New Zealand Mathematical Olympiad Committee

Camp Selection Problems 2010

Due: 27th August 2010

Junior division
J1. We number both the rows and the columns of an 8 × 8 chessboard with the numbers 1 to
8. Some grains of rice are placed on each square, in such a way that the number of grains
on each square is equal to the product of the row and column numbers of the square. How
many grains of rice are there on the entire chessboard?
Solution: 1296.
In the first row we have successively 1 × 1, 1 × 2, 1 × 3, 2 × (1 + 2 + 3 + · · · + 8) . . . , 1 × 8
grains of rice, for a total of 1 × (1 + 2 + 3 + · · · + 8) grains. Similarly, in the second row
we have 2 × 1, 2 × 2, 2 × 3, . . . , 2 × 8 grains of rice, for a total of 2 × (1 + 2 + 3 + · · · + 8)
grains. In the third row there is a total of 3 × (1 + 2 + 3 + · · · + 8) grains, and so on, up
until the eighth row, which has 8 × (1 + 2 + 3 + · · · + 8) grains. Adding these up, we see
that we have a total of

(1 + 2 + 3 + · · · + 8)2 = 362 = 1296

grains of rice.

J2. AB is a chord of length 6 in a circle of radius 5 and centre O. A square is inscribed in the
sector OAB with two vertices on the circumference and two sides parallel to AB. Find
the area of the square.
900
Solution: 109
.
Label points as in Figure 1, and let the square have sidelength 2a. Triangle ODZ is similar
to the 3-4-5 triangle OEB, so OD has length 4a/3. Now OF = OD + DF = 4a/3 + 2a =
10a/3, and applying the Theorem of Pythagoras to OF Y we get OF 2 = 25−a2 = 100a2 /9.
Solving we get a2 = (9 · 25)/109, so the area is 4a2 = 900/109.

J3. Find all positive integers n such that n5 + n + 1 is prime.

Solution: The only solution is n = 1, for which 15 + 1 + 1 = 3.
Let f (x) = x5 + x + 1. Working out the first few values we have

f (1) = 3 = 3 × 1, f (5) = 3131 = 31 × 101,

f (2) = 35 = 7 × 5, f (6) = 7783 = 43 × 181,
f (3) = 247 = 13 × 19, f (7) = 16815 = 57 × 295,
f (4) = 1029 = 21 × 49, f (8) = 32777 = 73 × 449.

1
 B

Z Y
a

O D E F

a
X
W

Figure 1: Diagram for problem J2.

Noticing that the left-hand factors above all have the form n2 + n + 1, we are led to
suspect that x2 + x + 1 is a factor of x5 + x + 1 (in all but two cases above both factors
are prime, so we don’t have to make a clever choice of factorisation of f (n) to see these
factors). This may be confirmed via long division, which gives

x5 + x + 1 = (x2 + x + 1)(x3 − x2 + 1).

For x > 1 we have 1 < x2 + x + 1 < x5 + x + 1, so f (n) is not a prime number for n ≥ 2.
Since f (1) = 3 is prime this is the only solution.
Remark: Prior to long division it is possible to check whether x2 + x + 1 is a factor using
a cube root of unity ω 6= 1. The root ω satisfies ω 3 = 1 and ω 2 + ω + 1 = 0 (because
ω 3 − 1 = (ω − 1)(ω 2 + ω + 1)), so ω 5 = ω 2 and we see immediately that f (ω) = 0.

J4. Find all positive integer solutions (a, b) to the equation

1 1 n 1
+ + =
a b lcm(a, b) gcd(a, b)
for

(i) n = 2007;
(ii) n = 2010.

Note: “lcm(a, b)” means the least common multiple of a and b, and “gcd(a, b)” means the
greatest common divisor of a and b. For example, lcm(18, 30) = 90, and gcd(18, 30) = 6.
Solution: There are no solutions for n = 2010, and the solutions are

{(2d, 2009d), (5d, 503d), (503d, 5d), (2009d, 2d) : d ≥ 1}

for n = 2007.

2
 z
B C
ψ φ
φ
y
x θ x

E
θ
θ
A φ ψ D
y

Figure 2: Diagram for problem J5.

Let d = gcd(a, b). Then there are positive integers x, y which are relatively prime such
that a = dx, b = dy. Using this in the original equation we find x + y + n = xy, that is,
n + 1 = (x − 1)(y − 1). Since x and y are relatively prime at least one of them is odd, so
at least one of x − 1, y − 1 is even, and so n + 1 must be even too. So there is no solution
for n = 2010.
For n = 2007 this gives (x − 1)(y − 1) = 2008 = 23 · 251, so the possible factorisations
are {x − 1, y − 1} = {1, 2008}, {2, 1004}, {4, 502} and {8, 251}. These give {2, 2009},
{3, 1005}, {5, 503}, and {9, 252} as the possibilities for {x, y}, and of these only the
pairs {2, 2009} and {5, 503} are relatively prime. This gives the list of solutions given
above.

