2024 ARML Local: Sponsored by Star League

Problems and Solutions

Team Round (45 minutes)
T-1 The sum of three prime numbers is 134. Compute the greatest possible value for the difference
between the greatest and the least of these three numbers.

Answer: 125

Solution: Because the sum of the three numbers is even, the three numbers cannot all be
odd. Because 2 is the only even prime, it follows that exactly one of the numbers is 2, and
this is the least of the three numbers. Let p and q be the other two primes, with p < q. Then
p + q = 134 − 2 = 132, and the greatest possible value of q − 2 will be achieved when p is as small
as possible. Note that 1 is not prime, so p ≥ 3. If p = 3, then q = 129 = 3 · 43, which is not
prime. If p = 5, then q = 127, which is prime. It can readily be verified that none of 2, 3, 5, 7, or
11 is a divisor of 127. If 127 were not prime, then it could be written in the form d · 127
d for some
integer d such that 1 < d < 127. But because d = 2, 3, 5, 7, 11, it follows that the least prime
divisor of both d and 127 127
d would be at least 13, and thus 127 = d · d ≥ 13 · 13 = 169, which is
impossible. Hence 127 is the greatest possible value of q, and the answer is 127 − 2 = 125.

T-2 In rectangle RECT , RE = 4 and EC = 8. Point

√ X is chosen at random from the interior of
RECT . Compute the probability that EX > 4 2.

6−π 3 π
Answer: or −
8 4 8

Solution: Consider
√ the diagram
√ below, which shows the midpoint M of RT and a segment
2 2
EM of length 4 + 4 = 4 2 inside RECT .

√
Every point X inside the sector centered at E satisfies EX < 4 2. This area is the sum of the
areas of a 45◦ −45◦ −90◦ triangle and a 45◦ sector of a circle. Thus the desired probability is
√
32 − 12 · 4 · 4 − 18 π(4 2)2
,
32
or 24−4π
32 = 6−π
8 .

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
T-3 Given that the numbers 11, 18, 25, . . . , m are in arithmetic progression and that their sum is
2024, compute m.

Answer: 165

Solution: Assume m is the nth term of the given sequence. It follows that 18 = 11 + 7 · 1,
25 = 11 + 7 · 2, . . . , m = 11 + 7(n − 1) and
n(m + 11) n(7n + 15)
11 + 18 + 25 + · · · + m = = = 2024.
2 2
Therefore n(7n + 15) = 4048 = 24 · 11 · 23, where n is an integer. If n ≤ 22, then 7n + 15 ≤ 169
and n(7n + 15) ≤ 3718 < 4048. If n ≥ 44, then 7n + 15 ≥ 323 and n(7n + 15) > 4048. Because
n divides 4048, the only value for n left to be checked is 23 and it works, and so the answer is
m = 11 + 7 · 22 = 165.

Alternate Solution: Notice that

k(k + 1)
11 + 18 + 25 + · · · + (11 + 7k) = 11(k + 1) + 7 · = 2024.
2
This implies 7k 2 + 29k − 4026 = 0, which implies (7k + 183)(k − 22) = 0. The only integer
solution to this equation is k = 22, and so m = 11 + 7 · 22 = 165.

T-4 Suppose that A, R, M , and L are distinct positive digits. Compute the least possible positive
value of
1 1
− .
AR ML

2
Answer:
6399

Solution: For fixed M L − AR = k > 0, observe that

1 1 k
− =
AR M L (M L − k) · M L
is minimized when M L is maximized. Consider the following casework on k.
• Note that k = 1 is impossible, because neighboring two-digit integers either share a tens
digit or contain a 0 as a digit.
• For k = 2, the maximum of M L is 81 (with AR = 79). Indeed, if M L ∈ {82, . . . , 89, 92, . . . , 99}
then A = M , and if M L = 91 then R = M = 9. For this M L, the expression evaluates to
2 2
79·81 = 6399 .
• For k = 3, the maximum of M L is 82 (with AR = 79). Indeed, if M L ∈ {83, . . . , 89, 93, . . . , 99}
then A = M , if M L = 91 then A = R = 8, and if M L = 92 then R = M = 9. For this
3 2
M L, the expression evaluates to 79·82 > 79·81 .
• For k ≥ 4, the expression is
k 4 2
> 2
> .
(M L − k) · M L 100 6399
2
Thus the least possible positive value of the expression is 6399 .

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
T-5 Let HEP T AGON be an equiangular octagon with all of its vertices on lattice points and four
sides parallel to the coordinate axes. Given that every side contains exactly six lattice points
(including the endpoints), compute the number of lattice points in the interior of HEP T AGON .

