.
9
Solutions
B-problems
1. 11
35
2. 2
4. 12
5. 500
C-problems
C1. (a) A matching combination is ‘BABC’ having a top row and bottom row
equal to 1101011101.
All other solutions are of the form ‘BABC BABC . . . BABC’.
(b) Suppose that we have a matching combination using only tiles of types
B, C, and D.
The top row of these tiles always starts with a 1. In a matching
combination, the bottom row must therefore start with a 1 as well.
This rules out the first tile being of type D. Type C is ruled out as
well, since otherwise the second tile must have a top row starting with
0. The first tile of a matching combination must therefore be of type
B.
The next tile must have a bottom row starting with 0. Hence, it must
be of type D. After that, we again require a tile with a bottom row
starting with a 0. That is, we again require a tile of type D. This
continues indefinitely. After depleting our supply of 1000 tiles of type
D, we still need another tile having a bottom row starting with a 0. It
follows that we cannot complete our row to a matching combination:
a contradiction.
We conclude that there is no matching combination using only tiles of
types B, C, and D.
The proof of part (c) will give an alternative proof for part (b).
(c) First, consider the number of 1’s in the top row and bottom row of
each type of tile. Tiles of type B, D, and E have equal numbers of
10
1’s in the top and bottom row. Tiles of type C have more 1’s in the
bottom row than in the top row. Since a matching combination must
have equal numbers of 1’s in both rows, it cannot contain tiles of type
C.
A matching combination can therefore only contain tiles of types B,
D, and E. Tiles of type D have an equal number of digits in the top
row and bottom row (three digits). In contrast, tiles of type B and
E have more digits in their top row than in their bottom row. Since
any matching combination has equal numbers of digits in both rows,
it can contain tiles of neither type B nor type E.
A matching combination can therefore only contain tiles of type D.
But that is also not possible since tiles of type D have a top row
starting with 1 and a bottom row starting with 0. We conclude that
no matching combination exists containing only tiles of types B, C, D,
and E.
11
the candidates x = 10, 11, 12 are ruled out. Since 12 = 1, 22 = 4,
and 32 = 9 have only one digit, also the candidates x = 1, 2, 3 are
ruled out.
We consider the remaining candidates x = 4, 5, 6, 7, 8, 9. In these
cases, x2 is a two-digit number, which implies that cd = 8 · x. This
is a square only in the case x = 8. Hence, we find one solution:
4082 = 166464.
ab = 25 Similarly to the previous case, we can write n = 500 + x,
where x is a positive integer. We have K = (500 + x)2 = 250000 +
1000x + x2 . Since K < 260000, it follows that 1000x < 10000 and
hence x ≤ 9.
We observe that for every possible choice of x, the digit d of K
will be equal to 0. Hence, there are no solutions.
ab = 36 We write n = 600 + x, where x is a positive integer. We
have K = (600 + x)2 = 360000 + 1200x + x2 . Since K < 370000,
it follows that 1200x < 10000 and hence x ≤ 8. Also, we have
x ≥ 4 as otherwise digit e will be equal to 0.
For the remaining candidates x = 4, 5, 6, 7, 8 we obtain cd = 12 · x.
However, this is not a square for any of the possible choices for x.
Hence, there are no solutions.
ab = 49 We write n = 700 + x, where x is a positive integer. We
have K = (700 + x)2 = 490000 + 1400x + x2 . Since K < 500000, it
follows that 1400x < 10000 and hence x ≤ 7. As in the previous
case, we also have x ≥ 4 because digit e cannot be 0.
For the remaining candidates x = 4, 5, 6, 7 we see that cd = 14 · x
is not a square. Hence, there are no solutions.
ab = 64 We write n = 800 + x, where x is a positive integer. We
have K = (800 + x)2 = 640000 + 1600x + x2 . Since K < 650000,
it follows that 1600x < 10000 and hence x ≤ 6. Again, we also
have x ≥ 4.
Of the remaining candidates x = 4, 5, 6, the number cd = 16 · x is
a square only for x = 4. Thus we find one solution 8042 = 646416.
ab = 81 We write n = 900 + x, where x is a positive integer. We
have K = (900 + x)2 = 810000 + 1800x + x2 . Since K < 820000,
it follows that 1800x < 10000 and hence x ≤ 5. Again we also
have x ≥ 4.
For the remaining candidates x = 4, 5, the number cd = 18 · x is
not a square. Hence, there are no solutions.
We conclude that, in total, there are two six-digit multi-squares that
are themselves a square: 166464 and 646416.
12
Final Round, September 2017
Problems
13
(a) Let n be a positive integer and let a and b be two positive integers
smaller than n. Prove: if you divide n by a and b, then the sum of
the two remainders never equals n.
(b) Determine all integers n > 229 having the property that if you divide
n by 99, 132, and 229, the sum of the three remainders is n.
5. The eight points below are the vertices and the midpoints of the sides of a
square. We would like to draw a number of circles through the points, in
such a way that each pair of points lie on (at least) one of the circles.
Determine the smallest number of circles needed to do this.
14
Solutions
1. An integer for which the digits (from left to right) are c1 , c2 , . . . , ck will be
denoted by c1 c2 . . . ck .
(a) Let n = c1 c2 . . . ck be an oddball integer greater than 9 (hence k ≥ 2).
We will show that n is indeed the sum of two oddball integers.
If c2 ≥ 1, then we can write n as the sum of the following two oddball
integers: 10 . . . 0 (k − 2 zeros) and c1 (c2 −1)c3 . . . ck .
If c2 = 0 and c1 ≥ 2, then we can write n as the sum of the following
two oddball integers: 10 . . . 0 (k − 1 zeros) and (c1 −1)c2 c3 . . . ck .
If n = 10 . . . 0, then we can write n as the sum of the following two
oddball integers: 1 and 9 . . . 9 (k − 1 nines).
The last case is the case in which c1 = 1, c2 = 0 and not all digits
c3 , . . . , ck are equal to 0. Let ct be a digit unequal to 0, with t ≥ 3 as
small as possible. Hence, n = 10 . . . 0ct . . . ck with ct ≥ 1.
Because n is oddball and ct−1 = 0 < 1 ≤ ct , we find that t must be
odd. We can now write n as the sum of the integers 10 . . . 0 (k − 1
zeros) and m = ct ct+1 . . . ck . Because t is odd, the digits at the odd
positions of m are also at odd positions of n. Therefore, these digits
are greater than or equal to their neighbouring digits (because n is
oddball), which yields that m is oddball.
(b) The integer n = 109 is oddball, but it is not the sum of two even-steven
integers. We will prove this by contradiction. Suppose that n = p + q
for some even-steven integers p and q. We will show that this leads to
a contradiction.
First observe that the integers 100 to 108 are not even-steven. Hence,
both p and q must be smaller than 100, and hence both are also greater
than 9. In other words, p and q have exactly two digits. Suppose
p = ab and q = cd. The equation p + q = 109 now yields b + d = 9
(because b + d < 19) and a + c = 10. Hence, b + d < a + c, which
implies that either b < a or d < c (or both). In the first case p is
not even-steven and in the second case q is not even-steven. This
contradicts the assumption that p and q are even-steven.
We conclude that the oddball integer 109 cannot be written as the
sum of two even-steven integers.
15
2. Since AED is an isosceles triangle, angles D C
∠EDA and ∠DAE are equal. In turn, these
angles are equal to angles ∠EBF and ∠BF E
(alternate interior angles). This implies that
triangle BF E is isosceles as well, with |BE| =
|EF |.
F
Comparing triangles ABE and DF E, we see
that |AE| = |DE| and |BE| = |F E|. Since E
∠BEA and ∠F ED are a pair of opposite
angles, they have the same size. It follows
that triangles ABE and DF E are congruent A B
(SAS).
From the fact that ABE and DF E are congruent it follows that |DF | =
|AB|. Since ABCD is a parallelogram, we also have |AB| = |CD|. It
follows that triangle CDF is isosceles as well (with apex D).
On the one hand, this implies that ∠F CD = ∠DF C. On the other hand,
we know that triangle DBC is isosceles (since |BD| = |AD| = |BC|), which
implies that ∠F CD = ∠CDB = 2 · ∠CDF since DF is the angle bisector
of ∠CDB.
Altogether, we have ∠DF C = ∠F CD = 2 · ∠CDF . Since the angles in any
triangle sum to 180 degrees, we also know that
16
If T = 30, we have g = 45 − 30 = 15. But then all games must have ended
in a draw and all teams must have the same score. Hence, the case T = 30
is ruled out.
If T = 36, then g = 45 − 36 = 9 and s = T −30 6 = 1. The six scores are
therefore 1, 3, 5, 7, 9, 11. The team that scored 1 point must have lost 4
games (and played one draw). The team that scored 3 points must have
lost at least 2 games (at most 3 games were not lost). The team that scored
11 points must have won at least 3 games (otherwise the score is at most
3 + 3 + 1 + 1 + 1 = 9), so apart from the teams with scores 1 and 3 at least
one other team has lost a game. In total, at least 4 + 2 + 1 = 7 games
ended in a loss for some team, contradicting the fact that 15 − 9 = 6 games
did not end in a draw. This rules out the case T = 36.
Finally, we consider that case T = 42. The six scores are 2, 4, 6, 8, 10, 12 and
we have g = 3. Since the total number of points obtained from won games
is a multiple of three, the six teams must have received at least 2, 1, 0, 2, 1, 0
points from draws, respectively. In total, exactly 2·3 = 6 points are awarded
in games that ended in a draw. Hence, since 2 + 1 + 0 + 2 + 1 + 0 = 6,
the six teams have received exactly the mentioned numbers of points from
draws. In particular, the team ending in the fourth place (with 6 points),
was involved in 0 draws and must have won exactly two games.
4. (a) Let r be the remainder upon dividing n by a. We will first prove that
r < n2 . If 2a ≤ n, this follows from the fact that r < a. If 2a > n, we
have r = n − a (since we already knew that a < n), which implies that
r = n − a < n − n2 = n2 .
For the same reasons, the remainder upon dividing n by b is smaller
than n2 .
It follows that the two remainders obtained by dividing n by a and b
add up to a number smaller than n.
(b) Let r, s, and t be the remainders upon dividing n by 99, 132, and 229,
respectively. The number n − t is a multiple of 229 and nonzero since
n > 229 > t. We know that r + s + t = n, and hence n − t = r + s.
We can conclude that r + s is a positive multiple of 229. Since
99 + 132 < 2 · 229, we have r + s < 2 · 229, which implies that we must
have r + s = 229.
Since r ≤ 98 and s ≤ 131, the fact that r + s = 229 implies that
r = 98 and s = 131. Therefore, the number n + 1 is divisible by both
99 and 132, and hence by their least common multiple lcm(99, 132) =
lcm(9 · 11, 3 · 4 · 11) = 4 · 9 · 11 = 396. Also, from n = 229 + t and
t < 229 we deduce that n + 1 ≤ 458. It follows that the only possibility
is n + 1 = 396, hence n = 395.
17
When n = 395 the three remainders are r = 98, s = 131, and t = 166,
and indeed satisfy the equation n = r + s + t.
5. Four of the eight points are coloured black and the other four points are
coloured white in the way indicated in the figure on the left. The circle
through the four black points is denoted C1 and the circle through the four
white points is denoted C2 . If two points lie on a circle C, we say that C
covers that pair of points.
Circle C1 covers all pairs of black points and circle C2 covers all pairs of
white points. It is easy to check that each of the 4 · 4 = 16 pairs consisting
of a white point and a black point is covered by one of the four circles in
the figure on the right. It follows that the six circles form a solution.
We will now prove that five or fewer circles do not suffice. First observe
that any circle passing through more than two black points must be equal
to C1 and that any circle passing through more than two white points must
be equal to C2 . Indeed, a circle is already determined by three points.
A circle passing through 2 or fewer black points covers at most one of the
4·3
2 = 6 pairs of black points. A solution consisting of only five circles must
therefore contain circle C1 (since otherwise at most 5 pairs of black points
are covered). In the same way we see that such a solution must contain
circle C2 .
Each of the remaining three circles in the (hypothetical) solution contains
at most 2 black points and at most 2 white points. Such a circle covers
at most 2 · 2 = 4 pairs consisting of a white and a black point. In total,
the five circles can therefore cover at most 0 + 0 + 3 · 4 = 12 such pairs,
while there are 16 to be covered. The five circles can therefore not form a
correct solution after all. We conclude that the smallest number of circles
in a solution is 6.
18
BxMO Team Selection Test, March 2018
Problems
2. Let 4ABC be a triangle of which the side lengths are positive integers
which are pairwise coprime. The tangent in A to the circumcircle intersects
line BC in D. Prove that |BD| is not an integer.
22 32 (n + 1)2 n(n + 3)
nx2 + + + ... + = nx + .
x+1 x+2 x+n 2
19
Solutions
1. Consider the circle in which the 25 balls of one colour are all next to each
other. To get at least 20 different colours, you have to take at least 18 of
these groups plus one ball one one side and one ball on the other side of
these groups. In total, you need at least 18 · 25 + 2 = 452 adjacent balls.
Hence, n ≥ 452.
Now we will prove that 452 balls is enough. Consider an arbitrary circle
of balls and all possible sets of consecutive balls having exactly 20 colours.
(There exists at least one such a set: take an arbitrary ball and add balls
on the left one by one, until you have exactly 20 colours.) Take such a set
having a minimum number of balls. Suppose that the first ball is white. If
there is another white ball in the set, then we could have removed the first
ball to get a smaller set with the same number of colours, but less balls in
total. This contradicts the minimality of our set. Therefore, no other ball
is white. In particular, also the last ball of the set is not white; suppose it
is black. In the same way, we find that no other ball in the set is black. So
there are only one white ball, one black ball, and balls in 18 other colours,
at most 25 of each colour. In total, there are at most 18 · 25 + 2 = 452 balls.
Indeed, it is always possible to find a set of 452 consecutive balls in at least
20 different colours. We conclude that the minimum n is 452.
20
3. Let bi = a1 a2 · · · ai , for 1 ≤ i ≤ p. We will prove that it is possible to
choose a permutation such that bi ≡ i mod p for all i. For i ≥ 2, the
congruence ai ≡ bi · b−1 i−1 mod p holds, if bi−1 6≡ 0 mod p. Therefore, we
now choose a1 = 1 and ai ≡ i · (i − 1)−1 mod p for 2 ≤ i ≤ p. Now it is
sufficient to prove that ai 6≡ 1 mod p for all 2 ≤ i ≤ p and ai 6≡ aj mod p
for all 2 ≤ j < i ≤ p.
Suppose the contrary. Then ai ≡ 1 mod p for certain 2 ≤ i ≤ p. Then
i · (i − 1)−1 ≡ 1 mod p holds, hence i ≡ i − 1 mod p, or 0 ≡ −1 mod p.
Because p ≥ 2, this is a contradiction. Now suppose that ai ≡ aj mod p
for certain 2 ≤ j < i ≤ p. Then we have i · (i − 1)−1 ≡ j · (j − 1)−1 mod p,
hence i(j − 1) ≡ j(i − 1) mod p, or ij − i ≡ ij − j mod p, or −i ≡ −j
mod p. However, as 2 ≤ j < i ≤ p, this is impossible.
We conclude that if we choose the ai as above, all ai are distinct, so that
they indeed form a permutation of 1, 2, . . . , p. Moreover, by definition
we have that a1 a2 · · · ai ≡ i mod p, therefore also the second condition is
satisfied.
5. For 1 ≤ i ≤ n we have
(i+1)2 (i+1)2 −(i+1)(x+i) (i+1)(1−x)
x+i =i+1+ x+i =i+1+ x+i ,
so we can rewrite the left hand side of the equation to
2(1−x) 3(1−x) (n+1)(1−x)
nx2 + 2 + 3 + . . . + (n + 1) + x+1 + x+2 + ... + x+n .
21
We have 2 + 3 + . . . + (n + 1) = 12 n(n + 3), hence this sum cancels against
n(n+3)
2 on the right hand side of the original equation. Moreover, we can
move nx2 to the other side, and write a separate factor 1 − x in all fractions.
The equation then becomes
(1 − x) · x+1 2 3
+ x+2 n+1
+ . . . + x+n = nx − nx2 .
The right hand side, we can factor as nx(1 − x). Now, we observe that
x = 1 is a solution to this equation. If there were another solutions x 6= 1,
then it would satisfy
2 3 n+1
x+1 + x+2 + ... + x+n = nx.
22
IMO Team Selection Test 1, June 2018
Problems
for all x ∈ R.
23
Solutions
1. Number the rows from bottom to top by 1, 2, . . . , 2m, and the columns
from left to right by 1, 2, . . . , 2n, so that the snail starts in square (1, 1) and
finishes in (2m, 2n). Colour the squares of the grid like a chessboard, where
(i, j) is coloured black if i + j is even and white if i + j is odd. Since 2m
and 2n are even, the number of black squares equals that of white squares.
The snail starts and finishes on a black square, and alternates between
black and white squares on its journey. Therefore, the number of black
squares it visits is one larger than that of white squares, so the pawn must
be on a white square (i.e. a square (i, j) with i + j odd) for the snail to be
able to finish its journey.
Let us show that if the pawn is on such a square, then the snail is always
able to undertake its journey. Let 1 ≤ k ≤ m and 1 ≤ l ≤ n be such the
pawn is either on row 2k − 1 or row 2k, and either on column 2l − 1 or
column 2l. Then have the snail move all the way to the right on odd rows
with a number less than 2k − 1 (and then move one square up, which puts
the snail on an even row), and all the way to the left on even rows with
a number less than 2k − 1 (and then move one square up, which puts the
snail on an odd row again). This puts the snail on square (2k − 1, 1). From
here, while the snail is in a column with number less than 2l − 1, have the
snail move through each 2 × 2-block in the order: bottom left, top left,
top right, bottom right. After moving one square to the right, this puts
the snail on the bottom left square of the next 2 × 2-block. This puts the
snail on square (2k − 1, 2l − 1), which cannot be occupied by the pawn as
2k − 1 + 2l − 1 is even. In this 2 × 2-block, the pawn is either on the top
left or the bottom right square; in the former case, move the snail in the
order: bottom left, bottom right, top right; in the latter case, move the
snail in the order: bottom left, top left, top right. So in both cases, the
snail ends up in the top right square of the 2 × 2-block. The remainder of
the 2 × 2-blocks in rows 2k − 1 and 2k can now be covered by moving in
each of them in the order: top left, bottom left, bottom right, top right.
This puts the snail on square (2k, 2n), with all required squares to the
bottom and left visited exactly once. The snail can now finish its journey
by repeatedly doing the following: move one square up, move all the way
to the left, move one square up, move all the way to the right. This puts
the snail on square (2m, 2n), as required.
Therefore, the snail is able to finish its journey if and only if the pawn is
put on a square (i, j) with i + j odd.
24
2. Let S be the intersection of the
tangent with AB. Then we need
to show that ∠BSC = 90◦ .
Since ∠BCD = 90◦ = ∠CED
we have 4BCD ∼ 4CED, so
|DC| |DE|
|DB| = |DC| . As |DA| = |DC|
|DA|
it follows that |DB| = |DE|
|DA| . To-
gether with ∠BDA = ∠EDA
this gives 4ADE ∼ 4BDA (SAS). Therefore ∠ABD = ∠EAD. Since
SC is a tangent to the circumcircle of 4AEC we also have that (by the
tangent chord angle theorem) ∠EAD = ∠EAC = ∠SCE. So ∠SCE =
∠ABD = ∠SBE. Hence SBCE is a cyclic quadrilateral. Therefore
∠BSC = ∠BEC = 90◦ .
√
3. Let m = b 2nc. We first show that the sequence of squares a0 + a1 , a1 +
a2 , . . . is the sequence (m + 1)2 , (m + 2)2 , . . . . After that, we will show that
the differences ai − ai−1 form a sequence of consecutive even numbers and
a sequence of consecutive odd numbers. We will use this to prove the given
statement.
Note that a0 + a1 is the smallest square greater than 2n. Therefore
a0 + a1 = (m + 1)2 . We show by induction on i that ai−1 + ai = (m + i)2 .
For i = 1, this is what we have just shown. Suppose that aj−1 +aj = (m+j)2 .
2
Then aj−1 ≥ (m+j−1) 2 (since otherwise aj could have been chosen in such
a way that aj−1 + aj = (m + j − 1)2 ), so
(m+j−1)2
aj = (m + j)2 − aj−1 ≤ (m + j)2 − 2
2m2 +4mj+2j 2 −(m2 +2mj+j 2 +1−2m−2j)
= 2
m2 +2mj+j 2 −1+2m+2j m2 +2mj+j 2 +1+2m+2j (m+j+1)2
= 2 < 2 = 2 .
25
for all i ≥ 1. Hence
4. Substituting x = 0 gives f1 f2 (y) = f3 (y) therefore the f3 corresponding
to f1 , f2 ∈ A is the composition f3 (x) = f1 (f2 (x)). Substituting x = −y
now gives that for all f1 , f2 ∈ A we have
f1 f2 (y) + y − 2y = f3 (0) = f1 (f2 (0))
for all y ∈ R.
Substituting x = f2 (y) now gives that for all f1 , f2 ∈ A, there exists f3 ∈ A
with
f1 (0) + 2f2 (y) = f3 f2 (y) + y
for all y ∈ R. We have already seen that we can write f3 (f2 (y) + y) as
2y + f3 (f2 (0)). Hence
26
IMO Team Selection Test 2, June 2018
Problems
2. Find all positive integers n for which there exists a positive integer k such
that for every positive divisor d of n, the number d − k is also a (not
necessarily positive) divisor of n.
3. Let ABC be an acute triangle, and let D be the foot of the altitude through
A. On AD, there are distinct points E and F such that |AE| = |BE| and
|AF | = |CF |. A point T 6= D satisfies ∠BT E = ∠CT F = 90◦ . Show that
|T A|2 = |T B| · |T C|.
27
Solutions
√
1. (a) We claim that (a, b, c) = (2, 2, −1 + 5) satisfies the given equalities.
