The Mathematical Association of America

American Mathematics Competitions

24th Annual

AMERICAN INVITATIONAL
MATHEMATICS EXAMINATION
(AIME)

SOLUTIONS PAMPHLET
Tuesday, March 7, 2006

This Solutions Pamphlet gives at least one solution for each problem on this year’s AIME
and shows that all the problems can be solved using precalculus mathematics. When more
than one solution for a problem is provided, this is done to illustrate a significant contrast in
methods, e.g., algebraic vs geometric, computational vs. conceptual, elementary vs. advanced.
The solutions are by no means the only ones possible, nor are they necessarily superior to others
the reader may devise.
We hope that teachers inform their students about these solutions, both as illustrations of the
kinds of ingenuity needed to solve nonroutine problems and as examples of good mathematical
exposition. Routine calculations and obvious reasons for proceeding in a certain way are often
omitted. This gives greater emphasis to the essential ideas behind each solution. Remember
that reproduction of these solutions is prohibited by copyright.
Correspondence about the problems and solutions for this AIME and orders
for any of the publications listed below should be addressed to:
American Mathematics Competitions
University of Nebraska, P.O. Box 81606
Lincoln, NE 68501-1606
Phone: 402-472-2257; Fax: 402-472-6087; email: [email protected]
The problems and solutions for this AIME were prepared by the MAA’s Committee on the
AIME under the direction of:
David Hankin, AIME Chair
207 Corbin Place, Brooklyn, NY 11235 USA

Copyright © 2006, The Mathematical Association of America

 2006 AIME SOLUTIONS 2

1. (Answer: 084)
Because AB 2 + BC 2 = AC 2 and AC 2 + CD2 = DA2 , it follows that DA2 = AB 2 +
BC 2 + CD2 = 182 + 212 + 142 = 961, so DA = 31. Then the perimeter of ABCD is
18 + 21 + 14 + 31 = 84.

2. (Answer: 901)
The least possible value of S is 1 + 2 + 3 + · · · + 90 = 4095, and the greatest possible
value is 11 + 12 + 13 + · · · + 100 = 4995. Furthermore, every integer between 4096 and
4994, inclusive, is a possible value of S. To see this, let A be a 90-element subset the sum
of whose elements is S, and let k be the smallest element of A such that k + 1 is not an
element of A. Because S ≤ 4994, conclude that k 6= 100. Therefore, for every 90-element
subset with sum S, where S ≤ 4994, a 90-element subset with sum S + 1 can be obtained
by replacing k by k + 1. Thus there are 4995 − 4095 + 1 = 901 possible values of S.

3. (Answer: 725)
The desired integer has at least two digits. Let d be its leftmost digit, and let n be the
integer that results when d is deleted. Then for some positive integer p, 10p ·d+n = 29n, and
so 10p · d = 28n. Therefore 7 is a divisor of d, and because 1 ≤ d ≤ 9, it follows that d = 7.
10p 100 · 10p−2
p
Hence 10 = 4n, so n = = = 25 · 10p−2 . Thus every positive integer with
4 4
the desired property must be of the form 7·10p +25·10p−2 = 10p−2 (7·102 +25) = 725·10p−2
for some p ≥ 2. The smallest such integer is 725.

OR

The directions for the AIME imply that the desired integer has at most three digits.
Because it also has at least two digits, it is of the form abd or cd, where a, b, c, and d are
digits, and a and c are positive. Thus bd · 29 = abd or d · 29 = cd. Note that no values of
c and d satisfy d · 29 = cd, and that d must be 0 or 5. Thus b0 · 29 = ab0 or b5 · 29 = ab5.
But b0 · 29 = ab0 implies b · 29 = ab, which is not satisfied by any values of a and b. Now
b5 · 29 = ab5 implies that b5 < 1000/29 < 35, and so b = 1 or b = 2. Because 15 · 29 = 435
and 25 · 29 = 725, conclude that the desired integer is 725.

4. (Answer: 124)
Let P = 1! 2! 3! 4! · · · 99! 100! . Then N is equal to the number of factors of 5 in P . For any
positive integer k, the number of factors of 5 is the same for (5k)!, (5k + 1)!, (5k + 2)!,
(5k + 3)!, and (5k + 4)!. The number of factors of 5 in (5k)! is 1 more than the number of
factors of 5 in (5k − 1)! if 5k is not a multiple of 25; and the number of factors of 5 in (5k)!
is 2 more than the number of factors of 5 in (5k − 1)! if 5k is a multiple of 25 but not 125.
 2006 AIME SOLUTIONS 3

Thus

N = 4·0+5(1+2+3+4+6+7+8+9+10+12+13+14+15+16+18+19+20+21+22)+24.

