South Carolina ARML Mathematics

1994 TEAM PROBLEMS
T1. The average grade of 20 students in a class on a test is 72.45 points. Two of the 20
students received 0 points on the test (because they didn't show up for the test). Calculate
the average grade of the remaining 18 students.
T2. Calculate the number of rational roots of the equation
x8 , 24x7 , 18x5 + 39x2 + 1155 = 0:
T3. Let g(x) = 1=x and h(x) = log2 x. Calculate the value of f (8037776589) if f (x) =
xg(h(x)).
T4. Observe that the set f2; 3; 4g has the property that each sum of 2 distinct elements
from the set is not divisible by 4 (in other words, 2 + 3 = 5, 2 + 4 = 6, and 3 + 4 = 7
are each not divisible by 4). On the other hand, it is possible to show that any set of 4
consecutive integers has the property that some sum of two of its elements is divisible by
4. Calculate the integer A, such that (i) and (ii) below both hold.
(i) There is a set S consisting of A , 1 consecutive integers such that each sum of 2
distinct elements of S is NOT divisible by 1994.
(ii) Every set of A consecutive integers is such that a + b is divisible by 1994 for some
choice of 2 distinct integers a and b from the set.
T5. A lattice point is a point (a; b) in the plane with a and b both integers. A lattice
point is \visible from the origin" if the line segment joining the origin to the point contains
no lattice points other than the origin and the point. For example, (1; 2) is visible from
the origin but (2; 2) is not (the point (1; 1) lies on the segment joining the origin to (2; 2)).
Find an integer b such that 0 < b 400 and
(2; b); (3; b); (4; b); (5; b); (6; b); (7; b); and (8; b)
are all NOT visible from the origin.
T6. Calculate the sum of the squares of the real roots of
x20 , x18 , x16 , x14 , , x4 , x2 , 2:
T7. Calculate the largest integer which is less than

p p p p p p
2500 , 2501 + 2502 , 2503 + , 2999 + 3000:
T8. Let a2; a3 ; : : : ; a100 be dened as follows:
a100 = 100
a99 = 99a100 = 99100
100
a98 = 98a99 = 9899
..
.
Thus, in general, an = nan+1 for 2 n 99. Calculate the units digit of a2.
T9. The graphs of y = 2x3 , 4x + 2 and y = x3 + 2x , 1 intersect in exactly 3 distinct
points. Calculate the slope of the line passing through 2 of these points.
T10. In the drawing below, the equation of the circle is x2 + y2 = 1. Also, \BDA = 17
and the distance from A to B is equal to the reciprocal of the distance from A to C .
Calculate the x,coordinate of C ? (Comment: The measurements in the drawing may not
be accurate.)
D A
Solutions to 1994 TEAM PROBLEMS
T1. The sum of the grades in the class is 20 72:45 = 1449. This is also the sum of the
18 remaining students' grades. Thus, their average is 1449=18 = 80:5.
T2. The Rational Root Test implies that the only \possible" rational roots are of the form
a=b where a divides 1155 and b divides 1. Since 1155 = 3 5 7 11, this means that there
are 32 possible rational roots (1; 3; 5; 7; 11; 15; 21; : : : ). Observe that all 32 of
these numbers are odd. But if x is an odd number, then x8 , 24x7 , 18x5 + 39x2 + 1155 is
odd and, therefore, not 0. It follows that there are no rational roots to the given equation.
The answer is 0.
T3. Use the equation 2log2 x = x to obtain
, 1= log2 x
f (x) = xg(h(x)) = x1= log2 x = 2log2 x = 2:
Thus, f (x) is 2 for all values of x, and the answer is 2.
T4. The answer is A = 999. The set S = 0; 1; 2; : : : ; 997 satises (i) since it has 998 =
A , 1 elements and since 1 a + b 996 + 997 = 1993 for any two elements a and b
from this set S . To see that (ii) holds, suppose we have a set T consisting of A = 999
consecutive integers. We can write T in the form T = fu; u + 1; u + 2; : : : ; u + 998g for
some integer u. Observe that u + (u + j ) = 2u + j for 1 j 998 so each of the
numbers 2u + 1; 2u + 2; 2u + 3; : : : ; 2u + 998 can be written in the form a + b for some
distinct a and b from T . Also, (u + j ) + (u + 998) = 2u + 998 + j for 1 j 997 so
2u + 999; 2u + 1000; 2u + 1001; : : :; 2u + 1995 can also be written in the form a + b for some
distinct a and b from T . In other words, there are 1995 consecutive integers from 2u + 1
to 2u + 1995 each of which can be written in the form a + b for some distinct a and b from
T . At least one of these must be divisible by 1994, so (ii) holds.
T5. An important observation here is that (a; b) is visible from the origin if and only if
the greatest common divisor of a and b is 1. Take b = 2 3 5 7 = 210. Then each of
2; 3; 4; : : : ; 8 has a common factor > 1 with b, so b satises the condition in the problem.
(Note: 210 is the only answer to this problem.)
T6. Descartes' Rule of Signs gives that there is exactly one positive
p real root and exactly
one negative real root of the given polynomial. Observe that 2 areproots of the given
polynomial. p 2 the only real roots to the given polynomial are 2, and the answer
p 2 Hence,
is ( 2) + (, 2) = 2 + 2 = 4.
T7. Observe that
p p
, 2501 + 2502 = p2501 +1 p2502 < 50 +1 50 = 100
1 :
p p p p p p
Similarly, each of , 2503 + 2504; , 2505 + 2506; : : : ; , 2999 + 3000 is < 1=100.
We get that
p p p p p p
2500 , 2501+ 2502 , 2503 + , 2999 + 3000
p 1 + + 1 = 50 + 250 = 52:5:
< 2500 + 100 100 100
We claim the answer is 52. We need only show now that
p p p p p p
2500 , 2501 + 2502 , 2503 + , 2999 + 3000 > 52:
In other words, we want
p p p p p p
3000 > 52 , 2500 + 2501 , 2502 + 2503 , + 2999:
An argument as above gives
p p p 1 = 54:5;

52 , 2500 + 2501 , + 2999 < 52 + 250 100
p
so we need only show that 3000 > 54:5 (this is correct { do you understand why?) which
follows since 3000 > 2970:25 = 54:52.
T8. The number a2 is a power of 2. Observe that the last digits of 2; 22; 23; 24 ; : : : are
2; 4; 8; 6; 2; 4; 8; 6; : : : . In other words, 24k+1; 24k+2; 24k+3; 24k+4 end with 2; 4; 8; 6, respec-
tively, for each non-negative integer k. Since a2 = 2a3 , we want to know in which of the
forms 4k +1; 4k +2; 4k +3; 4k +4 we can write a3 . Since a3 is a power of 3, we consider suc-
cessive powers of 3. Note that 3; 32; 33 ; 34; : : : are in the form 4k +3; 4k +1; 4k +3; 4k +1; : : :
(for dierent k's). In other words, if 3 is raised to an odd power, then the result is of the
form 4k + 3; and if 3 is raised to an even power, then the result is of the form 4k + 1. Since
a3 = 3a4 and a4 = 4a5 , a3 is 3 raised to an even power and, hence, of the form 4k + 1 for
some positive integer k. We get that the nal digit of a2 = 2a3 is 2.
T9. If two of the points are (x1 ; y1 ) and (x2 ; y2 ), then the slope of the line through the
two points is (y2 , y1 )=(x2 , x1 ). Also,
y1 = 2x31 , 4x1 + 2; y2 = 2x32 , 4x2 + 2;
y1 = x31 + 2x1 , 1; and y2 = x32 + 2x2 , 1:
Taking the dierence of the rst two of these, we get that
,
y2 , y1 = 2(x32 , x31) , 4(x2 , x1 ) = (x2 , x1 ) 2(x22 + x2 x1 + x21 ) , 4 :
Hence, the slope of the line through the two points is
y2 , y1 = 2(x2 + x x + x2 ) , 4:
x2 , x1 2 2 1 1
Taking another dierence of equations above gives
,
y2 , y1 = x32 , x31 + 2(x2 , x1 ) = (x2 , x1 ) x22 + x2 x1 + x21 + 2 :
Hence, the slope of the line through the two points is also
y2 , y1 = x2 + x x + x2 + 2:
x2 , x1 2 2 1 1
This means that
2(x22 + x2 x1 + x21) , 4 = x22 + x2 x1 + x21 + 2
so that x22 + x2 x1 + x21 = 6. Substituting this into either of the two formulas for the slope
above, we get that the slope is 8. (Comment: Since the answer does not depend on which
two points we consider, we can conclude that the three points of intersection all lie on a
straight line.)
T10. We use the notation PQ to repre-
sent the distance between two points P and C
Q. Let O ! represent the origin.

! Let E be a
point!on OA such that CE is perpendicular
to OA. Observe that \CEA = \DBA =
90 . One easily gets that CEA is similar B
to DBA. Hence, AB=AD = AE=AC so

that AB AC = AE AD. The given in- D O E A
formation implies that AB AC = 1 and

AD = 2. Therefore, AE = 1=2 which implies
OE = 1=2. It follows that the x,coordinate
of C is 1=2. (Note that the measurements
in the drawing are not correct. Also, 17 has
nothing to do with this answer.)
TEAM PROBLEMS 02/95
T1. The distance between the centers of two circles in a plane is 25. The radius of one
circle is 3 and the radius of the other circle is 4. A line ` is tangent to one circle at point P
and tangent to the other at point Q. There is a point on the line segment joining the two
centers of the circles that is also on `. What is the distance between P and Q? Simplify
your answer.
T2. What is the sum of all numbers of the form a b c where a is from the set f1; 2; 4; 8g,
b is from the set f1; 3; 17; 19g, and c is from the set f1; 7; 31; 61g?
T3. What is the greatest common divisor of the numbers 106577 and 1071089?
T4. For any two real numbers x and y, dene
f (x; y) = x2 + 13y2 , 6xy , 4y , 2:

For what value of the pair (x; y) is f (x; y) as small as possible?
T5. If the value of

X
98
1
n=1 n(n + 1)(n + 2)(n + 3)
is written as a reduced fraction, then what is the value of its denominator?
T6. Dave and Michael decide to play the following game. They begin with S = 0.
They take turns picking a number from the set f1; 2; 3; 4; 5; 6g. On each turn any of the
6 numbers can be chosen. The number is added to S and S is then replaced by the sum.
The rst person who chooses a number which when added to S totals 40 is the winner.
Dave begins by picking the number 6. What number is now the best choice for Michael to
pick?
T7. In how many points do the graphs of
y = x12 + x10 + x8 + x6 + x4 + x2 + 1 and y = x5 + x3 + x
intersect?
T8. Let A = (,3; 0) and B = (3; 0). Let C denote the circle x2 + y2 = 9 (so AB is
a diameter of C ). What is the probability that a random point P inside C is such that
\APB 3=4. (Here, \APB represents an angle with radian measure in [0; ].)
m
T9. For m a positive integer and k 2 f0; 1; 2; : : : ; mg, dene k = k!(mm,! k)! . It is
m
k m
known that the coecient of x in (x + 1) is k . Using this fact, nd positive integers
a and b each less than 1000 satisfying
502 502 502 502 a
0 + 1 + 2 + + 50 = b :
T10. Let a and b be integers for which

1 + 1 = a:
1 + 12 + 13 + + 1994 1995 b
Suppose that exactly one of a and b is even. Determine which of a and b is even and
determine the largest integer r such that 2r divides ab.
SOLUTIONS TO TEAM PROBLEMS 02/95
T1. Let C1 be the circle of radius 3 and C2
the circle of radius 4. Let A be the center of
C1 , and let B be the center!
of C2 . Let C be
the intersection of ` and AB . Set x = AC and
note BC = 25 , x. Then APC is similar to
BQC , and we get B
x = 25 , x :
3 4
Hence, AC = x = (3=7)25 = 75=7 and BC =
P C
Q
25 , x = (4=7)25 = 100=7. Therefore, the A
distance from P to Q is
p p
(3 25=7)2 , 9 + (4 25=7)2 , 16
p
= 73 252 , 72 + 47 252 , 72 = 252 , 72 = 576 = 24:
p p p
T2. (1 + 2 + 4 + 8)(1 + 3 + 17 + 19)(1 + 7 + 31 + 61) = 15 40 100 = 60000

T3. (Use the Euclidean Algorithm.) Let d = gcd(106577; 1071089). Then d divides
1071089 , (10 106577) = 5319 and, hence, 106577 , (20 5319) = 197. One checks that
197 divides 5319 and therefore both 106577 and 1071089. The answer is 197.
T4. Completing the square twice, f (x; y) = (x , 3y)2 + (2y , 1)2 , 3. Thus, f (x; y) ,3
for all pairs (x; y) and achieves this value when y = 1=2 and x = 3=2. The answer is
(x; y) = (1=2; 3=2).
T5. The sum can be evaluated via
98
X 1 98 1
X

1 1

= ,
n=1 n(n + 1)(n + 2)(n + 3) n=1
3 n(n + 1)(n + 2) (n + 1)(n + 2)(n + 3)