√
As all real numbers in this triple are positive, and 2 6= −1 + 5, the
answer to the first question is ”no”.
√ √ √ √
Note that c3 = (−1 + 5)3 = −1 + 3 √ · 5 − 3 · 5 + 5 5 = −16√+ 8 5.
Then we have c(a3 + b3 ) = (−1 + √5) · 2 · 8 = −16 + √ 16 5 and
a(b3 + c3 ) = b(c3 + a3 ) = 2 · (−16 + 8 5 + 8) = −16 + 16 5, so the
given equalities are indeed satisfied.
(b) We show that if the given equalities are satisfied, that then a = b = c.
Without loss of generality, we assume that a ≥ b, c. Then we have
or 1 − k = − 12 n.
In the latter case, we have k = 12 n + 1, so n − k = 12 n − 1. However, in
the same way as before, we see that for n > 6 no divisor of n is equal to
1 1 1 1
2 n − 1, as the next largest divisor after 2 n is at most 3 n < 2 n − 1. This
contradicts n − k being a divisor of n.
28
In the former case, we have 1 − k = −n, so k = n + 1. As n is composite,
there exists a divisor d with 1 < d < n. Then d − k = d − n − 1 is also
a divisor of n. However, d ≤ 12 n, so d − n − 1 ≤ − 12 n − 1, but the only
divisor that is at most − 12 n − 1 is −n. We deduce that d − n − 1 = −n, so
d = 1, which is a contradiction.
Therefore the n that have the given property are precisely those n for which
either n ≤ 6 or n is prime.
A
3. We consider the configuration
in the figure.
Let M be the midpoint of AB,
and let N be the midpoint of
AC. From what is given in the E
problem, it follows that E lies
on the perpendicular bisector of
T
AB, so ∠BM E = 90◦ . Since
we also have ∠BT E = 90◦ and B D
C
◦
∠BDE = 90 , by Thales’s theo-
rem it follows that BDT EM is
F
a cyclic pentagon. In the same
way, it follows that CF DT N is a cyclic pentagon. Now ∠N T M =
360◦ − ∠DT N − ∠M T D = (180◦ − ∠DT N ) + (180◦ − ∠M T D) = ∠DCN +
∠M BD = ∠BCA + ∠ABC = 180◦ − ∠CAB = 180◦ − ∠N AM . Therefore
AM T N is a cyclic quadrilateral (which also follows directly from Miquel’s
theorem).
Now let us chase some angles in order to show that 4T BA ∼ 4T AC. We
have ∠N T C = ∠N F C = 90◦ − ∠F CN since ∠F N C = 90◦ . Moreover,
we have ∠F CN = ∠F CA = ∠CAF since |AF | = |CF |, so ∠N T C =
90◦ − ∠CAF = 90◦ − ∠CAD = ∠DCA = ∠BCA. Therefore ∠ACT =
∠N CT = 180◦ − ∠N T C − ∠CN T = 180◦ − ∠BCA − ∠CN T . As M N
is a mid-parallel in triangle ABC, we have M N k BC and it follows
that 180◦ − ∠BCA = ∠CN M . Hence ∠ACT = ∠CN M − ∠CN T =
∠T N M = ∠T AM = ∠T AB (using the cyclicity of the quadrilateral
AM T N ). Similarly, we have ∠ABT = ∠T AC, so 4T BA ∼ 4T AC. It
follows that |T B| |T A| 2
|T A| = |T C| , so |T A| = |T B| · |T C|, as required.
4. Consider a group of students in which all students know each other, and
which has the maximal number of students possible. (Such a group exists,
as there are only finitely many students in the classroom, and there exists
a group of students in which all students know one another, e.g. a group
29
consisting of only one student.) Denote this group of students by X. We
show that by dividing the students in X and the group of students not in
X, we may achieve a division with the required properties.
Note that it suffices to show that in the group of students not in X, none
of the students know each other. Suppose for a contradiction that A and
B are students not in X who know each other. As X is chosen to have
the maximal number of students possible, there is a student A0 in X who
doesn’t know A, and a student B 0 who doesn’t know B. We first show that
we can take A0 and B 0 to be different. If not, then there is a unique student
C in X who doesn’t know A, and a unique student in X who doesn’t know
B (and they are the same student). In other words all other students
in X know both A and B, so by replacing C by A and B, we obtain a
group of students in which all students know one another; all students in
X know each other, A and B know everyone in X (aside from C), and A
and B know each other. However this group is larger than X, which is a
contradiction.
Therefore we may assume that A0 and B 0 are distinct. Note that A0 and B 0
are in X, so they know each other. Now if A, B, B 0 , A0 , in that order, take
seats around a round table, everyone knows precisely one of his neighbours;
A and B, and A0 and B 0 know each other, but A and A0 , and B and B 0
do not know each other. This contradicts the given property. Therefore
dividing the students into X and the group of students not in X gives a
division with the required properties.
30
IMO Team Selection Test 3, June 2018
Problems
1. A set of lines in the plan is called nice if every line in the set intersects an
odd number of other lines in the set.
Determine the smallest integer k ≥ 0 having the following property: for
each 2018 distinct lines `1 , `2 , . . . , `2018 in the plane, there exist lines
`2018+1 , `2018+2 , . . . , `2018+k such that the lines `1 , `2 , . . . , `2018+k are
distinct and form a nice set.
for all x, y ∈ R.
3. Determine all pairs (a, b) of positive integers such that (a + b)3 − 2a3 − 2b3
is a power of two.
31
Solutions
1. First we prove that the number of lines in a nice set is even. Suppose on
the contrary that the number of lines is odd. Then on each of the odd
number of lines, there is an odd number of intersection points, so the total
number of intersection points is odd. However, each intersection point is
counted twice (once for each of the lines on which is is lying), so the total
should be even, which is a contradiction. So the total number of lines in a
nice set has to be even. In particular, 2018 + k must be even, so k must be
even.
Now suppose that there are 1009 directions, such that in each direction
there are two of the original lines. Then each line is parallel to exactly
one other line. If k < 1010, then k ≤ 1008, so there must be a direction
in which we do not add a line. Consider one of the original lines ` in this
direction. Within the final set of all lines, it intersects all lines except itself
and the line it is parallel to. This is an even number. Hence, the resulting
set is not nice. Therefore, there is an example in which at least 1010 are
needed.
We will now show that it is possible to add exactly 1010 lines such that the
resulting set is nice. We consider all directions in which there is an even
number (greater than 0) of the original lines. There are at most 1009 such
directions. For each of these directions, we add one line. Each line is then
parallel to an even number (possibly 0) of other lines. First suppose that
the total number of lines is even. Then each lines intersects an odd number
of lines (the total set minus itself and an even number of other lines). Now
suppose that the total number of lines is odd. Then each line intersects
an even number of lines. We ass a line in a new direction, which therefore
intersects all lines (an odd number), so that each line after that intersects
an odd number of other lines.
Now we have a nice set, and we have added at most 1009 + 1 = 1010
lines. Possibly, there are less than 1010. In this case, the number of lines
added is even, because the total number of lines in a nice set is always
even. We choose a direction in which there is at least one line. We then
keep adding two lines in this direction. These lines intersect all lines in
the other direction, which is an odd number in total. These other lines
each got two new intersection points, so they still have an odd number
of intersection points. This also holds for the existing lines in the chosen
direction, because they get zero new intersection points. The set is still
nice. We keep adding pairs of lines until we have 1010 lines in total.
We conclude that the minimum k satisfying the conditions is k = 1010.
32
2. Substituting x = y = 0 yields 0 ≤ f (0) · −f (0). However, squares are
non-negative, so this yields f (0)2 = 0 and hence f (0) = 0. Now taking
x = 0 and y = t, we obtain −f (t2 ) ≤ t · −f (t), while taking x = t and y = 0
yields f (t2 ) ≤ f (t) · t. We find tf (t) ≤ f (t2 ) ≤ tf (t), hence f (t2 ) = tf (t)
for all t ∈ R. On the left hand side of the function inequality we can replace
f (x2 ) − f (y 2 ) by xf (x) − yf (y). By expanding on the right hand side, we
obtain xf (x) − yf (y) + xy − f (x)f (y), hence
Hence, equality must hold, which means that f (x)f (1) = x. In particular,
this yields f (1)2 = 1, hence f (1) = 1 or f (1) = −1. In the former case, we
get f (x) = x for all x and in the latter case, we get f (x) = −x for all x.
When we check f (x) = x with the original function inequality, we get
x2 − y 2 on the left hand side and (x + y)(x − y) = x2 − y 2 on the right
hand side, hence this function satisfies the inequality. When we check
f (x) = −x, the left hand side becomes −x2 + y 2 and the right hand side
becomes (−x + y)(x + y) = −x2 + y 2 , so also this function is a solution.
We conclude that the solutions are: f (x) = x for all x ∈ R and f (x) = −x
for all x ∈ R.
33
Suppose that a + b is at least 8. Because it is a power of two, it is also
divisible by 8, hence we can write b = 8m − a for a certain positive integer
m. Then we have
2 = −a2 − b2 + 4ab
= −a2 − (8m − a)2 + 4a(8m − a)
= −a2 − 64m2 + 16ma − a2 + 32ma − 4a2
= −6a2 + 48ma − 64m2 ,
hence
Modulo 8, this becomes 1 ≡ −3a2 . If a is even, then the right hand side
is even, and if a is odd, then the right hand side is −3 modulo 8. In both
cases, the equality does not hold modulo 8. We conclude that a + b is not
greater or equal to 8. Because a and b are both at least 1, a + b is equal to
2 or 4.
• If a + b = 2, then a = b = 1. Now we have −a2 − b2 + 4ab = 2, hence
the product of the two factors is a power of two (namely 4). Hence
(1, 1) is a solution.
• If a + b = 4, then because of gcd(a, b) = 1 we have either a = 1 and
b = 3 or the other way around. Now −a2 − b2 + 4ab = 2, so also the
product of the two factors is a power of two (namely 8). Hence (1, 3)
and (3, 1) are solutions.
34
4. We consider the configuration in which |AB| > |AC|, causing C to lie
between B and F . For other configurations, the proof is analogous.
Because ∠DAC = ∠BAD holds, D is the midpoint of the circle arc BC on
which A does not lie. Hence, the line DM is a diameter of the circle. Thales’s
theorem yields ∠DBM = 90◦ . Let K be the intersection point of BC and
DM . As DM is the segment bisector of BC, we have ∠BKM = 90◦ =
|M B| |M D|
∠DBM . Therefore, 4M KB ∼ 4M BD (AA). This yields |M K| = |M B| ,
hence |M D| · |M K| = |M B|2 . Because K is on the interior of M D, this
equality also holds using directed distances: M D · M K = M B 2 .
Because D is the midpoint of arc BC, we have |DB| = |DC|. Moreover,
∠CDI = ∠CDA = ∠CBA and ∠DCI = ∠DCB + ∠BCI = ∠DAB +
∠BCI = 12 ∠CAB + 12 ∠BCA. Using the sum of the angles in triangle
DCI, we now get that ∠DIC = 180◦ − ∠CBA − 21 ∠CAB − 12 ∠BCA =
1 1
2 ∠CAB + 2 ∠BCA = ∠DCI. Hence, triangle DCI is isosceles with
|DC| = |DI|. We conclude that D is the centre of the circle through B,
C and I. Now DB is the radius of this circle, and DB is perpendicular
to M B, which yields that M B is a tangent line. Using the secant-tangent
theorem, we get M B 2 = M I · M N . Together with the previous paragraph
we conclude that M D · M K = M I · M N . Hence, N DKI is a cyclic
quadrilateral because of the same theorem.
We also know that ∠F ID = 90◦ = ∠F KD, hence DKIF is a cyclic
quadrilateral. We conclude that N DKIF is a cyclic pentagon. This yields
∠DN F = 180◦ − ∠DIF = 90◦ . Hence, N F is perpendicular to the radius
DN of the circle through B, I and C; this means that F N is tangent to
this circle.
35
Junior Mathematical Olympiad, September 2017
Problems
Part 1
1. A positive three-digit number is called nice if the sum of the last two digits
equals the first digit. For example, 123 is not nice, because 1 is not equal
to 2 + 3. How many three-digit numbers are nice?
Note that a three-digit number cannot start with digit 0.
A) 45 B) 48 C) 50 D) 54 E) 55
2. The faces of a cube have different colours. In the figure you can see a net
for the cube. The points A and B in the net correspond to two vertices of
the same face of the cube.
What colour does that face have?
A) red B) blue C) green D) black E) yellow
B
red white blue
A
green black yellow
36
4. The number n2 + 21 is the square of an integer.
For how many positive integers n does this hold?
A) 0 B) 1 C) 2 D) 3 E) 4
6. Peter starts out with the numbers 1, 2, 3, and 4. He may take two of his
numbers and replace them by their sum, their product, or their difference.
He performs this replacement step three times, after which a single number
remains.
Example. He could replace the 2 and the 4 by 2 + 4 = 6, then replace the 1 and
the 3 by 3 − 1 = 2, and finally replace the 6 and the 2 by 8 = 2 + 6. Then, the
remaining number would be 8.
Which of the following five numbers cannot be the number that remains?
A) 28 B) 30 C) 32 D) 34 E) 36
37
8. You have a collection of hats. Each hat has three attributes: the colour
(red or blue), the shape (top hat or pointed hat), and the pattern (spots
or stripes). You put a number of gnomes in a room and put a hat on
each of them. For any two gnomes their hats must be different, yet share
at least one attribute (for example: both hats are blue). The gnomes
can see everyone’s hat, except their own. The gnomes are not allowed to
communicate with one another.
What is the minimum number of gnomes you have to put in the room in
order to be be sure that one of them can determine one of the attributes of
his own hat?
A) 3 B) 4 C) 5
D) 8 E) That is impossible for any number of gnomes.
Part 2
1. Stef has 18 coins of which 17 are identical, but one is slightly lighter than
the other coins. Together, the 18 coins weigh 214 grams. Stef removes
two of the coins and weighs the remaining 16 coins. Together the 16 coins
weigh 190 grams.
How much does the lighter coin weigh?
2. We say that a positive integer is balanced if the average of the first two
digits is 2, the average of the first three digits (if they all exist) is 3, the
average of the first four digits (if they all exist) is 4, et cetera.
What is the largest balanced number?
4. A clock has the numbers 1 to 12 for indicating the hours. Ernie has
interchanged these twelve numbers in such a way that any two adjacent
numbers on the clock differ by either 2 or 3. Fortunately, the number 12 is
still in the right place, but the 9 is where the 1 is supposed to be.
What number is in the place where the 9 is supposed to be?
38
5. The numbers 1 to 8 are assigned to the eight faces
of an octahedron. For each vertex, we compute
the sum of the four numbers on the faces meeting
in that vertex. For four of the vertices we get the
same outcome. For a fifth vertex the outcome is
16.
What is the outcome for the sixth vertex?
8. You may choose any number consisting of five different digits. The digit 0
cannot be used. The next step is to choose two adjacent digits and switch
their positions. You may perform this step five times in total. Finally, you
compute the difference between the initially chosen number and the final
number obtained after switching.
What is the largest possible difference that can be obtained?
Example. Suppose you choose the number 47632. Then you could switch
digits to obtain, in this order, the numbers 46732, 46372, 46327, 43627, and
34627. Then, the difference between the initial number and the final number is
47632 − 34627 = 13005.
39
Answers
Part 1
1. D) 54
2. D) black
3. C) 2 (statements 1 and 3)
4. C) 2
5. C) 665
6. D) 34
7. E) 7
Part 2
1. 10 grams 5. 20
2. 40579 6. 1296
3. 60 7. 60
4. 5 8. 85230
40
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
3 First Round, January 2017
7 Second Round, March 2017
13 Final Round, September 2017
19 BxMO Team Selection Test, March 2018
23 IMO Team Selection Test 1, June 2018
27 IMO Team Selection Test 2, June 2018
31 IMO Team Selection Test 3, June 2018
36 Junior Mathematical Olympiad, September 2017
MY
We eat problems
CMY
Olympiad 2018
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
sponsored by: Centrum Wiskunde & Informatica sterdam 2011
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
3 First Round, January 2018
7 Second Round, March 2018
12 Final Round, September 2018
19 BxMO Team Selection Test, March 2019
23 IMO Team Selection Test 1, May 2019
27 IMO Team Selection Test 2, May 2019
31 IMO Team Selection Test 3, May 2019
35 Junior Mathematical Olympiad, September 2018
The 1014 best students were invited to the second round, which was held in
March at twelve universities in the country. This round contained five open
questions, and two problems for which the students had to give extensive
solutions and proofs. The contest lasted 2.5 hours.
The 113 best students were invited to the final round. Also some outstanding
participants in the Kangaroo math contest or the Pythagoras Olympiad
were invited. In total about 150 students were invited. They also received
an invitation to some training sessions at the universities, in order to prepare
them for their participation in the final round.
The final round in September 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 57th edition 2018.
In February a team of four girls was chosen from the training group to
represent the Netherlands at the EGMO in Kyiv, Ukraine, from 7 until 13
April. The team brought home a silver medal, two bronze medals, and a
honourable mention; a very nice achievement. For more information about
the EGMO (including the 2019 paper), see www.egmo.org.
In March a selection test of three and a half hours was held to determine the
ten students participating in the Benelux Mathematical Olympiad (BxMO),
held in Valkenswaard, Netherlands, from 26 until 28 April. The Dutch team
received three silver medals, four bronze medals and a honourable mention.
For more information about the BxMO (including the 2019 paper), see
www.bxmo.org.
1
In May the team for the International Mathematical Olympiad 2019 was
selected by three team selection tests on 29, 30 and 31 May, each lasting
four hours. A seventh, young, promising student was selected to accompany
the team to the IMO as an observer C. The team had a training camp in
Bristol, United Kingdom, from 6 until 14 July.
For younger students the Junior Mathematical Olympiad was held in Oc-
tober 2018 at the VU University Amsterdam. The students invited to
participate in this event were the 100 best students of grade 2 and grade 3
of the popular Kangaroo math contest. The competition consisted of two
one-hour parts, one with eight multiple choice questions and one with eight
open 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.
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.
Dutch delegation
The Dutch team for IMO 2019 in the United Kingdom consists of
• Szabi Buzogány (19 years old) • Matthijs van der Poel (18 years old)
– silver medal at BxMO 2018, – bronze medal at BxMO 2016,
silver medal at BxMO 2019 bronze medal at BxMO 2017
– honourable mention at IMO – observer C at IMO 2016,
2018 silver medal at IMO 2017,
• Jesse Fitié (17 years old) silver medal at IMO 2018
• Jovan Gerbscheid (16 years old) • Richard Wols (16 years old)
– silver medal at BxMO 2018 – bronze medal at BxMO 2018,
– bronze medal at IMO 2018 silver medal at BxMO 2019
• Jippe Hoogeveen (16 years old) – observer C at IMO 2018
– bronze medal at IMO 2018
2
First Round, January 2018
Problems
A-problems
1. In a classroom there are chairs and stools. On each chair and on each stool
one child is seated. Each chair has 4 legs, each stool has 3 legs and each
child has 2 legs. Together, we have a total of 39 legs.
How many chairs are there in the classroom?
A) 3 B) 4 C) 5 D) 6 E) 9
2. On an island, there are knights and knaves. Knights always speak the truth
and knaves always lie. On the island you meet five people. You know that
four of them are knights and one of them is a knave, but you do not know
who is the knave. They make the following statements about the island
inhabitants:
• A: “All knaves have shoe size 40.”
• B: “All people with shoe size 40 have a goldfish.”
• C: “All people with a goldfish are knaves.”
• D: “I have shoe size 40.”
• E: “I have a goldfish.”
Which of them is the knave?
A) A B) B C) C D) D E) E
3
5. Nine people are at a party. While entering, some of them shook hands.
Quintijn is at the party and asks each of the others how many hands they
shook. He gets eight different answers.
How many hands did Quintijn shake?
A) 0 B) 1 C) 2 D) 3 E) 4
7. A frog starts in the point at coordinates (0, 0) in the plane. He can make
three kinds of jumps:
• from (x, y) to (x, y − 5);
• from (x, y) to (x − 2, y + 3);
• from (x, y) to (x + 4, y + 9).
Ahead, there are three juicy snacks that the frog would like to eat: a worm
at (2013, 2018), a beetle at (2018, 2019), and a snail at (2018, 2023).
Which of these snacks can the frog reach?
A) worm and snail B) beetle and snail C) worm and beetle
D) only the beetle E) only the snail
4
B-problems
The answer to each B-problem is a number.
1. Three years ago, Rosa’s mother was exactly five times as old as Rosa was
at that time. At that moment, Rosa’s mother was just as old as Rosa’s
grandmother was when Rosa’s mother was born. Now, Rosa’s grandmother
is exactly seven times as old as Rosa is.
How old is Rosa’s mother now?
2. Nanda and Mike both have a note containing the same five-digit number.
Nanda puts the digit 4 in front and the digit 8 at the end of her number to
obtain a seven-digit number. Mike puts one digit in front of his number.
Comparing their new numbers, Nanda’s number turns out to be exactly 6
times as large as Mike’s.
What was their starting number?
3. We consider a square, the circle through the vertices of the square and the
circle touching the four sides of the square (see the left-hand figure). The
ring-shaped area between the two circles is divided into four dark pieces
(inside the square) and four light pieces (outside the square). The area of
the square is 60.
What is the total area of two dark pieces and one light piece together as
depicted in the right-hand figure?
4. Elisa is making so-called dubious dice. Each face of a dubious die contains
one of the numbers 1 to 6, but not all these numbers need to occur and
some may occur more than once. However, from every direction it must
look like a real die. This means that in each corner three different numbers
must meet, no two of which add up to 7 (on a real die such pairs are always
on opposite faces). For example, the numbers 1, 2, and 4 may meet in a
corner, but 1, 2, and 5 may not as 2 + 5 = 7. Of course, a normal die is an
example of a dubious die as well.
Elisa is interested in the sum of the six numbers on a dubious die.