Answer: 156

Solution: Each interior angle of HEP T AGON measures 6 · 180◦ ÷ 8 = 135◦ . Thus it follows
that four sides are parallel to the axes and have √ length 5, while the other four sides are at a
45◦ angle to the horizontal and have length 5 2. Therefore HEP T AGON can be obtained by
removing four right triangles of area 12 · 5 · 5 from the corners of a 15 × 15 square, for a total area
of 225 − 50 = 175. There are 40 lattice points along the boundary, hence by Pick’s Theorem
there are n lattice points in the interior where n + 40 2 − 1 = 175. Thus n = 156.

T-6 Compute the sum of all three-digit positive integers whose squares end with the digits 29.

Answer: 19800

Solution: Note that (50 ± N )2 ≡ N 2 (mod 100). Check 12 , 22 , 32 , . . . , 252 modulo 100 to see
that only 232 ≡ 29 (mod 100). Thus look for squares equivalent to ±23 or 50 ± 23 modulo 100.
Every  block of 100 consecutive positive integers contains 4 numbers  that end with the digits
29: (100 · A + 23)2 , (100 · A + 27)2 , (100 · A + 73)2 , (100 · A + 77)2 . The sum for each block is
400 · A + 200. The possible values for A are 1, 2, . . . , 9, so the desired sum is 400 · (1 + 2 + 3 +
· · · + 9) + 9 · 200 = 400 · 45 + 1800 = 19800.

T-7 The graphs of the line y =√3x + k and the parabola y 2 = 6x intersect twice. The two intersection
points are a distance of 2 10 apart. Compute k.

Answer: −4

Solution: For the intersection points (x, y), it follows that (3x + k)2 = 6x. This implies
9x2 + (6k − 6)x + k 2 = 0, which has the two solutions
 √ √
6 − 6k ± (6k − 6)2 − 36k 2 6 − 6k ± 36 − 72k 1 − k ± 1 − 2k
x= = = .
18 18 3
√ √ √
Therefore the two intersection points are at ( 1−k−3 1−2k , 1 − 1 − 2k) and ( 1−k+3 1−2k , 1 +
√
1 − 2k). Applying the Pythagorean Theorem to calculate the distance between these points,
√
2· 1−2k
√
it follows that the horizontal change is and the vertical change is 2 · 1 − 2k, so the
3 √
square of the sums is 409 (1 − 2k). Because the distance between these points is 2 · 10, solve
40
9 (1 − 2k) = 40 to find that k = −4.

T-8 Point P lies inside square ABCD so that AP , BP , and DP form a geometric progression in
that order. Given that DP = 2 · AP , compute AB
AP .

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 √
Answer: 5

Solution: Because DP = 2√· AP = r2 · AP where r is the common ratio of the geometric

√
sequence, it follows that r = 2 and so without loss of generality AP = 1, BP = 2, and
DP = 2. The problem statement requires finding AB AP . Because AP = 1, it suffices to find
AB, which will be denoted by k. Denote by m and n the distances from P to AB and AD,
respectively. The diagram below shows the relevant information.

Note that m2 + n2 = 1 by the Pythagorean Theorem. Also by the Pythagorean Theorem, it is

true that m2 + (k − n)2 = 2 and n2 + (k − m)2 = 4. These imply m2 + n2 − 2kn + k 2 = 2 →
k 2 − 2kn = 1 and n2 + m2 − 2km + m2 = 4 → k 2 − 2km = 3. Therefore (k 2 − 2km) − (k 2 − 2kn) =
3−1 → k(n−m) = 1. Also note that k(k −2n) = 1, which implies k −2n = n−m → k = 3n−m.
Substitute into (k − m)2 + n2 = 4 to obtain (3n − 2m)2 + n2 = 4 → 4m2 + 10n2 − 12mn = 4,
which implies 6n2 − 12mn = 0 → n2 − 2mn = 0 and so n = 2m.
Because k(n − m) = 1, it follows that (3n − m)(n − m) = 1 → 5m · m = 1 and so m = √15 .
√ √
5
√
Therefore k = 5m = √55 = 5, and the answer is AB AP = 1 = 5.

T-9 In triangle ABC, AB = AC = 10 and BC = 12. Point D lies on AB and point E lies on AC
so that AD = DE = EC = x for some positive real number x. Compute x.

250
Answer:
39

Solution: Let M be the foot of the perpendicular from A to BC and let N be the foot of
the perpendicular from D to AC.

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 Because triangles ABC and ADE are isosceles, it follows that M and N are the midpoints
of BC and AE, respectively. Thus CM = 6 and AN = 10−x 2 . Let θ = m∠CAM . Because
√ 4
AM = 10 − 6 = 8, it follows that cos θ = 5 . Note that cos ∠BAC = cos 2θ = 2 cos2 θ − 1 =
2 2
 2
2 · 45 − 1 = 25
7
. Thus
 10−x 
2 7
cos ∠DAN = = .
x 25

250
Solving the above equation yields x = 39
.

T-10 Compute the number of ordered triples of composite positive integers (a, b, c) such that abc =
27000.