J5. The diagonals of quadrilateral ABCD intersect in point E. Given that |AB| = |CE|,
|BE| = |AD|, and ∠AED = ∠BAD, determine the ratio |BC|/|AD|.
√
1+ 5
Solution: 2
.
See Figure 2. We have ∠BEC = ∠AED = ∠BAD = θ, so triangles CEB and BAD
are congruent (side-angle-side). Therefore ∠BCA = ∠ABD = φ, ∠CBD = ∠BDA = ψ,
and |BC| = |BD|. The second equality implies that BC and AD are parallel, and from
this or from θ + ψ + φ = π we have ∠EAD = φ. It follows that triangles AED and CEB
are similar.
Let λ = |BC|/|AD|. Then |BC| = λ|AD| = λ|BE| = λ2 |ED|, and |BC| = |BD| =
2
|BE| + |ED| = (λ + 1)|ED|. Hence
√ λ satisfies the quadratic equation λ = λ + 1, and is
therefore the golden ratio (1 + 5)/2.

J6. At a strange party, each person knew exactly 22 others.

For any pair of people X and Y who knew one another, there was no other person at the
party that they both knew.
For any pair of people X and Y who did not know each other, there were exactly six other
people that they both knew.
How many people were at the party?

3
 x0 ≤ 4, x1 ≤ 4 x0 ≤ 4, x1 > 4 x0 > 4, x1 ≤ 4 x0 > 4, x1 > 4
16 4x1 16 4x1
x2 x0 x0 x0
64 16 16
x16
0

x3 x0 x1 x0 x1
max x0
, x161
16 16 4x0 4x0
x4 x1 x1 x1 x1
x5 x0 x0 x0 x0
x6 x1 x1 x1 x1

Table 1: Table of values for Problem S1.

Solution: 100.
Suppose there are n people at the party, p1 , p2 , . . . , pn . Fix i and count the number of
distinct ordered pairs (j, k) such that pi knows pj and pj knows pk . There are 22 pairs
where k = i. Suppose that k 6= i. Then pk is one of the n − 22 − 1 people that pi doesn’t
know, and there are 6 people pj such that we must include (j, k) in our count. So there
are 22 + 6(n − 23) such pairs.
On the other hand, there must be 222 = 484 such pairs altogether, because each person
knows 22 others. Hence 484 = 22 + 6(n − 23), and we solve to find n = 100.

Senior division
S1. For any two positive real numbers x0 > 0, x1 > 0, a sequence of real numbers is defined
recursively by

4 max{xn , 4}
xn+1 = for n ≥ 1.
xn−1
Find x2010 .
Note: “max{x, y}” means the maximum of x and y — that is, whichever of the two
numbers x and y is the larger. For example, max{2, 3} = 3.
Solution: All solutions are periodic with period five, as is easily checked by direct
computation (see Table 1; note that we must calculate as far as x6 to establish the
periodicity). Hence x2010 = x0 .

S2. In a convex pentagon ABCDE the areas of the triangles ABC, ABD, ACD and ADE
are all equal to the same value x. What is the area of the triangle BCE?
Solution: 2x.
We first observe that ABC and ABD have the same base and area, so they must have the
same altitude. Hence CD is parallel to AB. We next observe that ABD and ACD have
the same altitude and area, so they must have the same base; therefore |CD| = |AB|,
and ABCD is a parallelogram.

4
 Now observe that ABD and ADE have the same base AD, and the same area, so they
too must have the same altitude. In addition, E must be on the opposite side of AD to
B, since ABCDE is convex. Considering now the triangles BCD and BCE, we see that
they have the same base BC, but the altitude of BCE is twice that of BCD. It follows
that it has twice the area, as claimed.

S3. Let p be a prime number. Find all pairs (x, y) of positive integers such that

x3 + y 3 − 3xy = p − 1.

Solution: Since we want to use the fact that p is a prime, we add 1 to both sides and
try to factorise the left-hand side. Calculating x3 + y 3 − 3xy + 1 for lots of pairs (x, y)
gives you the idea that the expression is divisible by x + y + 1. It turns out that

x3 + y 3 − 3xy + 1 = (x + y + 1)(x2 + y 2 − xy − x − y + 1).