How many possible values are there for this sum?
5
Solutions
A-problems
1. B) 4 5. E) 4
2. C) C 6. E) 34
4. D) 7 8. E) A : C = D : B
B-problems
1. 33 years
2. 49998
3. 15
4. 19
6
Second Round, March 2018
Problems
B-problems
The answer to each B-problem is a number.
B1. Anouk, Bart, Celine, and Daan have participated in a math competition.
Each of their scores is a positive integer. The sum of the scores of Bart
and Daan is the same as the sum of the scores of Anouk and Celine. The
sum of the scores of Anouk and Bart is higher than the sum of the scores
of Celine and Daan. Daan’s score is higher than the sum of the scores of
Bart and Celine.
Write down the names of the four students in decreasing order of their
scores.
7
B5. A sawtooth number is a positive integer with the following property: for any
three adjacent digits, the one in the middle is either greater than its two
neighbours or smaller than its two neighbours. For example, the numbers
352723 and 314 are sawtooth numbers, but 3422 and 1243 are not.
How many 8-digit sawtooth numbers exist, for which each of the digits is
equal to 1, 2, or 3?
C-problems For the C-problems not only the answer is important; you also have to
describe the way you solved the problem.
C1. You have n balls that are numbered from 1 to n. You need to distribute
the balls over two boxes. The value of a box is the sum of the numbers of
the balls in that box. Your distribution of balls must obey the following
rules:
• Each box has at least one ball.
• The two boxes do not have the same number of balls.
• The value of the box with the least number of balls must be at least 2
more than the value of the box with the most balls.
Determine for which positive integers n this is possible.
(Prove that for those values of n it is indeed possible, and prove that it is
not possible for other values of n.)
8
Solutions
B-problems
1. A, D, B, C
2. 18 degrees
3. 2442
3
4. 4
5. 110
C-problems
C1. A distribution of the balls that follows all three rules will be called a correct
distribution. The box containing the larger number of balls will be called
the fullest box. It is clear that at least 1 + 2 = 3 balls are needed to obey
the first two rules.
We first consider the case that n is odd.
If n = 3, then the fullest box must have two balls. It therefore has a value
of at least 1 + 2 = 3, while the other box has one ball and hence a value of
at most 3. The third rule is broken so there is no correct distribution.
For n = 5 there is a correct distribution: put balls 1, 2, and 3 in one box
and put balls 4 and 5 in the other box. Since 4 + 5 is at least two more
than 1 + 2 + 3, this is indeed a correct distribution.
If there is a correct distribution for n balls, there is also one for n + 2 balls.
Indeed, we can simply add ball n + 1 to the fullest box and add ball n + 2
to the other box. The value of the fullest box increases by less than the
other box, hence we still follow rule 3.
Since we have a correct distribution for n = 5, we also have a correct
distribution for n = 7. Then, we also find a correct distribution for n = 9,
n = 11, et cetera.
Now we consider the case that n is even.
If n = 4, the fullest box must have at least three balls and hence has a
value of at least 1 + 2 + 3 = 6. The other box has only one ball and has a
9
value of at most 4. This means that the third rule is broken. There is no
correct distribution.
If n = 6, the fullest box must have at least four balls. It therefore has a
value of at least 1 + 2 + 3 + 4 = 10. The other box has at most two balls
and therefore has a value of at most 5 + 6 = 11. Since 11 < 2 + 10, there is
no correct distribution.
For n = 8, there is a correct distribution: put balls 1 to 5 into one box
and put balls 6 to 8 in the other box. The fullest box then has a value of
1 + 2 + 3 + 4 + 5 = 15, and the other box has a value of 6 + 7 + 8 = 21.
Since 21 ≥ 2 + 15, this is indeed a correct distribution.
Again, we see that a correct distribution of n balls gives a correct distribution
with n + 2 balls by adding ball n + 1 to the fullest box and adding ball
n + 2 to the other box. Since we have a correct distribution for n = 8, we
thus find correct distributions for n = 10, 12, 14, . . ..
We conclude: for n = 1, 2, 3, 4, 6 there is no correct distribution, but for all
other positive integers n a correct distribution does exist.
C2. (a) Yes, such a number a exists. For example, take a = 33. Then we have:
16 + a = 72 , 3 + a = 62 , and 16 · 3 + a = 92 .
A suitable a can be easily found by trying 48 + a = 49, 64, 81.
(b) No, such a number a does not exist. The numbers 20 + a and 18 + a
cannot both be squares since the difference of two squares is never
equal to 2. Indeed, suppose that m2 − n2 = 2. Then we would have
(m + n)(m − n) = 2, where m > n. Since 1 and 2 are the only positive
divisors of 2, we would have m + n = 2 and m − n = 1. But this would
imply that 2m = (m + n) + (m − n) = 3, which is impossible since
2m is even whereas 3 is odd.
(c) Given an odd integer n, we will show that there exist integers a, x, y,
and z such that
2018 + a = x2 , (1)
2
n+a = y , (2)
2018n + a = z2. (3)
10
We therefore choose x and y such that x + y = 2018 − n and x − y = 1.
Addition of these two equations gives us 2x = 2019 − n, subtraction
gives us 2y = 2017 − n. We therefore choose
2019 − n 2017 − n
x= and y= .
2 2
Since n is odd, these are integers. Indeed, we now have x − y = 1 and
x + y = 2018 − n, and hence that
2018 − n = x2 − y 2 . (4)
2018n + a = 2018n + y 2 − n
(2017 − n)2
= 2017n +
4
4 · 2017n + 20172 − 2 · 2017n + n2
=
4
20172 + 2 · 2017n + n2
=
4
2
(2017 + n)2
2017 + n
= = .
4 2
11
Final Round, September 2018
Problems
2. The numbers 1 to 15 are each coloured blue or red. Determine all possible
colourings that satisfy the following rules:
3. Determine all triples (x, y, z) consisting of three distinct real numbers, that
satisfy the following system of equations:
x2 + y 2 = −x + 3y + z,
2 2
y +z = x + 3y − z,
2 2
x +z = 2x + 2y − z.
12
4. In triangle ABC, ∠A is smaller than ∠C. Point D lies on the (extended)
line BC (with B between C and D) such that |BD| = |AB|. Point E lies
on the bisector of ∠ABC such that ∠BAE = ∠ACB. Line segment BE
intersects line segment AC in point F . Point G lies on line segment AD
such that EG and BC are parallel.
Prove that |AG| = |BF |.
C
E
F
5. At a quiz show there are three doors. Behind exactly one of the doors, a
prize is hidden. You may ask the quizmaster whether the prize is behind
the left-hand door. You may also ask whether the prize is behind the
right-hand door. You may ask each of these two questions multiple times,
in any order that you like. Each time, the quizmaster will answer ‘yes’
or ‘no’. The quizmaster is allowed to lie at most 10 times. You have to
announce in advance how many questions you will be asking (but which
questions you will ask may depend on the answers of the quizmaster).
What is the smallest number you can announce, such that you can still
determine with absolute certainty the door behind which the prize is hidden?
13
Solutions
1. First observe that a shuffle number can only contain the digits 2, 4, 6, and
8. Indeed, if we place the digits in any order, we obtain an even number
(since it is divisible by 12) because of property (3). Since the last digit of
an even number is also even, the digit must be 2, 4, 6, or 8 since it cannot
be 0 because of property (1). Since we can put any of the digits in the last
position, this holds for each digit of a shuffle number.
Next, observe that a shuffle number can only contain the digits 4 and 8.
Indeed, suppose that we had a shuffle number containing the digit 2. We
could reorder the digits so that the last digit is 2. The last two digits would
then be 22, 42, 62, or 82. But then the number would not be divisible by
4 (and hence also not divisible by 12), contradicting property (3). In the
same way we see that a shuffle number cannot contain digit 6 because a
number ending in digits 26, 46, 66, or 86 is not divisible by 4.
A shuffle number is divisible by 3 (since it is divisible by 12), hence
the sum of its digits is divisible by 3 as well. If our 10-digit number
has k eights and 10 − k fours, then the sum of its digits is equal to
8k + 4(10 − k) = 40 + 4k = 36 + 4(k + 1). This is divisible by 3 if and only
if k + 1 is divisible by 3. That is, if k = 2, k = 5, or k = 8. We see that a
shuffle number must have 2 eights and 8 fours, or 5 eights and 5 fours, or 8
eights and 2 fours. Note that each of those numbers satisfy (1) and (3). It
remains to examine which of these numbers are divisible by 11 (prop. (2)).
For this we use the 11-criterion: a number is divisible by 11 if the alternating
sum of the digits is divisible by 11. By alternating sum we mean that
instead of adding them, we alternately add and subtract. Since in our
case all digits are equal to 4 or to 8, the alternating sum must even be
divisible by 4 × 11 = 44. Since the alternating sum cannot be greater than
8 + 8 + 8 + 8 + 8 − 4 − 4 − 4 − 4 − 4 = 20 (and cannot be smaller than −20),
it must be equal to 0. In other words: the sum of the five digits in the odd
positions must equal the sum of the five digits in the even positions. This
means that the number of digits 8 in the odd positions must be equal to
the number of digits 8 in the even positions. We examine the three cases
that we found before:
• Suppose exactly 2 digits are eights. The only requirement is now that
there is exactly 1 eight in the odd positions and exactly 1 eight in the
even positions. There are 5 × 5 = 25 ways to do this.
• Suppose exactly 5 digits are eights. Since we cannot divide the eights
into two equal groups, there are no solutions in this case.
14
• Suppose exactly 8 digits are eights. The only requirement is now that
we have 4 eights in the odd positions and 4 eights in the even positions.
In other words: 1 odd position must be a four and 1 even position
must be a four. Again we find 5 × 5 = 25 possibilities.
We conclude that the total number of 10-digit shuffle numbers is 25+25 = 50.
2. We first consider the case that 1 is red. Then all numbers from 1 to 15
are red. Indeed, suppose that some number k is blue. Then 1 and k have
different colours, hence by the third rule the number 1 · k = k must be
red. But this contradicts the assumption that it was blue. Observe that
colouring all numbers red indeed gives a correct colouring.
Now we consider the case that 1 is blue. Observe that when two numbers
sum to 15, those numbers must have the same colour by the second rule.
We get the following pairs of numbers of the same colour: 1 and 14, 2 and
13, 3 and 12, 4 and 11, 5 and 10, 6 and 9, 7 and 8.
The number 2 is blue. Indeed, suppose that 2 is red. From the second rule,
we derive that 3 = 1 + 2 is blue. Repeatedly applying the same rule we find
that 5 = 3 + 2 is blue, that 7 = 5 + 2 is blue, and finally that 15 is blue.
Since 15 is in fact not blue, 2 cannot be red.
The number 7 is blue. Indeed, suppose that 7 is red. It then follows from
the second rule that 8 = 1 + 7 is blue. However, 7 and 8 must have the
same colour, so this cannot be the case.
The number 4 is blue. Indeed, suppose that 4 is red. It then follows from
the second rule that 11 = 4 + 7 is blue. But 4 and 11 have the same colour,
so this cannot be the case.
Recall, that 3 and 12 have the same colour, and so do 6 and 9. In fact,
all four numbers must have the same colour. Indeed, otherwise 9 = 3 + 6
would be blue by the second rule and also 12 = 3 + 9 would be blue by the
second rule.
So far, we know that 15 is red, that 1, 2, 4, 7, 8, 11, 13, 14 are blue, that
3, 6, 9, 12 have the same colour, and that 5, 10 have the same colour. The
numbers 3 and 5 cannot both be red, since if 3 is red, 5 = 2 + 3 is blue by
the second rule. The three remaining colour combinations for the numbers
3 and 5 result in the following three colourings:
(1) Only 15 is red.
(2) Only 5, 10, 15 (the numbers divisible by 5) are red.
(3) Only 3, 6, 9, 12, 15 (the numbers divisible by 3) are red.
15
It is easy to check that all three colourings are indeed correct. We write this
out for the third colouring, the other two can be checked in a similar way.
That the sum of a red and a blue number is always blue follows from the
fact that the sum of a number divisible by 3 and a number not divisible by
3 is itself not divisible by 3. That the product of a blue and a red number
is always red follows from the fact that the product of a number divisible
by 3 and a number not divisible by 3 is itself divisible by 3.
We conclude that there are 4 correct colourings in total: the colouring in
which all numbers are red, and colourings (1), (2), and (3).
x2 − z 2 = −x + z − (x − z),
(x − z)(x + z + 2) = 0.
y 2 − x2 = x + 3y − (2x + 2y),
(y − x)(y + x − 1) = 0.
2x2 − 2x + 1 = −5x + 1.
16
This is a quadratic equation 2x2 + 3x = 0, whose roots are x = 0 and
x = − 32 . In both cases we can deduce the values of y and z from the formulas
y = 1 − x and z = −x − 2. This gives the solutions (x, y, z) = (− 32 , 25 , − 21 )
and (x, y, z) = (0, 1, −2).
For these solutions we know that they satisfy the first equation. Since the
difference between the first and the second equation gives zero on both
sides (since we chose z = −x − 2), the second equation is satisfied as well.
Because of the choice y = 1 − x, also the third equation is satisfied. We
conclude that both solutions indeed satisfy all three equations.
C
E
F
B
A
G
17
Using the fact that |AE| = |AF |, which follows from the fact that 4AEF
is isosceles, we can now conclude that 4ABF and 4EGA are congruent.
This immediately implies |AG| = |BF | as required.
5. The smallest number of questions that suffices is 32. First we will show a
strategy to locate the prize in only 32 questions.
Start by asking the quizmaster if the prize is behind the left-hand door.
Repeat this until you are sure whether the prize is there or not. You
can only be 100% sure when you have received the same answer 11 times,
because the quizmaster can lie a maximum of 10 times. After doing this,
suppose that the quizmaster has lied n times. This means that you have
thus far asked 11 + n questions and the quizmaster is entitled to 10 − n
more lies.
Now ask the quizmaster whether the prize is behind the right-hand door,
and keep asking this until you are 100% sure of the true answer. Since the
quizmaster can lie only 10 − n more times, you need at most 2(10 − n) + 1 =
20 − 2n + 1 questions for this. In total you have now asked 32 − n questions.
We conclude that with this strategy, 32 questions always suffice.
Having only 31 questions, it is not always possible to locate the prize. We
will show how the quizmaster can make sure of that. At the beginning,
as long as you have asked about both doors at most 10 times, he always
answers ‘no’. To keep it simple, we assume that the left-hand door is the
first door that you query for the eleventh time (the other case is completely
analogous). We consider the situation that the prize is not behind the left-
hand door (which may happen). In this case, we show that the quizmaster
can make sure that after 31 questions, you cannot tell whether the prize is
behind the middle door or behind the right-hand door.
So far, the quizmaster has been speaking the truth about the left-hand
door, and he will continue to do so: when asked about the left-hand door,
he will always answer ‘no’. For the right-hand door, the quizmaster will
keep saying ‘no’ up to and including the tenth time he is asked about this
door. The next ten times he is asked about this door, he will answer ‘yes’.
Since you ask at most 20 questions about the right-hand door (since you
already queried the left-hand door at least 11 times), the quizmaster needs
to lie at most 10 times. Whether the prize is behind the middle door or
behind the right-hand door, in both cases the quizmaster gives the same 31
answers. Therefore, you cannot determine the door behind which the prize
is located.
18
BxMO Team Selection Test, March 2019
Problems
1. Prove that for each positive integer n there are at most two pairs (a, b) of
positive integers with following two properties:
(i) a2 + b = n,
(ii) a+b is a power of two, i.e. there is an integer k ≥ 0 such that a+b = 2k .
2. Let ABC be a triangle and let I be the incentre of this triangle. The line
through I perpendicular to AI intersects the circumcircle of 4ABC in
points P and Q, where P lies on the same side of AI as B. Let S be the
second intersection point of the circumcircles of 4BIP and 4CIQ. Prove
that SI is the angle bisector of ∠P SQ.
5. In a country, there are 2018 cities, some of which are connected by roads.
Each city is connected to at least three other cities. It is possible to travel
from any city to any other city using one or more roads. For each pair of
cities, consider the shortest route between these two cities. What is the
greatest number of roads that can be on such a shortest route?
19
Solutions
1. Suppose there are three or more of such pairs for one particular n. Then
by the pigeonhole principle, there are at least two such pairs, say (a, b) and
(c, d), with a ≡ c mod 2. Write a + b = 2k and c + d = 2` where k and `
are positive integers. Without loss of generality, assume that ` ≥ k. We
see that
2` − 2k = (c + d) − (a + b)
= (c + n − c2 ) − (a + n − a2 ) = c − a − c2 + a2
= (a + c − 1)(a − c).
As ` ≥ k, we have that 2k is a divisor of 2` − 2k . Because a ≡ c mod 2,
the integer a + c − 1 is odd, hence 2k is a divisor of a − c. As 2` − 2k ≥ 0
and a + c − 1 > 0, we have a − c ≥ 0. On the other hand, we have
a − c < a + b = 2k , hence 0 ≤ a − c < 2k , while 2k | a − c. We conclude
that a − c = 0, hence a = c. Now we have b = n − a2 = n − c2 = d, hence
the two pairs (a, b) and (c, d) are identical, which gives a contradiction.
Therefore, there at most two pairs (a, b) having the two properties.
2. Write ∠CAB = 2α, ∠ABC = 2β, and ∠BCA = 2γ. As the angles in
triangle 4AIB add up to 180◦ , we have
∠BIA = 180◦ − ∠IAB − ∠ABI = 180◦ − α − β = 90◦ + γ,
hence
∠BIP = ∠BIA − ∠AIP = 90◦ + γ − 90◦ = γ = ∠BCI.
Moreover, as BP QC is a cyclic quadrilateral, we have
If we now consider the sum of the angles in triangle BP I, and use both
previous results, we find
Using the fact that BSIP and CQIS are cyclic quadrilaterals, we get
∠P SI = ∠P BI = ∠ICQ = ∠ISQ,
21
5. The greatest number of roads that can be in such a shortest route, is
1511. We first describe a country for which this number is attained. Divide
the cities in 504 groups: two groups of five cities (A0 , B0 , C0 , D0 , E0 ) and
(A503 , B503 , C503 , D503 , E503 ), and groups of four cities (Ai , Bi , Ci , Di ) for
1 ≤ i ≤ 502. For all i with 0 ≤ i ≤ 503: connect Ai to Bi , connect Ai to
Ci , connect Bi to Di , and connect Ci to Di . Moreover, connect E0 to A0 ,
B0 , and C0 . Connect E503 to B503 , C503 , and D503 . For each 1 ≤ i ≤ 502:
connect Bi to Ci . Finally, for each 0 ≤ i ≤ 502: connect Di to Ai+1 . Now
every city is connect two exactly three other cities.
If we now want to travel from A0 to D503 , then we have to travel through
all Ai and Di , because the only connections between the different groups
are there, and each group is only connected to the previous and the next
group. Moreover, for 0 ≤ i ≤ 503, the city Ai is not connected to Di , which
causes the route within group i to go through either Bi or Ci . At least
three of the four or five cities of each group must lie on the route. In total,
this route visits at least 3 · 504 = 1512 cities and has at least 1511 roads.
We will now prove that the shortest route between two cities cannot have
more than 1511 roads. Consider such a shortest route which visits the cities
A0 , A1 , . . . , Ak consecutively. Each of the cities Ai has one more neighbour
besides Ai−1 and Ai+1 , which we will call Bi (note that the cities B0 , B1 ,
. . . , Bk do not have to be distinct). Moreover, A0 and Ak are connected
to a third city, say C0 6= B0 and Ck 6= Bk respectively. If one of the cities
Bi or Ci equals one of the cities Aj , then we could have found a shorter
route by going directly from Ai to Aj (or vice versa), which would be a
contradiction. Hence, the cities Bi and Ci are not equal to any of the cities
Aj .
If one of the cities Bi is connected to four cities Aj , say Bi is connected to
Ai , Am , An , and Ap with i < m < n < p, then we could shorten the route
by going from Ai to Bi and then to Ap . This makes us skip at least two
cities of the original route (Am and An ) and in their place we only visit Bi ,
hence this new route is shorter, which would be a contradiction. Therefore,
within the cities Bi with 3 ≤ i ≤ k − 3 there are at least k−5 3 distinct cities.
For B0 and C0 , the route can only be shortened if one of these cities equals
Bi for certain i ≥ 3. That would be a contradiction, hence B0 and C0 are
two distinct cities, not equal to Bi for i ≥ 3. In the same way, Bk and
Ck are two distinct cities, not equal to Bi for i ≤ k − 3. Altogether, we
have k + 1 cities on the route itself and at least k−5 3 + 2 + 2 other cities.
Therefore, k−53 + k + 5 ≤ 2018, hence 4k − 5 + 15 ≤ 3 · 2018 = 6054, hence
4k ≤ 6044, hence k ≤ 1511.
We conclude that the greatest number of roads occurring in a shortest route
is 1511.
22
IMO Team Selection Test 1, May 2019
Problems
1. Let P (x) be a quadratic polynomial with two distinct real roots. For all
real numbers a and b satisfying |a|, |b| ≥ 2017, we have P (a2 + b2 ) ≥ P (2ab).
Show that at least one of the roots of P is negative.
2. Write Sn for the set {1, 2, . . . , n}. Determine all positive integers n for
which there exist functions f : Sn → Sn and g : Sn → Sn such that for
every x exactly one of the equalities f (g(x)) = x and g(f (x)) = x holds.
4. We are given a triangle ABC. On edge AC there are points D and E such
that the order of points on this line is A, E, D, C. The line through E
parallel to BC intersects the circumcircle of 4ABD in a point F , with E
and F lying on opposite sides of AB. The line through E parallel to AB
intersects the circumcircle of 4BCD in a point G, with E and G lying on
opposite sides of BC. Prove that DEF G is a cyclic quadrilateral.