Answer: 424

Solution: Note that 27000 = 23 · 33 · 53 , and the only possible non-composite values of
a, b, c are 1, 2, 3, and 5. Ignoring the composite condition, choosing a, b, and c is equivalent to
choosing how to distribute three factors
  of each of 2, 3, and 5 among the three numbers. The
factors of 2 can be distributed in 52 = 10 ways, in what is essentially a “sticks and stones”
problem: choosing 3 instances of factors of 2, and 2 dividers. The factors of 3 and 5 can also be
distributed in 10 ways each, for a total of 1000 ordered triples.
Now count the number of such triples in which a is non-composite. If a = 1, then there are
 43  2  
1 = 64 choices for b and c. If a = 2, 3, 5, then there are 41 · 31 = 48 choices, for a total of
64 + 3 · 48 = 208. Similarly, there are 208 triples in which b is non-composite, and 208 in which
c is non-composite.
Finally, count the number of triples in which a and b are both non-composite. In this case,
each of a and b can be independently chosen from {1, 2, 3, 5}, and c is then determined. Thus
there are 16 such triples, and 48 total in which two of the three numbers are non-composite.
There are no triples in which all three are non-composite.
Each of the triples with two non-composite numbers has been counted twice, so the actual
number of triples with one or more numbers non-composite is 3 · 208 − 48 = 576. Thus, the
number of triples in which all three of a, b, c are composite is 1000 − 576 = 424.

T-11 Let S(n) denote the sum of the digits of n. Compute the least three-digit prime p for which
S(p) = S(p2 ).

Answer: 199

Solution: Because S(n) ≡ n (mod 9), it follows that p ≡ p2 (mod 9). For p(p − 1) to be a
multiple of 9, it must be that p − 1 is a multiple of 9 and also of 2. Thus p = 18k + 1 for some
integer k. Listing such three-digit primes, it follows that p is one of 109, 127, 163, 181, 199, 271, ....
For the first four, S(p) = 10 but S(p2 ) exceeds 10. However S(199) = 19 and because (200−1)2 =
40000 − 400 + 1 = 39601, it follows that S(1992 ) = S(199) = 19. The answer is 199.

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
T-12 Let ABC be an equilateral triangle with side length 1. Say that a point P on the same plane
as ABC is judicious if there exists a right triangle with side lengths P A, P B, and P C. The set
of all judicious points is the union of several curves with total combined length L. Compute L.

Answer: 6π
←→
Solution: Let D be the reflection of A across BC. The key claim is that for any point P in
the plane,
P A2 + P D2 = P B 2 + P C 2 + 1.
√
3
This can be proven using coordinates. Let A = (0, 2 ), B = (− 12 , 0), and C = ( 12 , 0) so that
√
3
D = (0, − 2 ). Then if P = (x, y),
√3
P A2 = x2 + y 2 − 3y +
4
√ 3
P D2 = x2 + y 2 + 3y +
4
2 2 2 1
PB = x + y − x +
4
1
P C 2 = x2 + y 2 + x +
4
so the claim is true.
It follows that P A2 = P B 2 + P C 2 if and only if P lies on the circle of radius 1 centered at D.
By symmetry, the locus of judicious points is three circles of radius 1, for a total length of 6π.

T-13 Compute the least four-digit positive integer that divides the number
N = 9 · 10 · 23 · 24 + 196 · 23 + 169 · 24 + 10.

Answer: 1022

Solution: Let a = 23, b = 24, c = 10. Write

N = 9 · 10 · 23 · 24 + 196 · 23 + 169 · 24 + 10
= 9 · a · b · c + 142 · a + 132 · b + 12 · c
= 9abc + a(b − c)2 + b(c − a)2 + c(a − b)2
= ab2 + ac2 + bc2 + ba2 + ca2 + cb2 + 3abc
= (a + b + c)(bc + ca + ab)
= 57 · 1022.
If there is an integer d between 1000 and 1021 which divides N , then
57 · 1022 N 57 · 1022
57 < ≤ ≤ < 59,
1021 d 1000
so Nd = 58. But it is easy to check that 58 does not divide N , so 1022 is the least positive
four-digit divisor of N .