This sum is equal to

µ ¶
22 · 23
5· − (5 + 11 + 17) + 24 = 1124,
2

so the required remainder is 124.

OR

Let P = 1! 2! 3! 4! · · · 99! 100! . Then N is equal to the number of factors of 5 in P . When

the factorials in P are expanded, 5 appears 96 times (in 5!, 6!, . . . , 100!), 10 appears 91
times, and, in general, n appears 101 − n times. Every appearance of a multiple of 5 yields
a factor of 5, and every appearance of a multiple of 25 yields an additional factor of 5. The
number of multiples of 5 is 96 + 91 + 86 + · · · + 1 = 970, and the number of multiples of 25
is 76 + 51 + 26 + 1 = 154, so P ends in 970 + 154 = 1124 zeros. The required remainder is
124.

5. (Answer: 936)√ √ √ √ √ √
Expand (a 2 + b 3 + c 5)2 to obtain 2a2 + 3b2 + 5c2 + 2ab 6 + 2ac 10 + 2bc 15, and
conclude that 2a2 + 3b2 + 5c2 = 2006, 2ab = 104, 2ac = 468, and 2bc = 144. Therefore
ab = 52 = 22 · 13, ac = 234 = 2 · 32 · 13, and bc = 72 = 23 · 32 . Then a2 b2 c2 = ab · ac · bc =
26 · 34 · 132 , and abc = 23 · 32 · 13 = 936.
abc 23 · 32 · 13
Note that a = = = 13 and similarly that b = 4 and c = 18. These values
bc 23 · 32
yield 2a2 + 3b2 + 5c2 = 2006, as required.

6. (Answer: 360)
For any digit x, let x0 denote the digit 9 − x. If 0.abc is an element of S, then 0.a0 b0 c0 is
also in S, is not equal to 0.abc, and

0.abc + 0.a0 b0 c0 = 0.999 = 1.

It follows that S can be partitioned into pairs so that the elements of each pair add to 1.
Because S has 10 · 9 · 8 = 720 elements, the sum of the elements of S is 21 · 720 = 360.

OR
 2006 AIME SOLUTIONS 4

Recall that 0.abc = abc/999, so the requested sum is 1/999 times the sum of all numbers of
the form abc. To find the sum of their units digits, notice that each of the digits 0 through
9 appears 720/10 = 72 times, and conclude that their sum is 72(0+1+2+· · ·+9) = 72·45.
Similarly, the sum of the tens digits and the sum of the hundreds digits are both equal to
72 · 45. The requested sum is therefore 72 · 45(100 + 10 + 1)/999 = 360.

7. (Answer: 408)
..............
Without loss of generality, choose a unit of length
.........................
.........................
.........................
................
.......... .
. ......................... equal to the distance between two adjacent parallel
.......................................
....... K K
... ........................1 2 K 3 lines. Let x be the distance from the vertex of the
K 0 ............
............
............
............
............
............
angle to the closest of the parallel lines that intersect
............
............
............
.............
the angle. Denote the areas of the regions bounded
............
.
by the parallel lines and the angle by K0 , K1 , K2 , . . . ,
as shown. Then for m > 0 and n > 0,
Km K0 + K1 + K2 + · · · + Km K0 + K1 + K2 + · · · + Km−1
= −
K0 K0 K0
µ ¶2 µ ¶2
x+m x+m−1 2x + 2m − 1
= − = , so
x x x2
Km Km /K0 2x + 2m − 1
= = .
Kn Kn /K0 2x + 2n − 1
K4 2x + 7
Thus the ratio of the area of region C to the area of region B is = , and so
K2 2x + 3
2x + 7 11
= . Solve this equation to obtain x = 1/6. Also, the ratio of the area of region
2x + 3 5
K6 2x + 11
D to the area of region A is = . Substitute 1/6 for x to find that the requested
K0 x2
ratio is 408.

8. (Answer: 089)
A Y..•...... Draw the diagonals of rhombus T . Let Z be their B
..•
.. .. .. .. •.... .....
.... .... .....
...... point of intersection, and let X and Y be the shared
.....
.
..... . .....
..... ... ........
.....
.....
.....
P .....
.....
..
.
.....
.....
vertices of rhombuses P and T , with √
..
.
Y on AB. Let ...
.....
.....
.....
.....
.....
.....
.... .... ..
.... ........ ....... ........ ....... ....... ......... .....
F .
.
•.......
..
Z C Y Z = x and XY
.
.
•.......
..
.
= z. Consequently 2006 = [P] =
•... .....•
..... .....
...
...•
X .....
W
.....
.....
.....
√
.....
.....
.....
.....
.
...
.
..
.
.....
.... ....
.....
. ..
..