23
Solutions
1. Write P (x) = c(x − d)(x − e), where d and e are the distinct roots of P ,
so d =6 e. We also have c 6= 0, since otherwise P is not quadratic. We
distinguish two cases. First suppose that c > 0. Substituting b = −a =
−2017 in P (a2 + b2 ) ≥ P (2ab), we obtain P (2a2 ) ≥ P (−2a2 ), and therefore
Dividing both sides by the positive real number c, expanding, and then
canceling the terms 4a4 and de, we get
or equivalently,
4a2 (d + e) ≤ 0.
Dividing by 4a2 = 4 · 20172 , we see that d + e ≤ 0. As the roots d and e
are distinct, one of them must be negative.
Now consider the remaining case c < 0. Then the graph of P opens up
to the bottom, and therefore is descending on the right of the apex of the
parabola. Choose any a 6= b with a, b ≥ 2017 and a, b both to the right
side of the apex. Then by AM-GM we have a2 + b2 > 2ab. Since a, b are
positive, at least 1, and to the right of the apex, we have 2ab > a, so 2ab
(and therefore a2 + b2 as well) also lies to the right of the apex. Since P
is descending on the right side of the apex, we have P (a2 + b2 ) < P (2ab),
contrary to the given condition. We deduce that the case c < 0 is impossible
and therefore that c > 0, from which we conclude that at least one of the
roots of P is negative.
First note that all values lie in Sn , so these indeed define functions Sn →
Sn . The range of f is equal to {1, 2, . . . , m} and that of g is equal to
{m + 1, m + 2, . . . , 2m}. So f (g(x)) 6= x if x ≥ m + 1 and g(f (x)) 6= x if
x ≤ m. Moreover, for x ≤ m we have f (g(x)) = f (x + m) = x + m − m = x,
and for x ≥ m + 1 we have g(f (x)) = g(x − m) = x − m + m = x. Therefore
for every x ∈ Sn , exactly one of f (g(x)) = x and g(f (x)) = x holds.
24
Now suppose that n is odd, say equal to 2m + 1, and suppose that f and
g are functions satisfying the given conditions. Without loss of generality,
we assume that f (g(x)) = x for (at least) m + 1 elements of Sn , say for
x = x1 , . . . , xm+1 . Suppose that for some i, j with i ≤ i, j ≤ m + 1 we
have g(xi ) = xj . Then f (xj ) = f (g(xi )) = xi , so g(f (xj )) = g(xi ) = xj .
But now we have both f (g(x)) = x and g(f (x)) = x for x = xj , which
gives a contradiction. So we see that for all i with 1 ≤ i ≤ m + 1 the
number g(xi ) is not equal to any of the xj with 1 ≤ j ≤ m + 1. Since
there are m numbers in Sn not equal to any of the xj with 1 ≤ j ≤ m + 1,
and since there are m + 1 possible values of i, it follows that two of the
g(xi ) must be equal, say g(xk ) = g(xl ) with 1 ≤ k < l ≤ m + 1. But then
xk = f (g(xk )) = f (g(xl )) = xl , which gives a contradiction. Therefore no
such f and g can exist if n is odd.
Hence n satisfies the given conditions if and only of n is even.
G ≤ a + b + a ≤ a + b + c ≤ 5n.
25
4. Let G0 be the intersection of the line BF
and the circumcircle 4BCD. We show that
DEF G0 is a cyclic quadrilateral and that G =
G0 .
Since EF k BC we have 180◦ − ∠DEF =
∠AEF = ∠ACB, and since BDCG0 is a cyc-
lic quadrilateral we have ∠ABC = ∠DCB =
∠DG0 B = ∠DG0 F , therefore 180◦ −∠DEF =
∠DG0 F . Therefore DEF G0 is a cyclic quad-
rilateral. It follows that we have ∠DEG0 =
∠DF G0 = ∠DF B. Since ADBF is a cyclic quadrilateral we have ∠DF B =
∠DAB, so ∠DEG0 = ∠DAB, and therefore EG0 k AB. As G0 lies on the
circumcircle of 4BCD, we deduce that G = G0 .
26
IMO Team Selection Test 2, May 2019
Problems
1. In each of the different grades of a high school there are an odd number of
pupils. Each pupil has a best friend (who possibly is in a different grade).
Everyone is the best friend of their best friend. In the upcoming school
trip, every pupil goes to either Rome or Paris. Show that the pupils can be
distributed over the two destinations in such a way that
(i) every student goes to the same destination as their best friend;
(ii) for each grade the absolute difference between the number of pupils
that are going to Rome and that of those who are going to Paris is
equal to 1.
27
Solutions
1. We call a pupil a singleton if their best friend is in a different grade.
First give all singletons a destination as follows. Start by choosing any
singleton A1 and send them to Paris together with their best friend A2 .
Suppose the destinations for pupils A2m+1 , A2m+2 , . . . , A2k , with m ≤ 0
and k ≥ 1 are chosen. If possible, first choose in the grade of A2k a
singleton A2k+1 who doesn’t have a destination yet, and send them to the
city different from the destination of A2k , together with their best friend
A2k+2 . If possible, then choose in the grade of A2m+1 a singleton A2m
who doesn’t have a destination yet, and send them to the city different
from the destination of A2m+1 , together with their best friend A2m−1 .
Continue this process until pupils A2m+1 , A2m+2 , . . . , A2k with m ≤ 0 and
k ≥ 1 have destinations assigned to them and there are no singletons in the
grades of A2m+1 and A2k who don’t have a destination yet. These pupils
A2m+1 , A2m+2 , . . . , A2k are by construction such that A2i−1 and A2i are
best friends and A2i and A2i+1 have different destinations.
In every grade except that of A2m+1 and A2k the same number of pupils
is assigned to each destination. In the grades of A2m+1 and A2k (which
are different: otherwise no other singletons are in this grade, and all other
pupils form pairs of best friends, making the number of pupils in that class
even; contradiction!) one destination is assigned one more pupil than the
other, and these grade no longer have any singletons without destination.
Repeat this process until every singleton has a destination. All remaining
pupils now are in the same grade as their best friends, so every grade has
an even number of pupils without destination and therefore an odd number
of pupils with destination. So every grade contains a pupil on an end of
exactly one of the sequences of pupils (which then must be the last time a
pupil of this grade occurs in such a sequence). Therefore for each grade,
one destination is assigned exactly one more pupil of that grade than the
other.
Now consider any grade, and suppose that the number of singletons assigned
to Paris is one more than that assigned to Rome. Give the pupils in this
grade whose best friends are also in this grade a destination as follows. If
the number of pairs of best friends is odd, first send one pair to Rome;
otherwise do nothing. From the remaining even number of pairs of best
friends, send half of them to Paris and half of them to Rome.
Now note that in the first step, if the number of pairs was odd, then Rome
is assigned one more pupil than Paris. Therefore in any grade, the absolute
difference of the number of students going to Paris and the number going
to Rome is equal to 1.
28
2. We assume that a + b ≤ c + d. As we can simultaneously interchange a, b
and c, d without changing the problem, this assumption will not give any
loss of generality. We also assume that a ≤ b and c ≤ d; no generality is
lost here as we can interchange a and b (resp. c and d) without changing
2 2
the problem. Now we have a3 ≤ b3 , so ab ≤ ba , and analogously we have
2 2
c d
d ≤ c . The given equation is now equivalent to
b2 d2
a · c = (a + b)4 .
2 2
As ba ≥ b and dc ≥ d, the left hand side is at least bd. We now show that
bd ≤ (a + b)4 , from which follows that equality must hold everywhere.
Since a ≤ b we have b ≥ 12 (a + b), and analogously we have d ≥ 12 (c + d).
Since c+d ≥ a+b, it follows that d ≥ 21 (a+b). Moreover, from a+b+c+d = 1
and a + b ≤ c + d we deduce that a + b ≤ 12 , so
Q
3. Let T be the intersection of the cir-
cumcircle of 4BOC with AQ. We
show that T lies on N M . Write α = C
∠BAC. Then we have ∠BOC =
M
2α by the inscribed angle theorem.
N
Since |OB| = |OC| and ∠OCQ = O
90◦ = ∠OBQ by Thales’s theorem, B
we have 4OBQ ∼ = 4OCQ. Hence
A
29
As AT CM is a cyclic quadrilateral, we have ∠AT M = ∠ACM = 180◦ −
∠QCB − ∠BCA, using supplementary angles at C. Recall that ∠QCB =
∠QOB = α = ∠CAB, so using the sum of the angles in triangle ABC,
we obtain ∠AT M = 180◦ − ∠CAB − ∠BCA = ∠ABC. Since AT N B is a
cyclic quadrilateral, we see that ∠ABC = ∠ABN = 180◦ − ∠AT N , so we
have ∠AT M = 180◦ − ∠AT N . Therefore M , T , and N are collinear, as we
have set out to prove.
for all integers x and all prime numbers p by Fermat’s Little Theorem.
It follows that f (x) = x is the unique function Z → Z satisfying both
relations.
30
IMO Team Selection Test 3, May 2019
Problems
31
Solutions
32
3. The function f (x) = x for all x, satisfies the conditions. Now suppose that
f (x) = x does not hold for all x. As of (i), there is both a value of x for
which f (x) > x and a value of x for which f (x) < x. Let a ∈ Z be such
that f (a) < a and consider an arbitrary x with x 6≡ a mod 2. Then we
have f (x) + f (a) ≥ x + a, hence f (x) ≥ x + a − f (a) > x. Write w = f (x),
then we have f (w) = f (f (x)) = x < f (x) = w. If w 6≡ a mod 2, then we
would have f (w) > w, which gives a contradiction. Hence f (x) = w ≡ a
mod 2 and also f (x) 6≡ x mod 2.
Let x1 and x2 both be of different parity from a. Then they both do
not have the same parity as their function values, hence x1 + f (x2 ) and
x2 + f (x1 ) are odd. By substituting x = x1 , y = f (x2 ), and also x = x2 ,
y = f (x1 ), we find
f (x1 )+x2 = f (x1 )+f (f (x2 )) ≥ x1 +f (x2 ) = f (f (x1 ))+f (x2 ) ≥ f (x1 )+x2 .
33
4. We will prove that the maximum value of n is 200. We will first give an
example with n = 200. Consider players A1 , A2 , . . . , A200 and B1 , B2 ,
. . . , B100 . These are 300 players in total. Let Bi play a game against the
players Aj with 1 ≤ j ≤ i + 100, and assume no other contestants play a
game against each other. Then Bi has exactly 100 + i opponents, so for
101 ≤ m ≤ 200 there is a contestant playing exactly m games of chess.
Moreover, for j > 100, the player Aj is playing against the contestants Bi
with i ≥ j − 100; these are 100 − (j − 101) = 201 − j players. Because
j can vary from 101 up to and including 200, the number of opponents
varies from 100 up to and including 1. Hence, for 1 ≤ m ≤ 100 there is
also a contestant playing exactly m games of chess. Lastly, for j ≤ 100,
contestant Aj plays against all Bi ; these are 100 opponents. Hence, there
is no contestant playing more than 200 games of chess. We see that this
example meets the requirements for n = 200.
Now we will prove that n > 200 cannot be realised. We prove this by
contradiction, so assume that n > 200. Then in any case there is a
contestant A playing exactly 201 games of chess, against players B1 , B2 ,
. . . , B201 . In total, there are 300 contestants, so besides them and A,
there are another 300 − 1 − 201 = 98 contestants, who we will call C1 , . . . ,
C98 . If a contestant Bi is playing a game against another contestant Bj ,
then together with A they form a triple that plays three games among
each other; this is not allowed. Hence Bi does not play against any of the
other Bj . Therefore, they play at most 1 + 98 = 99 games. This means
that the contestants who play against exactly m other contestants with
100 ≤ m ≤ 200, must all be among the contestants Ci . However, these are
only 98 distinct contestants, which gives a contradiction. We conclude that
n > 200 is impossible and that n = 200 is the maximum.
34
Junior Mathematical Olympiad, September 2018
Problems
Part 1
1. When writing down the date 12 August 2018 using eight digits, each digit
occurs exactly twice: 12-08-2018. There are more dates in 2018 having the
same property. How many dates in 2018 have this property, including 12
August?
A) 5 B) 6 C) 7 D) 8 E) 9
3. Sophie likes to wear red and blue T-shirts. She decided to wear either a
red or a blue T-shirt each day, starting from 1 January 2019. She does not
want to say which colour she will be wearing on 1 and 2 January. From 3
January on, she will choose the colour of her T-shirt each day according to
the following rule: she chooses red if she wore two different colours the last
two days, and she chooses blue if she wore the same colour the last two
days.
By following this rule, she will wear a blue T-shirt on her birthday, 14
January. Is it possible to determine with certainty the colours of the T-
shirts she will be wearing on 28 and 29 January?
35
A) 28th: red, 29th: blue D) 28th: could be either 29th: blue
B) 28th: blue 29th: blue E) 28th: blue 29th: could be either
C) 28th: blue 29th: red
Pay attention: the right two squares are not considered to be the same. You
could obtain the right square from the middle one by reflection, but not by
using a rotation!
In how many different ways can you put the two crosses?
A) 21 B) 30 C) 32 D) 34 E) 36
36
7. Using 27 small cubes, each coloured black or white, we build a 3×3×3 cube.
This large cube has six views: a front, rear, left, right, top, and bottom
view. In the figure, five of the views of the large cube are depicted.
A) B) C) D) E)
8. How many distinct pairs of digits a and b are there such that 5a68 × 865b
is divisible by 824?
Pay attention: we are counting pairs, so for example if for a = 0 the values
b = 0, b = 1, and b = 2 are all valid, then these are counted as three
different pairs.
A) 10 B) 11 C) 15 D) 19 E) 21
Part 2
x x
2. What is the smallest positive integer x for which the outcome of 2 + 3 +
x x x
4 + 5 + 6 is an integer?
37
3. For each of the two fractions 2018 2054
2011 and 2019 we subtract the same integer a
from both the numerator and the denominator. The two fractions we get,
are equal.
What integer is a?
4. The sides of the square ABCD have length 10. The points P and Q lie on
the line connecting the midpoints of AD and BC. If we connect P with
A and C, and also Q with A and C, then the square is divided into three
parts having equal area.
What is the length of P Q? (The figure is not drawn to scale.)
D C
P
Q
A B
5. Jan has written the number A in two different ways as a fraction. Unfortu-
nately, two numbers have become invisible due to spilled ink spots:
+3 15
A= = .
12 26 −
Lia has found Jan’s note and wants to find all possibilities for the number
A. She knows that the numbers underneath the ink spots must be positive
integers, but the number A does not have to be an integer. Because she
does not know which numbers are hidden below the ink spots, she searches
for all combinations for which the equality holds. In this way, she finds
multiple possible values for A. The largest one she calls Amax and the
smallest one she calls Amin .
Determine Amax − Amin .
38
6. Tim draws five straight lines in a plane. The lines are continuing indefinitely,
and there are no three lines going through the same point. For each
intersection point of two lines he gets one sweet, and for each set of two or
more parallel lines he gets one sweet as well. For instance, in the example
below, he gets 8 sweets because firstly there are 7 intersection points, and
secondly there is 1 set of three parallel lines, which is worth another sweet.
What are all possible numbers of sweets that Tim can get?
7. Six equilateral and equally sized triangles are glued to form a parallelogram
ABCD, see the figure.
D C
Y
X
A B
8. In a class room there are a number of students. They find out that for each
triplet of students, the following two statements are both true:
• Two of them never wrote a report together.
• Two of them did once write a report together.
What is the maximum possible number of students in the class room?
39
Answers
Part 1
1. B) 6
2. E) 6
1
4. A) 72
5. C) 920
6. C) 32
7. D)
8. D) 19
Part 2
57
1. 5 years 5. 4 (= 14 14 )
2. 20 6. 1, 5, 8, 10
5 10
3. 2009 7. 6 (= 12 )
20
4. 3 (= 6 23 ) 8. 5
40
We thank our sponsors
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
3 First Round, January 2018
7 Second Round, March 2018
12 Final Round, September 2018
19 BxMO Team Selection Test, March 2019
23 IMO Team Selection Test 1, May 2019
27 IMO Team Selection Test 2, May 2019
31 IMO Team Selection Test 3, May 2019
35 Junior Mathematical Olympiad, September 2018
Olympiad 2019
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2019
8 Second Round, March 2019
14 Final Round, September 2019
19 BxMO Team Selection Test, March 2020
24 IMO Team Selection Test 1, June 2020
28 IMO Team Selection Test 2, June 2020
32 IMO Team Selection Test 3, June 2020
36 Junior Mathematical Olympiad, September 2019
The 940 best students were invited to the second round, which was held in
March at twelve universities in the country. This round contained five open
questions, and two problems for which the students had to give extensive
solutions and proofs. The contest lasted 2.5 hours.
The 118 best students were invited to the final round. Also some outstanding
participants in the Kangaroo math contest or the Pythagoras Olympiad
were invited. In total about 150 students were invited. They also received
an invitation to some training sessions at the universities, in order to prepare
them for their participation in the final round.
The final round in September 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 58th edition 2019.
In February a team of four girls was chosen from the training group to
represent the Netherlands at the EGMO, which was supposed to be held in
the Netherlands. Being the host country, a second team of another four girls
was allowed to participate. Unfortunately, due to the Covid-19 pandemic
the EGMO could not take place in Egmond aan Zee. Instead, a virtual
event was held, where the Dutch team managed to win three bronze medals.
For more information about the EGMO (including the 2020 paper), see
www.egmo.org.
In March a selection test of three and a half hours was held to determine the
ten students participating in the Benelux Mathematical Olympiad (BxMO).
This contest was turned into a virtual event as well, held on 2 May. The
Dutch team achieved an outstanding result: one gold medal, three silver
1
medals and five bronze medals. For more information about the BxMO
(including the 2020 paper), see www.bxmo.org.
In June the team for the International Mathematical Olympiad 2020 was
selected by three team selection tests on 10, 11 and 12 June, each lasting
four hours. Since the IMO had been postponed to September, we decided to
participate in the new virtual event Cyberspace Mathematical Competition
(CMC) in July with the same six students. Three young, promising students
were selected to accompany the team to the IMO/CMC training camp,
which was held from 4 until 11 July in Egmond aan Zee.
For younger students the Junior Mathematical Olympiad was held in Oc-
tober 2019 at the VU University Amsterdam. The students invited to
participate in this event were the 100 best students of grade 2 and grade 3
of the popular Kangaroo math contest. The competition consisted of two
one-hour parts, one with eight multiple choice questions and one with eight
open 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.
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
Dutch delegation
The Dutch team for the virtual IMO 2020 consists of
Also part of the IMO/CMC selection, but not officially part of the IMO
team, are:
3
First Round, January 2019
Problems
A-problems
1. Arthur has written down five distinct positive integers smaller than 10. If
you add any two of these five numbers, then the result will never be equal
to 10.
Which number did Arthur write down for sure?
A) 1 B) 2 C) 3 D) 4 E) 5
4
4. There are 13 distinct multiples of 7 that consist of two digits. You want
to create a longest possible chain consisting of these multiples, where two
multiples can only be adjacent if the last digit of the left multiple equals
the first digit of the right multiple. You can use each multiple at most once.
For example, 21 – 14 – 49 is an admissible chain of length 3.
What is the maximum length of an admissible chain?
A) 6 B) 7 C) 8 D) 9 E) 10
5. In a table with two rows and five columns, each of the squares is coloured
black or white according to the following rules:
• Two adjacent columns may never have the same number of black
squares.
• Two 2×2-squares that overlap in one column may never have the same
number of black squares.
How many possible colourings of the table comply with these rules?
A) 6 B) 8 C) 12 D) 20 E) 24
6. Which of the following numbers is the largest number you can get by
separating the numbers 1, 2, 3, 4, and 5 by using each of the operations
+, −, :, and × exactly once, where you may use parentheses to indicate
the order in which the operations should be executed? For example:
(5 − 3)×(4 + 1) : 2 = 5.
53 69
A) 21 B) 2 C) 33 D) 2 E) 35
7. Agatha, Isa and Nick each have a different kind of bike. One of them has
an electric bike, one has a racing bike, and one has a mountain bike. The
bikes have different colours: green, blue and black. The three owners make
two statements each, of which one is true and the other is false:
• Agatha says: “I have an electric bike. Isa has a blue bike.”
• Isa says: “I have a mountain bike. Nick has an electric bike.”
• Nick says: “I have a blue bike. The racing bike is black.”
Exactly one of the following statements is certainly true. Which one?
A) Agatha has a green bike. D) Isa has a mountain bike.
B) Agatha has a mountain bike. E) Nick has an electric bike.
C) Isa has a green bike.
5
8. Quadrilateral ABCD has right angles at A and D. A D C
circle of radius 10 fits neatly inside the quadrilateral
and touches all four sides. The length of edge BC
is 24. The midpoint of edge AD is called E and the E F
midpoint of edge BC is called F .
What is the length of EF ?
√ √
A) 43 B) 13 11 C) 33 11 D) 22 E) 45 A B
2 2 5 2
B-problems
The answer to each B-problem is a number.
1. Every day, Maurits bikes to school. He can choose between two different
routes. Route B is 1.5 km longer than route A. However, because he
encounters fewer traffic lights, his average speed along route B is 2 km/h
higher than along route A. This makes that travelling along the two routes
takes exactly the same amount of time.
How long does it take for Maurits to bike to school?
4. Alice has a number of cards. Each card contains three of the letters A
to I. For any choice of two of those letters, there is at least one card that
contains both letters.
What is the smallest number of cards that Alice can have?
6
Solutions
A-problems
1. E) 5 5. D) 20
2. E) 10000 6. E) 35
4. B) 7 8. D) 22
B-problems
1. 45 minutes
2. 1836
10
3. 9
4. 12
7
Second Round, March 2019
Problems
B-problems
The answer to each B-problem is a number.