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
T-14 The measures of the angles of ABC are in the ratio m∠A : m∠B : m∠C = 1 : 2 : 4. Compute
tan2 A tan2 B tan2 C
+ + .
tan2 B tan2 C tan2 A

Answer: 83

Solution: Because the measures of the angles are in the ratio 1 : 2 : 4, it follows that
m∠A = π7 , m∠B = 2π 4π π
7 , and m∠C = 7 . Consider a = 7 . By de Moivre’s Theorem, it follows
that cos(7a) + i sin(7a) = (cos a + i sin a)7 = −1 + 0i. By the Binomial Theorem, and considering
only the imaginary parts of the expansion, it follows that
7i · cos6 a · sin a − 35i · cos4 a · sin3 a + 21i · cos2 a · sin5 a − i · sin7 a = 0,
which implies i · tan a(7 − 35 tan2 a + 21 tan4 a − tan6 a) = 0. Let x1 = tan2 a; then the cubic
polynomial x3 − 21x2 + 35x − 7 has x1 as a root.
A similar argument can be made for 2a and 4a, and so the three roots of x3 − 21x2 + 35x − 7
are x1 = tan2 a, x2 = tan2 (2a), and x3 = tan2 (4a).
Because tan(8a) = tan a, it follows that the desired sum is cyclic. Consider the double-angle
 2 2 x)2
2 tan x
formula tan(2x) = 1−tan 2x ; rearranging and squaring obtains tan x
tan(2x) = (1−tan
4 . Thus it
3
(1−xk )2
follows that the desired sum is given by 4 . This can be found using Vieta’s formulas.
k=1
Notice that (1 − x1 )2 + (1 − x2 )2 + (1 − x3 )2 = (x21 + x22 + x23 ) − 2(x1 + x2 + x3 ) + 3. By Vieta’s
formulas, it follows that x21 +x22 +x23 = (x1 +x2 +x3 )2 −2(x1 x2 +x2 x3 +x3 x1 ) = 212 −2·35 = 371,
3
(1−xk )2
and so the desired sum is 4 = 371−2·21+3
4 = 332
4 = 83.
k=1

T-15 The three circles and three squares on the edges of the triangle below are filled in with distinct
two-digit numbers with nonzero digits such that reading in order along each edge of the triangle
– from any circle to the next circle – the following hold:
• the three numbers form a geometric progression;
• the tens digits of the three numbers form an arithmetic progression; and
• the units digits of the three numbers form a geometric progression.

Compute the sum of the numbers in the three squares.

Answer: 191

Solution: The first step is to characterize all three-term geometric progressions of positive
integers. The key characterization is that they are all of the form (ka2 , kab, kb2 ) for positive
integers a, b, k with k squarefree. In particular, the two endpoints have the same squarefree part.

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
In addition, the numbers in the circles must be {ka2 , kb2 , kc2 } and the numbers in the squares
must be {kbc, kca, kab} for positive integers a, b, c, k with k squarefree and a, b, c distinct.
To prove the characterization, let the geometric progression be (x, y, z). Then xy = yz , so
xz = y 2 . Write x = ka2 for positive integers a, k with k squarefree so that ka2 z = y 2 . Then
a divides y, so let w = ay to rewrite the equation as kz = w2 . Then k divides w because k is
√
squarefree, so let b = wk to rewrite the equation as z = kb2 . Then y = xz = kab as desired.
This makes it easy to find all possible three-term geometric progressions of digits. The
endpoints must either be the same or one of {1, 4}, {1, 9}, {4, 9}, or {2, 8}. But in the triangle
setup, if two circles have the same units digit, then the square on that edge has the same unit
digit. Then the numbers along that edge form an arithmetic progression as well as a geometric
progression, which is impossible if they are distinct. Thus no two circles can have the same units
digit, so the units digits in the circles must be {1, 4, 9}.
It follows that k must be odd and not 5 – otherwise, no circle could have units digit 1. Also,
if k > 11, then k · 32 > 99, so there are not enough two-digit numbers with squarefree part k.
Furthermore, if 7 ≤ k < 10, then k · 12 < 10 while k · 42 > 99, so again there are not enough
two-digit numbers with squarefree part k. Thus k ∈ {1, 3, 11}.
• The two-digit numbers with squarefree part 11 are {11, 44, 99}. This corresponds to the
three edges being (11, 22, 44), (11, 44, 99), and (44, 66, 99). But the tens digits do not form
an arithmetic progression in any of these, contradiction. So there are no solutions when
k = 11.
• The two-digit numbers with squarefree part 3 are {12, 27, 48, 75}. None of these have units
digit in {1, 4, 9}, so there are no solutions when k = 3.
• The two-digit numbers with squarefree part 1 are {16, 25, 36, 49, 64, 81}. Of these, {49, 64, 81}
have appropriate units digits. This corresponds to the three edges being (49, 56, 64),
(49, 63, 81), and (64, 72, 81). Each of these edges satisfies the criterion on tens digits, so
this is a solution.
Thus the only solution is when the circles are {49, 64, 81} and the squares are {56, 63, 72}, so
the answer is 56 + 63 + 72 = 191.

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 Individual Round (10 minutes per pair)
I-1 The original price of an item is N dollars. A store carrying this item has a sale in which every
item is priced at a 10% discount. Gerry incorrectly calculates the item’s sale price by subtracting
10 dollars from the item’s original price. Gerry’s price is 1 dollar more than the discounted price.
Compute N .

Answer: 110

Solution: The discounted price is N − 0.1N = 0.9N . The given information implies that
N − 10 = 0.9N + 1, hence N = 110.