..... . ........ . .....

.....
F X · Y Z = zx, so z = 2006/x. Thus
.....
.....
..... . .....
............. ..
. .....
.....
• .. ... .
•..
E D ³ p ´ r
1 1 2006 p
K = (2 · Y Z)(2 · XZ) = (2x) 2 z 2 − x2 = 2x − x2 = 8024 − 4x4 .
2 2 x2
¥√ ¦
There are 8023 = 89 positive values of x that yield a positive square for the radicand,
so there are 89 possible values for K.

OR
 2006 AIME SOLUTIONS 5

Define X, Y , and Z as in the first solution, and let W be the shared vertex of rhombuses
Q and T , W 6= Y , let α = m6 AY X, let β = m6 XY W , and let z be the length of the
sides of the rhombuses.
√ Then β + 2α = 180◦ , and the area of each of the four rhombuses
2
is z sin α = 2006. Therefore
√
K = z 2 sin β = z 2 sin 2α = 2z 2 sin α cos α = 2 2006 cos α.
√ √
Thus 1 ≤ K < 2 2006 = 8024, so 1 ≤ K ≤ 89, and there are 89 possible values for K.

9. (Answer: 046)
Note that
¡ ¢ ¡ ¢
log a1 + log a2 + · · · + log a12 = log(a1 a2 · · · a12 ) = log a · ar · · · · · ar11 = log a12 r66 ,

where the base of the logarithms is 8. Therefore a12 r66 = 82006 = 23·2006 , so a2 r11 = 21003 .
Because a and r are positive integers, each must be a factor of 21003 . Thus a = 2x and
r = 2y for nonnegative integers x and y. Hence 2x + 11y = 1003, and each ordered pair
(a, r) corresponds to exactly one ordered pair (x, y) that satisfies this equation. Because
2x is even and 1003 is odd, y must be odd, so y has the form 2k − 1, where k is a positive
integer. Then 1003 = 2x + 11y = 2x + 22k − 11, so x = 507 − 11k. Therefore 507 − 11k ≥ 0,
and so 1 ≤ k ≤ b507/11c = 46. Thus there are 46 possible ordered pairs (a, r).

10. (Answer: 065)

y..... Such a line is unique. This is because when a line with equa-
.. ..
..................... ........................
...
.... tion y = 3x + d intersects R, then as d decreases, the area
......
...
...
..
.... . ..
......
...
B ..... of the part of R above the line is strictly increasing and the
.
.. .
.
........................ ...........................
.
...
.. ...... ...... • ...... ........................
.....
...
.....
area of the part of R below the line is strictly decreasing. The
......
....
......
....
...
...
... ........ ........ ...
...
.....
symmetry point of the two circles that touch at A(1, 1/2) is
....................
.
. ..
. .
. .....
. .... .
. .....
. ....
...... .............................. ..............................
..... ...... ...... ...
...
.....
A ...
A, and so any line passing through A divides their region into
•
.....
.
.
.....
.....
.
.
.....
...
...
.
..... ... .
. ... .
. ...
x two regions of equal area. Similarly, any line passing through
................... ...................... ...................... ....
...

B(3/2, 2) divides the region consisting of the two circular regions that touch at B into two
regions of equal area. Of the remaining four circles, two of them are on either side of line
AB. Thus line AB divides R into two regions of equal area. Because the slope of line AB
is 3, line AB is line `, and it has equation y − (1/2) = 3(x − 1) or 6x = 2y + 5. Then
a2 + b2 + c2 = 36 + 4 + 25 = 65.
 2006 AIME SOLUTIONS 6

11. (Answer: 458)

Let S(n) be the number of permissible towers that can be constructed from n cubes, one
each with edge-length k for 1 ≤ k ≤ n. Observe that S(1) = 1 and S(2) = 2. For n ≥ 2, a
tower of n + 1 cubes can be constructed from any tower of n cubes by inserting the cube
with edge-length n + 1 in one of three positions: on the bottom, on top of the cube with
edge-length n, or on top of the cube with edge-length n − 1. Thus, from each tower of n
cubes, three different towers of n + 1 cubes can be constructed. Also, different towers of
n cubes lead to different towers of n + 1 cubes, and each tower of n + 1 cubes becomes a
permissible tower of n cubes when the cube with edge-length n + 1 is removed. Hence, for
n ≥ 2, S(n + 1) = 3S(n). Because S(2) = 2, it follows that S(n) = 2 · 3n−2 for n ≥ 2.
Hence T = S(8) = 2 · 36 = 1458, and the requested remainder is 458.