B1. After breakfast, the sisters Anna and Birgit depart for school, each going
to a different school. Their house is next to a bicycle path running between
the two schools. Anna is cycling with a constant speed of 12 km per hour
and Birgit is walking in the opposite direction with a constant speed of 4
km per hour. They depart at the same time. Shortly after their departure,
mother notes that the girls have forgotten their lunch and decides to go
after them. Exactly 10 minutes after Anna and Birgit have left, mother
departs on her electric bike. First, she catches up with Anna. She hands
her a lunch box, immediately turns around, and goes after Birgit. When
she catches up with Birgit, she hands her a lunch box and immediately
rides back home. Mother always rides at a constant speed of 24 km per
hour.
How many minutes after the departure of Anna and Birgit does mother
return home?
B2. In a tall hat there are one hundred notes, numbered from 1 to 100. You
want to have three notes with the property that each of the three numbers
is smaller than the sum of the other two. For example, the three notes
numbered 10, 15, and 20 would be suitable (as 10 < 15 + 20, 15 < 10 + 20,
and 20 < 10 + 15), but the notes numbered 3, 4, and 7 would not (as 7 is
not smaller than 3 + 4). You may (without looking at the numbers on the
notes) take some notes from the hat.
What is the smallest number of notes you have to take to be sure to have
three notes that meet your wish?
8
B4. If you try to divide the number 19 by 5, you will get a remainder. The
number 5 fits 3 times in 19 and you will be left with 4 as remainder. There
are two positive integers n having the following property: if you divide n2
by 2n + 1, you will get a remainder of 1000.
What are these two integers?
D C
B5. In a square ABCD of side length 2 we draw lines from
each vertex to the midpoints of the two opposite sides.
For example, we connect A to the midpoint of BC and to 2
C-problems For the C-problems not only the answer is important; you also have to
describe the way you solved the problem.
1 + (−2) + 17 = 2 + 6 + 8 = 3 + 13 = 4 + 12.
9
C2. A year is called interesting if it consists of four distinct digits. For example,
the year 2019 is interesting. It is even true that all years from 2013 up
to and including 2019 are interesting: a sequence of seven consecutive
interesting years.
10
Solutions
B-problems
1. 42
2. 11
3. −12
2
5. 3
C-problems
a1 + a4 = a2 + a5 = a3 + a6 and a1 + a3 + a5 = a2 + a4 + a6 .
11
We are looking for a sequence a1 , a2 , a3 , a4 , a5 , a6 , a7 that is 2-, 3-, and
4-composite. Hence, we need that
a1 + a3 + a5 + a7 = a2 + a4 + a6 , (1)
a1 + a4 + a7 = a2 + a5 = a3 + a6 , (2)
a1 + a5 = a2 + a6 = a3 + a7 = a4 . (3)
a1 = a4 − a5 , a2 = a4 − a6 , a3 = a4 − a7 .
2a4 + a7 − a5 = a4 + a5 − a6 = a4 + a6 − a7 .
a4 + a7 − a5 = a5 − a6 = a6 − a7 .
Hence, we obtain
a2 = 27 − 10 = 17, and a1 = 27 − 19 = 8.
Hence, we have found a solution.
(c) The largest k for which a k-composite sequence of 99 distinct integers
exists, is k = 50. An example of such a sequence is
The 99 integers in the sequence are indeed distinct and we see that
1 + 99 = 2 + 98 = . . . = 48 + 52 = 49 + 51 = 100, so this sequence is
50-composite.
12
Now suppose that k > 50 and that we have a k-composite sequence
a1 , a2 , . . . , a99 . Consider the group that contains the number a49 .
Since 49 − k < 0 and 49 + k > 99, this group cannot contain any other
number beside a49 . Next, consider the group containing the number
a50 . Since 50 − k < 0 and 50 + k > 99, this group cannot contain any
other number beside a50 . Hence, the numbers a49 and a50 each form
a group by themselves and must therefore have the same value. But
this is not allowed since the 99 numbers in the sequence had to be
distinct.
C2. (a) The years 2103 to 2109 are seven consecutive interesting years. If
there is an earlier sequence of seven, then it must start before 2100.
We shall now prove that this is not possible.
Because an interesting year cannot end with the digits 99, the first two
digits are the same for all years in a sequence of consecutive interesting
years (of four digits). Now suppose we have seven consecutive years
starting with digits 20. The seven final digits are consecutive and
unequal to 0 and 2, and therefore also unequal to 1. The seven final
digits must be the digits 3 to 9, in this exact order. Hence, the third
digit must be the only remaining digit, namely digit 1. We conclude
that 2013 to 2019 is the only sequence of seven consecutive interesting
years between 2000 and 2100.
(b) Suppose that there is a sequence of eight consecutive interesting years
between 1000 and 9999. Because an interesting year cannot end with
99, all eight years have the same first two digits. If also the third digit
does not change, then there are only 7 possibilities for the last digit,
which is not enough. Therefore, there are two consecutive years in our
sequence of the shape abc9 and abd0 with d = c + 1. Because there
are eight possible final digits, these must be the eight digits unequal
to a and b. Hence, both c and d = c + 1 must occur as final digit.
Because the numbers abcc and abdd cannot occur, this means that in
our sequence both abcd and abdc must occur. The difference between
these two numbers is 9, and our sequence consists of eight consecutive
numbers. This is also not possible. We have obtained a contradiction,
and conclude that the assumption that there exists a sequence of eight
consecutive interesting years between 1000 and 9999 is wrong.
13
Final Round, September 2019
Problems
2. There are n guests at a party. Any two guests are either friends or not friends.
Every guest is friends with exactly four of the other guests. Whenever a
guest is not friends with two other guests, those two other guests cannot
be friends with each other either.
What are the possible values of n?
F2 = F0 + F1 = 2, F3 = F1 + F2 = 3, F4 = F2 + F3 = 5, F5 = 8.
a0 + a1 + a2 + · · · + am < 1.
14
5. Thomas and Nils are playing a game. They have a number of cards,
numbered 1, 2, 3, et cetera. At the start, all cards are lying face up on
the table. They take alternate turns. The person whose turn it is, chooses
a card that is still lying on the table and decides to either keep the card
himself or to give it to the other player. When all cards are gone, each of
them calculates the sum of the numbers on his own cards. If the difference
between these two outcomes is divisible by 3, then Thomas wins. If not,
then Nils wins.
(a) Suppose they are playing with 2018 cards (numbered from 1 to 2018)
and that Thomas starts. Prove that Nils can play in such a way that
he will win the game with certainty.
(b) Suppose they are playing with 2020 cards (numbered from 1 to 2020)
and that Nils starts. Which of the two players can play in such a way
that he wins with certainty?
Solutions
15
2. We first consider the friends of one guest, say Marieke. We know that
Marieke has exactly four friends at the party, say Aad, Bob, Carla, and
Demi. The other guests (if there are any other guests) are not friends
with Marieke. Hence, they cannot have any friendships among themselves
and can therefore only be friends with Aad, Bob, Carla, and Demi. Since
everyone has exactly four friends at the party, each of them must be friends
with Aad, Bob, Carla, and Demi (and with no one else).
Since Aad also has exactly four friends (including Marieke), the group of
guests that are not friends with Marieke can consist of no more than three
people. If the group consists of zero, one, or three people, we have the
following solutions (two guests are connected by a line if they are friends):
D C D C B A
M
A D
M
B C A B M
Now we will show that it is not possible for this group to consist of two
people. In that case, Aad would have exactly one friend among Bob, Carla,
and Demi. Assume, without loss of generality, that Aad and Bob are friends.
In the same way, Carla must be friends with one of Aad, Bob, and Demi.
Since Aad and Bob already have four friends, Carla and Demi must be
friends. However, since they are both not friends with Aad, this contradicts
the requirement in the problem statement.
We conclude that there can be five, six, or eight guests at the party. Hence,
the possible values for n are 5, 6, and 8.
16
C
I
B
A
In this sum all terms cancel, except the first and last. In this way, we get
1 1 1
a0 + a1 + a2 + · · · + am = − =1− < 1.
F0 F1 Fm+1 Fm+2 Fm+1 Fm+2
5. (a) Thomas and Nils both make 1009 moves and Nils makes the last move.
Nils can make sure that the last card on the table contains a number
that is not divisible by 3. Indeed, he could start taking cards with
numbers that are divisible by 3, until all these cards are gone. Because
there are only 672 such cards, he has enough turns to achieve that.
17
We now consider the situation before the last move of Nils. Let k
be the number on the last card, and let the sums of the numbers
of Thomas and Nils at that very moment be a and b. Nils has two
options. If he gives away the last card, the difference between the
outcomes becomes (a + k) − b, and if he keeps the card, the difference
becomes a − (b + k). Nils is able to win, unless both numbers are
divisible by 3. But in that case (a + k − b) − (a − b − k) = 2k would
also be divisible by 3. Because k is not divisible by 3, the number 2k
is also not divisible by 3 and hence Nils can win with certainty.
(b) Nils can win. We distinguish three types of cards, depending on
the number on the card: type 1 (the number has remainder 1 when
dividing by 3), type 2 (the number has remainder 2 when dividing by
3), and type 3 (the number is divisible by 3). Because 2019 = 3 · 673
and the card 2020 is of type 1, there are 674 cards of type 1, 673 cards
of type 2, and 673 cards of type 3.
In order to win, Nils chooses a card of type 3 in his first turn (and
gives it to Thomas). Then there are 674 cards of type 1 left, 673 of
type 2, and 672 of type 3. In the next turns he responds to Thomas’s
move in the following way (as long as he is able to).
(i) If Thomas chooses a card of type 1, then Nils chooses a card of
type 2 and gives it to the same person that got Thomas’s card.
(ii) If Thomas chooses a card of type 2, then Nils chooses a card of
type 1 and gives it to the same person that got Thomas’s card.
(iii) If Thomas chooses a card of type 3, then Nils does the same (and
gives the card to Thomas).
As long as Nils keeps this up, the sum of each player’s cards is divisible
by 3 after his turn (because a number of type 1 and a number of type
2 add up to a number which is divisible by 3).
Because the number of cards of type 3 is always even after Nils’s turn,
Nils can always execute his planned move in case (iii). Because the
number of cards of type 1 is always 1 greater than that of type 2 after
Nils’s turn, he can also always execute his planned move in case (ii).
Only at the moment when all cards of type 2 are gone and Thomas
takes the last card of type 1 (case (i)), Nils cannot execute his planned
move. However, in that case Nils cannot lose anymore. Indeed, after
Thomas’s turn the sum of the cards of one player is still divisible by
3, but the sum of the cards of the other player is not divisible by 3
anymore. Because there are only cards of type 3 left now, this will stay
the same until all cards are gone. At the end, the difference between
the sums of both players is not divisible by 3 and Nils wins.
18
BxMO Team Selection Test, March 2020
Problems
2. In an acute triangle ABC the foot of the altitude from A is called D. Let D1
and D2 be reflections of D in AB and AC, respectively. The intersection of
BC and the line through D1 parallel to AB, is called E1 . The intersection
of BC and the line through D2 parallel to AC, is called E2 . Prove that
D1 , D2 , E1 , and E2 lie on a circle whose centre lies on the circumcircle of
4ABC.
for all x, y ∈ R.
4. On a circle with centre M there are three distinct points A, B, and C such
that |AB| = |BC|. The point D lies inside the circle in such a way that
4BCD is isosceles. The second intersection point of AD and the circle is
called F . Prove that |F D| = |F M |.
19
Solutions
A
D2
D1
E2
E1 B D C
2. Let K be the midpoint of DD1 , and let L be the midpoint of DD2 . Then
K lies on AB and L lies on AC. Because ∠AKD = 90◦ = ∠ALD, the
20
quadrilateral AKDL is cyclic. Hence, ∠DLK = ∠DAK = ∠DAB =
90◦ − ∠ABC. Moreover, KL is a midsegment in triangle DD1 D2 , hence
∠DLK = ∠DD2 D1 . We conclude that ∠DD2 D1 = 90◦ − ∠ABC.
Because AC ⊥ DD2 en D2 E2 k AC, we have ∠DD2 E2 = 90◦ . Hence,
∠D1 D2 E2 = ∠D1 D2 D + ∠DD2 E2 = 90◦ − ∠ABC + 90◦ = 180◦ − ∠ABC.
On the other hand, as D1 E1 k AB, we have ∠D1 E1 E2 = ∠ABC, hence
we get ∠D1 D2 E2 = 180◦ − ∠D1 E1 E2 . We conclude that D1 E1 E2 D2 is a
cyclic quadrilateral.
Let M be the point such that AM is a diameter of the circumcircle of
4ABC. Thales’ theorem yields ∠ACM = 90◦ . Hence, CM ⊥ AC, which
yields CM ⊥ D2 E2 and CM k DD2 . Moreover, L is the midpoint of
DD2 and LC k D2 E2 , hence LC is a midsegment in triangle DD2 E2 .
This means that C is the midpoint of DE2 . Because CM k DD2 , we
get that CM is also a midsegment, hence CM intersects D2 E2 in the
middle. As CM ⊥ D2 E2 , the line CM is the perpendicular bisector of
D2 E2 . Analogously, we get that BM is the perpendicular bisector of D1 E1 .
Hence, M is the intersection point of the perpendicular bisectors of two
of the chords of the circle through D1 , D2 , E1 , and E2 . Hence, M is the
centre of this circle.
2f (y 2 ) = f (y)2 . (4)
Using this, we can cancel the 2f (y 2 ) on the left hand side of the original
functional equation against the f (y)2 on the right hand side:
f (x2 y) = x2 f (y).
21
• c = 2, d = 2: then f (x) = 2x for all x;
• c = 2, d = −2: then f (x) = 2x for x ≥ 0, and f (x) = −2x for x < 0,
or, in other words, f (x) = 2|x| for all x.
Using the first function, both sides of the functional equation become 0, so
this function is a solution. Using the second function, we get 2x2 y + 4y 2
on the left hand side, and (x2 + 2y) · 2y = 2x2 y + 4y 2 on the right hand
side, so this function is a solution as well. Using the third function, we get
2|x2 y| + 4|y 2 | = 2x2 |y| + 4y 2 on the left hand side, and (x2 + 2|y|) · 2|y| =
2x2 |y| + 4|y|2 = 2x2 |y| + 4y 2 on the right hand side, so also this function is
a solution.
Altogether, we found the three solutions: f (x) = 0, f (x) = 2x, and
f (x) = 2|x|.
C
A
M
22
4CM F is isosceles with an angle of 60◦ , which yields that the triangle is
equilateral. This means that |F C| = |F M |.
This concludes the proof that |F D| = |F C| = |F M |.
For each pi , there are 201 numbers in S that are divisible by pi (namely,
pi and all multiples of P ). Of these, at most one has two factors pi ; the
rest has only one factor pi . If we now take 100 numbers from S, then their
product has at most 101 factors pi . The other numbers contain at least
101 numbers which are divisible by pi , hence their product has at least 101
factors pi . Because this holds for any pi , and the numbers in S do not have
any other prime factors, this implies that S has the desired property.
We now prove that S cannot contain more than 1819 primes. Consider a
prime divisor q of a number in S. Suppose that at most 199 numbers S
are divisible by q. Then we take the 100 elements of S having the most
factors q; these always have more factors q in total than the other elements,
which contradicts the condition in the problem statement. Hence, there
are at least 200 numbers in S which are divisible by q. If there are exactly
200, then we also get that the number of factors q in all of these numbers
must be equal, otherwise we get a contradiction again by taking the 100
elements having the most factors q.
We see that S contains at least 199 non-primes, because a prime p in S
divides at least 199 other elements of S. Suppose that S contains exactly
199 non-primes. Then the prime factor p in each of these 199 non-primes
occurs exactly once (namely, equally often as in the prime number p).
Moreover, the numbers in S cannot be divisible by a prime r that is not
contained in S, because then there would be at least 200 multiples of r
inside S, and these would be 200 non-primes, which is a contradiction.
We get that each of the 199 non-primes in S must be the product of the
primes in S. In particular, these 199 numbers are not distinct. This is a
contradiction, hence S must contain at least 200 non-primes, and hence at
most 1819 primes.
23
IMO Team Selection Test 1, June 2020
Problems
1. In an acute triangle ABC, the centre of the incircle is I, and |AC| + |AI| =
|BC|. Prove that ∠BAC = 2∠ABC.
Solutions
1. Let D be a point on BC such that C
|CD| = |AC|. Because |BC| =
|AC| + |AI|, the point D lies on
the interior of side BC, and we have
|BD| = |AI|. Because triangle ACD
I
is isosceles, the angle bisector CI
is also the perpendicular bisector
of AD, hence A is the reflection of B
D in CI. Hence, we get ∠CDI = A
∠CAI = ∠IAB, hence 180◦ −∠BDI =
∠IAB, which means that quadrilateral ABDI is cyclic. In this cyclic quad-
rilateral BD and AI have the same length. Therefore, AB and ID are
parallel. Hence, ABDI is an isosceles trapezium, which has equal angles at
the base. Hence, ∠CBA = ∠DBA = ∠BAI = 12 ∠BAC, which proves the
statement.
24
2. We rewrite the equation as
2
P (x2 ) − 1 = P (x) − 1 .
Q(x2 ) = Q(x)2 .
By comparing the coefficients in front of x2n on the left and right hand
side, we get b = b2 . As b =
6 0, we must have b = 1. Subtracting x2n on
both sides, yields
R(x2 ) = 2xn · R(x) + R(x)2 .
If R is non-zero, then it has degree m ≥ 0. We have m < n. Then the
left hand side of this equation has degree 2m, and the right hand side has
degree m + n, as m + n > 2m. This is a contradiction. Hence, R must
be the zero polynomial, which yields Q(x) = xn . This polynomial indeed
satisfies the polynomial equation for Q.
Hence, we find the following solutions: P (x) = 1, P (x) = 2, and P (x) =
xn + 1 with n ≥ 1.
3. The number of red squares must equal the total number of squares covered
by the n tiles, hence the tile are only put on top of red squares. Consider a
colouring of the board and the corresponding horizontal covering by the
tiles (where all tiles are places horizontally) and the vertical covering. We
will deduce a number of properties for the colouring, and then count how
many colourings there are. The tile whose size is 1 × k is called the k-tile.
Because the horizontal covering contains an n-tile, each column has at
least one red square. In the vertical covering, each column must therefore
contain at least one tile; because there are exactly n tiles, this means that
there must be exactly one tile in each column. In the same way, each row
must contain exactly one tile in the horizontal covering. Now number the
rows and columns depending on the number of the tile that has been put
there: so row i is the row containing the i-tile in the horizontal covering,
and analogously for the columns.
25
We will now prove that the square in row i and column j (which will be
called (i, j)) is red if and only if i + j ≥ n + 1. We prove this using induction
on i. In row 1, there is only one red square, so that must be in the column
containing the n-tile in the vertical covering, i.e. column n. Hence, the
square (1, j) is red if and only if j = n, or if and only if 1 + j ≥ n + 1. Now
let k ≥ 1 and suppose the statement has been proved for all i ≤ k. We want
to prove the statement for i = k + 1, i.e. that the square (k + 1, j) is red if
and only if k + 1 + j ≥ n + 1, or j ≥ n − k. Consider a column j ≥ n − k.
Because of the induction hypothesis, we know exactly how many red squares
this column has in rows 1, 2, . . . , k: namely, the square (i, j) is red if and
only if i+j ≥ n+1, or i ≥ n+1−j; these are k −(n−j) = j +k −n squares.
In the other n − k rows, this column needs another j − (j + k − n) = n − k
red squares. Hence, this column has a red square in each of these rows, in
particular in the row i = k + 1. In row i = k + 1, the squares (i, j) with
j ≥ n − k are all red, and these are k + 1 squares. Hence, these are exactly
all red squares in row i = k + 1, hence the square (i, j) is red if and only if
j ≥ n − k, or if and only if i + j ≥ n − k + k + 1 = n + 1. This finishes the
proof by induction.
Now consider two adjacent rows with row numbers a and b, with a > b.
In column n − b, there is a red square row a (because a + n − b > n), but
not in row b. In the row directly on the other side of row b (if this row
exists), there cannot be a red square in column n − b, because the red
squares in column n − b would otherwise not be consecutive, and then the
tile with number n − b cannot lie there. The row number of this row must
therefore be smaller than b. We conclude that the row numbers cannot
decrease first and then increase. Above and below row n, there must be a
row with a smaller number (or no row at all), and the row numbers must
descend in both directions from there. We see that the row numbers from
top to bottom must first ascend until we get to row n, and then they must
descend. The same can be proved for the column numbers.
Vice versa, we have to prove that if the row and columns numbers are first
ascending and then descending, then the horizontal and vertical tiles can be
put. To prove this, we colour the square (i, j) red if and only if i + j ≥ n + 1.
For a fixed i, the red squares are the squares (i, j) with j ≥ n + 1 − i;
because of the order of the column numbers, these columns are adjacent.
Hence, in each row, the red squares are adjacent. The horizontal tiles can
be put exactly on top of the red squares. In the same way, this can be done
for the vertical tiles. For these row and column numbers, there was not
other way to choose the colouring, because we already know that for each
suitable colouring the square (i, j) is red if and only if i + j ≥ n + 1.
Altogether, we are looking for the number of ways to choose the row
26
and column numbers such that the numbers are first ascending and then
descending; corresponding to each of these choices, there is exactly one
way to colour the squares so that they satisfy the conditions. The number
of ways to put the numbers 1 to n in an order that is first ascending and
then descending, equals the number of subsets of {1, 2, . . . , n − 1}. Namely,
each ordering corresponds to the subset of numbers that appear before the
number n; these can be sorted in a unique way (ascending), and the rest of
the numbers must be sorted descending and put after the n. The number
of subsets is 2n−1 . Hence, the total number of colourings satisfying the
conditions, is (2n−1 )2 = 22n−2 .
a
4. We will first show that it is possible to choose c and d such that b − dc = 1
bd .
Because gcd(a, b) = 1, there exists a multiplicative inverse b−1 of b modulo
a. Now let c with 1 ≤ c ≤ a be such that c ≡ −b−1 mod a. Then we
have bc ≡ −1 mod a, hence a | bc + 1. Define d = bc+1 a , which is a
bc+1 ba+1 1
positive integer. We have d = a ≤ a = b + a . Because a ≥ 2 and
d is an integer, we get d ≤ b. All conditions are met. Hence, we have
a c ad−bc
b − d = bd = bc+1−bc
bd = bd1
.