1 1 1
= 3 6 , 99 100 101 :
Since 6 divides 99 100 101, the least common denominator of the dierence above is
99100101 = 999900. It is easy to see that the resulting numerator will not be divisible by
any of the prime divisors of 99 100 101 including 3. The answer is 3 999900 = 2999700.
T6. The answer is 6. After Dave and Michael choose 6, the value of S is 12. No matter
what number k Dave chooses after this point, Michael can choose 7 , k. Every turn by
both Dave and Michael will increase the sum by 7. After Michael has four more turns,
S = 12 + (4 7) = 40, so Michael wins with this strategy. No other choice for Michael's
rst turn will do (if Michael picks ` 2 f1; 2; 3; 4; 5g, Dave can choose 6 , ` for his second
move and continue by choosing 7 , k for each integer k Michael chooses; Dave will be able
to win).
T7. It suces to count the number of real roots of
, ,
f (x) = x12 + x10 + x8 + x6 + x4 + x2 + 1 , x5 + x3 + x
= x12 + x10 + x8 + x6 , x5 + x4 , x3 + x2 , x + 1:
Clearly, f (0) 6= 0. Since
f (,x) = x12 + x10 + x8 + x6 + x5 + x4 + x3 + x2 + x + 1;
Descartes Rule of Signs gives that f (x) has no negative roots. Observe that
,
(x + 1)f (x) = (x + 1) x12 + x10 + x8 + x6 , x5 + x4 , x3 + x2 , x + 1
= x13 + x12 + x11 + x10 + x9 + x8 + x7 + 1:
By Descartes Rule of Signs, (x + 1)f (x) and, hence, f (x) has no positive real roots. The
answer is 0.
T8. Consider the circle C 0 centered at D =
(0; ,3) and passing through A and B. The
measure of angle \ADB is =2, from which
it follows that for any point P on the smallest
arc (along C 0 ) from A to B, the measure of
angle \APB is 3=4. Similarly, any point P
on the smallest arc from A to B along a circle A B
centered at (0; 3) is such that the measure of

angle \APB is 3=4. Let R be the region
strictly between these two arcs. Then P is
a point in R if and only if the measure of
\APB is > 3=4. The answer is therefore
the ratio of the area of C minus the area of R D
p of C . The area of C is 9. Since

to the area
DB = 3 2 and the measure \ADB = =2, the area of R is

p 2 1 p 2
1 (3 2) , (3 2) = 18 , 1 = 9 , 18:

2
2 2 2 2
After a little simplication, we deduce that the probability is 2=.
, ,
T9. Observe that mk = mm,k . Thus,
X50 50 ! X 50 50 ! X50 50 ! X50 50 !
(x +1)50(x +1)50 = k xk k xk = k xk k x50,k :
k=0 k=0 k=0 k=0
The coecient of x50 in this last product (when expanded) is
50 2 + 50 2 + 50 2 + + 50 2:

0 1 2 50
,
On the other hand, (x + 1)50 (x + 1)50 = (x + 1)100 , so the coecient of x50 must be 100
50 .
An answer is a = 100 and b = 50.
T10. Let t be the positive integer for which 2t divides 1995! and 2t+1 does not divide
1995!. So we may write 1995! = 2t m where m is an odd integer. Let
1995
X 1995!
S= j :
j =1
Here, S is a sum of integers. Observe that the term 1995!=1024 = 2t,10m is divisible by
2t,10 and not by 2t,9. Also, since no other denominator in the sum S above is divisible
by 1024, every other term is divisible by 2t,9. Write
S , 1995!
1024 = 2t,9 k;
where k is an integer. Then

a = 1 S = 1 ,2t,9 k + 2t,10m = 1 2t,10 (2k + m) = 2k + m :
b 1995! 2tm 2t m 210m
Hence,
210ma = b(2k + m):
Since m is odd, so is 2k + m, and we deduce that b is even. Only one of a or b is even, so
a is odd. The above equation implies therefore that r = 10.
TEAM PROBLEMS February, 1996
T1. The number which is four more than the square of 625 has exactly two prime factors.
Determine what they are.
T2. A circle of radius 1 rolls (without sliding) along the x-axis so that its center is of
the form (t; 1) with t increasing. A certain point P touches the x-axis at the origin as the
circle rolls. As the circle rolls further, the point P touches the point (x; 1=2). Given that
x 3, what is the value of x?
T3. Let 1 , 2 , and 3 denote the roots of x3 , x + 1 = 0. Calculate the value of
41 + 42 + 43 :
Simplify your answer. It should be an integer.
T4. An experiment consists of choosing two points P and Q at random from the circum-
ference of a given circle of radius one. The average value of the distance from P to Q is
4=. What is the average value of the square of the distance from P to Q?
T5. A pentagon (not necessarily a regular pentagon) has vertices A, B , C , D, and E .

Given that
the midpoint of edge AB is (1; 2)
the midpoint of edge BC is (3; 0)
the midpoint of edge CD is (8; 1)
the midpoint of edge DE is (5; 4)
the midpoint of edge EA is (3; 4),
nd the coordinates of A.
T6. Suppose
(x2 , x + 1)3(x3 + 2x2 + 2x + 1)5 = a21 x21 + a20 x20 + + a1 x + a0 :
What is the value of a1 + a2 + + a10 ?
T7. For k 1, dene ak by ak = 2k + 3. Thus, a1 = 5; a2 = 7; a3 = 11; : : : . It is not

hard to see that every ak is not divisible by 2 and every ak is not divisible by 3. Find the
next two primes p having the property that every ak is not divisible by p.
T8. Two honest people, named Bert and Ernie, each make you an oer as indicated
below. Determine who is making the better oer and determine an appropriate statment
you might make. The statement should clarify why the particular person you claim is
making the better oer is the person who is making the better oer.
Bert's Oer: Make a statement. If it is true, then I will give you exactly ten dollars. If
it is false, then I will give you either less than ten dollars or more than ten dollars but not
exactly ten dollars.
Ernie's Oer: Make a statement. If it is true, then I will give you more than ten dollars.
If it is false, then I will give you more than ten dollars. (That's right, either way I will
give you more than ten dollars.)
T9. A chessboard contains 64 one-by-one squares, but it also contains some two-by-two
squares, some three-by-three squares, and so on. In particular, there is one eight-by-eight
square on a chessboard. How many total squares are there on a chessboard? Simplify your
answer.
T10. How many total rectangles are there on a chessboard (see Problem T9)? (To clarify,
a \rectangle" must have positive area. In other words, we do not allow here for one of its
dimensions to be zero.) Simplify your answer.
T1. Since x4 + 4 = (x2 , 2x + 2)(x2 + 2x + 2), we obtain
, ,
6252 + 4 = 254 + 4 = 252 , 2 25 + 2 252 + 2 25 + 2 = 577 677:
The primes must be 577 and 677.
T2. When P is located at A = (x; 1=2), suppose the center of the circle is located at
B = (u; 1). Then u will equal the distance the center of the circle has traveled since P
was at the origin which will also be equal to the arclength from (x; 1=2) to C = (u; 0) (the
point where the circle touches the x-axis). It follows that \ABC = u (in radians). Hence,
, x = u , x and
sin u = uAB 1=2 = 1 :
cos u = AB 2
p
Hence, u = =3 and x = u , sin u = (=3) , 3=2. (Note that cos u = 1=2 for other values
of u but each of these will result in a value of x > 3.)
T3. Use the \elementary symmetric functions." The sum of the roots of an x + an,1 x
n n,1 +
an,2xn,2 + + a0, where an 6= 0, is 1 = ,an,1=an. The sum of the roots taken two at
a time is 2 = an,2=an. Also, note that the sum of the roots squared is 12 , 22 . In the
present problem, we deduce
1 + 2 + 3 = 0 and 21 + 22 + 23 = (0)2 , 2(,1) = 2:
Since each j satises 3j , j + 1 = 0 (so that 3j = j , 1), we deduce 4j = 2j , j for
each j . Therefore,
41 + 42 + 43 = (21 + 22 + 23) , (1 + 2 + 3) = 2 , 0 = 2:
T4. Note that the problem did not ask
for the average of the values of the square
of the distance from P to Q. This would be
a considerably easier question with unfortu-
nately the same answer (the average of the
distances being 4= is given to clarify this
distinction). To answer the question, imag-
ine choosing rst the point P and then the
point Q. The distance from P to Q is de-
termined then simply by the placement of Q
relative to P . In other words, we can view P as being a xed point on the circle. For each
point Q on the circle, consider the point Q0 such that QQ0 is a diameter of the given circle.
Then \QPQ0 = =2 and Q0 Q = 2 so that QP 2 + Q0 P 2 = Q0 Q2 = 4. Thus, the average of
QP 2 and Q0 P 2 is 2. As P varies, we can deduce that the average of the distances squared
is 2.
T5. Suppose A = (x; y ). The midpoint of AB being (1; 2) implies that B = (2 , x; 4 , y ).
Now the midpoint of BC being (3; 0) implies that C = (4 + x; ,4 + y). Continuing in this
manner, we deduce that D = (12 , x; 6 , y), E = (,2 + x; 2 + y), and A = (8 , x; 6 , y).
Comparing our two expressions for A, we deduce x = 4 and y = 3. Hence, A = (4; 3).
T6. Call a polynomial f (x) palindromic if it satises x
deg f f (1=x) = f (x) (the coecients
read the same in either direction). Observe that if f (x) and g(x) are both palindromic
then so is their product (just use the denition). It follows that the polynomial F (x) =
(x2 , x + 1)3 (x3 + 2x2 + 2x + 1)5 is palindromic. Since a0 = 1, we deduce
a1 + a2 + + a10 = 12 (a0 + a1 + + a21 , 2a0 ) = 12 (F (1) , 2) = 12 (65 , 2) = 3887:
T7. The given information and the fact that a4 = 19 imply that the rst three primes to
consider are 13, 17, and 23. One checks directly that 2k + 3 modulo 13 is 0 when k = 10
(i.e., 24 3 (mod 13) =) 28 9 (mod 13) =) 210 + 3 39 0 (mod 13)). On the
other hand, a direct computation gives that 2k +3 modulo 17 is never 0 and 2k +3 modulo
23 is never 0. The primes are 17 and 23.
T8. Bert's oer is better. A corresponding statement would be, \You will not give me
ten dollars and you will not give me a billion dollars."
T9. There are 8 one-by-one squares, 7 two-by-two squares, 6 three-by-three squares,
2 2 2
and so on. The answer is
12 + 22 + + 82 = 204:
T10. A rectangle will be determined by the endpoints of one of its diagonals. There
are 9 possible endpoints. Choose one of these at random. The opposite endpoint of a
2
diagonal cannot lie directly to the North, South, East, or West of the chosen endpoint
(where we picture here the chessboard being oriented in an obvious manner). Thus, there
are 82 possibilities for the second endpoint of a diagonal. This gives us 92 82 ways of
choosing the endpoints. Each rectangle will be counted four times in the process, as the
rst endpoint chosen could correspond to any one of the four corners. This gives a total
of 92 82=4 = 1296 rectangles.
By the way, there is a connection here with the answer to T9. Note that
13 + 23 + + 83 = (1 + 2 + + 8)2 = 1296:
TEAM PROBLEMS* February, 1997
T1. How many divisors does the number 30! (thirty factorial) have?
T2. What is the area of the region between the two
circles x + y = 16 and (x , 16) + y = 100 that
2 2 2 2
p p
lies under the line joining (2 3; 2) to (11; 5 3) and
p p
above the line joining (2 3; ,2) to (11; ,5 3)? (See
the shaded region in the picture, not drawn to scale.)
p
Express your answer in the form a + b 3 + c where a,
b, and c are integers.
T3. Beginning with the origin (0; 0) in the xy-plane, a path is made by taking steps
where each step consists of either (i) adding 1 to both coordinates or (ii) adding 1 to the
x-coordinate and subtracting 1 from the y-coordinate. For example, an acceptable path
would be one going from (0; 0) to (1; 1) and then to (2; 2); another acceptable path would
be one going from (0; 0) to (1; ,1) to (2; 0) and then to (3; 1). Each point occurring on such
a path, we call a thisforlackofabettername point. For example, each of the points (0; 0),
(1; 1), (2; 2), (1; ,1), (2; 0), and (3; 1) mentioned above is a thisforlackofabettername point.
On the other hand, (1; 0) and (1; 3) are not. How many thisforlackofabettername points
(x; y) are there satisfying 0 x 10 and 0 y 10?
T4. Consider the points A = (0; 1), B = (0; 0), and C = (2; 0) in the Cartesian plane.
How many points P in the plane have the property that PA , PB , PA , PC , and PB , PC
are all integers? Here PA represents the distance from the point P to the point A and
similarly for PB and PC . Do not assume that the coordinates of P are necessarily integers.
T5. The number 2 cos(10 ) is a root of a polynomial f (x) satisfying: (i) the degree of
f (x) is 6, (ii) the coecient of x is 1, and (iii) the coecients of f (x) are integers. What
6
is f (x)?
*Calculators are allowed to be used during this competition. The problems are written accordingly.
Also, note that solutions to the problems will be available through the World Wide Web at the URL
http://www.math.sc.edu/~laseta/contests.html.
T6. Three dice (cubes) are thrown simultaneously, one with 1 garnet face and 5 black
faces, one with 2 garnet faces and 4 black faces, and one with 3 garnet faces and 3 black
faces. Two of the dice turn up garnet (i.e., with a garnet face on top). What is the
probability that the third die also turns up garnet? (Note that you know that two of the
dice turn up garnet but you don't know which ones they are.)
T7. A reciprocal polynomial is a polynomial f (x) =
Pn a xj satisfying
j j =0
a = an ; a = an, ; a = an, ; : : : ; an = a :
0 1 1 2 2 0
For example, x , 2x , 2x + 1 and 3x , 4x + 3 are reciprocal polynomials. Consider all

5 4 2
the reciprocal polynomials with integer coecients that are factors of x , x , x + 1. 1234 3
Find the factor which has the largest degree. (Hint: Your answer should have degree 4.)
X 1 1997
T8. Let S = 2n . Suppose that the decimal expansion of S is given by S =