This needs to be equal to a prime number. Since x + y + 1 > 1, we must have x2 + y 2 −

xy − x − y + 1 = 1. As x2 + y 2 ≥ 2xy, we find

1 = x2 + y 2 − xy − x − y + 1 ≥ xy − x − y + 1 = (x − 1)(y − 1).

This can only be true if x = 1 or y = 1 or x = y = 2. Suppose x = 1, then 2 + y 3 − 3y = p,

but now the left-hand side is even for any y, so we must have p = 2. However, that
means y 3 = 3y, which does not have any integer solutions. Similarly y = 1 leads to a
contradiction. The only possible solution is x = y = 2 and this is a solution if and only
if p = 5.

S4. A line drawn from the vertex A of an equilateral triangle ABC meets the side BC at D
and the circumcircle at P . Show that
1 1 1
= + .
|P D| |P B| |P C|

B A

C
D

Figure 3: Diagram for problem S2.

5
 A

B D
C

Figure 4: Diagram for problem S4.

Solution: See Figure 4. Because ∠P AC = ∠P BC, ∠AP C = ∠ABC = 60◦ and

∠BP A = ∠BCA = 60◦ the triangles AP C and BP D are similar. Thus
|P A| |P C|
= ,
|P B| |P D|
so
|P A| · |P D| = |P B| · |P C|. (1)

Since ABP C is a cyclic quadrilateral, |P A| · |BC| = |P B| · |AC| + |P C| · |AB|. But the

triangle ABC is equilateral, so it follows that

|P A| = |P B| + |P C|. (2)

From equations (1) and (2), it follows that

|P B| · |P C| = |P D| · (|P B| + |P C|),

and we divide by the product |P B| · |P C| · |P D| to get the desired equality

1 1 1
= + .
|P D| |P B| |P C|

S5. Determine the values of the positive integer n for which

r
9n − 1
A=
n+7
is rational.

6
 Solution: n = 1 or 11.
It is enough to determine for which n there exist positive integers a, b with gcd(a, b) = 1
such that
9n − 1 a2
= 2.
n+7 b
From this relation we get

7a2 + b2 7(a2 − 9b2 ) + 64b2 64b2

n= = = −7 + .
9b2 − a2 9b2 − a2 9b2 − a2
Since gcd(a, b) = 1 it follows that gcd(a2 , b2 ) = 1 and gcd(9b2 − a2 , b2 ) = 1, so n is an
integer if and only if 9b2 − a2 is a divisor of 64. Moreover 9b2 − a2 must be positive, in
order for n to be positive.
Now 9b2 − a2 = (3b + a)(3b − a). If a = b = 1 then 9b2 − a2 = 8; otherwise, 9b2 − a2 ≥
3b + a ≥ 5, so 9b2 − a2 ≥ 8. So the possible values for 9b2 − a2 are 8, 16, 32, 64. The
factors 3b + a and 3b − a differ by a multiple of 2, sum to a multiple of 6, and satisfy
3b + a > 3b − a, so the possibilities for (3b + a, 3b − a) are (4, 2), (16, 2), and (8, 4).
The corresponding possibilities for (a, b) are (1, 1), (7, 3) and (2, 2), and we discard (2, 2)
because gcd(2, 2) = 2 6= 1. Substituting the remaining pairs into n = (7a2 + b2 )/(9b2 − a2 )
we get n = 1 or n = 11.

S6. Suppose a1 , a2 , . . . , a8 are eight distinct integers from {1, 2, . . . , 16, 17}. Show that there
is an integer k > 0 such that there are at least three different (not necessarily disjoint)
pairs (i, j) such that ai − aj = k.
Also find a set of seven distinct integers from {1, 2, . . . , 16, 17} such that there is no integer
k > 0 with that property.
Solution: Without loss of generality, assume that a1 < a2 < . . . < a8 . Consider the
numbers s1 = a2 − a1 , s2 = a3 − a2 , . . . , s7 = a8 − a7 , and s8 = a3 − a1 , s9 = a4 − a2 , . . . ,
s13 = a8 − a6 . We have
13
X
si = (a8 − a1 ) + (a8 + a7 − a2 − a1 ) = 2a8 + a7 − a2 − 2a1 ≤ 2 · 17 + 16 − 2 − 2 · 1 = 46.
i=1

On the other hand, suppose that for any k > 0 there are at most two values of i for which
si = k, then we have
X13
si ≥ 2 · (1 + 2 + · · · + 6) + 7 = 49.
i=1

So this is impossible, and hence there must be a k > 0 such that for at least three values
of i we have si = k.
For the second part, take {1, 2, 3, 5, 8, 12, 17}.

July 2010
www.mathsolympiad.org.nz