I-2 Suppose that DBZ is drawn with ZB = 3, DB = 4, and DZ = 5, and suppose that point A
←→
is drawn on the opposite side of DZ as B with AD = 12 and AZ = 13, as shown in the diagram.
Compute the area of ABZ.

84 4
Answer: or 16 or 16.8
5 5

Solution: Note that DBZ and ADZ are right triangles by the converse of the Pythagorean
Theorem. By elementary calculations, [DBZ] = 12 · 3 · 4 = 6 and [ADZ] = 12 · 5 · 12 = 30 and
thus [ADBZ] = 6 + 30 = 36. Note also that ∠ZDB has a cosine of 45 , so by the prop-
erty of addition on sines of angles it follows that sin(∠ADB) = 1 · 45 + 0 · 35 = 45 . Thus
84
[ADB] = 12 · 4 · 12 · sin(∠ADB) = 96 96
5 . Therefore the answer is [ABZ] = 36 − 5 = 5 .

I-3 A standard, fair six-sided die is rolled three times, obtaining the three values a, b, and c.
Compute the probability that the polynomial ax2 + bx + c has two distinct real roots.

19
Answer:
108

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 Solution: A quadratic function ax2 + bx + c has two distinct real roots if its discriminant
2
b2− 4ac is positive, or b4 > ac. Given that ac ≥ 1, it is only necessary to consider when b ≥ 3.
When b is 3, 4, 5, or 6, the product ac must be strictly less than 94 , 4, 254 , and 9, respectively.
There are 3, 5, 14, and 16 pairs of values for a and c, respectively, that satisfy the inequality, so
19
the answer is 3+5+14+16
63
= 108 .

I-4 Call a number digit-prime if it is not divisible by 2, 3, 5, or 7. Compute the number of three-digit
digit-prime numbers whose digits are chosen (possibly with repetition) from {2, 3, 5, 7}.

Answer: 21

Solution: By inspection, three-digit prime numbers cannot end in a 2 or 5; otherwise they

would be divisible by 2 or 5, respectively. Therefore consider only the 4 · 4 · 2 = 32 cases that
do not end in a 2 or 5. From least to greatest, these are: 223, 227, 233, 237, 253, 257, 273, 277,
323, 327, 333, 337, 353, 357, 373, 377, 523, 527, 533, 537, 553, 557, 573, 577, 723, 727, 733, 737,
753, 757, 773, 777.
There are four cases to consider.
If 3 divides X Y 3, then it follows that X + Y ≡ 0 (mod 3), which means that 273, 333, 573,
723, and 753 are multiples of 3.
If 3 divides X Y 7, then it follows that X + Y ≡ 2 (mod 3), which means that 237, 327, 357,
537, and 777 are multiples of 3.
If 7 divides X Y 3, then it follows that 100X + 10Y + 3 ≡ 2X + 3Y + 3 ≡ 0 (mod 7), which
implies 2X + 3Y ≡ 4 (mod 7). If X = 2, then Y = 7, but 273 had already been excluded. If
X = 3, then Y = 4, and 343 is not on the list anyway. If X = 5, then Y = 5, so it follows that
553 is a multiple of 7. If X = 7, then Y = 6, and 763 is not on the list anyway.
If 7 divides X Y 7, then it follows that X Y ≡ 0 (mod 7), and no new numbers are excluded by
this criterion.
There are therefore 32 − 11 = 21 digit-prime three-digit integers.

I-5 Sixty-four teams are assigned unique integer values between 1 and 64 inclusive and are placed
into a 6-round single elimination tournament where the team with the lesser value always defeats
a team with a greater value (e.g., 1 would defeat 2, 63 would defeat 64). Let Li be the maximum
value of a team that could possibly win a game in round i. Compute L1 + L2 + L3 + L4 + L5 + L6 .

Answer: 264

Solution: Because 64 cannot win, L1 ≤ 63. The winner of any bracket will be its least
element, so to maximize the winner, place the largest possible numbers in each bracket. Thus
the size 2k bracket is {64, 63, 62, . . . , 64 − 2k + 1} and so Lk = 65 − 2k . Thus the answer is
6
Lk = 63 + 61 + 57 + 49 + 33 + 1 = 264.
k=1

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
I-6 Suppose that regular octagon ABCDEF GH is drawn. Diagonals AD, BE, CF , DG, EH,
F A, GB, and HC intersect in a total of 16 points. These points are the vertices of two regular
octagons, the smaller of which has area 1. Compute the area of ABCDEF GH.
√
Answer: 3 + 2 2

Solution: Consider the diagram below.

Let ST = k, the side length of the smaller octagon. Denote by P and Q the intersections of CH
with BE and AF , respectively. Because P U T is an isosceles right triangle, it follows that P T =
√
2
√
2
√
2 k, and similarly, SQ = 2 k. Thus P Q = (1 + 2)k. Furthermore, P BAQ is a rectangle, so
√  2 √ 2 √
AB = (1 + 2)k. The area of ABCDEF GH is equal to AB ST = (1 + 2) =3 + 2 2.