!
n
0:d d d : : : (so that each of the dj 's represents a digit). What is the value of the sum
=1
1 2 3
d +d +d +d +d ?
20 21 22 23 24
T9. Let A = (1; 9) and B = (13; ,4) be points in the xy-plane. For each point P on the
x-axis, consider the distance from A to P (denoted AP ) and the distance from B to P
(denoted BP ). What is the maximum value of the dierence AP , BP ?
T10. Given that
= 3:141592653589793238462643383279502884197169399375 : : :
is accurate to the digits shown and that
cos x = 1 , x2 + E (x) where jE (x)j x24 if jxj 1;
2 4
calculate the sum of the 49th and 50th digits after the decimal in the decimal expansion
of 628318530717958647692528
cos 10 : 23
Throughout this problem, the angles are to be considered to be measured in radians. Note
that the numerator in the last expression above contains 24 digits.
T1. The number of times a prime p divides n! is given by

n + n + n + ;
p p p 2 3
where [x] denotes the greatest integer x (note the sum above is nite since [n=pk ]
is 0 if k is large). We deduce that if 2r is the highest power of 2 dividing 30!, then
r = [30=2] + [30=4] + = 15 + 7 + 3 + 1 = 26. Similarly, if 3s is the highest power of 3
dividing 30!, then s = [30=3] + [30=9] + = 10 + 3 + 1 = 14. Continuing, we obtain
30! = 2 3 5 7 11 13 17 19 23 29:
26 14 7 4 2 2
The number of divisors is obtained by multiplying one more than each of the exponents
together, so the number of divisors is 27 15 8 5 3 3 2 2 2 2 = 2332800 .
p p
T2. Draw line segments joining the (0; 0) to each of (2 3; 2) and (2 3; ,2) and joining
p p
(16; 0) to each of (11; 5 3) and (11; ,5 3). The answer is easily obtained by summing
the appropriate areas of triangles and trapezoids and subtracting the sum of the areas of
p
two sectors. The answer is ,8 + 80 3 , 36 .
T3. The thisforlackofabettername points are seen to be those of the form (x; y) where
x jyj and x + y is even. The condition x jyj comes from noting that (0; 0) is the initial
point of a path and other points are obtained by adding at least as much to the x-coordinate
as the absolute value of what is added to the y-coordinate. The condition x + y is even
is seen, for example, by induction. That each of these points can be obtained is also easy
to see (to get such an (x; y), simply consider the path from (0; 0) to ((x , y)=2; (y , x)=2)
along a straight line and then from ((x , y)=2; (y , x)=2) to (x; y) along a straight line).
The number of such (x; y) satisfying 0 x 10 and 0 y 10 is 36 .
T4. Since PA , PC = (PA , PB )+(PB , PC ), it suces to consider the case that only
PA , PB and PB , PC are integers. The sum of the lengths of two sides of a triangle
is at least as big as the length of the third side so that PA + AB PB (by considering
ABP ) and PB + BC PC (by considering BCP ). Note that equality could occur if
A, B , and P are collinear of if B , C , and P are collinear, respectively. Similarly, we get
PB + AB PA and PC + BC PB . The conditions in the problem imply AB = 1 and
BC = 2. It follows that ,1 PA , PB 1 and ,2 PB , PC 2. Since PA , PB and
PB , PC are integers, we have PA , PB 2 f,1; 0; 1g and PB , PC 2 f,2; ,1; 0; 1; 2g.
The condition PA , PB = 1 occurs precisely when P is on the line passing through
A and B . The condtion PA , PB = 0 occurs precisely when P is on the perpendicular
bisector of line segment AB . The condition PB , PC = 2 occurs precisely when P is on
the line passing through B and C . The condition PB , PC = 1 occurs precisely when
P is on a specic hyperbola with foci at B and C (that the absolute value of the dierence
in the distances PB and PC is constant denes geometrically a hyperbola). The condtion
PB , PC = 0 occurs precisely when P is on the perpendicular bisector of line segment
BC . We need both PA , PB 2 f,1; 0; 1g and PB , PC 2 f,2; ,1; 0; 1; 2g to occur so
that in the end we are interested in the number of intersection points between one of the
two lines determined by PA , PB 2 f,1; 0; 1g and one of the two lines and one hyperbola
determined by PB , PC 2 f,2; ,1; 0; 1; 2g. One counts the number of intersection points
directly. For example, the line passing through A and B and the hyperbola dened above
intersect in two points. It is of relevance when counting to note that the point Q = (0; 1=2)
is not on the hyperbola (this can be seen by computing QB , QC and verifying that it is
not 1). We deduce that there are 6 points P as in the problem.
T5. Using the identities
cos(A+B ) = cos(A) cos(B ),sin(A) sin(B ) and cos(A,B ) = cos(A) cos(B )+sin(A) sin(B )
with A = n and B = , one obtains
() cos((n + 1)) + cos((n , 1)) = 2 cos() cos(n):
We take g (x) = x and g (x) = 2x , 1 so that cos() = g (cos()) and cos(2) =
2
g (cos()). By setting gn (x) = 2xgn (x) , gn, (x) for n 2, we get from () that
1 2 1
cos(m) = gm (cos()) for all positive integers m. The formula gn (x) = 2xgn (x),gn, (x)
2 +1 1
for n 2 (together with the values given for g (x) and g (x)) allows one to compute
+1 1
g (x) = 32x , 48x + 18x , 1. Letting u = cos() with = 10 , we deduce that
1 2
6 4 2
6
1 = cos(60 ) = cos(6) = g (cos()) = 32u , 48u + 18u , 1:

6 4 2
2 6
Thus, for u = cos() with = 10 , we have

0 = 64u , 96u + 36u , 3 = (2u) , 6(2u) + 9(2u) , 3:
6 4 2 6 4 2
It follows that 2 cos(10 ) = 2u is a root of x , 6x + 9x , 3 . (The answer is unique.)

6 4 2
T6. Let A denote the die with one garnet face, and let R denote the garnet face. Let
B denote the die with two garnet faces; denote them by S and S . Let C denote the die
1 2
with three garnet faces; denote them by T , T , and T . The possibilities for the roll of
1 2 3
the three dice are:

Two Faces Known Third Face # of Cases
1. R and an Si T , T , or T
1 2 3 6
2. R and an Si one of 3 blue faces of die C 6
3. R and a Tj S or S
1 2 6
4. R and a Tj one of 4 blue faces of die B 12
5. an Si and a Tj R 6
6. an Si and a Tj one of 5 blue faces of die A 30
There are 66 cases above exactly 18 of which occur with the third face turning up garnet.
The probability is 18=66 or (simplied) 3=11 .
T7. If g(x) is a reciprocal polynomial of degree r (and non-zero constant term), then
xr g(1=x) = g(x). Suppose g(x) is a reciprocal polynomial of degree r dividing u(x) =
x1234
, x , x + 1 so that u(x) = g(x)h(x) for some polynomial h(x) of degree 1234 , r.
3
Then
x 1234
,x1233
,x 1231
+ 1 = x u(1=x) = x g(1=x)h(1=x)
1234 1234
,
= xr g(1=x)x ,r h(1=x) = g(x) x ,r h(1=x) :
1234 1234
Thus, g(x) is a factor of w(x) = x , x 1234 1233

,x 1231
+ 1. Since g(x) divides each of u(x)
and w(x), we deduce that it divides
() x w(x),(x , x , 1)u(x) = x + (x , x , 1)(x + x , 1)
3 3 2 3 3 2 3
= x , x + x , 2x + x , x + 1 = (x , 1) (x + 1)(x + x + 1):
6 5 4 3 2 2 2 2
It follows that g(x) must divide this last expression. In particular, the only possible factors
of g(x) are x , 1, x +1, and x + x +1. Each of these is a factor of u(x) = x , x , x +1
2 2 1234 3
which can be veried as follows. First, u(1) = 0 and u(i) = 0 imply that x , 1 and x + 1 2
are factors. Since x + x + 1 is a factor of x , 1 and x , 1 is a factor of

2 3 3
, ,
xx 3 411
,1 , x ,1 =x 3 1234
, x , x + 1;
3
we get x + x + 1 is a factor of u(x).

2
Next, we note that (x , 1) does not divide u(x). Two ways to see this are:
2
(i) Use that if (x , 1) is a factor of u(x), then u0 (1) = 0 but u0 (1) = 1230 6= 0. (This
2
is a Calculus approach; a non-Calculus approach follows.)

(ii) Use that if (x , 1) is a factor of u(x), then u(x) = (x , 1) v(x) for some polynomial
2 2 2
v(x) so that u(x + 1) = x v(x + 1). This implies that the coecient of x in u(x + 1)
2
is 0. The Binomial Theorem implies, however, that the coecient of x in u(x + 1) =

(x + 1) , (x + 1) , (x + 1) + 1 is 1230.
1234 3
Either way, we deduce that (x , 1) is not a factor of u(x). It follows that g(x) divides
2
(x , 1)(x +1)(x + x +1). Since g(x) is reciprocal, the leading coecient and constant term
2 2
must be equal. It follows that x,1 is not a factor of g(x). (Note that if (x,1) = x ,2x+1 2 2
could be a factor of g(x), the situation would be dierent.) The reciprocal factor of u(x)
of largest degree is now seen to be (x + 1)(x + x + 1) = x + x + 2x + x + 1 .
2 2 4 3 2
1
1 = 1, so that the decimal representation of S in the problem
X
Observe that S < n
n 2
T8.
=1
is justied (i.e., there are no non-zero digits to the left of the decimal). We work with
10 S = d d d : : : d :d d : : : :
24
1 2 3 24 25 26
For n 5, the value of 10 =2n is less than or equal to 10 =2
24 ! 24 5!+(n, so that
5)
1X
1997
10 X 1 10 1 1 10
X 24 1997
10

24 24 24
1024
n < =
2 n 2n, 2 n 2n 2 < (10 = )
< 1:
n 2
! 5! 5 120 119 3 10 119
=5 =5 =0
On the other hand, for n 4, the value of 10 =2n is an integer, namely 2 ,n 5 .

24 ! 24 ! 24
Observe that for n 3, the number 10 =2n = 2 ,n 5 is divisible by 10 , = 10 so

24 ! 24 ! 24 24 3! 18
that the right 18 most digits of 10 =2n are all 0. The sum of the three terms 10 =2n
24 ! 24 !
for 1 n 3 in 10 S therefore contribute nothing to the rst 18 digits to the left of the
24
decimal in 10 S . It remains to consider 10 =2n with n = 4. We have 10 =2 = 5 .

24 24 ! 24 4! 24
We compute the right most ve digits of 5 . This can be done easily with a calculator.
24
For example, 5 = 390625. We are only interested in the last 5 right most digits, so to
8
calculate the last 5 digits of 5 we can simply compute the value of 90625 625 which
12
ends with 40625. If the calculator can calculate 40625 40625 exactly, then we can get the
last 5 digits of 10 by computing the last 5 digits in this product. Alternatively, one can
24
multiply by 625 three times (each time dropping all but the last 5 digits). This gives that
the last 5 digits of 5 are 90625. The information above now implies that d = 9, d = 0,
24
20 21
d = 6, d = 2, and d = 5. Hence, the sum in the problem is 9 + 0 + 6 + 2 + 5 = 22 .

22 23 24
0
T9. Let B = (13; 4), and let P denote the intersection of the x-axis with the line passing
through A and B 0 . We claim that with P so chosen, AP , BP gives the desired maximum.
To see this observe that if Q is anyppoint on the x-axis, then AQ , BQ = AQ , B 0 Q.
In particular, AP , BP = AB 0 = 12 + 5 = 13. On the other hand, if Q 6= P , then
2 2
since the sum of the lengths of two sides of a triangle is greater than the third, we obtain
AB 0 + B 0 Q > AQ so that AQ , BQ = AQ , B 0 Q < AB 0 = 13. Hence, the answer is 13 .
T10. Let = 628318530717958647692528=10 , and observe that 2. Since >
23
1, we cannot use the estimate given in the problem for cos x directly. However, since
cos = cos(2 , ), we can use the estimate by setting x = 2 , . Since 2 10 = 23
628318530717958647692528:6766559 : : : , we obtain that

2
cos = cos(2 , ) = 1 , 12 0:6766559
10
:::
23
+ E 0:6766559
10
::: :
23
From the information given in the problem, E (x) < 1=10 . Since (0:6766559 : : : ) = 92 1 2
0:22893 : : : , it follows that cos = 1 , u + v where u < 1, the decimal expansion of u

2
consists of 46 zeroes to the right of the decimal followed by the digits 22893, and the
decimal expansion of v consists of at least 92 zeroes to the right of the decimal. Hence,
cos = 0:999 : : : 9977106 : : : , where 46 nines occur between the decimal and the rst digit
seven. The answer is 1 .
TEAM PROBLEMS January, 1998
(Calculators are permitted on this competition.)
T1. Let A = ( 1; 7), B = (11; 3), C = ( 2; 0), and D = (14; 0). There are three points
P = (x0 ; 0) on the x-axis such that \AP C + \BP D = 90Æ . What are the three possible
values of x0? Simplify your answers.
T2. Given 10 points (labelled A; B; : : : ; J ) and
18 edges (line segments) joining them as shown, a
path is made by beginning with one of the points,
traversing an edge to another point, traversing an
edge again to a point, and so on. Consider only
paths where each edge is traversed at most once
(each point may occur more than once). In such a
path, what is the maximum number of edges that
can be traversed?
T3. There are positive integers a and b for which
1000000
X 1p pa p
p
k+ k+1
= b:
k =2
What are a and b?