If this is the smallest possible outcome, then we are done, because 1r = bd
would be an integer. We will show that no smaller positive outcome is
achievable. Let c and d be as above, and suppose there are positive integers
0
c0 ≤ a and d0 ≤ b such that 0 < ab − dc0 < bd
1
. We will derive a contradiction.
0
Let x = ad0 − bc0 , then we have ab − dc0 = bdx0 , hence xbd < bd0 , which yields
xd < d0 . We also know that x > 0. Hence, 0 < xd < d0 ≤ b, which means
that xd and d0 are two distinct numbers whose difference is less than b.
Moreover, from x = ad0 − bc0 we get that x ≡ ad0 mod b. On the other
hand, we know that ad − bc = 1, hence ad ≡ 1 mod b, hence xad ≡ x
mod b. Combining this, we get ad0 ≡ xad mod b. Because gcd(a, b) = 1
we may divide by a, hence d0 ≡ xd mod b. However, we already saw that
d0 and xd are distinct numbers whose difference is smaller than b, hence
this is impossible.
We conclude that the c and d we found indeed give the smallest possible
outcome, and hence 1r is an integer.
27
IMO Team Selection Test 2, June 2020
Problems
if it exists. Assume that an+1 exists for all n ≥ 2020. Prove that an+1 ≤ an
for all n ≥ 2021.
4. Let ABC be an acute triangle and let P be the intersection of the tangents
in B and C to the circumcircle of 4ABC. The line through A perpendicular
to AB and the line through C perpendicular to AC intersect in a point
X. The line through A perpendicular to AC and the line through B
perpendicular to AB intersect in a point Y . Prove that AP ⊥ XY .
28
Solutions
1. If x = α is a root of Pn , then x = −α is root of Pn as well, as all terms of
Pn have even degree. The minimal root of Pn therefore cannot be positive.
Therefore an ≤ 0 for all n > 2020. We have Pn+1 (x) = x2 · Pn (x) + an+1 .
Substitute x = an+1 ; as that is a root of Pn , we have Pn+1 (an+1 ) =
0 + an+1 ≤ 0.
As the maximal degree term in Pn (x) is x2n , there exists an N < 0 such
that Pn (x) > 0 for all x < N . Taking for example −N = max(2, |a1 | +
|a2 | + . . . + |an |), we see for x < N that x2i−2 ≤ x2n−2 for all 1 ≤ i ≤ n
and therefore that
a1 x2n−2 + a2 x2n−4 + . . . +an−1 x2 + an
2. Gabrielle can always win using the following strategy: during each of her
turns, she picks the largest two numbers on the sheet as a and b. Using
induction on k, we will prove that she is always allowed to do so, and that
after her k-th turn, the sheet contains the number 2k + 1 and 998 − 2k
ones.
In his first turn, Ward can only pick a = b = 1, after which the sheet
contains the number 2 and 997 ones. Gabrielle then picks the two largest
numbers, a = 2 and b = 1, after which the sheet contains 3 and 996 ones.
This finishes the basis k = 1 of the induction.
Now suppose that for some m ≥ 1 after Gabrielle’s m-th turn the sheet
contains the number 2m + 1 and 998 − 2m ones. If 998 − 2m = 0, then
Ward cannot make a move. If not, then Ward can do one of two things,
either pick a = b = 1 or pick a = 2m + 1 and b = 1. We consider these two
cases separately:
29
numbers, so a = 2m + 1 and b = 2 (which is allowed since their gcd is
1). After her turn the sheet contains the numbers 2m+3 = 2(m+1)+1
and 996 − 2m = 998 − 2(m + 1) ones.
• If Ward picks a = 2m + 1 and b = 1, then the sheet contains the
number 2m+2 and 997−2m ones. Gabrielle then picks the two largest
numbers, so a = 2m + 2 and b = 1 (which is allowed since their gcd is
1). Note that there is a one left, as 997 − 2m is odd, so not equal to 0.
After her turn the sheet contains the numbers 2m + 3 = 2(m + 1) + 1
and 996 − 2m = 998 − 2(m + 1) ones.
3. First suppose that a = 1. Then ϕ(1) = 1. For all positive integers b we have
gcd(a, b) = 1. Therefore in this case the equation is 1 + b = 1 + ϕ(b) + 1, or
equivalently, ϕ(b) = b − 1. This is equivalent to the statement that there
exists a unique integer from {1, 2, . . . , b}; which then has to be b itself (since
unless b = 1, we have gcd(b, b) > 1, but if b = 1 we have ϕ(b) = b). In other
words, this is equivalent to b being a prime number. Hence the solutions
for a = 1 are precisely the pairs (1, p) with p a prime number. Similarly,
the solutions for b = 1 are precisely the pairs (p, 1) with p a prime number.
Now assume that a, b ≥ 2. As gcd(b, b) > 1 we have ϕ(b) ≤ b − 1. Therefore
Let p be the minimal prime divisor of a (which exists as a ≥ 2). Since for all
multiples tp ≤ a of p, we have gcd(tp, a) > 1, it follows that a − ϕ(a) ≥ p1 · a.
Therefore we have
a
gcd(a, b) ≥ a − ϕ(a) + 1 ≥ + 1.
p
30
m = ϕ(m). Therefore m = 1, and a = b = 2k . Indeed, the equation holds
for all pairs (2k , 2k ) with k ≥ 1.
Therefore the solutions of the equation are: the pairs (1, p) and (p, 1) for
prime numbers p, and (2k , 2k ) for all positive integers k.
Y
B C
4. Let M the circumcentre of 4ABC and let α = ∠BAC. We first show that
4BY A ∼ 4BM P and then that 4Y BM ∼ 4ABP . By the inscribed
angle theorem we have ∠BM C = 2∠BAC = 2α. Quadrilateral P BM C is a
kite with axis of symmetry P M (by the equality of radii |M B| = |M C| and
equality of tangent segments |P B| = |P C|), so M P bisects angle ∠BM C.
Therefore ∠BM P = 12 ∠BM C = α. Moreover, we have ∠P BM = 90◦
(tangent to a circle is perpendicular to its radius), so by the sum of angles
of a triangle we have ∠M P B = 90◦ − α.
On the other hand, we are given that ∠ABY = 90◦ and we also have
∠Y AB = ∠Y AC − ∠BAC = 90◦ − α. Therefore ∠ABY = ∠P BM and
∠Y AB = ∠M P B, from which follows that 4BY A ∼ 4BM P . From this
similarity it follows that |Y B| |M B|
|AB| = |P B| . Combining this with the equality of
angles
∠Y BM = ∠Y BA+∠ABM = 90◦ +∠ABM = ∠ABM +∠M BP = ∠ABP,
we see that 4Y BM ∼ 4ABP .
Let T now be the intersection of AP and Y M , then we have
∠BY T = ∠BY M = ∠BAP = ∠BAT,
from which it follows that BY AT is a cyclic quadrilateral. Therefore
∠AT Y = ∠ABY = 90◦ , so AT ⊥ Y M . Analogously, AP ⊥ XM . But
from this it now follows that Y M and XM coincide and we get that
AP ⊥ XY .
31
IMO Team Selection Test 3, June 2020
1. For a positive integer n, let d(n) be the number of positive divisors of n.
Determine the positive integers k for which there exist positive integers a
and b satisfying
k = d(a) = d(b) = d(2a + 3b).
2. Let a triangle ABC such that |AC| < |AB| be given, together with its
circumcircle. Let D be a varying point on the short arc AC. Let E be the
reflection of A in the internal angular bisector of ∠BDC. Prove that the
line DE passes through a fixed point, independent of where D lies.
for all x, y ∈ Z.
32
Solutions
1. For i ≥ 0, let a = 2 · 5i and b = 3 · 5i . Then both a and b have 2(i + 1)
divisors. Moreover, we have 2a + 3b = 4 · 5i + 9 · 5i = 13 · 5i , which also has
2(i + 1) divisors. Therefore all even values of k satisfy the condition in the
problem.
Now suppose that k is odd. Then a has an odd number of divisors and
therefore is a square, say a = x2 . The same reasoning shows that b is also a
square, say b = y 2 , and 2a + 3b is also a square, say 2a + 3b = z 2 . Therefore
we have
2x2 + 3y 2 = z 2 .
We show that this equation has no positive integer solutions.
Suppose for a contradiction that this equation does have a positive integer
solution. Let (x, y, z) = (u, v, w) be the solution with minimal x + y + z. So
2u2 + 3v 2 = w2 . Modulo 3 this equation is 2u2 ≡ w2 . If u is not divisible
by 3, then u2 ≡ 1 mod 3, so w2 = 2u2 ≡ 2 mod 3, however, that isn’t
possible. Therefore u is divisible by 3, from which it follows that w is
divisible by 3 as well. Now 2u2 and w2 are both divisible by 9, so 3v 2 is
divisible by 9. It follows that v is divisible by 3 as well. However, now
(x, y, z) = ( u3 , v3 , w3 ) also satisfies the equation, and it is a solution with
smaller x + y + z than the supposed minimal one. This is a contradiction.
Therefore the equation 2x2 + 3y 2 = z 2 has no positive integer solutions.
It follows that no odd k satisfies the condition in the problem. The positive
integers k that satisfy that condition are therefore the even integers.
A. We have ∠BDM = ∠M DC as
DM is the internal angular bisector
of ∠BDC, so arcs BM and CM have
B
the equal lengths. Hence M is inde- A
pendent of D.
Let S be the intersection of DE and
the circumcircle of 4ABC. We show
that S is independent of D. As S and M lie on the circumcircle of 4ABC,
we have ∠AM D = ∠ASD = ∠ASE. As E is the reflection of A in DM ,
we now see that ∠AM E = 2∠AM D = 2∠ASE.
33
Consider the circle with centre M passing through A. As E is the reflection
of A in DM , we have |M A| = |M E|, so this circle also passes through E.
By the inscribed angle theorem, from ∠AM E = 2∠ASE it follows that S
is also on this circle. Therefore S is the second intersection point of the
circumcircle of 4ABC and the circle with centre M passing through A.
This is a description of S independent of D. As DE passes through S, the
point S is the point required.
4. We show that the largest m Merlijn can pick is min(n, 3(n − k) + 1). First
we show that m ≤ min(n, 3(n − k) + 1). If Merlijn asks for n + 1 tiles,
Julian can instruct Merlijn to place them all horizontally. As n ≤ 2k − 1,
it is impossible to place more than one tile horizontally on a single row, so
Merlijn would need at least n + 1 rows, this is a contradiction. Therefore
m ≤ n.
Now suppose that Merlijn asks for 3(n − k)+2 tiles. Julian can now instruct
Merlijn to place n − k + 1 tiles vertically and 2n − 2k + 1 tiles horizontally.
34
Note that vertical tiles always cover the k − (n − k) = 2k − n ≥ 1 rows in
the middle of the board. Therefore the vertical tiles together cover at least
n − k + 1 squares in each of these rows in the middle of the board, leaving at
most k − 1 squares for the horizontal tiles. Therefore no horizontal tiles fit
in these middle rows, and the 2n − 2k + 1 horizontal tiles need to fit in the
remaining n − (2k − n) = 2n − 2k rows. This is a contradiction. Therefore
we must have m ≤ 3(n − k) + 1. It follows that m ≤ min(n, 3(n − k) + 1).
Now we show that it is always possible to place a pile of min(n, 3(n + k) + 1)
tiles on the board. We first consider two special configurations. Put n − k
horizontal tiles into a (n − k) × k-rectangle at the bottom left. To the right
of that, we can fit n − k more vertical tiles into a k × (n − k)-rectangle in
the bottom right. Above that rectangle, we can fit n − k more horizontal
tiles into a (n − k) × k-rectangle in the top right. Finally, to the left of
that, we can fit n − k more vertical tiles into a k × (n − k)-rectangle in the
top left. In this way, we can fit 2(n − k) horizontal and 2(n − k) vertical
tiles on the board.
One other way to cover the board is as follows. Place n vertical tiles in
the top left, covering a k × n-rectangle. Below that there is room for n − k
horizontal tiles to fit in an (n − k) × k-rectangle. In this way, we can fit n
vertical and n − k horizontal tiles on the board.
n−k n−k
k k
Suppose that Merlijn receives A horizontal tiles and B vertical tiles from
Julian. Then we have A + B ≤ n and A + B ≤ 3(n − k) + 1. Without
loss of generality we assume that A ≤ B. If A ≤ n − k, then Merlijn
uses the second special configuration, omitting tiles he doesn’t have. As
B ≤ n, this works. Else, A ≥ n − k + 1, so B ≤ 3(n − k) + 1 − A ≤
3(n − k) + 1 − (n − k + 1) = 2(n − k). As A ≤ B, we also have A ≤ 2(n − k),
and Merlijn can use the first configuration, omitting tiles he doesn’t have.
Therefore it is always possible for Merlijn to place min(n, 3(n − k) + 1) tiles
on the board.
35
Junior Mathematical Olympiad, September 2019
Problems
Part 1
3. In the puzzle below, a, b, c, d, and e are nonzero digits such that the two
calculations are correct. The digits need not be distinct.
How many solutions are there for which a < b?
A) 1 B) 2 C) 3 D) 4 E) 5
36
5. In a class, 30 students did a test. Every student got a grade that was an
integer from 1 to 10. The grade 8 was given more often than any of the
other grades.
What is the smallest possible average grade of the students?
8
A) 3 15 B) 3 23 C) 3 56 D) 4 11
30
8
E) 4 15
How many truly different colourings meet the requirement (including the
two from the figure)?
A) 2 B) 3 C) 4 D) 6 E) 12
8. Five smart students are sitting in a circle. The teacher gives one or more
marbles to each of them. He explains that he has handed out a total of
37
18 marbles, and that everyone got a different number of marbles. Each
student is allowed to see his own number of marbles, as well as the number
of marbles of his neighbour on the left and his neighbour on the right.
Using only this information, each student must try to logically deduce the
difference between the numbers of marbles of the two students opposite to
him. The teacher has distributed the marbles in such a way as to minimise
the number of students that are able to do this. How many students can
do it?
A) 0 B) 1 C) 2 D) 3 E) 5
Part 2
2. The numbers abcd and dcba consist of the same four digits a, b, c, and d,
but in opposite orders. When we add the two numbers, we get 13552.
Determine a + b + c + d.
3. One hundred students wear shirts numbered from 1 to 100. The students
are arranged in a square of ten rows by ten columns. It turns out that
adding the ten shirt numbers of the students in any row or any column
always yields the same outcome.
Determine that outcome.
4. The four points A, B, C, and D lie on a common line (in this order). There
is a point T not on the line such that |AB| = |BT | and |CD| = |CT |. Also,
angle T of triangle BT C is 54 degrees. In the figure you can see a sketch
of the situation; angles and sizes are not necessarily accurate.
Determine angle T of triangle AT D.
T
54◦
A B C D
38
5. Mieke has a stack of 21 cards. Mieke repeats the following operation:
She takes the top two cards from the stack, changes their order, and
then puts them at the bottom of the stack (so the top card becomes
the bottom card).
Mieke repeats this operation until the cards are back in their original order.
How many times does Mieke perform the operation?
6. The number
· · · 93} 919
2222 |9393{z
100 times 93
is divided by 2019.
Determine the sum of the digits of the resulting number.
7. On a white strip that is 100 mm long and 10 mm wide, ten black squares
are drawn that, from left to right, have sides of length 1, 2, . . . , 10 mm. The
centre of each black square is in the middle of the strip and 5, 15, . . . , 95
mm from the start (left edge) of the strip.
A transparent square is moving along the strip from left to right (indicated
in grey). In the figure, two possible positions of the transparent square are
depicted: in the first, its left edge is 21 mm from the start of the strip, and
in the second, it is 54 mm from the start of the strip. In both cases, less
than half of the part underneath the square is coloured black (only 9% in
the first case). There is one position in which exactly half of the part of
the strip underneath the square is coloured black.
Determine, for that position, how far the left edge of the square is from
the start of the strip.
39
8. We are given a triangle with an additional two points on each side. So in
total, there are nine points (see figure).
We want to choose three of the nine points that are not on one line. For
example, we could choose (1) the three vertices of the triangle, or (2) the
left vertex and the two additional points on the opposite side.
How many possible choices are there in total, including the two examples
given?
Answers
Part 1
1. C) 368 5. B) 3 23
16
2. D) 49 6. B) 3
4. D) 42 8. D) 3
Part 2
1. 9090 5. 110
2. 26 6. 103
3. 505 7. 61 18 mm
4. 117 graden 8. 72
We thank our sponsors
Olympiad 2020
de VU de eerste Junior Wiskunde Olympiade gehouden
voor de 100 beste deelnemers aan de Kangoeroewedstrijd.
International
De JWO wordt een jaarlijks terugkerend evenement. Mathematical
Zie ook: www.wiskundeolympiade.nl/junior Olympiad Am
NEDERLANDSE
WISKUNDE
OLYMPIADE
Contents
1 Introduction
4 First Round, January 2020
9 Second Round, March 2020
14 Final Round, September 2020
21 BxMO Team Selection Test, March 2021
27 IMO Team Selection Test 1, June 2021
31 IMO Team Selection Test 2, June 2021
36 IMO Team Selection Test 3, June 2021
The 944 best students were invited to the second round, which was held in
March at twelve universities in the country. This round contained five open
questions, and two problems for which the students had to give extensive
solutions and proofs. The contest lasted 2.5 hours.
The 128 best students were invited to the final round. Also some outstanding
participants in the Kangaroo math contest or the Pythagoras Olympiad
were invited. In total about 150 students were invited. They also received
an invitation to some training sessions at the universities, in order to prepare
them for their participation in the final round.
The final round in September 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 59th edition 2020.
In February a team of four girls was chosen from the training group to repre-
sent the Netherlands at the EGMO in Georgia, from 9 until 15 April. At this
virutal event the Dutch team won one bronze medal. For more information
about the EGMO (including the 2021 paper), see www.egmo.org.
In March a selection test of 3.5 hours was held to determine the ten students
participating in the Benelux Mathematical Olympiad (BxMO), also a virtual
event held on 1 and 2 May. The Dutch team achieved an outstanding result:
three gold medals, two silver medals and three bronze medals. For more
information about the BxMO (including the 2021 paper), see www.bxmo.org.
In June the team for the International Mathematical Olympiad 2021 was
selected by three team selection tests on 2, 3 and 4 June, each lasting 4
1
hours. A seventh, young, promising student was selected to accompany
the team to the IMO as an observer C. The team had a training camp in
Egmond aan Zee from 10 until 18 July.
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
Dutch delegation
The Dutch team for the virtual IMO 2021 consists of
Also part of the IMO selection, but not officially part of the IMO team, is:
3
First Round, January 2020
Problems
A-problems
2. Each of Kwik, Kwek, and Kwak is lying on two consecutive days of the week
and is telling the truth on the other five days. No two of them are lying
on the same day. Uncle Donald wants to know who of his nephews ate his
sweets. The three nephews know all too well who did it. On Sunday, Kwik
says that Kwek ate the sweets. On Monday, Kwik says that it actually was
not Kwek who ate the sweets, while Kwak claims that Kwik is innocent.
On Tuesday, however, Kwak says that it was Kwik who ate the sweets.
Who ate the sweets?
A) It was Kwik.
B) It was Kwek.
C) It was Kwak.
D) It was either Kwik or Kwek, but you cannot determine who of the
two.
E) It was either Kwik or Kwak, but you cannot determine who of the
two.
3. We consider numbers with two digits (the first digit cannot be 0). Such a
number is called vain if the sum of the two digits is greater than or equal
to the product of the two digits. For example, the number 36 is not vain,
as 3 + 6 is smaller than 3 · 6.
How many numbers with two digits are vain?
A) 17 B) 18 C) 26 D) 27 E) 37
4
4. A box measuring 4 dm by 15 dm is shoved against
6
the wall. On top of it, a second box, measuring 12
dm by 6 dm, is placed. A ladder exactly touches the 12
ground, the two boxes and the wall. See the figure 4
(which is not drawn to scale). 15
What is the length of the ladder in dm?
√ √ √
A) 30 B) 8 15 C) 31 D) 22 2 E) 18 3
5. On a 4×4 board, there are 16 grass hoppers, each on its own square. At a
certain time, each grass hopper jumps to an adjacent square: to the square
above, below, left, or right of its current square, but not diagonally and
not leaving the board.
What is the maximum number of squares that can be empty after the grass
hoppers have jumped?
A) 8 B) 9 C) 10 D) 11 E) 12
6. In the table below each of the three rows is a correct calculation (the symbol
÷ denotes division). Also each of the three columns (read from top to
bottom) is a correct calculation. However, the digits in the table have been
replaced by letters. Different letters represent different digits and no digits
are 0.
ABC − ADF = F
+ − −
ADD ÷ GC = C
= = =
CEF ÷ GD = D
5
7. We consider figures consisting of six squares whose
sides have length 1. The radius of such a figure is
the radius of the smallest circle containing the whole
figure. On the
√ right, there is an example of a figure
with radius 5.
Which of the following five figures has the smallest radius?
A) A B) B C) C D) D E) E
A B C D E
8. Lieneke is making bracelets with beads. Each bracelet has six beads: two
white, two grey, and two black beads. Some bracelets look different on first
sight, but are actually not different: by turning or flipping the first one
over, it looks the same as the other one. For example, the following three
bracelets are the same.
6
B-problems
The answer to each B-problem is a number.
B C
3. Annemiek and Bart each have a note on which they have written three
different positive integers. It appears that there is exactly one number that
is on both their notes. Moreover, if you add any two different numbers
from Annemiek’s note, you get one of the numbers on Bart’s note. One
of the numbers on Annemiek’s note is her favourite number, and if you
multiply it by 3, you get one of the numbers on Bart’s note. Bart’s note
contains the number 25, his favourite number.
What is Annemiek’s favourite number?