12. (Answer: 906)

The given equation implies that

cos3 3x + cos3 5x = (2 cos 4x cos x)3

3
= (cos(4x + x) + cos(4x − x))
= (cos 5x + cos 3x)3 .

Let y = cos 3x and z = cos 5x. Then y 3 + z 3 = (y + z)3 . Expand and simplify to obtain
0 = 3yz(y + z). Thus y = 0 or z = 0 or y + z = 0, that is, cos 3x = 0 or cos 5x = 0 or
cos 5x + cos 3x = 0. The solutions to the first equation are of the form x = 30 + 60j, where
j is an integer; the second equation has solutions of the form x = 18 + 36k, where k is
an integer. The third equation is equivalent to cos 4x cos x = 0, so its solutions are of the
form x = 22 12 + 45m and x = 90 + 180n, where m and n are integers. The solutions in the
interval 100 < x < 200 are 150, 126, 162, 198, 112 12 , and 157 12 , and their sum is 906.

13. (Answer: 899)

Note that Sn is defined as the sum of the greatest powers of 2 that divide the 2n−1
consecutive even numbers 2, 4, 6, . . . , 2n . Of these, 2n−2 are divisible by 2 but not 4, 2n−3
are divisible by 4 but not 8, . . . , 20 are divisible by 2n−1 but not 2n , and the only number
not accounted for is 2n . Thus

Sn = 2 · 2n−2 + 22 · 2n−3 + · · · + 2n−1 · 20 + 2n = (n + 1)2n−1 .

In order for Sn = 2n−1 (n+1) to be a perfect square, n must be odd, because if n were even,
then the prime factorization of Sn would have an odd number of factors of 2. Because n
is odd, n + 1 must be a square, and because n + 1 is even, n + 1 must be the square of
an even integer. The greatest n < 1000 that is 1 less than the square of an even integer is
302 − 1 = 899.
 2006 AIME SOLUTIONS 7

14. (Answer: 183)

The feet of the unbroken tripod are the vertices of an equilateral triangle ABC, and the
foot of the perpendicular from the top to the plane of this triangle is at the center of the
triangle. By the Pythagorean Theorem, the distance from the center to each vertex of the
triangle is 3. Place a coordinate system so that the coordinates
√ of the top T are (0, √
0, 4) and
the coordinates of A, B, and C are (3, 0, 0), (−3/2, 3 3/2, 0), and (−3/2, −3 3/2, 0),
respectively. Let the break point A0 be on T A. Then T A0 : A0 A = 4 : 1. Thus the
coordinates of A0 are
4 1
(3, 0, 0) + (0, 0, 4) = (12/5, 0, 4/5).
5 5

Note that the coordinates of M , the midpoint of BC, are (−3/2, 0, 0). The perpendicular
from T to the plane of 4A0 BC will intersect this plane at a point on M A0 . This segment
lies in the xz-plane and has equation 8x−39z +12 = 0 in this plane. Then h is the distance
from T to line M A0 , and is equal to

| 8 · 0 − 39 · 4 + 12| 144
p =√ ,
82 + (−39)2 1585
√
so bm + nc = 144 + 39 = 183.

OR
z....
.... ....

....... ....... •
T (0, 4)
............ ....... ....... ......
...
.
.
... ...
... ......
.
...
.. Place a coordinate system as in the first solution. Note
... . ...
...
.
.
.
.
.
.
.
...
...
....
...
.. that 4A0 M T is in the xz-plane. In this plane, circum-
... ... ... ...
.
... ....
..
.
.
. ....
... .
... ..
. scribe a rectangle around 4A0 M T with its sides parallel
. ... ... ..
... ....
.
. ... .
... to the axes. Then
. .. ........•
................ ..
..
A0 (2.4, 0.8)
M• ..
.......................................
............... ..... [A0 M T ] = 4(3.9) − 21 (4)(1.5) − 12 (3.2)(2.4) − 21 (3.9)(0.8)
.... x
.. .
(−1.5, 0) = 7.2.
2[A0 M T ] 14.4 144 √
Thus h = 0
=√ =√ , so bm + nc = 144 + 39 = 183.
AM 15.85 1585
 2006 AIME SOLUTIONS 8