T4. A number is chosen at random from the set f1; 2; 3; : : : ; 1998g. What is the probability
that it has no prime factors in common with 10! (i.e., 10 factorial)?
T5. An ellipse is symmetric about the origin and passes through the points (1; 0), (0; 1),
and (1; 1). Determine the point (x; y) on the ellipse with the largest possible value of y.
Give explicit values for both x and y and simplify your answers.
T6. Let f (x) be the polynomial one obtains by expanding the product
(x 1)(x 2)(x 3) (x 1000):
Let k be the coeÆcient of x125 in f (x). There is a non-negative integer n such that 2n
divides k and 2 +1 does not divide k. What is the value of n?

n
T7. Find the sum of the four smallest distinct primes which divide the number
15 15
15 + 15:
99999
X
2j .
T8. Calculate and simplify cos4 99999
j =1
p p p p
ANSWERS: T1: 5 15 and 5 57 T5: (1= 3; 2= 3)
T2: 16 T6: 500
T3: a = 1000001 and b = 2 T7: 39
T4: 457=1998 T8: 299997=8
T1. Suppose rst that P is on segment CD. The condition \AP C + \BP D = 90Æ
implies that \AP B = 90Æ . Let Q = (5; 5) and note that Q is the midpoint of segment
AB . Since \AP B = 90Æ , we deduce that P is on the circle centered at Q of radius QB .
Hence, QP = QB and we get
p
(x0 5)2 + 52 = 62 + 22 =) x0 = 5 15:
p
Note that the points (5 15; 0) are on CD.
Now, suppose P is on the x-axis but not on CD. Let B 0 = (11; 3). The condition
\AP C + \BP D = 90Æ implies that \AP B 0 = 90Æ . Let Q0 = (5; 2), the midpoint of AB 0 .
Then P is on the circle centered at Q0 of radius Q0 B 0 . Thus,
p
(x0 5)2 + 22 = 62 + 52 =) x0 = 5 57:
p p
The point (5 + 57; 0) is on CD and, hence, is not an answer. The point (5 p57; 0) is
on CD and, hence, is an answer. Thus, the possible values for x0 are 5 15 and
not p
5 57.
T2. We describe a path by listing the points it traverses in order. A path (as described)
traversing 16 edges is F ABCDEF GJ F IJ DHJ BG. It remains to show that more than
16 of the 18 edges cannot be traversed. Each of the 6 points B , D, F , G, H , and I has an
odd number of edges emanating from it. Suppose we are given a point in a path and the
given point is neither the rst nor the last point in the path. The number of edges in the
path containing the given point as an endpoint must be even since for every edge traversed
going to the point there must be another edge traversed going from the point. Thus, in
every path, at least 4 of the 6 points B , D, F , G, H , and I must have the property that
there is an edge not in the path that has the point as an endpoint. Since an edge has two
endpoints, it follows that there must be at least two edges not traversed in every path.
Therefore, each path as described in the problem can traverse at most 16 edges.
p p p p
T3. We use that ( k + 1 + k)( k + 1 k ) = 1 to rewrite the sum as a \telescoping"
sum. We obtain
1000000
X 1
1000000
X p p p p
p p = k+1 k = 1000001 2:
k=2
k+ k+1 k=2
Thus, apparently a = 1000001 and b = 2. It can easily be shown that this choice for a and
b is unique, but this is not required by the problem.
T4. The primes dividing 10! are 2, 3, 5, and 7. For d a positive integer, dene A(d) to be
the number of integers in S = f1; 2; 3; : : : ; 1998g which are divisible by d. Then A(d) is the
number of positive integral multiples of d which are 1998. It follows that A(d) = [1998=d]
where [x] denotes the greatest integer x. By the principle of inclusion-exclusion, the
number of integers not divisible by any of 2, 3, 5, and 7 is
A(1) A(2) A(3) A(5) A(7) + A(6) + A(10) + A(14)

+ A(15) + A(21) + A(35) A(30) A(42) A(70) A(105) + A(210):
Given A(d) = [1998=d], the above can be computed quickly with a calculator. It's value is
457. Therefore, the probability that an element of S chosen at random is not divisible by
any of 2, 3, 5, and 7 is 457=1998.
T5. The equation of the ellipse must be of the form ax2 + bxy + cy 2 + dx + ey + f = 0
for some numbers a, b, c, d, e, and f . Since the ellipse is symmetric about the origin and
passes through the points (1; 0), (0; 1), and (1; 1), it must also pass through the points
( 1; 0), (0; 1), and ( 1; 1). Plugging these 6 points into the equation of the ellipse
above, we obtain the equations:
(i) a + d + f = 0 (iv) a d + f = 0
(ii) c + e + f = 0 (v) c e + f = 0
(iii) a + b + c + d + e + f = 0 (vi) a + b + c d e + f = 0
The equations (i) and (iv) imply d = 0. The equations (ii) and (v) imply e = 0. Now, (i)
and (ii) imply a = f and c = f . Plugging in these values for d, e, a, and c into (iii)
gives b = f . Thus, the equation of the ellipse is f x2 + f xy f y 2 + f = 0. This equation
would not describe an ellipse unless f 6= 0. Thus, we can divide by f to deduce that the
ellipse can be described by the equation x2 xy + y 2 1 = 0. Hence, if (x; y ) is a point
on the ellipse, then
2
y 3
() x + y 2 = x2 xy + y 2 = 1:
2 4
p
The two terms on the left-hand side are clearly non-negative. Since their sum is 1, each
must be 1. Thus, the largestppossible value for y is 2= 3 which occurs when the second
term is p x = y=2 = 1= 3, () clearly holds so that the desired point on the ellipse
p 1. When
is (1= 3; 2= 3).
T6. The number k which is the coeÆcient of x125 in f (x) can be determined by taking
minus the sum of all products of 875 roots of f (x). The roots of f (x) are the integers from
1 to 1000. The product of 875 roots of f (x) divisible by the least power of 2 is the one
obtained by using the 500 odd positive integers 1000, the 250 positive integers 1000
that are divisible by 2 and not 4, and the 125 positive integers 1000 that are divisible
by 4 and not 8. The product of these 875 roots of f (x) is divisible by 2500 and not by 2501 .
Every other product of 875 roots of f (x) must be divisible by a larger power of 2 (think
about it if this is not clear) and, hence, by 2501 . We deduce that there are integers s and
t with s odd such that
k = 2500 s + 2501 t = 2500 (s + 2t):
Since s is odd, we deduce s + 2t is odd and the answer is n = 500.

Let N = 15 15 + 15. Clearly, N is divisible by both 3 and 5, so these must be two
15
T7.
of the four least primes dividing N . Also, N is even, so N is divisible by 2. It remains to
determine the next least prime dividing N . For each prime p > 5, we compute the smallest
positive integer, say e(p), such that 15e(p) 1 (mod p). We then compute the value of 1515
modulo e(p), call it r(p). Then there is an integer q (p) such that 1515 = q (p)e(p) + r(p).
(One can make use of a calculator for these computations.) It follows that
q (p) r(p)
N = 15 15 + 15 15r(p) + 15 (mod p):
15
+ 15 = 15q(p)e(p)+r(p) + 15 = 15e(p) 15
We compute 15r(p) + 15 modulo p directly to determine if p divides N . For example, if
p = 13, then we compute e(13) = 12 and r (13) = 3 (the latter may we require two steps
with a calculator, rst computing the value of 153 modulo 12 and then raising this value to
the fth power modulo 12). Then we compute the value of 15r(p) + 15 = 153 + 15 modulo
p = 13. For p = 13, we obtain the number 10 so N 10 (mod 13) which implies N is not
divisible by 13. Letting p vary, we derive:
e(7) = 1; r (7) = 0 =) N 150 + 15 2 (mod 7);
e(11) = 5; r (11) = 0 =) N 150 + 15 5 (mod 11);
e(13) = 12; r (13) = 3 =) N 153 + 15 10 (mod 13);
e(17) = 8; r (17) = 7 =) N 157 + 15 6 (mod 17);
e(19) = 18; r (19) = 9 =) N 159 + 15 14 (mod 19);
e(23) = 22; r (23) = 1 =) N 151 + 15 7 (mod 23);
e(29) = 28; r (29) = 15 =) N 1515 + 15 0 (mod 29):
We deduce that the least prime p > 5 dividing N is 29. Thus, the answer is 2+3+5+29 =
39. There are ways of simplifying the computations, for example by taking advantage of
Fermat's Little Theorem which implies that 15p 1 1 (mod p) when p > 5.

T8. We use that cos = ei + e i
=2 (this can be derived from ei = cos + i sin ).
Thus,
1 i 4 1 4i
() cos4 = e + e i = e + 4e2i + 6 + 4e 2i + e 4i :
16 16
Suppose that k is a non-zero integer and n is a positive integer for which n does not divide
k . Then
X n
e2 ik 1
e2 ikj=n = e2 ik=n 2 ik=n
j =1
e 1
where the latter is obtained by taking x = e2ik=n in the polynomial identity

x xn 1
x+x +x +
2 3
+ x n
=
x 1
:
Since k is an integer, e2ik 1 = cos(2k) + i sin(2k) 1 = 0. Hence, the numerator

above is 0. (Note that the condition that n does not divide k implies the denominator
e2 ik=n 1 6= 0.) Thus,
n
X
( ) e2 ikj=n = 0 provided n does not divide k:
j =1
Observe that if k < n, then n cannot divide k. We consider n > 4. We take = 2j=n and
n
X 2j
sum from j = 1 to n on both sides of (). The left-hand side becomes cos 4
.
n
j =1
The right-hand side can be expressed as
n n n n n
1 X 2i4j=n X
2 i2j=n
X X X
e +4 e +6 1+4 e2 i( 2)j=n
+ e
2 i ( 4)j=n
:
16 j =1 j =1 j =1 j =1 j =1
The middle sum has value n and, using (), we deduce that each of the remaining sums
is 0. Hence,
X n
2j 1 3n
cos 4
= (6n) = for n > 4:
j =1
n 16 8
Taking n = 99999, we derive the answer to the problem, namely 299997=8.

TEAM PROBLEMS* February, 1999
(Calculators are permitted on this competition.)
T1. A fair coin is to be ipped at random 6 times. What is the probability that the coin
will land on heads face-up more often than on tails face-up?
n
T2. If n and j are integers with 0 j n, then the binomial coeÆcient j
(which is

read \n choose j ") is dened by nj = j !(nn! j )! . What is the value of
X
j 16
16
2 j ?
j =0
T3. Two points A and B are on a circle of radius one. The distance from A to B is one.
If C is a point on the circle such that the center of the circle is inside ABC , then what
is the measure of \C ?
T4.
1 , what are the rst four digits (in order) after the
In the decimal expansion of 1999
1 = 0:d1d2 d3 : : : where the d
rst 100 digits after the decimal? In other words, if 1999 j
denote digits, then what is d101d102d103d104?
T5. The graphs of y = 4x3 x2 +3x 1 and y = 8x2 5 intersect in exactly three points.
If these points are (x1; y1), (x2; y2), and (x3 ; y3), then what is the value of x1 + x2 + x3?
T6. Let f (x) = cos x + 3 sin x. Suppose t is a real number such that f (t) f (x) for all
real x. Calculate the exact value of sin t.
*For the solutions, see http://www.math.sc.edu/~filaseta/contests.html.
T7. A 2 10 board
is to be completely covered by 10 dominoes (i.e., 2 1 or 1 2 boards shaped or

, respectively). In how many ways can this be done? For example, there are 3
dierent ways to cover a 2 3 board with dominoes illustrated below.
T8. The polynomial x2 4x +5 can be multiplied by a non-zero polynomial g(x) (having

integer coeÆcients) with the resulting product having non-negative integer coeÆcients
(and, hence, degree at least two). What is the smallest possible degree for such a g(x)?
ANSWERS: T1: 11=32 T5: (p+ 8)=4

T2: 43046721 T6: 3= 10
T3: 30Æ or =6 radians T7: 89
T4: 5547 T8: 5
T1. There are 26 = 64 possible outcomes for the 6 coin ips. The number of these with
heads landing face-up more often than tails is the number of times heads lands face-up
on four ips of the coin plus the number of times heads lands face-up on 5 ips plus the
number of times heads lands face-up on all 6 ips. This is the same as
6 6 6
4 + 5 + 6 = 15 + 6 + 1 = 22:
Therefore, the probability of heads being face-up more often than tails is 22=64 = 11=32.
T2. The binomial theorem asserts that
n
X
(x + y) =
n n
j
xj y n j :
j =0
Take n = 16, x = 2, and y = 1 to get
16 16
X
2j j = (2 + 1)16 = 43046721:
j =0
T3. The measure of \C is one-half of the measure of \AOB where 0 denotes the center
of the circle. On the other hand, each side of AOB has length one, so AOB is an
equilateral triangle. Thus, \AOB = 60Æ so that \C = 30Æ.
Another solution is as follows. The measure of \C does not change! if the location of
C is moved to another point on the circle on the same side of line AB . We move C so
that line segment AC passes through the center of the circle. It follows that AC = 2 (the
diameter of the circle). Also, \ABC = Æ90Æ. Since we are given that AB = 1, we deduce
that sin \C = 1=2 and, hence, \C = 30 .
T4. Observe100that if q is the quotient and r is the remainder when 10100 is divided by
1999, then 10 = 1999q + r so that
1 =q+ r :
10100 1999 1999
It follows that r=1999 = 0:d101d102d103d104 : : : . We determine r by computing 10100 mod-
ulo 1999. With the help of a calculator, we obtain
104 5 (mod 1999) =) 1020 55 1126 (mod 1999)
=) 1040 11262 510 (mod 1999) =) 1080 5102 230 (mod 1999)
=) 10100 1020 1080 1126 230 1109 (mod 1999):
Thus, r = 1109. Since 1109=1999 = :554777 : : : , we deduce d101d102d103d104 = 5547.
T5. The x-coordinates where the graphs intersect satisfy 4x3 x2 + 3x 1 = 8x2 5
so that 4x3 ( + 8)x2 + 3x + 4 = 0. The sum of the roots of this equation is ( + 8)=4.
We deduce that x1 + x2 + x3 = ( + 8)=4.
p p
T6. Let be such that cos = 1= 10 and sin = 3= 10 (you should justify that such a
exists). We rewrite the function f (x) as follows:
p p
( ) = cos x + 3 sin x = 10 cos cos x + sin sin x = 10cos(
f x )
x :
Taking t = , we see that p p