7
Solutions
A-problems
1. E) 13 31 5. C) 10
2. C) It was Kwak. 6. E) 9
3. D) 27 7. B) B
4. A) 30 8. B) 11
B-problems
1. 2
2. 36
3. 5
4. 10
8
Second Round, March 2020
Problems
B-problems
The answer to each B-problem is a number.
B1. The digit sum of a number is obtained by adding all digits of the number.
For example, the digit sum of 1303 is 1 + 3 + 0 + 3 = 7.
Find the smallest positive integer n for which both the digit sum of n and
the digit sum of n + 1 are divisible by 5.
J
D F C
B2. Rectangle ABCD is subdivided into four rectangles as in
the figure. The area of rectangle AEIG is 3, the area of G
I
H
rectangle EBHI is 5, and the area of rectangle IHCF A E B
is 12.
What is the area of the parallelogram AHJF ?
B3. A square sheet of paper lying on the table is divided into 8×8 = 64 equal
squares. These squares are numbered from a1 to h8 as on a chess board
(see fig. 1). We now start folding, in such a way that square a1 always stays
in the same spot on the table. First we fold along the horizontal midline
(fig. 1). This will cause square a8 to fold on top of square a1. Then we fold
along the vertical midline (fig. 2). Next, we fold along the new horizontal
midline (fig. 3), et cetera. After folding six times, we have a small package
of paper in front of us (fig. 7) that we can consider as a stack of 64 square
pieces of paper.
8
7
6
5
4 4 4
3 3 3
2 2 2 ···
1 1 1 1
a b c d e f g h a b c d e f g h a b c d a
fig. 1 fig. 2 fig. 3 fig. 7
The squares in this stack are numbered from bottom to top from 1 to 64.
So square a1 gets number 1.
Which number does square f6 get?
9
B4. One hundred brownies (girl scouts) are sitting in a big circle around the
camp fire. Each brownie has one or more chestnuts and no two brownies
have the same number of chestnuts. Each brownie divides her number of
chestnuts by the number of chestnuts of her right neighbour and writes
down the remainder on a green piece of paper. Each brownie also divides
her number by the number of chestnuts of her left neighbour and writes
down the remainder on a red piece of paper. For example, if Anja has 23
chestnuts and her right neighbour Bregje has 5, then Anja writes 3 on her
green piece of paper and Bregje writes 5 on her red piece of paper.
If the number of distinct remainders on the 100 green pieces of paper equals
2, what is the smallest possible number of distinct remainders on the 100
red pieces of paper?
10
C-problems For the C-problems not only the answer is important; you also have to
describe the way you solved the problem.
C1. Given a positive integer n, we denote by n! (‘n factorial’) the number we get
if we multiply all integers from 1 to n. For example: 5! = 1 · 2 · 3 · 4 · 5 = 120.
(a) Determine all integers n with 1 ≤ n ≤ 100 for which n! · (n + 1)! is a
perfect square. Also, prove that you have found all solutions n.
(b) Prove that no positive integer n exists such that n! · (n + 1)! · (n + 2)! ·
(n + 3)! is a perfect square.
C2. Three consecutive vertices A, B, and C of a regular octagon (8-gon) are the
centres of circles that pass through neighbouring vertices of the octagon.
The intersection points P , Q, and R of the three circles form a triangle
(see figure).
B
A C A C
P
Q R
P
Q R
11
Solutions
B-problems
1. 49999
2. 24 51
3. 43
4. 100
5. 02
C-problems
C1. (a) We observe that (n + 1)! = (n + 1) · n!, and therefore that n! · (n + 1)! =
(n!)2 · (n + 1). That product is a perfect square if and only if n + 1 is
a perfect square, since (n!)2 is a perfect square. For 1 ≤ n ≤ 100 this
is the case for n = 3, 8, 15, 24, 35, 48, 63, 80, 99 (perfect squares minus
one that are below 100).
(b) We rewrite the product n! · (n + 1)! · (n + 2)! · (n + 3)! as follows:
Since (n!)2 and ((n + 2)!)2 are both perfect squares, the above product
is a perfect square if and only if (n + 1)(n + 3) is a perfect square.
However, (n + 1)(n + 3) cannot be a perfect square. Indeed, suppose
that (n + 1)(n + 3) = k 2 were a perfect square. Since (n + 1)2 <
(n + 1)(n + 3) < (n + 3)2 we would have n + 1 < k < n + 3, so k = n + 2.
This is impossible because (n + 1)(n + 3) = (n + 2)2 − 1, which is not
equal to (n + 2)2 .
C2. An octagon can be subdivided into six triangles (see figure on the left).
Together, the angles of those six triangles add up to the same number of
degrees as the eight angles of the octagon. Since the angles of any triangle
add up to 180 degrees, this means that the eight angles of the octagon add
up to 6 · 180◦ = 1080◦ . Hence, each of the angles of the regular octagon is
1 ◦ ◦
8 · 1080 = 135 .
12
We now consider the figure from the problem statement (see figure on the
◦
right). Line segment BP bisects angle ABC, so ∠ABP = ∠P BC = 67 12 .
Since triangles ABP and BCP are isosceles (as |AB| = |AP | and |BC| =
◦
|CP |), we also have ∠AP B = ∠BP C = 67 12 and ∠BAP = ∠BCP =
180◦ − 135◦ = 45◦ .
In triangles ABQ and BCR all sides have the same length. These triangles
are therefore equilateral and all angles are 60◦ . From this, we deduce that
∠P AQ = ∠BAQ − ∠BAP = 15◦ . In the same way, we find ∠P CR = 15◦ .
Furthermore, triangles P AQ and P CR are isosceles (since |AP | = |AQ|
◦ ◦
and |CP | = |CR|), so ∠AP Q = 12 (180◦ − 15◦ ) = 82 12 and ∠CP R = 82 12 .
By mirror symmetry, P Q and P R have the same length, so P QR is an
isosceles triangle with apex P . We have already determined all angles at
P , except ∠QP R. We deduce that
From this and the fact that P QR is isosceles, we directly conclude that
P QR is equilateral.
A C
P
Q R
13
Final Round, September 2020
Problems
14
3. Given is a parallelogram ABCD with ∠A < 90◦ and |AB| < |BC|. The
angular bisector of angle A intersects side BC in M and intersects the
extension of DC in N . Point O is the centre of the circle through M , C,
and N .
Prove that ∠OBC = ∠ODC.
D C N
O
M
A B
4. Determine all pairs of integers (x, y) such that 2xy is a perfect square and
x2 + y 2 is a prime number.
5. Sabine has a very large collection of shells. She decides to give part of her
collection to her sister.
On the first day, she lines up all her shells. She takes the shells that are in
a position that is a perfect square (the first, fourth, ninth, sixteenth, etc.
shell), and gives them to her sister. On the second day, she lines up her
remaining shells. Again, she takes the shells that are in a position that is a
perfect square, and gives them to her sister. She repeats this process every
day.
The 27th day is the first day that she ends up with fewer than 1000 shells.
The 28th day she ends up with a number of shells that is a perfect square
for the tenth time.
What are the possible numbers of shells that Sabine could have had in the
very beginning?
15
Solutions
2. (a) First, we determine for what starting values a1 > 0 the inequalities
4 3
3 ≤ a2 ≤ 2 hold. Then, we will prove that for those starting values,
the inequalities 43 ≤ an ≤ 32 are also valid for all n ≥ 2.
First, we observe that a2 = aa11 +2
+1 and that the denominator, a1 + 1, is
positive (since a1 > 0). The inequality
4 a1 + 2 3
≤ a2 = ≤ ,
3 a1 + 1 2
is therefore equivalent to the inequality
4
3 (a1 + 1) ≤ a1 + 2 ≤ 32 (a1 + 1),
16
only with a1 replaced by a2 . Since a2 also satisfies 1 ≤ a2 ≤ 2, the
same argument now shows that 43 ≤ a3 ≤ 32 .
We can repeat the same argument to show this for a4 , a5 , etcetera.
Hence, we find that 43 ≤ an ≤ 32 holds for all n ≥ 2. The formal proof
is done using induction: the induction basis n = 2 has been shown
above. For the induction step, see the solution of part (b) of the
version for klas 5 & klas 4 and below. The result is that all inequalities
hold if and only if 1 ≤ a1 ≤ 2.
(b) Let’s start by computing the first few numbers of the sequence in
terms of a1 . We see that
a1 − 3
a2 =
a1 + 1
and
a1 −3
a2 − 3 a1 +1 −3 a1 − 3 − 3(a1 + 1)
a3 = = a1 −3 =
a2 + 1 +1
a1 +1
a1 − 3 + (a1 + 1)
−2a1 − 6 −a1 − 3
= = .
2a1 − 2 a1 − 1
Here, it is important that we do not divide by zero, that is, a1 = 6 −1
and a2 6= −1. The first inequality follows directly from the assumption.
For the second inequality we consider when a2 = −1 holds. This is
the case if and only if a1 − 3 = −(a1 + 1), if and only if a1 = 1. Since
we assumed that a1 6= 1, we see that a2 6= −1. The next number in
the sequence is
−a1 −3
a3 − 3 a1 −1 −3 −a1 − 3 − 3(a1 − 1) −4a1
a4 = = −a1 −3 = = = a1 .
a3 + 1 a1 −1 +1 −a1 − 3 + (a1 − 1) −4
3. As an intermediate step, we first show that triangles OCM and OCN are
congruent. Since AD and BC are parallel, we have (F angles): ∠CM N =
17
∠DAM = 12 ∠DAB. Since DN and AB are parallel, we have (Z angles):
∠CN M = ∠N AB = 12 ∠DAB. It follows that ∠CM N = ∠CN M , so
triangle CM N is isosceles with apex C. We obtain |CM | = |CN |. Line
segments OC, ON , and OM are radii of the same circle, and therefore of
equal length. Triangles OCM and OCN are therefore congruent (three
pairs of equal sides).
To show that ∠OBC = ∠ODC, we will show that triangles OBC and
ODN are congruent. We will do this using the ZHZ-criterion. We will
show that ∠ON D = ∠OCB, and |ON | = |OC|, and |DN | = |BC|.
The equality |ON | = |OC| follows since ON and OC are radii of the
same circle. In part (a), we saw triangles OCM and OCN are congruent.
Furthermore, these two triangles are isosceles (|OC| = |OM | and |OC| =
|ON |). Hence, the four base angles ∠ON C, ∠OCN , ∠OM C, and ∠OCM
are equal. We see that ∠ON D = ∠OCB. The only thing we still need to
show is that |DN | = |BC|.
Observe that ∠BM A = ∠DAM (Z angles) and ∠DAM = ∠M AB (as AM
is the angular bisector of A). We find that ∠BM A = ∠M AB. Triangle
AM B is therefore isosceles and we have |AB| = |BM |. We previously
saw that |CM | = |CN |, and we also have |AB| = |CD| as ABCD is a
parallelogram. We therefore obtain
18
Adding these two equations, we get 2x + 2y = p + 1. We also know that
x2 + y 2 + 1 = p + 1, so 2x + 2y = x2 + y 2 + 1. By bringing all terms to the
right-hand side and adding 1 to both sides, we obtain
1 = x2 + y 2 − 2x − 2y + 2 = (x − 1)2 + (y − 1)2 .
We now have two perfect squares that add up to 1. This implies that one
of the squares is 0 and the other is 1. So (x − 1)2 = 0 and (y − 1)2 = 1, or
(x − 1)2 = 1 and (y − 1)2 = 0. As x and y are positive, we find two possible
solutions: x = 1 and y = 2, or x = 2 and y = 1. In both cases 2xy = 4 is a
perfect square and x2 + y 2 = 5 is a prime number. It follows that both are
indeed solutions.
Adding the solutions obtained by replacing (x, y) by (−x, −y), we obtain a
total of four solutions (x, y), namely
(1, 2), (2, 1), (−1, −2), (−2, −1).
5. Suppose that on a given day, Sabine is left with n2 shells, where n > 1.
Then the next day, she will give n shells to her sister and will be left with
n2 − n shells. This is more than (n − 1)2 , since
(n − 1)2 = n2 − 2n + 1 = (n2 − n) − (n − 1) < n2 − n
as n > 1. The next day, she therefore gives n − 1 shells to her sister and is
left with n2 − n − (n − 1) = (n − 1)2 shells, again a perfect square. We see
that the numbers of shells that Sabine is left with are alternately a perfect
square and a number that is not a perfect square.
Let d be the first day that Sabine is left with a number of shells that is
a perfect square, say n2 shells. Then days d + 2, d + 4, . . . , d + 18 are the
second to tenth day that the remaining number of shells is a perfect square
(namely (n−1)2 , (n−2)2 , . . . , (n−9)2 shells). We conclude that d+18 = 28,
and hence d = 10.
On day 26 the number of remaining shells is at least 1000, but on days 27
and 28 this number is less than 1000. We see that (n−9)2 < 1000 ≤ (n−8)2 .
As 312 < 1000 ≤ 322 , we see that n − 8 = 32, and hence n = 40. We find
that day 10 is the first day that the number of remaining shells is a perfect
square, and that this number is 402 .
In the remainder of the proof, we will use the following observation.
Observation. On any day, starting with more shells, means that Sabine
will have more (or just as many) shells left after giving shells to her sister.
19
Indeed, suppose that Sabine starts the day with x shells, say n2 ≤ x <
(n + 1)2 . After giving away shells, she will be left with x − n shells. If she
had started with x + 1 shells instead of x, she would have been left with
x + 1 − n > x − n or x + 1 − (n + 1) = x − n shells.
We see that the case x = 1679 is ruled out because it would imply that
fewer than 402 = 1600 shells are left on day 10. By the above observation,
this also rules out the case x < 1679. The case x = 1682 is ruled out
because it would imply that more than 402 shells will be left on day 10.
Hence, also x > 1682 is ruled out. The number of shells left on day 8 must
therefore be 412 − 1.
To follow the pattern back in time, we consider the case that the number
of remaining shells is just shy of a perfect square. Suppose that on a
given day the number of remaining shells is n2 − a, where 1 ≤ a < n.
Then the following day, the number of remaining shells is n2 − a − (n − 1).
Since a < n, we have n2 − a − (n − 1) > n2 − n − (n − 1) = (n − 1)2 .
The day after that, the number of remaining shells must therefore be
n2 − a − (n − 1) − (n − 1) = (n − 1)2 − (a − 1).
So if Sabine originally had 452 − 5 shells, then the number of remaining
shells on days 2, 4, 6, and 8 are 442 − 4, 432 − 3, 422 − 2, and 412 − 1,
respectively. This gives us a solution.
If Sabine originally had 452 − 4 shells, then she would be left with too
many shells on day 8, namely 412 − 0. The original number of shells could
therefore not have been 452 − 4 or more.
If Sabine originally had 452 − 6 shells, then she would be left with too
few shells on day 8, namely 412 − 2. The original number of shells could
therefore not have been 452 − 6 or fewer.
We conclude that the only possibility is that Sabine started with a collection
of 452 − 5 = 2020 shells.
20
BxMO Team Selection Test, March 2021
Problems
1. Let ABCD be a cyclic quadrilateral with |AB| = |BC|. Point E lies on the
arc CD which does not contain A and B. The intersection of BE and CD
is denoted by P , the intersection of AE and BD is denoted by Q. Prove
that P Q k AC.
x2 − yz = |y − z| + 1,
y 2 − zx = |z − x| + 1,
z 2 − xy = |x − y| + 1.
(a) Suppose Patricia is using the number p at least once. How often does
she need to use the number p at least?
(b) Suppose that Patricia does not use the number p. In how many ways
can she assign the numbers? (Two ways are considered to be different
if to at least one black dot distinct numbers have been assigned. The
figure is not being turned or mirrored.)
4. Jesse and Tjeerd are playing a game. Jesse has n ≥ 2 stones. There are
two boxes: in the black box there is space for half of the stones (rounded
21
down) and in the white box there is space for half of the stones (rounded
up). Jesse and Tjeerd alternate turns, with Jesse as first player. In his
turn, Jesse takes one new stone, writes a positive real number on the stone
and puts it in one of the boxes which is not full yet. Tjeerd can see all
the numbers on the stones in each of the boxes and is allowed to move one
stone of his choice to the other box, if that other box is not full yet, but he
is also allowed to choose to do nothing. The game stops when both boxes
are full. If the total value of the stones in the black box is greater than
the total value of the stones in the white box, Jesse wins; otherwise Tjeerd
wins. Determine for each n ≥ 2 who can always win this game (and give a
winning strategy).
5. A triangle ABC has the property that |AB| + |AC| = 3|BC|. Let T be
the point on line segment AC satisfying |AC| = 4|AT |. Let K and L be
points on the interior of line segments AB and AC, respectively, such that
KL k BC, and KL is tangent to the incircle of 4ABC. Let S be the
|SL|
intersection of BT and KL. Determine the ratio |KL| .
22
Solutions
1. Because |AB| = |BC|, we have ∠AEB = ∠BDC, hence ∠QEP =
∠AEB = ∠BDC = ∠QDP , which yields that QP ED is a cyclic quadri-
lateral. Therefore, ∠QP D = ∠QED = ∠AED = ∠ACD. From this, we
get that QP and AC are parallel.
E
Q
P
A C
x2 − yz = y − z + 1,
y 2 − zx = x − z + 1,
z 2 − xy = x − y + 1.
23
but does not satisfy our assumption x ≥ y ≥ z. Thus, the only remaining
possibility is x = 43 , which gives the triple ( 43 , 43 , − 53 ). We verify that this
is indeed a solution.
Now consider the case x 6= y. Then we have x + y + z = −1, hence we
cannot have x + y + z = 1, and we see that y = z. Now x + y + z = −1
yields x + 2z = −1, hence x = −1 − 2z. Now the first equality becomes
(−1 − 2z)2 − z 2 = 1, which can be simplified to 3z 2 + 4z = 0. From this, we
conclude that z = 0 or z = − 43 . With z = 0, we find y = 0, x = −1, which
does not satisfy our assumption x ≥ y ≥ z. Hence, the only remaining
possibility is z = − 43 , and this gives rise to the triple ( 53 , − 34 , − 43 ). We
verify that this is indeed a solution.
By also considering the permutations of these two solutions, we find all
six solutions: ( 43 , 34 , − 53 ), ( 43 , − 53 , 43 ), (− 35 , 43 , 43 ), ( 53 , − 43 , − 43 ), (− 43 , 53 , − 34 ),
and (− 34 , − 43 , 53 ).
3. (a) From now on, with ‘lines’ we mean the six lines and the circle. As soon
as the number p is used somewhere, there is a ‘line’ whose product
is divisible by p, and hence has remainder 0 upon division by p. All
‘lines’ must give remainder 0 in that case, hence all ‘lines’ contain at
least one p. This can be achieved by assigning the number p to the
bottom three dots. Suppose that it is already possible with at most
two times the number p. Each dot is lying on exactly three ‘lines’,
hence there are at most six ‘lines’ containing a p. There are seven
‘lines’ in total, however, hence this is impossible. We conclude that
she needs the number p at least three times in total.
(b) Denote the bottom three numbers from left to right by a, b, and c.
Denote the number in the middle on the left side of the triangle by d.
Now we can compute all numbers modulo p; note that all numbers are
invertible modulo p because p itself does not appear, and p is prime.
The bottom line has product abc. The left side also needs to have this
product, hence the top number in the triangle must be congruent to
abc(ad)−1 = bcd−1 modulo p. If we consider the line from the right
bottom to the left middle, we see that the middle number must be
congruent to abc(cd)−1 = abd−1 . By considering the circle, we find
the number in the middle on the right side: abc(bd)−1 = acd−1 .
On the right side, we now get the equation bcd−1 · acd−1 · c ≡ abc
mod p, hence c2 ≡ d2 mod p. Similarly, the vertical line yields b2 ≡
d2 , and the line from the left bottom to the right middle yields a2 ≡ d2
mod p. We conclude that the numbers a, b, c, and d all must have
the same square. From x2 ≡ y 2 we obtain (x − y)(x + y) ≡ 0 mod p,
hence p | x − y or p | x + y, because p is prime. Therefore, we have
24
x ≡ y or x ≡ −y. These are two possibilities, because if y ≡ −y
were to hold, then 2y ≡ 0, hence p | 2y, hence p | y since p > 2; a
contradiction. Thus the numbers b, c, and d are all congruent to a or
−a. If these conditions are met, then we do have a2 ≡ b2 ≡ c2 ≡ d2
and from the preceding argument, we get that the product of the three
numbers on all six lines and the circle is congruent to abc.
There are p − 1 possibilities for the number a, and after choosing a,
there are 2 possibilities for each of the numbers b, c, and d. In total,
there are 8(p − 1) ways to assign the numbers.
4. We will show that the capicity of the two boxes does not matter, as long
as the total capacity is n (and at least 1 for each box). Jesse can always
win this game, and can do that by first playing the power 20 = 1 of two,
and then in each following turn the next power of two that is smaller or
greater. That means: if he played the numbers
at a certain moment, he will play either 2−(i+1) or 2j+1 in his next turn.
By playing cleverly, Jesse can make sure that the greatest power of two
among the stones played so far is always contained in the black box. We
will prove this by induction. In his first move, he puts the stone with
value 20 in the black box and the claim is true; this is the base case of the
induction. When it is his turn again, and Tjeerd moved the greatest power
of two so far, which according to the induction hypothesis was contained
in the black box, to the white box, then the black box actually has a free
space, and Jesse can put a new greater power of two in there, and the claim
is true. If Tjeerd moved some other stone or did nothing, then the greatest
power of two so far is still in the black box, and Jesse can play a smaller
power of two; it does not matter where he puts it. Also in this case, the
claim is true. This proves the induction step, and the claim is proved.
Therefore, after playing the last stone, the greatest power of two is in the
black box. It is greater than the sum of all smaller powers of two played
(2j > 2j − 2−i = 2j−1 + 2j−2 + · · · + 2−(i−1) + 2−i ), hence it is certainly
greater than the sum of the powers of two in the white box. Therefore, the
total value inside the black box is greater than the total value in the white
box.