I-7 Compute the number of ways to arrange in a line five red and five blue balloons so that there
are no three consecutive balloons of the same color.

Answer: 84

Solution: Suppose there are 2m balloons. Then to arrange balloons according to the condi-
tions of the problem is equivalent to partitioning m into blocks of 1 and 2 only. The number k of
ways to partition m = a1 + a2 + a3 + · · · + ak into k blocks with 1 ≤ ai ≤ 2 is equal to m−k
because m − k = (a1 − 1) + (a2 − 1) + · · · + (ak − 1) and it is necessary only to choose which of
the k numbers ai − 1 are equal to 1.
If the number of blocks of red balloons is the same as the number of blocks of blue balloons,

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 then the number of balloon arrangements is
2 2 2
5 4 3
2 + + = 52,
0 1 2

where the doubling occurs because either the red balloons or the blue balloons could go first.
If the number of blocks of colors of balloon differs by 1, then the number of balloon arrangements
is
5 4 4 3
2 + = 32,
0 1 1 2
where the doubling occurs because either color could have more blocks of balloons than the
other color.
Thus the total number of arrangements is 52 + 32 = 84.

(1 + x)4 81
I-8 The solutions to the equation 4
= are mapped on the complex plane. The four
(1 − x) 16
solutions are the vertices of convex quadrilateral ARM L. Compute the area of ARM L.

288 28
Answer: or 4
65 65

Solution: The given equation implies 16(1+x)4 = 81(1−x)4 , which implies 2+2x = k(3−3x)
where k ∈ {±1, ±i}. Solving the four resulting linear equations yields 2 + 2x = 3 − 3x → x = 15 ,
5
2 + 2x = 3x − 3 → x = 5, 2 + 2x = 3i − 3ix → x = 13 + 12 5 12
13 i, and 2 + 2x = 3ix − 3i → x = 13 − 13 i.
Thus ARM L is a kite with one diagonal running from 15 to 5 and the other diagonal running
5 288
from 13 − 12 5 12 1 1 12
13 i to 13 + 13 i. The area of ARM L is 2 · 2 · (5 − 5 ) · 13 = 65 .

I-9 The solution set for x3 −9x2 −36x = 0 is {0, 12, −3}. The solution set for x3 −42x2 +525x−2024 =
0 is {a, 12 + a, −3 + a}. Compute a.

Answer: 11

Solution: The sum of the roots of x3 − 42x2 + 525x − 2024 = 0 is − −42

1 = 42. This implies
that (a + 12) + a + (−3 + a) = 3a + 9 = 42, so a = 11.

I-10 Consider the equation

(cos x + cos 5x)2 = 2 cos2 3x.
Let S be the set of real values of x that satisfy this equation, with 0 ≤ x ≤ 2π. The sum of
2
the squares of all values of S can be written in the form aπb , where a and b are relatively prime
positive integers. Compute the ordered pair (a, b).

Answer: (1337, 72)

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 Solution: Note that by a sum-to-product identity, cos x+cos 5x is equivalent to 2(cos 3x)(cos 2x).
Thus 4(cos2 3x)(cos2 2x) = 2 cos2 3x, which implies either cos2 3x = 0 or cos2 2x = 12 .
First consider the case where cos 3x = 0. The inequality 0 ≤ x ≤ 2π is equivalent to
0 ≤ 3x ≤ 6π. In the interval [0, 6π], cos z = 0 when z = π2 , 3π 5π 7π 9π 11π
2 , 2 , 2 , 2 , 2 . The sum of the
squares of the corresponding values of x is
π2 2 2 2 2 2 2 143π 2
(1 + 3 + 5 + 7 + 9 + 11 ) = .
22 · 32 18
√
2
Now consider the case where cos 2x = ± 2 .√ The inequality 0 ≤ x ≤ 2π is equivalent to
2 3π 5π 15π
0 ≤ 2x ≤ 4π. In the interval [0, 4π], cos z = ± 2 when z = π4 , 4 , 4 , ..., 4 . The sum of the
squares of the corresponding values of x is
π2 85π 2
(12 + 32 + · · · + 152 ) = .
·22 42 8
Thus the sum of the squares of the solutions to the given equation in the interval [0, 2π] is
143π 2 85π 2 1337π 2
18 + 8 = 72 , and the desired ordered pair is (1337, 72).

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 Relay Round (6, 8, 10 minutes)

R1-1 The roots of the quadratic polynomial x2 + ax + b are a2 and 2b for some a = b. Compute
the numerical value of the positive difference between the roots of the quadratic polynomial
x2 + 2ax + 2b.
√
Answer: 4 10

Solution: The product of the two roots of x2 + ax + b is b = a2 · 2b so b = 0 or a = 4.