( ) = 10 10cos( x)
f t
for every x. Hence, t = givespthe value of t in the statement of the problem. Therefore,
the answer is sin t = sin = 3= 10. (Note that there is more than one t as in the problem,
but sin t is uniquely determined.)
T7. Let Dn denote the number of ways of covering a 2 n board with dominoes. One
checks directly that D1 = 1 and D2 = 2. If n 3, then every covering of a 2 n board must
either have its last column covered by a 2 1 domino or its last two columns covered by two
1 2 dominoes. It follows that each covering of a 2 n board can be constructed from either
a covering of a 2 (n 1) board followed by a 2 1 domino or a covering of a 2 (n 2)
board followed by two 1 2 dominoes. In other words, Dn = Dn 1 + Dn 2 for each n 3.
A direct computation now gives D3 = D2 + D1 = 2 + 1 = 3, D4 = D3 + D2 = 3 + 2 = 5,
D5 = D4 + D3 = 5 + 3 = 8, : : : , D10 = 55 + 34 = 89.
T8. There are a variety of dierent polynomials g(x) asn in the problem. Wenbegin by
explaining one way to obtain such a g(x). Let g(x) = an x + an 1 x + an 2 x + ,
n 1 2
and let f (x) = (x2 4x + 5)g(x). We consider an = 1 and integers an 1 ; an 2; : : :
successively as small as possible so that the the leading coeÆcients in f (x) are nonnegative.
Observe that
(x2 4x + 5)g(x) = (x2 4x + 5)(xn + an 1 xn 1 + an 2 xn 2 + )
= xn+2 + (an 1 4)xn+1 + (an 2 4an 1 + 5)xn
+ (an 3 4an 2 + 5an 1 )xn 1 + :
We take an 1 = 4, an 2 = 11, and then an 3 = 24. We continue in this manner; but to
help with the computations, note that the coeÆcient of xn j is an j 2 4an j 1 +5an j
for j 1 so that an j 2 4an j 1 5an j . Hence, we take
an 4 = 4 24 5 11 = 41 and an 5 = 4 41 5 24 = 44:
Since an 6 4 44 5 41 = 29, a negative number an 6 will make the coeÆcient of

xn 4 in f (x) nonnegative. This suggests that perhaps an 6 is unnecessary (it can be 0),
so we attempt to take n = 5 and g (x) = x5 + 4x4 + 11x3 + 24x2 + 41x + 44. One obtains
that (x2 4x + 5)g(x) = x7 + 29x + 220 in this case, so that (x2 4x + 5)g(x) does indeed
have nonnegative coeÆcients.
It remains to show that g(x) 2cannot be replaced by a polynomial of degree < 5. It
suÆces to show that f (x) = (x 4x + 5)g(x) cannot be of degree < 7. Recall that
deg f 2 since g(x) is not identically 0. Suppose f (x) is a polynomial of degree 2 and
6 with nonnegative
p icoeÆcients. Note1 that the roots of x2 4x +5 are 2 i. We consider
the root 2 + i = 5e where = tan (1=2) 2 (0; 0:46365). It follows that
p p
(2 + i)k = 5k eik = 5k (cos(k) + i sin(k)):
Observe that for 1 k 6,k we have 0 < k < 6 0:46365 = 2:7819Pn< . Hence,
the imaginary part of (2 + i) is positive for 1 k 6. If f (x) = k=0 ck xk with
cn 6= 0, 2 P nn 6, and ck 0 for each k, then it follows that the imaginary part of
f (2 + i) = k=0 ck (2 + i)k is > 0. This implies that f (x) cannot have 2 + i as a root and,
therefore, cannot have x2 4x +5 as a factor. Thus, if f (x) = (x2 4x +5)g(x) with g(x)
as in the problem, then deg f 7.
Instructions: Answer as many of the problems below as you can. At the end of the time
allotted, turn in a list of your answers. Your answers should be expressed in simplest form.
Exact answers are required on all problems (not numerical approximations).
√ √
q q
1. Simplify 17 + 273 − 17 − 273.
2. What is the value of the coefficient of x2000999 in the expansion of
(x + 1)(x − 2)2 (x + 3)3 (x − 4)4 · · · (x + 1999)1999 (x − 2000)2000 ?
3. Let A, B, and C be points such that AB = 9, AC = 8, and ∠ABC = 60◦ . There are two
different possibilities for the value of BC. Suppose ` is the largest of these and s is the
smallest. Then what is the value of ` − s?
4. If a thin 10 foot ladder leans against a wall, a right triangle is formed with the hypotenuse
being the ladder and the legs being formed by the ground and the wall. If the highest
point of the ladder is taken at random between 0 and 10 feet, what is the probability that
the area of the right triangle formed is greater than 15?
5. Let P denote the point (4, 2), and consider the circle x2 + y 2 = 2. A line which passes
through P and intersects the circle in exactly one point is said to be tangent to the circle.
There are two such tangent lines. At what two points do they intersect the circle?
6. If 952000 is written in base 2, then what is the sum of its 10 right-most bits (binary digits)?
7. Observe that 6492 − 13 × 1802 = 1. There is one pair (a, b) 6= (649, 180) with each of a
and b a positive integer < 108 and with a2 − 13b2 = 1. Determine (a, b).
8. The polynomial xn + 2x2 + 3x − 4 has exactly one positive real root. Denote it by αn . As
n gets larger and larger (as n approaches infinity), the value of αn gets closer and closer
to some real number β (αn approaches β) in the sense that the value of |αn − β| can be
made arbitrarily small by taking n to be sufficiently large. What is the value of β?
Solutions are located at the website http://www.math.sc.edu/~filaseta/contests/contests.html

SOLUTIONS TO TEAM PROBLEMS
February, 2000
√ √
Answers: 1. 26 4. 10/5 7. (842401,
√ 233640)
2. −2001000
√ 5. (−1/5, 7/5) and (1, −1) 8. (−3 + 41)/4
3. 13 6. 2
√
1. Squaring the expression in question gives 26. So it must equal 26.
2. The polynomial in question is monic and has degree 1 + 2 + 3 + · · · + 2000 = 2001000. The
coefficient of x2000999 is, therefore, minus the sum of the roots counted to their multiplicity.
Noting that n2 − (n + 1)2 = −n − (n + 1), we deduce this coefficient is
12 − 22 + · · · + 19992 − 20002 = −1 − 2 − · · · − 1999 − 2000 = −2001000.
3. If x is one of the possible values of BC, then by the Law of Cosines we obtain AC 2 =
AB 2 +x2 −2xAB cos 60◦ . Simplifying, we deduce x2 −9x+17 = 0. The positive difference √
of these solutions is the value of ` − s. By the quadratic formula, we see that ` − s = 13.
4. Let h denote the height and b the base of the right triangle. Then h2 + b2 = 102 so that
b2 = 100 − h2 . We want to know the probability that (1/2)bh > 15 (where 0 ≤ h ≤ 10).
Since (1/2)bh > 15 precisely when
100 − h2 h2 = b2 h2 > 900,

we deduce that (1/2)bh > 15 precisely when h is between the two positive real roots of
x4 −√100x2 + 900.
√ Since
4 2 2 2
√ x − 100x + 900 = (x − 90)(x − 10), the two positive real roots
are 10 and 90 = 3 10 (both between 0 and 10). The probability is therefore
√
1 √ √ 10
3 10 − 10 = .
10 5
5. If (x, y) is such a point, then x2 + y 2 = 2. Furthermore, the points (0, 0), (x, y), and (4, 2)
form the vertices of a right triangle so that
x2 + y 2 + (x − 4)2 + (y − 2)2 = 42 + 22 .
This last equation is equivalent to x2 + y 2 − 4x − 2y = 0. Since x2 + y 2 = 2, we obtain

4x + 2y = 2 so that y = 1 − 2x. Using x2 + y 2 = 2 again, we obtain x2 + (1 − 2x)2 = 2 so
that 5x2 − 4x − 1 = 0. Since 5x2 − 4x − 1 = (5x + 1)(x − 1) and y = 1 − 2x, we deduce
that the two points are (−1/5, 7/5) and (1, −1).
Alternatively, one can note that such a point (x, y) must lie on the circle centered at the
midpoint of (0, 0) and (4, 2) and passing through (0, 0). It follows that (x−2)2 +(y −1)2 =
22 + 12 . Hence, x2 − 4x + y 2 − 2y = 0. Using x2 + y 2 = 2, one obtains y = 1 − 2x as before
(and follows the argument above).
6. Observe that 95 = 3 × 25 − 1. Therefore, by the binomial theorem,

2000
952000 = 3 × 25 − 1 = (−1)2000 + 2000(−1)1999 (3 × 25 ) + 210 × m,
where m is some integer (the point here is that 210 is a factor of each of the remaining
2000
terms in the binomial expansion of 3×25 −1 ). Since (−1)2000 +2000(−1)1999 (3×25 ) =
−191999, we obtain
952000 = 1024 × 188 − 191999 + 210 (m − 188) = 513 + 210 (m − 188).
It follows that the 10 right-most bits of the binary expansion of 952000 are the same as
that of 513. Since 513 = 29 + 1 = (1000000001)2 , the sum of the 10 right-most bits is 2.
7. Since 6492 − 13 × 1802 = 1, we see that

√ √
649 + 180 13 649 − 180 13 = 1.
Squaring both sides, we obtain

√ 2 √ 2
649 + 180 13 649 − 180 13 = 1.
√ 2 √ √ 2 √
Since 649+180 13 = 842401+233640 13 and 649−180 13 = 842401−233640 13,
we obtain
√ √
8424012 − 13 × 2336402 = 842401 + 233640 13 842401 − 233640 13 = 1.
Since each of 842401 and 233640 is < 108 , the answer is (a, b) = (842401, 233640).
8. Observe that if x ≥ 0.9, then xn + 2x2 + 3x − 4 > 2x2 + 3x − 4 ≥ 2(0.9)2 + 3(0.9) − 4 > 0.
Thus, any real root of xn + 2x2 + 3x − 4 must be < 0.9. In particular, 0 < αn < 0.9. It
follows that αnn approaches 0 as n approaches infinity. Since αn is a root of xn +2x2 +3x−4,
we deduce that 2αn2 +3αn −4 is approaching 0. By considering the graph of y = 2x2 +3x−4,
one sees that 2αn2 + 3αn − 4 cannot
√
2
√ αn is near a root of 2x + 3x − 4. The
be near 0 unless
2
roots of 2x +3x−4 are (−3+ 41)/4 and (−3− 41)/4, and the latter of these √ is negative
(so αn cannot be near it). It follows that αn is approaching the root (−3 + 41)/4 as n
approaches infinity.
Exact answers are required on all problems.
1. What is the maximum value of the function
3 cos θ − 2 sin θ ?
2. There are 6 people in a room. Each person randomly chooses a positive integer ≤ 20.
What is the probability that some two of the people choose the same number? Express
your answer as an exact decimal.
3. The largest known explicit example of a prime number is currently 26972593 −1. How many
digits does it have?
4. Let P be a point in an equilateral triangle with each side of length 1. Let h1 , h2 , and
h3 be the distances from P to the three sides of the triangle. What are all the possible
values for h1 + h2 + h3 ?
5. Suppose that f (x) is a polynomial of degree 5 and with leading coefficient 2001. Suppose
further that
f (1) = 1, f (2) = 3, f (3) = 5, f (4) = 7, and f (5) = 9.
What is the value of f (6)?
6. Suppose x and y are real numbers such that
2x2 + y 2 − 2xy + 12y + 72 ≤ 0.
What is the value of x2 y?

1
7. A triangle has two sides of length 5 and one side of length 6. A rectangle R is formed
with one edge on the side of length 6 and a vertex (or corner point) on each of the other
two sides. What is the maximum possible value for the area of R?
8. Let
N = 1096 − 1080 + 1064 − 1048 + 1032 − 1016 + 1.
Then
1
= 0.d1 d2 d3 . . . dr
N
(so that the block of digits d1 d2 d3 . . . dr repeats indefinitely). There is more than one
value of r for which such digits d1 , d2 , . . . , dr exist. If the smallest such r is k, then it is
known (and not so hard to show) that every such value of r is divisible by k. What is the
value of k?
2
February, 2001
√ √
Answers: 1. 13 4. 3/2 7. 6
2. 0.56395 (or 56.395%) 5. 240131 8. 224
3. 2098960 6. −432
√ 2 √ 2 √
1. Since
√ (3/ 13) + (2/ 13) = 1, there is an angle φ such that cos φ = 3/ 13 and sin φ =
2/ 13. Hence,
√

3 2
3 cos θ − 2 sin θ = 13 √ cos θ − √ sin θ
13 13
√ √
= 13 cos φ cos θ − sin φ sin θ = 13 cos(φ + θ).
√ of 3 cos θ − 2 sin θ is apparently maximized then when θ = −φ. This maximum

The value
value is 13. (An alternative approach would be to use Calculus.)
2. The probability that the 6 people choose different numbers is

19 18 17 16 15
× × × × = 0.43605
20 20 20 20 20
as the first person can choose any number, then the second person would have 19 numbers
to choose, the third 18 numbers to choose, etc. Hence, the answer is 1−0.43605 = 0.56395.
3. Let N = 26972593 . Note that N has a non-zero unit’s digit (since 5 does not divide N ).
Hence, N − 1 and N have the same number of digits. The number of digits of N is the
smallest integer larger than
log10 N = 6972593 × log10 2 = 2098959.64 . . . .
The answer is 2098960.
4. Let A, B, and C be the vertices of the equilateral triangle. Note that AB = AC = BC =

1. It follows that the areas of ∆AP B, ∆AP C, and ∆BP C are h1 /2, h2 /2, and√ h3 /2 (in
some order). Hence, h1 + h2 + h3 must be twice the area of ∆ABC, which is 3/2.
5. Let g(x) = f (x) − 2x + 1. The given information about the values of f (x) implies
g(1) = g(2) = g(3) = g(4) = g(5) = 0. Also, g(x) is a fifth degree polynomial with
leading coefficient 2001. It follows that
g(x) = 2001(x − 1)(x − 2)(x − 3)(x − 4)(x − 5).
Hence, g(6) = 2001 × 5! = 240120. Since g(6) = f (6) − 11, we deduce that f (6) = 240131.
6. Observe that
2x2 + y 2 − 2xy + 12y + 72 = (x + 6)2 + (x − y − 6)2 .
It follows that 2x2 +y 2 −2xy +12y +72 ≤ 0 if and only if x+6 = 0 and x−y −6 = 0 (strict
inequality never holds). Solving, we deduce x = −6 and y = −12. Hence, x2 y = −432.
7. The picture to the right represents one-half of

the triangle and one-half of R placed in a xy-
6
coordinate system. The point (x, y) is on the line 4
y = (−4x + 12)/3. The area A of one-half of R is S
S
therefore x(−4x + 12)/3. As x varies from 0 to 3, S
the equation St(x, y)
S
x(−4x + 12) −4x2 + 12x
S
A= = S
3 3 S
S
represents part of a parabola (in the variables S
S
x and A). Its maximum value is at its vertex t S -
(3/2, 3). The answer is therefore 2 × 3 = 6. 3
8. The given information implies

10r − 1 10r 1
= − = d1 d2 . . . dr
N N N
is an integer. Hence, N divides 10r − 1. The converse is also true; in other words, if N
divides 10r − 1, then 1/N can be written as 0.d1 d2 . . . dr (this follows by noting that the
digits after the decimal of 10r /N and 1/N must cancel in order for their difference to be
an integer). In particular, we deduce that k is the minimal positive integer such that N
divides 10k − 1 and (from the information given in the problem) if N divides 10r − 1, then
k divides r. Since
1016 + 1 N = 1016 + 1 1096 − 1080 + 1064 − 1048 + 1032 − 1016 + 1 = 10112 + 1,

we obtain that N divides 10112 + 1. Also, we deduce that N divides
10224 − 1 = 10112 + 1 10112 − 1 .