OR

The feet of the unbroken tripod are the vertices of an equilateral triangle ABC, and the
foot of the perpendicular from the vertex to the plane of this triangle is at the center of
the triangle. Let T be the top of the tripod, let O be the center of 4ABC, let A0 be the
break point on T A, and let M be the midpoint of BC. Apply the Pythagorean √ Theorem
to 4T OA to conclude that OA = 3. Therefore 4ABC has sides of length 3 3. Notice
that A0 , M , and T are all equidistant from B and C, so the plane determined by 4T A0 M
is perpendicular to BC, and so h is the length of the altitude from T in 4T A0 M . Because

1 0 1
A M · h = [T A0 M ] = A0 T · T M sin 6 A0 T M,
2 2
it follows that
A0 T · T M sin 6 A0 T M
h= .
A0 M
√ p √
The length of A0 T is 4, and T M = T B 2 − BM 2 = 25 − (27/4) = 73/2. To find A0 M ,
note that A0 M 2 = A0 C 2 − CM 2 , and that A√0 C 2 = A0 T 2 + T C 2 − 2A0 T · T C cos 6 A0 T C.
52 + 52 − (3 3)2 23
But cos 6 A0 T C = cos 6 AT C = = ,
2·5·5 50 r r
0 2 2 2 0 113 27 317
so A C = 4 + 5 − 2 · 4 · 5 · (23/50) = 113/5, and A M = − = .
5 4 20
16 + 73/4 − 317/20 23
Now cos 6 A0 T M = √ = √ ,
2 · 4 · 73/2 5 73
2
23 36
so sin2 6 A0 T M = 1 − , and sin 6 A0 T M = √ . Thus
25 · 73 5 73
√
73 36
4· 2 · √
5 73 144
h= q =√ ,
317 1585
20

√
so bm + nc = 144 + 39 = 183.
 2006 AIME SOLUTIONS 9

15. (Answer: 027)

The condition |xk | = |xk−1 + 3| is equivalent to x2k = (xk−1 + 3)2 . Thus

n+1 n+1 n
à n
! Ã n
!
X X X X X
x2k = (xk−1 + 3)2 = (xk + 3)2 = x2k + 6 xk + 9(n + 1), so
k=1 k=1 k=0 k=0 k=0
n+1 n
à n
!
X X X
x2n+1 = x2k − x2k = 6 xk + 9(n + 1), and
k=1 k=0 k=0
n
X 1£ 2 ¤
xk = xn+1 − 9(n + 1) .
6
k=0

¯2006 ¯
¯X ¯ 1 ¯ ¯
¯ ¯
Therefore ¯ xk ¯ = ¯x22007 − 18063¯ . Notice that xk is a multiple of 3 for all k, and that
¯ ¯ 6
k=1
xk and k have the same parity. The requested sum will be a minimum when |x22007 −18063|
is a minimum, that is, when x2007 is the multiple of 3 whose square is as close as possible
to 18063. Check odd multiples of 3, and find that 1292 < 16900, 1412 > 19600, and
1352 = 18225. The requested minimum is therefore 16 |1352 − 18063| = 27, provided there
exists a sequence that satisfies the given conditions and for which x2007 = 135. An example
of such a sequence is (
3k for k ≤ 45,
xk = −138 for k > 45 and k even,
135 for k > 45 and k odd.

Problem Authors
1. David Hankin 9. Chris Jeuell
2. Elgin Johnston 10. Zuming Feng
3. Jacek Fabrykowski 11. Elgin Johnston
4. Gregory Galperin 12. David Hankin
5. Chris Jeuell 13. Zuming Feng
6. Elgin Johnston 14. Elgin Johnston
7. Jonathan Kane 15. Jacek Fabrykowski
8. Chris Jeuell
 The
American Mathematics Competitions
are Sponsored by
The Mathematical Association of America
University of Nebraska-Lincoln
The Akamai Foundation
Contributors
American Mathematical Association of Two Year Colleges
American Mathematical Society
American Society of Pension Actuaries
American Statistical Association
Art of Problem Solving
Canada/USA Mathcamp
Canada/USA Mathpath
Casualty Actuarial Society
Clay Mathematics Institute
Institute for Operations Research and the Management Sciences
L. G. Balfour Company
Mu Alpha Theta
National Council of Teachers of Mathematics
National Assessment & Testing
Pedagoguery Software Inc.
Pi Mu Epsilon
Society of Actuaries
U.S.A. Math Talent Search
W. H. Freeman and Company
Wolfram Research Inc.