Furthermore, the sum of the two roots is −a = a2 + 2b so b = −3a. Therefore a = 4 and b = −12
so let r and s be the two roots of x2 + 2ax + 2b = x2 + 8x − 24. By Vieta’s
√ Formulas, r + s = −8
and rs = −24 so (r − s)2 = (r + s)2 − 4rs = 160, hence |r − s| = 4 10.

R1-2 Let T be the number you will receive. Suppose sin θ = T4 . Compute the minimum possible value
of the infinite geometric series tan θ − tan2 θ + tan3 θ − · · · .

1
Answer: − or −0.5
2

tan θ
Solution: If |tan θ| < 1, the value of the infinite geometric series becomes =
1 + tan θ
1
1− , so it suffices to minimize tan θ. Using sin2 θ + cos2 θ = 1 yields:
1 + tan θ

 √
2 T 2 − 16 4
cos θ = ± 1 − sin θ = ± =⇒ tan θ = ± √ .
T T 2 − 16

√
With T = 4 10, it follows that tan θ = ± 13 , and indeed, |tan θ| < 1. Taking tan θ = − 13 , the
1 1
minimum possible value of the given infinite geometric series is 1 − 1 = −2.
1− 3

R2-1 A team has won exactly 50% of the games they played this season (there are no ties). If they
win their next 2 games, then they will have won exactly 60% of their games. Compute the
number of games they have already won.

Answer: 4

Solution: Let x be the current number of wins. Then the team has played 2x games.
x+2
Winning the next 2 games means 2x+2 = 0.6, so x + 2 = 1.2x + 1.2, which implies 0.2x = 0.8,
thus x = 4.

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
R2-2 Let T be the number you will receive. A sequence is defined by a0 = 9, and for n ≥ 0,

an /2 if an is even,
an+1 =
3an + 1 if an is odd.

Compute the T -th greatest term that appears in the sequence.

Answer: 28

Solution: Computation can be made easier by noticing that if an is even, then the next few
terms result from continuing to divide by 2 until the greatest odd factor of an is revealed. Thus
the sequence is
9, 28, 14, 7, 22, 11, 34, 17, 52, 26, 13, 40, 20, 10, 5, 16, 8, 4, 2, 1.
At this point, the sequence cycles 1 → 4 → 2 → 1 → . . . . With T = 4, the 4th greatest integer
that appears is 28.

R2-3 Let T be the number you will receive. In right triangle ABC, the line segment connecting the
midpoints of the two legs is drawn, dividing the triangle into two polygons, one with perimeter
52 and one with perimeter T . Compute the area of triangle ABC.

Answer: 112
√
Solution: Let the triangle have legs a and b and hypotenuse c = a2 + b2 . Then the two
polygons are:
• a right triangle with sides a2 , 2b , and 2c ; and
• a trapezoid with sides 2c , 2b , c, and a2 .
Thus the two perimeters are a+b+c
2 and a+b+3c
2 . Let {x, y} = {T, 52} with x < y. Then a+b+c
2 =x
and a+b+3c
2 = y, so c = y − x and a + b = 2x − c = 3x − y. Thus
ab 1  1 
[ABC] = = (a + b)2 − (a2 + b2 ) = (3x − y)2 − (y − x)2 = x · (2x − y).
2 4 4
With T = 28, this gives x = 28 and y = 52, thus [ABC] = 28 · 4 = 112.

R3-1 A standard, fair six-sided die is rolled four times, resulting in the values a, b, c, and d. Suppose
that the maximum and minimum possible values of a + ab − cd + d are M and N , respectively.
Compute M − N .

Answer: 70

Solution: Rewriting the expression as a(b + 1) + d(1 − c), given that 1 ≤ a ≤ 6 and
2 ≤ b + 1 ≤ 7, then 2 ≤ a(b + 1) ≤ 42. Because 1 ≤ d ≤ 6 and 0 ≤ (c − 1) ≤ 5, it

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 follows that 0 ≤ d(c − 1) ≤ 30 → −30 ≤ d(1 − c) ≤ 0. Summing the two inequalities gives
−28 ≤ a(b + 1) + d(1 − c) ≤ 42, so N = −28, M = 42, and M − N = 70.

R3-2 Let T be the number you will receive. The angles of a convex n-gon form an arithmetic progression
with largest angle 2T ◦ . Compute the maximum possible value of n.

Answer: 9

Solution: The largest angle must be at least the size of the angles of a regular n-gon, namely
◦
180◦ − 360 360 360
n . Thus 2T ≥ 180 − n , so n ≤ 180−2T (because n and 180 − 2T are both positive).
With T = 70, the answer is 360
40 = 9.

R3-3 Let T be the number you will receive. Compute the real number x such that 2ln(T ) = xln(2) .