Therefore, k must divide 224. Note that N > 1095 . Hence, it is not possible for N to
divide 10r − 1 for any positive integer r ≤ 95. This implies that k = 112 or k = 224.
Since N divides 10112 + 1, we obtain that N cannot divide 10112 − 1 (otherwise, it would
divide 2 the difference between 10112 + 1 and 10112 − 1). Thus, k = 224.
Exact answers are required on all problems unless specified otherwise.
1. In the drawing (not to scale) at the right,

DC = DE, ∠ABC = 28◦ , and the points A,
B, C, and D all lie on the same circle (not
drawn). What is the value of x (in degrees)?
2. Find the largest and smallest values of

sin (2n )◦ where n is an integer satisfying
0 ≤ n ≤ 2002. Express your answers accu-
rate to five decimal digits.
3. Describe with one word the graph of x2 + xy + y 2 = 0 in the xy-plane. Be precise. For
example, “square” is precise enough but wrong, “graph” is not precise enough (and is
otherwise correct), and “round” is not precise enough (and may or may not be otherwise
correct).
4. The number 102002 − 1 is divisible by 2003. What are the 11111111th and 11111112th
digits after the decimal in the decimal expansion of 1/2003?
5. The product
2 1001 1002 1003 2002
x − 22 x − 22 · · · x − 22 x + 22 x + 22 · · · x + 22
is a polynomial of degree 2002. The first 1001 factors above contain a negative term, and
n n
the last 1001 factors do not. Interpret 22 as 2(2 ) . When the product is expanded, it
has 2003 coefficients (counting the constant term). How many of these coefficients are
positive?
1
6. The polynomial x4 + x3 + x2 + x + 1 has one root of the form a + bi where a and b are
positive real numbers and i2 + 1 = 0. If
ar + bs = cos(3◦ ) and r = cos(75◦ ),
what is the exact value of r + s? Express your answer in a form that does not use
trigonometric functions.
7. It is known that the largest integer n such that each prime factor of n(n + 1) is ≤ 11 is
n = 9800. What is the largest integer y having each prime factor ≤ 11 and satisfying the
equation x2 − 2y 2 = 1 for some integer x?
8. A person is said to be n years old, where n is a nonnegative integer, if the person has
lived at least n years and has not yet lived n + 1 years. At some point in time, Tammy
is 4 years old and John is three times as old as Martha (so if John is j years old and
Martha is m years old, then j and m are nonnegative integers with j = 3m). At some
other time, Martha is twice as old as Tammy and John is five times as old as Tammy. At
yet another time, John is twice as old as Martha and Tammy is T years old. There are
different possibilities for what T can be. What is the largest possible value of T ?
2
February, 2002
Answers: 1. 76◦ 4. 2 and 3 7. 13860

2. 0.89879 . . . and −0.99939 . . . √
5. 1502 √ r 8. 22
3. “point” or “origin” or (0, 0) 6 3 3
6. or √ or
2 2 2
1. From the given information, ∠DCE = ∠DEC = x. Hence, ∠ADC = 2x. Since A, B, C,
and D all lie on the same circle, the angles ∠ADC and ∠ABC are supplementary angles.
Therefore, 2x = ∠ADC = 180◦ − ∠ABC = 180◦ − 28◦ = 152◦ , so x = 76◦ .
2. The numbers 23 and 215 are congruent modulo 360 (that is, their difference is divisible
n ◦ n−12 ◦
by 360). Hence, the value
of sin (2 ) will be the same as sin (2 ) for n ≥ 15. In
n ◦
other words, sin (2 ) only takes on different values for 0 ≤ n ≤ 14. The maximum
value of sin (2n )◦ occurs when n = 6 and it is 0.89879 . . . . The minimum value occurs
when n = 13 and it is −0.99939 . . . .
√ √
3. The graph is a point (the origin). Set α = (−1 + −3)/2 and β = (−1 − −3)/2
(the roots of x2 + x + 1 = 0). Then x2 + xy + y 2 = (x − αy)(x − βy). It follows that
x2 + xy + y 2 = 0 if and only if x = αy or x = βy. Since α and β are imaginary, either of
these happens for real x and y only in the case that x = y = 0.
4. The problem is equivalent to finding the first two digits after the decimal in the decimal
expansion of 1011111110 /2003. The number 11111110 has a remainder of 10 when divided
by 2002. Hence, there is an integer k such that 11111110 = 2002k + 10. Since 102002 − 1
is divisible by 2003 and 102002 − 1 is a factor of 102002k − 1, 2003 divides 102002k − 1. Thus,
102002k − 1 = 2003m for some integer m. We deduce
1011111110 102002k+10 (102002k − 1) · 1010 1010

= = + = 1010 m + 4992511.23315 . . . .
2003 2003 2003 2003
Hence, the 11111111th and 11111112th digits after the decimal in the decimal expansion
of 1/2003 are 2 and 3, respectively.
5. Let f (x) = a2002 x2002 + a2001 x2001 + · · · + a1 x + a0 be the expansion of

2 1001 1002 1003 2002
x − 22 x − 22 · · · x − 22 x + 22 x + 22 · · · x + 22 .
Observe that each aj can be expressed as a sum of distinct terms of the form ε2w where w
is a sum of n − j distinct powers of 2 (appearing in the exponents of the product above)
and ε = ±1. For any integer m > 0,
j w sign
2m > 2m−1 + 2m−2 + · · · + 2 + 1.
2002 0 +
2002
It follows that the sign of aj is deter- 2001 2 +
2002
mined only by the sign of the term 2000 2 + 22001 +
ε2w in aj with w as large as pos- ..
.
..
.
..
.
sible. The table to the right indi-
1001 22002 + 22001 + · · · + 21002 +
cates the term in aj with w as large
as possible. The sign of each term 1000 22002 + 22001 + · · · + 21002 + 21001 −
and the value of w are given. The 999 22002 + 22001 + · · · + 21001 + 21000 +
first 1001 rows correspond to posi- 998 22002 + 22001 + · · · + 21000 + 2999 −
tive terms. The next 1002 rows al- .. .. ..
. . .
ternate in sign so that they include
501 positive terms and 501 negative 1 22002 + 22001 + · · · + 23 + 22 +
terms. The total number of positive 0 22002 + 22001 + · · · + 22 + 2 −
coefficients is therefore 1502.
6. The number a + bi is a root of x5 − 1 = (x − 1)(x4 + x3 + x2 + x + 1). The root of

x5 − 1 with positive real and imaginary parts is cos(72◦ ) + i sin(72◦ ). So a = cos(72◦ ) and
b = sin(72◦ ). From the given information,
cos(72◦ ) cos(75◦ ) + sin(72◦ )s = cos(3◦ ).
This uniquely determines s and from the identity cos(A − B) = cos A cos B + sin A sin B
with A = 75◦ and B = 72◦ , we see that s = sin(75◦ ). Hence,
r + s = cos(75◦ ) + sin(75◦ )
√
◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ 6
= cos(45 ) cos(30 )− sin(45 ) sin(30 )+ sin(45 ) cos(30 )+ sin(30 ) cos(45 ) = .
2
7. The equation x2 − 2y 2 = 1 implies that x is odd. Write x = 2n + 1. Then
2y 2 = (2n + 1)2 − 1 = 4n2 + 4n = 4n(n + 1).
Since y has all of its prime factors ≤ 11, so does n(n + 1). The information in the problem
implies now that
y 2 ≤ 2n(n + 1) ≤ 2 · 9800 · 9801 = 24 · 34 · 52 · 72 · 112 .
Hence, y ≤ 22 ·32 ·5·7·11 = 13860. With y = 13860 and x = 2n+1 = 2·9800+1 = 19601,
it is easy to see that x2 − 2y 2 = 1 is in fact satisfied. Hence, the answer is 13860.
8. We use numbers θ1 , θ2 , . . . to represent parts of years so that each θj ∈ [0, 1). We will use
ε1 , ε2 , . . . to represent numbers that are either 0 or 1. At some point, Tammy has lived
4 + θ1 years, John has lived j + θ2 years, and Martha has lived m + θ3 years where j and
m are integers satisfying j = 3m. At another time, say a + θ4 years later where a is an
integer, Martha is twice as old as Tammy and John is five times as old as Tammy. Since
Tammy will have lived 4 + a + θ1 + θ4 years, she will be 4 + a + ε1 years old (where, as
noted earlier, εj ∈ {0, 1}). Similarly, John will be j + a + ε2 years old and Martha will
be m + a + ε3 years old. Hence,
(∗) m + a + ε3 = 2(4 + a + ε1 ) and j + a + ε2 = 5(4 + a + ε1 ).
At yet another time, say b + θ5 years after Tammy lived 4 + θ1 years, we obtain Tammy is
T = 4+b+4 years old and j +b+ε5 = 2(m+b+ε6 ). This last equation and j = 3m imply
b = j − 2m + ε5 − 2ε6 = m + ε5 − 2ε6 . From the first equation in (∗), m = a + 8 + 2ε1 − ε3 .
From the second equation in (∗) and j = 3m, we obtain 3m = 4a+20+5ε1 −ε2 . Therefore,
m = 4m − 3m = 4(a + 8 + 2ε1 − ε3 ) − (4a + 20 + 5ε1 − ε2 ) = 12 + 3ε1 − 4ε3 + ε2 .
Combining the above, we deduce T = 16 + 3ε1 − 4ε3 + ε2 + ε5 − 2ε6 + 4 . Since each

εj ∈ {0, 1}, we obtain that T ≤ 16 + 3 + 1 + 1 + 1 = 22. One checks that t = 4.6, j = 48.6,
m = 16, a = 6.5, and b = 17.5 lead to the conditions in the problem holding with T = 22,
so 22 is the answer.
Team Problems January, 2003
Exact answers are required on all problems unless specified otherwise.
1. Let
√
√ 2
√ 2
√ 2
√ 2 √
√ 2 √ 3
√ 2 √ 3
√ 2 √ 3
√ 2 √ 3
√ √
√ 2 √ 3
√ 2
2
10100 22
22
2
√ 3
3
A= 2 , B = 10 , C=2 , and D = 3 .
Order these from least to greatest. Express your answer in the form A < B < C < D.
4
(Note that exponentiation is done from the top down. For example, 23 = 281 .)
2. Three married couples, Alan and Alice, Bernice and Bob, and Calvin and Cathy, attend
a dinner. They shake hands. No one shakes hands with him-or-herself. No one shakes
hands with his-or-her spouse. And no two people shake hands more than once. Bob
doesn’t like Calvin, so the two of them don’t shake hands with each other. On the other
hand, Alan does shake hands with Calvin. After everyone finishes shaking hands, Alan
asks everyone else, including his wife, how many hands they shook. Each person gives him
a different answer. Assuming everyone has given a correct response to Alan, determine
how many hands each person shook.
3. If p is a prime, a an integer > 1, and q a prime dividing ap − 1, then it is known that

either p divides q − 1 or q divides a − 1. What is the smallest integer > 1 dividing 229 − 1?
4. How many times does the graph of
y = (x − 1)(x − 2)(x − 3) · · · (x − 2002)(x − 2003) + 1
cross the x-axis in the interval [1000, 1005]?

1
√
5. The base 10 number 2 can be converted to base 2 as 1.0110101 . . . . This means that
√ 1 1 1 1
2=1+ 2
+ 3 + 5 + 7 + ··· .
2 2 2 2
√
The base 2 digits of the base 10 number 2 are, therefore, 1, 0, 1, 1, 0, 1, 0, 1, . . . . Suppose
we wish to convert a base 10 number to base 26. Let the digit 0 be represented by A, the
digit 1 be represented by B, the digit 2 be represented by C, and so on. In other words,
the digits in base 26 are to be represented by the alphabet in its usual order. What are
√
the first four letters (digits) when the base 10 number 29085 is converted to base 26?
6. Two circles C0 and C1 , each with radius 1, are

tangent to each other and to line `. Circle C2
is tangent to C0 , C1 , and to ` as shown. Circle C0 C1
C3 is tangent to C1 , C2 , and ` as shown. What
C2
is the radius of C3 ? C3
7. For how many a ∈ {1, 2, . . . , 200} does the polynomial
x2003 + ax10 + 200
have a root (real or imaginary) with absolute value ≥ 1.003?
8. For n a positive integer, define

1 1 1 1
Sn = 1 + + + ··· + +
2 3 n−1 n
and
1 1 1 1
Tn = + + ··· + + .
n+1 n+2 2n − 1 2n
It is known that Sn / log n approaches 1 as n tends to infinity. Here, the base of the
logarithm is the number e (which you do not need to have prior knowledge of for this
problem). It is also known that Tn approaches a constant as n tends to infinity. What is
that constant?
2
January, 2003
Answers: 1. A < B < C < D 3. 233

2. Alan shakes 2 hands 4. 6
Alice shakes 2 hands 5. GOOD or GO.OD or good or go.od
Bernice shakes 4 hands 6. 1/9
Bob shakes 0 hands 7. 3
Calvin shakes 3 hands 8. log 2 or loge 2 or ln 2
Cathy shakes 1 hand
1. Let f (x) = log log x. The base of the logarithm is not important. We will use the natural
logarithm here. Since f (x) is an increasing function, we may consider f (A), f (B), √ f (C),
and f (D) instead of A, B, C, and D. The number A is a tower (say) of twelve 2’s. It
√ √ 2
is increased by replacing the highest exponent 2 (at the top) with 2. Since 2 = 2, we
deduce A ≤ 2 so that f (A) ≤ log log 2 < 0. Note that

f (B) = log log B = log 10100 log 10 = 100 log 10 + log log 10 = 231.09 . . . .
Similarly,
22
f (C) = 22 log 2 + log log 2 = 216 log 2 + log log 2 = 45425.727 . . . .
√ t √ √ √
Also, f (D) = 3 log 3 + log log 3 where t consists of a tower of five 3’s. The value
of t can be estimated with a calculator by considering
√
√
√ 3
√ 3
3 √
log t = 3 log 3 = 5.359 . . . .
Thus,
√ e5.359 √ √
f (D) > 3 log 3 + log log 3 > 2 × 1050 .
Thus, f (A) < f (B) < f (C) < f (D). The answer is A < B < C < D (no wonder the
problem said to put the answer in this form).
2. Let h(x) denote the number of hands person x shakes. The most hands that anyone
shakes in this problem is 4. Since Alice, Bernice, Bob, Calvin and Cathy shook a different
number of hands, they shook 0, 1, 2, 3, and 4 hands in some order. Also, the given
information in the problem implies
h(Bob) ≤ 3, h(Alan) ≥ 1, and 1 ≤ h(Calvin) ≤ 3.