25
5. Denote the radius of the incricle of 4ABC by r. Then the area of triangle
ABC is
1 1 1 1
|AB| · r + |BC| · r + |AC| = r · (3|BC| + |BC|) = 2r|BC|.
2 2 2 2
On the other hand, the area of ABC equals 12 h|BC|, where h is the altidude
from A. Hence, h = 4r. Because the distance from KL to BC is exactly
2r, the distance from A to KL is also 2r. Triangles AKL and ABC are
similar, because KL k BC, and the altitudes from A have lengths 2r and
4r, respectively, giving a multiplication factor of exactly 2. Hence, K is the
midpoint AB, and L is the midpoint of AC.
For the point T , we have |AC| = 4|AT |, hence |AT | = 14 |AC| = 12 |AL|,
hence T is the midpoint of AL. Now consider triangle ABL. In this triangle,
the segment BT is a median, because T is the midpoint of AL. Also LK
is a median as K is the midpoint AB. Their intersection point S is the
|SL|
centroid, from which we get that |KL| = 32 .
A K
B
26
IMO Team Selection Test 1, June 2021
Problems
an+1 − an = n(an − 1)
27
Solutions
1. The sequence is given by the formula an = 2 · n! + 1 for n ≥ 0. (We use the
usual definition 0! = 1, which satisfies 1! = 1 · 0!, in the same way we have
n! = n · (n − 1)! for other positive integers n.) We will prove the equality
by induction. We have a0 = 3, which equals 2 · 0! + 1. Now suppose for
certain k ≥ 0 that ak = 2 · k! + 1, then
2. Subtracting the second equation from the first yields x2 − x3 = x23 − x22 +
6x4 (x3 − x2 ), which we can factor as 0 = (x3 − x2 )(x3 + x2 + 1 + 6x4 ). We
see that x2 = x3 or x2 + x3 + 1 + 6x4 = 0. Similarly, we also have either
x2 = x3 or x2 + x3 + 1 + 6x1 = 0. Hence, if x2 6= x3 , the second equality
must hold in both cases; subtracting one from the other, we obtain x1 = x4 .
We conlude that either x2 = x3 or x1 = x4 . Analogously, we get for each
permutation (i, j, k, l) of (1, 2, 3, 4) that either xi = xj or xk = xl .
We will prove that at least three of the xi must be equal. If all four are equal,
this is true of course. Otherwise, there are two unequal ones, say x1 6= x2
without loss of generality. Then we have x3 = x4 . If also x1 = x3 holds,
then there are three equal elements. Otherwise, we have x1 = 6 x3 , hence
x2 = x4 and we also get three equal elements. Up to order, the quadruple
(x1 , x2 , x3 , x4 ) is thus equal to a quadruple of the shape (x, x, x, y), where
we could have that x = y.
Substituting this in the equations gives x + y = 8x2 and 2x = x2 + y 2 + 6xy.
Adding these two equations: 3x + y = 9x2 + y 2 + 6xy. The right hand
side can be factored as (3x + y)2 . Defining s = 3x + y, the equation
28
becomes s = s2 , from which we get either s = 0 or s = 1. We have
s = 3x + y = 2x + (x + y) = 2x + 8x2 . Hence, 8x2 + 2x = 0 or 8x2 + 2x = 1.
In the first case, we have x = 0 or x = − 14 . We find y = 0 − 3x = 0 and
y = 0 − 3x = 34 , respectively. In the second case, we get the factorisation
(4x − 1)(2x + 1) = 0, hence x = 14 or x = − 12 . We find y = 1 − 3x = 14 or
y = 1 − 3x = 52 , respectively.
Altogether, we found the following quadruples: (0, 0, 0, 0), (− 14 , − 14 , − 14 , 34 ),
( 14 , 14 , 41 , 14 ) and (− 12 , − 12 , − 12 , 25 ), and permutations thereof. It is a simple
computation to verify that all these quadruples are indeed solutions to the
equations.
O H
A B
where we use that |OA| = |OS|. Now we conclude that 4OHS ∼ = 4OHD
(SAA). This yields that D and S are each others reflection images in OH.
29
Therefore, if we reflect the circumcentre K of 4OHS in OH, we get the
circumcentre L of 4OHD. Now we must prove that L lies on BC.
Point D is the reflection of H in BC. This is a known fact, which we can
prove as follows: ∠DBC = ∠DAC = ∠HAC = 90◦ − ∠ACB = ∠HBC
and analogously ∠DCB = ∠HCB, hence 4DBC ∼ = 4HBC (ASA). Hence,
D is indeed the reflection of H in BC, from which we get that BC is the
perpendicular bisector of HD. Becuase L lies on the perpendicular bisector
of HD, we get that L lies on BC, which is what we wanted to prove.
4. Assign each empty square to the domino directly right of this square (unless
the square is on the right edge of the board). Now suppose that two empty
squares are assigned to the same domino, then this domino must be placed
vertically and both squares left of this domino are empty. However, that
would mean that another domino could fit, which is a contradiction. Hence,
no two empty squares are assigned to the same domino.
The empty squares on the right edge of the board have not been assigned a
domino yet. We try to assign these squares to dominos that do not have
an empty square directly left of them (i.e. dominos which have not been
assigned yet). First suppose that we succeed in assigning all empty squares
on the right edge of the board in this way to different dominos. In that case,
we assigned each empty square to a domino, where no domino has been
assigned more than one square. Because each domino covers two squares of
the board, there are two covered squares for each empty square, and hence
at most 13 of the squares is uncovered. In this case, we are done.
Now we will show that this assignment always works. Let k be the number
of empty squares on the right edge, and ` the number of empty squares on
the left edge. The empty squares on the left edge cannot be adjacent, hence
there are at least ` − 1 dominos on the left edge and these all do not have
an empty square left of them. If ` > k, then there are enough dominos on
the left edge to assign to all empty squares on the right edge. If ` < k, we
could turn everything around and assign all empty squares to the domino
left of them and we could also prove that at most 13 of the squares on the
board are uncovered. The only remaining situation is when ` = k and both
on the left and right exactly k − 1 dominos are on the edge. For both edges,
we have that there must be an empty square between each two dominos,
and there must also be empty squares in the corners. This, however, is
in contradiction with the condition that not all corners are empty. Hence,
this situation cannot occur.
30
IMO Team Selection Test 2, June 2021
Problems
3. Show that for every positive integer n there exist positive integers a and b
with
n | 4a2 + 9b2 − 1.
4. Determine all positive integers n with the following property: for every
triple (a, b, c) of positive real numbers there exists a triple (k, `, m) of
non-negative integers such that ank , bn` , cnm are the lengths of sides of a
(non-degenerate) triangle.
31
Solutions
D
B C
2. We use the convention that m is the number of rows, and that n is the
32
number of columns. If m is even, then we pair the squares of the board as
follows: in every column we pair the top two squares, then squares 3 and 4,
etc. As the number of rows is even, this pairs the squares of every column
completely. Sting can use the following strategy: whenever Prickle places
a pawn into one of a pair of squares, Sting places a pawn into the other,
which is necessary adjacent to the pawn that Prickle just placed. After
each move of Sting, all pairs contain either zero or two pawns, so Sting
can always place a pawn according to this strategy. Therefore Sting can
make sure that the entire board is eventually completely filled with pawns,
so Sting wins. Analogously, if n is even, then Sting has a similar winning
strategy.
If m = n = 1, Sting wins after Prickle’s first move. If m = 1 and n = 3 (or
the other way around), Sting can place his first pawn on a square adjacent
to the square in which Prickle has just placed his; such an adjacent (empty)
square always exists. The board will then be completely filled after Prickle’s
next move, so Sting has a winning strategy in these cases as well.
Now consider the case m = n = 3. Prickle can follow the following strategy.
He places his first pawn in the centre square. Sting must place his pawn
either in the same row or in the same column. Without loss of generality,
we assume that Sting’s pawn is placed in the same column as that of Prickle.
Prickle then places a pawn in the remaining square in the middle column.
At that point, the left and right columns are completely empty. Sting must
place his pawn into one of these columns. Prickle then chooses any square
in the other column, all three squares of which are still empty. This forces
Sting to place a pawn in that column as well, as the middle column was
already completely filled. Now Prickle places a pawn in the only remaining
square of that column, after which Sting can no longer place a pawn, while
there are two empty squares remaining. Therefore Prickle has a winning
strategy if m = n = 3.
The remaining case is that m and n are both odd, and that at least one
of m and n is at least 5. We consider the case that n (the number of
columns) is at least 5; the other case is analogous. Prickle can follow the
following strategy. Prickle places his pawns into the centre column until it
is completely filled with pawns. When it is Prickle’s turn again, there is an
even number of pawns on the board, all in the three columns in the centre
of the board, of which at least the middle one is completely filled. The
first and the last column are completely empty. As there is an odd number
of squares remaining, either the area to the left of the middle column has
an odd number of empty squares remaining, or the area to the right has.
Prickle chooses the area which contains an odd number of empty squares,
and places his pawns there until that area is full. As the centre column is
33
completely filled, Sting must also place all of his pawns in that area. As
this area contained an odd number of empty squares, Prickle is the last
player (to be able) to place a pawn in this area. Sting can no longer place a
pawn, while the other area contains a column of empty squares, so Prickle
wins. (Of course, if Sting runs out of moves earlier, Prickle also wins.)
We conclude that Prickle wins if m and n are odd and m ≥ 5, if m and n
are odd and n ≥ 5, and if m = n = 3. Sting wins in all other cases: if m
and n are both even, if m = n = 1, if m = 1 and n = 3, and if m = 3 and
n = 1.
4. It is clear that n = 1 does not satisfy the property, as not every three
positive real numbers a, b, and c are the lengths of the sides of a triangle.
We first show that any n ≥ 5 cannot satisfy the property by considering
the triple (a, b, c) = (1, 2, 3). Suppose that there exist k, `, m such that
nk , 2n` , 3nm are the lengths of the sides of a triangle. As n 6= 2, 3, no three
of these are equal. By repeatedly removing common factors n if they exist,
we can and do assume that one of k, `, m is equal to 0. Suppose that the
other two of those three integers are positive, then their corresponding
34
side lengths both are multiples of n. Their difference then is at least n,
while the third side has length of at most 3. This contradicts the triangle
inequality. Hence of k, `, m, at least two must be zero. The corresponding
two side lengths then sum to at most 5, so the third side must have length
less than 5. As n ≥ 5, this third side cannot have a factor n either, so
k, `, m are all equal to 0. However, then 1, 2, 3 should be the lengths of
the sides of a triangle, while 3 = 2 + 1. This is a contradiction. We deduce
that n ≥ 5 does not satisfy the property.
Now consider n = 2, 3, 4. We construct (k, `, m) as follows. Take a triple
(a, b, c). If its entries already are the lengths of the sides of a triangle, we
take k = ` = m = 0. Otherwise there exists a triangle inequality that is not
satisfied. Without loss of generality, we assume that a ≥ b + c. Multiply
the lesser of b and c with n. If the right hand side still isn’t larger than a,
then take the new summands, and again multiply the lesser of the two by
n. So: if a ≥ ni b + nj c, then we multiply the lesser of ni b and nj c by n,
and repeat if the inequality still holds. This process always stops at some
point, as there exists i such that ni > a.
Consider the i and j such a ≥ ni b + nj c, such that applying the process
makes the right hand side larger than a. We assume without loss of
generality that ni b ≤ ni c, so we have a < ni+1 b + nj c. We claim that we
can take (k, `, m) to be equal to either (0, i + 1, j) or (0, i + 1, j + 1).
By definition, we have a < ni+1 b + nj c, and we have nj c < ni b + nj c ≤ a <
a+ni+1 b. Therefore, if (0, i+1, j) doesn’t satisfy the property, then we must
have ni+1 b ≥ a + nj c. Moreover ni b ≤ nj c, so ni+1 b ≤ nj+1 c < a + nj+1 c.
We also have a < ni+1 b + nj+1 c, so if (0, i + 1, j + 1) does not satisfy
the property, then we must have nj+1 c ≥ a + ni+1 b. We will derive a
contradiction in case both triples do not satisfy the property, i.e. in case
both ni+1 b ≥ a + nj c and nj+1 c ≥ a + ni+1 b.
Adding these inequalities, and subtracting ni+1 b on both sides, yields
nj+1 c ≥ 2a + nj c, or equivalently (n − 1)nj c ≥ 2a. Therefore
2a
ni+1 b ≥ a + nj c ≥ a + = 1 + n−12
a,
n−1
from which follows that
(n − 1) + 2 + 2n 3n + 1
a ≥ ni b + nj c ≥ n1 1 + n−1
2
a+ 2
n−1 a = a= a.
n(n − 1) n(n − 1)
So 3n + 1 ≤ n(n − 1). For n = 2, 3, 4, this inequality reads 7 ≤ 2, 10 ≤ 6,
and 13 ≤ 12, all of which are false. This is a contradiction.
Hence n = 2, 3, 4 are precisely the values which satisfy the property.
35
IMO Team Selection Test 3, June 2021
Problems
2. Let ABC be a right angled triangle with ∠C = 90◦ and let D be the foot of
the altitude from C. Let E be the centroid of triangle ACD and let F be
the centroid of triangle BCD. Let P be the point satisfying ∠CEP = 90◦
and |CP | = |AP |, and let Q be the point satisfying ∠CF Q = 90◦ and
|CQ| = |BQ|. Show that P Q passes through the centroid of triangle ABC.
f x + yf (x + y) = y 2 + f (x)f (y)
for all x, y ∈ R.
4. Let p > 10 be a prime number. Show that there exist positive integers m
and n with m + n < p for which p is a divisor of 5m 7n − 1.
36
Solutions
1. First suppose that n is odd. Then we must have V ≥ 1, as the difference
must be odd. We show that V = 1 is always possible. Colour the vertical
domino tiles in the odd numbered columns red and the vertical domino
tiles in the even numbered columns blue. As in every row, every horizontal
domino tile covers a square in an even numbered column and one in an odd
numbered column, every row contains one more square covered by a red
domino tile than squares covered by a blue domino tile. Now colour the
horizontal domino tile in each row alternatingly blue and red (starting with
blue). If the number of horizontal domino tiles is even, then at the end,
the number of red squares will be one more than that of blue squares; if
the number of horizontal domino tiles is odd, then at the end, the number
of blue squares will be one more than that of red squares. The difference
will therefore always be equal to 1.
Now suppose that n ≡ 2 mod 4. Then we have V ≥ 2 if Jetze places
every domino tile horizontally; then every row contains an odd number of
horizontal domino tiles. We show that V = 2 is always possible. Use the
same strategy as in the odd case. After colouring the vertical domino tiles,
the numbers of red and blue squares are equal. If we alternatingly colour
the horizontal domino tiles in each row blue and red again, we see that in
the end, in every row the difference between the number of red and blue
squares is 0 or 2.
Finally, suppose that n ≡ 0 mod 4. We show that V = 0 is always possible.
Number the rows from top to bottom from 1 up to m, and let bi be the
number of vertical domino tiles of which the top square is in row i. By
induction on i, we easily show that bi is even, using the fact that a horizontal
domino tile always covers an even number of squares in a row. We now
colour the vertical domino tiles in rows i and i + 1 as follows: if bi ≡ 0
mod 4, we colour half of them red, and the other half blue, and if bi ≡ 2
mod 4 we colour two more domino tiles red than we colour blue if i is even,
and we colour two more domino tiles blue than we colour red if i is odd.
We show that we can now colour the horizontal domino tiles in each row k
in such a way that every row has the same number of red and blue squares.
If bk−1 ≡ bk ≡ 0 mod 4, then vertical domino tiles in row k cover the same
number of red squares as blue squares. Moreover, the number of horizontal
domino tiles in row k is even, so we can simply colour half of them red and
half of them blue. If bk−1 ≡ bk ≡ 2 mod 4, then vertical domino tiles in
row k again cover the same number of red squares as blue squares, since of
k − 1 and k, one is odd and one is even. Again, the number of horizontal
domino tiles in row k is even, so we can again simply colour half of them
red and half of them blue. If bk−1 6≡ bk mod 4, then the difference in the
37
number of squares covered by red vertical domino tiles and blue vertical
ones is equal to 2. The number of horizontal domino tiles is odd, so we can
colour those in such a way that in the end, the number of red and blue
squares are equal.
Hence the minimal values for V are: V = 1 if n is odd, V = 2 if n ≡ 2
mod 4, and V = 0 if n ≡ 0 mod 4.
E
Z F
Q
A B
D
2. Let M , N , R, S be the midpoints of line segments BC, CA, BD, AD. Let
Z be the centroid of 4ABC. Quadrilateral QF M C is cyclic as ∠QF C =
90◦ = ∠QM C. Note that therefore CQ is a diameter of the circumcircle of
QF M C. Analogously, we see that P N EC is a cyclic quadrilateral, with
diameter CP .
We show that Z also lies on the circumcircles of these cyclic quadrilaterals.
The similarity transforming triangle BCA into triangle BDC transforms
triangle CZM into DF R, as C is mapped to D, the centroid Z is mapped
to the centroid F , and the midpoint M of BC is mapped to the midpoint of
BD, which is R. So 4CZM ∼ 4DF R, in particular ∠CZM = ∠DF R =
∠CF M (opposite angles). Therefore Z lies on the circumcircle of the cyclic
quadrilateral QF M C. Analogously, we have ∠CZN = ∠DES = ∠CEN ,
from which follows that Z lies on the circumcircle of the cyclic quadrilateral
P N EC.
We can now show that Z lies on P Q. As CQ is a diameter of the circle
through Q, F , M , C, Z, we have ∠QZC = 90◦ . As CP is a diameter of
the circle through P , N , E, Z, C, we also have ∠CZP = 90◦ . Hence P , Z,
and Q are collinear.
38
3. Note that the function f (x) = 0 for all x does not satisfy the condition.
Hence there is some a with f (a) 6= 0. Substituting x = a and y = 0, we
get f (a) =f (a)f (0), so f (0) = 1. Substituting x = 1 and y = −1, we get
f 1 − f (0) = 1 + f (1)f (−1). As f (0) = 1, this reads 1 = 1 + f (1)f (−1),
so f (1) = 0 or f (−1) = 0. We consider these two cases separately.
First suppose that f (1) = 0. Substituting x = t and y = 1 − t, and x = 1 − t
and y = t then gives
or equivalently
f (t)2 + (2t − 2)f (t) + (1 − t)2 = 0,
2
or equivalently, f (t) − (1 − t) = 0. Therefore f (t) = 1 − t for all t.
We verify this function in the original equation: the left hand side is
1 − (x + y(1 − x − y)) = 1 − (x + y − xy − y 2 ) = 1 − x − y + xy + y 2 , and
the right hand side is y 2 + (1 − x)(1 − y) = y 2 + 1 − x − y + xy, which is
equal to the left hand side. Therefore the function f (x) = 1 − x for all x is
a solution to the original equation.
Now suppose that f (−1) = 0. Substituting x = t and y = −1 − t, and
x = −1 − t and y = t, then gives
or equivalently
f (t)2 − (2t + 2)f (t) + (t + 1)2 = 0,
2
or equivalently, f (t) − (t + 1) = 0. We deduce that f (t) = t + 1 for all t.
39
We verify this function satisfies the original equation. The left hand side
is x + y(x + y + 1) + 1 = x + xy + y 2 + y + 1 and the right hand side is
y 2 + (x + 1)(y + 1) = y 2 + xy + x + y + 1, which is equal to the left hand side.
Hence the function f (x) = x + 1 for all x satisfies the original equation.
We conclude that there are exactly two solutions of the original equation:
f (x) = 1 − x for all x, and f (x) = x + 1 for all x.
4. By Fermat’s Little Theorem, we have ap−1 ≡ 1 mod p for all a such that
p - a. As p > 10, p is odd, so p − 1 is even. We have
p−1 p−1
a 2 − 1 a 2 + 1 = ap−1 − 1 ≡ 0 mod p.
p−1 p−1
So p | (a 2 − 1)(a 2 + 1), so p is a divisor of at least one of the factors.
p−1
Hence a 2 is congruent to 1 or −1 modulo p.
We apply this to a = 5 and a = 7. Note that p > 10, so p 6= 5, 7. If
p−1 p−1
5 2 ≡ 1 mod p and 7 2 ≡ 1 mod p, we choose m = n = p−1 2 , which
p−1 p−1
satisfy the condition. The same holds if 5 2 ≡ −1 mod p and 7 2 ≡ −1
mod p.
The remaining case is that one is congruent to 1 and the other is congruent
p−1 p−1
to −1. Assume that 5 2 ≡ 1 mod p and 7 2 ≡ −1 mod p. The case in
which it is the other way around is analogous.
If there exists an n such that 0 < n < p−1 n
2 and 7 ≡ 1 mod p, then we
p−1
choose this n and m = 2 , which satisfy the condition. If not, then no
n such that p−1 n
2 < n < p − 1 and 7 ≡ 1 mod p can exist either, since
otherwise 7 p−1−n
≡ 7 (7 ) ≡ 1 mod p, while 0 < p − 1 − n < p−1
p−1 n −1
2 ,
which is a contradiction. Moreover, no i, j such that 1 ≤ i < j ≤ p − 1
and 7i ≡ 7j mod p can exist, since otherwise 7j−i ≡ 1 mod p with 1 ≤
j − i < p − 1. Hence 7i assumes distinct values for 1 ≤ i ≤ p − 1 modulo
p, which are all non-zero. So 7i assumes all non-zero values modulo p for
1 ≤ i ≤ p − 1.
In particular, there exists an n such that 7n ≡ 5−1 mod p. Then n ≤ p − 2,
since 7p−1 ≡ 1 6≡ 5−1 mod p. Choosing this n and m = 1 gives 7n · 5m ≡
5−1 · 5 ≡ 1 mod p.
Therefore it is always possible to find m and n satisfying the conditions.
We thank our sponsors
Sectorgelden
Outreach Bèta
59th Dutch Mathematical Olympiad 2020
and the team selection for IMO 2021 Russia
MY
CMY