Answer: 9

Solution: Assuming T > 0 so that the equation is well-defined, taking the natural log of
both sides and using the fact that ln(ab ) = b ln(a) on each side yields ln(T ) · ln(2) = ln(2) · ln(x),
thus ln(T ) = ln(x) and x = T = 9.

R3-4 Let T be the number you will receive. Real numbers x and y satisfy 20x + 24y − 22T = 36 and
23x − 25y + T = −169. Compute x + y.

Answer: 10

Solution: Moving the T term to the right side and dividing the first equation by 4 yields
5x + 6y = 9 + 11T
2 . Multiplying this equation by 23 and the second by 5 yields 115x + 138y =
207 + 253T
2 and 115x − 125y = −845 − 5T , and subtracting yields 263y = 1052 + 263T
2 . Dividing
by 263 yields y = 4 + 2 and it follows that x = −3 + 2 , so x + y = T + 1. With T = 9, the
T T

answer is 10.

R3-5 Let T be the number you will receive. Franklin’s Fabulous Foods serves coffee and sandwiches.
Each customer buys one cup of coffee for $2 and at most one sandwich for $K. Given that
T % of customers buy a sandwich, and 60% of the restaurant’s revenue comes from sandwiches,
compute K.

Answer: 30

Solution: The restaurant’s revenue from sandwiches is 60% of the total, so the revenue from
sandwiches is 1.5 times as much as the revenue from coffee (which is 40% of the total). Every

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 100 customers buy 100 cups of coffee for $200, so they must spend $300 on sandwiches. Those
100 customers buy T sandwiches, so KT = 300 → K = 300 T . With T = 10, the answer is 30.
Author’s Note: This problem was inspired by observing a striking statistic that half of all
the water usage by Hawai’i’s tourism industry comes from irrigation for golf courses.

R3-6 Let T be the number you will receive. The figure below has T boxes, with T − 2 boxes in the
top row and 2 boxes in the bottom row. Compute the number of ways to arrange the numbers
1 through T , with one number in each box, so that the numbers in each row and each column
are in increasing order (the rectangle with the three dots represents T − 6 boxes).

Answer: 405

Solution: The top left corner must get the number 1. If 2 is immediately to the right of 1,
then the bottom two numbers can be any two from 3 to T (in order), then the rest is forced,
  2
so there are T −2 2 = T −5T
2
+6
such arrangements. If 2 is below 1, then the number to the
right of 1 must be 3 and there are T − 3 choices for the number to the right of 2 and below
3, with the rest forced, so there are T − 3 such arrangements. The total number is therefore
T 2 −5T +6
or T (T2−3) . With T = 30, the answer is 405.
2
2 + 2T2−6 = T −3T
2

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.
 Tiebreaker (10 minutes)
TB Compute the least value of the constant c such that the inequality
4t3 + ct2 − 3t + 9 ≥ 0
is true for any positive real number t.

Answer: −8

Solution: Rewrite the given inequality as (4t3 − 12t2 + 9t) + (4t2 − 12t + 9) + (c + 8)t2 ≥ 0.
This is equivalent to t(2t − 3)2 + (2t − 3)2 + (c + 8)t2 ≥ 0. Note that both t(2t − 3)2 and (2t − 3)2
are minimized when t = 32 . Thus (c + 8)t2 ≥ 0, and the least value of c for which this is true is
−8.

Alternate Solution: Suppose that f (t) = 4t3 + ct2 − 3t + 9 has zeroes z1 , z2 , z3 . From
Vieta’s Formulas, it follows that z1 z2 z3 = − 94 and z1 z2 + z2 z3 + z3 z1 = − 34 . Note that f (t)
has at least one negative real zero because the graph of f tends to negative infinity as t tends
to negative infinity and also the y-intercept of the graph is at (0, 9). Note that f cannot have
three negative real zeroes because that would imply z1 z2 + z2 z3 + z3 z1 is positive, contradicting
the fact that it is − 34 < 0, as established above. Thus when the value of c varies, the cubic
polynomial eventually will change from one that has three real roots (two positive, one negative)
to a cubic polynomial with one negative real root and two nonreal roots. It is worth examining
how the graph of the cubic function changes when the value of c changes.
The limiting case of interest happens when the graph of the cubic function becomes tangent to
the x-axis, in which case there is a double root. Suppose 4t3 + ct2 − 3t + 9 = 4(t − a)2 (t − b); then
using Vieta’s Formulas, it follows that 4a2 b = −9 and 4(a2 + 2ab) = −3. Thus 4a3 + 8a2 b = −3a,
which implies 4a3 + 3a − 18 = 0 → (2a − 3)(2a2 + 3a + 6) = 0. Hence a = 32 and b = −1, yielding
c = −8.

ARML encourages the reproduction of our contest problems for non-commercial, educational purposes.
Commercial usage of ARML problems without permission and posting entire contests or contest books are prohibited.