Observe that Alice cannot shake 4 hands since then nobody shakes 0 hands. Also, if
Cathy shakes 4 hands, then nobody shakes 0 hands. It follows that Bernice shakes 4
hands. From this, we deduce that Bob shakes 0 hands (since Alice, Calvin and Cathy
shake hands with Bernice). Since Calvin shakes hands with both Alan and Bernice,
h(Calvin) ≥ 2. Now, h(Alice) 6= 1 since otherwise each of Calvin and Cathy do not shake
hands with either Alice or Bob so that no one shakes 3 hands. Thus, Cathy shakes 1 hand,
apparently Bernice’s. This means that Alice cannot shake 3 hands so that h(Alice) = 2
and h(Calvin) = 3. One now checks that h(Alan) = 2 (he shakes hands with Bernice and
Calvin and nobody else).
3. If an integer m > 1 divides 229 − 1, then so does each of m’s prime factors. It follows that
the smallest integer m > 1 that divides 229 − 1 is a prime. Taking this prime to be q and
taking a = 2 and p = 29 in the information given in the problem, we deduce that q − 1 is
divisible by 29. Since q is prime, we only need consider odd q. The odd numbers that are
one more than a multiple of 29 are 59, 117, 175, 233, . . . . One checks directly that 229 − 1
is not divisible by 59. The numbers 117 and 175 are not prime (the first is divisible by 3
and the second by 5). Finally, one checks directly that 229 − 1 is divisible by 233, so this
is the answer.
4. Let
f (x) = (x − 1)(x − 2)(x − 3) · · · (x − 2002)(x − 2003) + 1.
Since f (x) has degree 2003, the graph of y = f (x) crosses the x-axis at most 2003 times.
Observe that f (k) > 0 for each k ∈ {1, 2, . . . , 2003}. Also,
f (0.5) < 0, f (2.5) < 0, f (4.5) < 0, . . . , f (2002.5) < 0.
Therefore, the graph of y = f (x) must cross the x-axis at least once in each of the intervals
(0.5, 1), (2, 2.5), (2.5, 3), (4, 4.5), (4.5, 5), . . . , (2002, 2002.5), (2002.5, 2003).
Since this consists of 2003 intervals, we deduce that the graph of y = f (x) crosses the
x-axis exactly once in each of the intervals above and nowhere else. It follows that the
graph crosses the x-axis 6 times in the interval [1000, 1005].
5. We want to write the number

√
29085 = 170.5432496 . . .
in base 26. Since 170 = 6 · 26 + 14, the first two digits come from converting 6 and 14
to letters, so they are G and O, respectively. Since 14/26 = 0.53846 . . . and 15/26 =
0.57692 . . . , we obtain that the next digit corresponds to 14 and, hence, is the letter O.
Next, we use that

14
170.5432496 · · · − 6 · 26 + 14 + = 0.004788 . . . .
26
Since 3/262 = 0.0044 . . . and 4/262 = 0.0059 . . . , the fourth digit corresponds to 3 and,
hence, is the letter D. Thus, the answer is GOOD.
6. Consider the two circles tangent to each other and a
line as shown to the right. Suppose the smaller circle has
radius r and the larger circle radius R. Then the shaded
triangle is a right triangle with hypotenuse R + r and
one leg of length R − r. It follows that if the “horizontal
distance between these circles” (that is the length of the
remaining leg) is h, then
√
h2 = (R + r)2 − (R − r)2 = 4Rr =⇒ h = 2 Rr.
In the problem, let h(i, j) denote the horizontal dis-

tance between Ci and Cj . Let r denote the radius
of√C2 . Since
√ h(0, 1) = h(0, 2) + h(1, 2), we deduce from the formula for h above that 2 =
0
2 r+2 r so that r = 1/4. Let r denote the radius √ of C3 . √
Since h(0, 2) = h(0, 3)+h(2, 3),
we deduce from the formula for h above that 1 = 2 r0 + r0 so that r0 = 1/9.
7. Let f (x) = x2003 + ax10 + 200 where a ∈ {1, 2, . . . , 200}. If z is a complex number with
|z| ≥ 1.003, then
|f (z)| ≥ |z|2003 − a|z|10 − 200 = |z|10 |z|1993 − a − 200 ≥ 1.03 391.525 − a − 200.

For a ≤ 197, this last expression is > 0 so that f (z) 6= 0. In other words, f (x) has no
roots with absolute value ≥ 1.003 unless a > 197. For a ∈ {198, 199, 200}, one checks
directly that the value of f (−1.003) > 0 and f (−2) < 0 so that f (x) in fact has a negative
real root between −1.003 and −2. The answer, therefore, is 3.
8. One can use that if n is large, then Sn ≈ log n and S2n ≈ log(2n) = log n + log 2 so that
Tn = S2n − Sn ≈ log 2. This idea is not a true explanation, though, as we are not given
sufficient information in the problem to determine how good an approximation log n is
to Sn . We give a more precise explanation as follows. Let c denote the constant that Tn
approaches as n tends to infinity. Let k be a positive integer. We use that
1 + T1 + T2 + T4 + T8 + · · · + T2k−1 = S2k .
We rewrite this as
1 + T1 + T2 + T4 + T8 + · · · + T2k−1 k + 1 1 S2k S2k
· · = = .
k+1 k log 2 k log 2 log 2k
Observe that the first fraction on the left is simply the average of 1, T1 , T2 , T4 , . . . , T2k−1
and, as k gets large, this average approaches c. Also, as k gets large, (k + 1)/k and the
last fraction above approach 1 (the latter by what is given). We deduce that c/ log 2 = 1
so that c = log 2.
Team Problems November 15, 2003
1. Compute
(log2 3)(log3 4)(log4 5) · · · (log2002 2003)
accurate to 4 digits after the decimal point.
2. In the figure to the right (not drawn to scale), F

A
4ABC and 4DEF are equilateral triangles
each having side lengths equal to one unit. If
D is in the center of 4ABC and E is on line
←→ D
BC, then what is the distance from C to E?
Give an answer accurate to 6 digits after the
decimal point. B C E
3. Two players play a game on the board below as follows. Each person takes turns moving
the letter A either downward at least one rectangle or to the left at least one rectangle
(so each turn consists of moving either downward or to the left but not both). The first
person to place the letter A on the rectangle marked with the letter B wins. How
should the first player begin this game if we wants to assure that he wins? Answer with
the number given on the rectangle that he should move the letter A to.
1 2 3 4 5 6 7 8 A
9
10
11
B 12
1
4. Find every positive integer N for which 2003 ≤ N ≤ 2500 and x4 − y 4 = N holds for
some integers x and y.
5. Two positive integers a and b each ≤ 100 are chosen independently at random (so it is
possible that a = b). What is the probability that gcd(a, b) (the greatest common divisor
of a and b) is divisible by at least one of 11, 13, 23, and 31? Give an exact answer written
in decimal notation.
6. The polynomial
x2003 − x2002 − 2003 x2001 − x2 − 2003
has one positive real root. What is the value of this root accurate to 5 digits after the
decimal point?
7. How many consecutive zeroes appear after the decimal point and before the first non-zero
digit after the decimal point in the decimal expansion of
√
22004 + 1 ?
8. The equation

1! 2! 3! 4! 2003! 2004! 1 2005!
+ + + + ··· + + =A· − ,
2003! 2004! 2005! 2006! 4005! 4006! 2002! 4006!
holds for some rational number A. Find the value of A and express your answer in the
form a/b where a and b are positive integers < 4000.
2
November 15, 2003
Answers: 1. 10.9679 5. 0.0155

2. 0.457427 6. 45.25768 . . .
3. 5 7. 301
4. 2145, 2320, 2385, 8. 1/2001
2400, 2401, & 2465
1. As logb a = log a/ log b (where log x can be the logarithm to any fixed base), we deduce
that
log 3 log 4 log 5 log 2002 log 2003
(log2 3)(log3 4) · · · (log2002 2003) = × × × ··· × ×
log 2 log 3 log 4 log 2001 log 2002
log 2003
= = 10.9679467 . . . .
log 2
←→ ←→
2. Let G be the midpoint of segment BC so line DG is perpendicular to line BC.√ Since
4BDG is a 30−60−90 degrees triangle with BG = 1/2, we √ see that DG = p
1/(2 3). Since
DE
p = 1 and 4DGE is a right triangle, we deduce
p GE = DE 2 − DG2 = 1 − (1/12) =
11/12. Since GC = 1/2, we obtain CE = 11/12 − 1/2 = 0.457427 . . . .
3. The first person should begin by moving A to the rectangle numbered 5. In fact, the first
person can win by, on each turn, putting A on a rectangle along the diagonal from the
rectangle labelled B to the rectangle numbered 5. The second player will have to move
off this diagonal and then the first player can continue to move on the diagonal. Since
the rectangle labelled with B is on the diagonal, the first person will eventually win (with
this strategy). Observe that if first player does not put A on the rectangle numbered 5,
then the second player can force a win by using the above strategy (putting A along the
diagonal).
4. We may suppose that x ≥ 0 and y ≥ 0 and do so. Clearly, we must have y < x. Observe
that 64 < 2003 < 74 , so x ≥ 7. If x ≥ 10, then x4 − y 4 ≥ 104 − 94 > 3000. So x ∈ {7, 8, 9}.
Since 74 − 54 < 2000, if x = 7, we must have 0 ≤ y ≤ 4. Trying these values of y, we see
that x4 − y 4 can equal 2145, 2320, 2385, 2400, and 2401. Since 84 − 74 = 1695 < 2003
and 84 − 64 = 2800 > 2500, there are no N satisfying the conditions in the problem
when x = 8. Since 94 − 84 = 2465 and 94 − 74 > 4000, 2465 is the only N satisfying the
conditions in the problem when x = 9. The answer is 2145, 2320, 2385, 2400, 2401, and
2465.
5. There are b100/11c = 9 positive multiples of 11 that are ≤ 100, b100/13c = 7 positive
multiples of 13 that are ≤ 100, b100/23c = 4 positive multiples of 23 that are ≤ 100, and
b100/31c = 3 positive multiples of 31 that are ≤ 100. We deduce that gcd(a, b) is divisible
by 11 precisely when each of a and b is divisible by 11 which occurs for 92 = 81 pairs (a, b).
Similarly, gcd(a, b) is divisible by 13 for exactly 72 = 49 pairs (a, b), gcd(a, b) is divisible
by 23 for exactly 42 = 16 pairs (a, b), and gcd(a, b) is divisible by 31 for exactly 32 = 9
pairs (a, b). Observe that there are no positive integers ≤ 100 divisible by two of 11, 13,
23, and 31. Hence, there are exactly 81 + 49 + 16 + 9 = 155 pairs (a, b) with gcd(a, b)
divisible by at least one of 11, 13, 23, and 31. Thus, the probability is 155/1002 = 0.0155.
6. Let f (x) be the polynomial in the problem. Thus, f (x) = (x2 −x−2003)x2001 −x2 −2003.
Let √
α = (1 + 8013)/2 = 45.25768 . . .
so that α is a root of the quadratic x2 − x − 2003. Clearly, f (α) < 0. One checks directly
that f (45.257681) is positive (and, in fact, > 103000 ). It follows that f (x) must have a
root between 45.25768 and 45.257681. Thus, the answer is 45.25768 . . . .
7. Observe that

√ 1 1 1
22004 + 1 − 21002 = √ ∈ ,
22004 + 1 + 21002 2.1 · 21002 2 · 21002

⊆ 1.1110281 . . . · 10−302 , 1.1665795 . . . · 10−302 .
Thus, the answer is 301.
8. For 2 ≤ n ≤ 2005, we use that

(n − 1)! 1
=
(n + 2001)! n(n + 1)(n + 2) · · · (n + 2001)

1 1 1
= − .
2001 n(n + 1) · · · (n + 2000) (n + 1)(n + 2) · · · (n + 2001)
The sum in the problem is therefore
2005
1 X 1 1
− .
2001 n=2 n(n + 1) · · · (n + 2000) (n + 1)(n + 2) · · · (n + 2001)
This is a telescoping sum with the second part of each term cancelling with the first part
of the next (except in the case of the last term). Hence, the sum is

1 1 2005!
− .
2001 2002! 4006!
The answer is therefore 1/2001 (and this answer is unique).

Team Problems March 5, 2005
1. Circles of radius one unit each

unit circles
are placed around the outside and
large circle
tangent to a circle of radius 2005
units with no two of the smaller
circles overlapping. What is the
maximum number of unit circles
that can be used?
2. The equation
cos(18θ) = 131072 cos18 θ − 589824 cos16 θ + 1105920 cos14 θ − 1118208 cos12 θ

+ 658944 cos10 θ − 228096 cos8 θ + 44352 cos6 θ − 4320 cos4 θ + 162 cos2 θ − 1,
Y
9

is a valid trigonometric identity. Determine the exact value of cos (2k − 1) · 5◦ .
k=1
3. What is the greatest common divisor of the 2010 digit and 2005 digit numbers below?
2222222
| . . . 22222}
{z 77777
| . . . 7777}
{z
2010 2’s 2005 7’s
4. What is the coefficient of x2005 in
(x+1)7 (x2 +1)4 (x4 +1)5 (x8 +1)(x16 +1)(x32 +1)(x64 +1)(x128 +1)(x256 +1)(x512 +1)(x1024 +1) ?
5. Find the smallest positive angle θ in degrees satisfying
sin2 (2004 θ) + cos2 (2005 θ) = 1.
Express your answer as a reduced rational number of degrees.

1
6. Let A = (0, 0) and B = (100, 0). For each point C in the plane with the area of 4ABC
←→
equal to 2005, consider the point D on line BC with segment AD an altitude to 4ABC.
The set of all such points D together with the point A enclose a region in the plane. What
is the area of this region? Give an exact answer.
7. Consider the powers of 2, beginning with 2, which have a leading digit 1. The first few
are 24 = 16, 27 = 128, 210 = 1024, and 214 = 16384. How many such powers of 2 are
there ≤ 20052005 ?
8. In 4ABC (not drawn to scale), the altitude from A, the angle bisector of ∠BAC, and
the median from A to the midpoint of BC divide ∠BAC into four equal angles. What is
the measure in degrees of angle ∠BAC?
B C
2
March 5, 2005
Answers: 1. 6302
√ √ √ 5. (180/4009)◦ or 180/4009
2. 1/ 131072 or 1/(256 2) or 2/512 6. 2500π
3. 77777 7. 6620
4. 8192 8. 90◦
1. Consider a tight fitting of the smaller unit circles around the larger circle of radius 2005
so that each smaller circle is tangent to two adjacent circles and the larger circle. Let
A and B denote the centers of some two adjacent circles, and let C denote the center of
the circle with radius 2005. Then 4ABC is isosceles. Letting D be the point on AB
with AD an altitude of 4ABC, we see that 4ADC is a right triangle, and sin ∠ACD =
1/2006. We deduce that ∠ACD = sin−1 (1/2006) = 0.0285622◦ . It follows that ∠ACB =
0.0571244 . . .◦ . With this tight fit, the number of smaller circles around the larger circle is,
therefore, the greatest integer ≤ 360/0.0571244 · · · = 6302.035557 . . . . This implies that
we can fit 6302 unit circles and no more around the larger circle of radius 2005 with some
room to spare (so that, in particular, if you do not want the nonoverlapping unit circles
to even touch at a point, we have room to move them ever so slightly so that this is the
case). The answer is 6302. (Note that the computations above are approximations that
should be dealt with carefully. The number 6302.035557 . . . should really be replaced by
π/ sin−1 (1/2006) = 6302.0346 . . . .)
2. Letting θ = (2k −1)·5◦ for k an integer in [1, 9], we see that 0 < θ < 90◦ and
cos(18θ) = 0.
◦
From the information given, we deduce that the 9 numbers cos (2k−1)·5 with 1 ≤ k ≤ 9
are distinct positive real roots of
f (x) = 131072x18 − 589824x16 + 1105920x14 − 1118208x12 + · · · − 4320x4 + 162x2 − 1.
The polynomial f (x) above is merely the expression on the right-hand side of the display
in the problem with cos θ replaced with x so that, in particular, f (x) isa polynomial in
x2 (i.e., each exponent of x is even). We deduce that − cos (2k − 1) · 5◦ with 1 ≤ k ≤ 9
are distinct negative real roots
of f (x). As f (x) is of degree 18, we obtain that the 18
numbers cos (2k − 1) · 5 and − cos (2k − 1) · 5◦ with 1 ≤ k ≤ 9 are all the roots
◦
of f (x). If P denotes the product in the problem, we see that −P 2 is the product of
these 18 roots. On the other hand, the product of these roots must also be the constant
2
coefficient of f (x)
√ divided by the√leading√coefficient of f (x). Hence, −P = −1/131072
so that P = 1/ 131072 = 1/(256 2) = 2/512.
3. Let A denote the 2010 digit number that is a string of 2’s, and let B denote the 2005 digit
number that is a string of 7’s. We consider also the numbers A0 = A/2 and B 0 = B/7,
each of which is a string of 1’s. Observe that B is odd, so 2 does not divide B. On the
other hand, the divisor 7 of B is a divisor of A as can be seen as follows. The number
111111 is divisible by 7 so that any string of a multiple of 6 number of 1’s is also divisible
by 7 (for example, 111111111111111111 = 111111 · 1012 + 111111 · 106 + 111111 is divisible
by 7 since each term on the right is). As 2010 is a multiple of 6, we deduce A0 and,
hence, A is divisible by 7. Thus, 7 is a common factor of A and B. Now, the equation
A0 − 105 B 0 = 11111 implies any divisor of A0 and B 0 must also be a divisor of 11111.
Arguing in a manner similar to before, we see that since 2010 and 2005 are both multiples
of 5, the number 11111 divides both A0 and B 0 . Thus, the greatest common divisor of A0
and B 0 is 11111. Note that 7 does not divide 11111. We obtain from the definition of A0
and B 0 that the greatest common divisor of A and B is 7 · 11111 = 77777.
4. Let g(x) = (x + 1)6 (x2 + 1)3 (x4 + 1)4 , and let
h(x) = (x + 1)(x2 + 1)(x4 + 1) · · · (x256 + 1)(x512 + 1)(x1024 + 1)

= 1 + x + x2 + x3 + x4 + x5 + · · · + x2046 + x2047 ,
where the last equation follows by using that each positive integer can be written uniquely
as a sum of powers of 2. If f (x) is the polynomial in the problem, then f (x) = g(x)h(x),
so we want to know the coefficient of x2005 in the product g(x)h(x). Writing g(x) =
b28 x28 + b27 x27 + · · · + b2 x2 + b1 x + b0 , we see that the coefficient of x2005 in g(x)h(x)
can be obtained by taking each bk (the coefficient of xk in g(x)) times 1 (the coefficient
of x2005−k in h(x)) and adding these numbers together. In other words, the answer to
the problem is b28 + b27 + · · · + b2 + b1 + b0 . On the other hand, this number is just
g(1) = 26 · 23 · 24 = 213 = 8192.
5. Using that 1 − sin2 (2004 θ) = cos2 (2004 θ), the given equation can be rewritten as
cos2 (2005 θ) = cos2 (2004 θ) which is equivalent to cos(2005 θ) = ± cos(2004 θ). As cos x is
decreasing for x ∈ [0, π], we have cos(2005 θ) 6= + cos(2004 θ) for θ ∈ (0, π/2005]. We find
the θ ∈ (0, π/2005] for which cos(2005 θ) = − cos(2004 θ). Equality occurs if and only
if 2005 θ = π/2 + t and 2004 θ = π/2 − t for some positive t (in the appropriate range).
Solving for θ (for example, by adding these equations) we obtain θ = π/4009. Hence, the
answer (converting to degrees as requested) is (180/4009)◦ .
6. For each D constructed in the problem, ∠ADB = 90◦ so that the point D lies on the
circle with diameter AB. As the line through A and B rotates about the point B, there
are two points C on the line for which 4ABC has area 2005 and a unique point D on
←→
line BC (independent of the C chosen) such that AD is an altitude to 4ABC. As the
line rotates, the curve traced out by D is a complete circle except the point A is missing
(when the line being rotated is on the x-axis). The set of points D together with the point
A, therefore, form a circle with radius 50. The area of the circle, and the answer to the
problem, is π · 502 = 2500π. Note that 2005 may be replaced by any positive real number
without affecting the answer to the question; the described region is still the same circle
independent of the fixed value for the area.
7. The positive integers having leading digit 1 are precisely the integers in the intervals
[10k , 2 · 10k ) where k ∈ {0, 1, 2, . . . }. For each such k, there is exactly one power of 2 lying
in the interval [10k , 2 · 10k ) (to see this for k ≥ 1, consider the largest power of 2 that is
< 10k and double it; convince yourself that this number is a power of 2 in [10k , 2 · 10k )
and that the next power of 2 is outside this interval). As the problem concerns powers of
2 beginning with 2, we consider only k ≥ 1. We are interested in the number of intervals
[10k , 2 · 10k ) with 2 · 10k ≤ 20052005 but also need to consider whether 20052005 itself lies
in such an interval. As

log10 20052005 = 2005 log10 2005 = 6620.7393 . . . ,
we see that 20052005 = 106620 100.7393... = 5.486 . . . · 106620 . We deduce that 20052005 does
not have leading digit 1 (so it is not in an interval of the from [10k , 2 · 10k )) and that the
k of interest to us are the positive integers ≤ 6620. Hence, the answer is 6620.
8. Label the points along BC that are on the altitude from A, on the angle bisector of
∠BAC, and on the median from A as D, E and F , respectively. Let θ = ∠BAD so
that we are interested in the value of 4θ. The triangle 4BAD is a right triangle so that
∠ABC = 90◦ − θ. Also, ∠ACB = 90◦ − 3θ. We apply the Law of Sines to the triangles
4ABF and 4ACF . Letting s = AF and t = BF = F C, we deduce
cos θ sin(90◦ − θ) sin ∠ABF sin ∠BAF sin(3θ)

= = = =
s s s t t
and
cos(3θ) sin(90◦ − 3θ) sin ∠ACF sin ∠CAF sin θ

= = = = .
s s s t t
We obtain that cos θ/ sin(3θ) = s/t = cos(3θ)/ sin θ so that
sin(2θ) = 2 sin θ cos θ = 2 sin(3θ) cos(3θ) = sin(6θ).
As 4θ = ∠BAC < 180◦ , we have θ < 45◦ . It follows that sin(2θ) = sin(6θ) only if
2θ = 180◦ − 6θ. Therefore, 4θ = 90◦ .

South Carolina ARML Mathematics

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

South Carolina ARML Mathematics

Hochgeladen von

Copyright:

Verfügbare Formate

1994 TEAM PROBLEMS

T7. Calculate the largest integer which is less than

Q. Let O ! represent the origin.

to DBA. Hence, AB=AD = AE=AC so

formation implies that AB AC = 1 and

T4. For any two real numbers x and y, de ne

f (x; y) = x2 + 13y2 , 6xy , 4y , 2:

T5. If the value of

y = x12 + x10 + x8 + x6 + x4 + x2 + 1 and y = x5 + x3 + x

T10. Let a and b be integers for which

25 , x = (4=7)25 = 100=7. Therefore, the A

T2. (1 + 2 + 4 + 8)(1 + 3 + 17 + 19)(1 + 7 + 31 + 61) = 15 40 100 = 60000

centered at (0; 3) is such that the measure of

p of C . The area of C is 9. Since

where k is an integer. Then

T3. Let 1 , 2 , and 3 denote the roots of x3 , x + 1 = 0. Calculate the value of

Simplify your answer. It should be an integer.

T5. A pentagon (not necessarily a regular pentagon) has vertices A, B , C , D, and E .

(x2 , x + 1)3(x3 + 2x2 + 2x + 1)5 = a21 x21 + a20 x20 + + a1 x + a0 :

What is the value of a1 + a2 + + a10 ?

T7. For k  1, de ne ak by ak = 2k + 3. Thus, a1 = 5; a2 = 7; a3 = 11; : : : . It is not

a1 + a2 + + a10 = 12 (a0 + a1 + + a21 , 2a0 ) = 12 (F (1) , 2) = 12 (65 , 2) = 3887:

For example, x , 2x , 2x + 1 and 3x , 4x + 3 are reciprocal polynomials. Consider all

T8. Let S = 2n . Suppose that the decimal expansion of S is given by S =

1 = cos(60 ) = cos(6) = g (cos()) = 32u , 48u + 18u , 1:

Thus, for u = cos() with  = 10 , we have

It follows that 2 cos(10 ) = 2u is a root of x , 6x + 9x , 3 . (The answer is unique.)

the three dice are:

Thus, g(x) is a factor of w(x) = x , x 1234 1233

are factors. Since x + x + 1 is a factor of x , 1 and x , 1 is a factor of

we get x + x + 1 is a factor of u(x).

is a Calculus approach; a non-Calculus approach follows.)

is 0. The Binomial Theorem implies, however, that the coecient of x in u(x + 1) =

On the other hand, for n  4, the value of 10 =2n is an integer, namely 2 ,n 5 .

Observe that for n  3, the number 10 =2n = 2 ,n 5 is divisible by 10 , = 10 so

decimal in 10 S . It remains to consider 10 =2n with n = 4. We have 10 =2 = 5 .

d = 6, d = 2, and d = 5. Hence, the sum in the problem is 9 + 0 + 6 + 2 + 5 = 22 .

628318530717958647692528:6766559 : : : , we obtain that

0:22893 : : : , it follows that cos = 1 , u + v where u < 1, the decimal expansion of u

What are a and b?

divides k and 2 +1 does not divide k. What is the value of n?

A(1) A(2) A(3) A(5) A(7) + A(6) + A(10) + A(14)

Since k is an integer, e2ik 1 = cos(2k) + i sin(2k) 1 = 0. Hence, the numerator

Taking n = 99999, we derive the answer to the problem, namely 299997=8.

denote digits, then what is d101d102d103d104?

is to be completely covered by 10 dominoes (i.e., 2 1 or 1 2 boards shaped or

T8. The polynomial x2 4x +5 can be multiplied by a non-zero polynomial g(x) (having

ANSWERS: T1: 11=32 T5: (p+ 8)=4

Taking t = , we see that p p

Since an 6  4 44 5 41 = 29, a negative number an 6 will make the coeÆcient of

2. What is the value of the coefficient of x2000999 in the expansion of

(x + 1)(x − 2)2 (x + 3)3 (x − 4)4 · · · (x + 1999)1999 (x − 2000)2000 ?

Solutions are located at the website http://www.math.sc.edu/~filaseta/contests/contests.html

12 − 22 + · · · + 19992 − 20002 = −1 − 2 − · · · − 1999 − 2000 = −2001000.

100 − h2 h2 = b2 h2 > 900,

This last equation is equivalent to x2 + y 2 − 4x − 2y = 0. Since x2 + y 2 = 2, we obtain

6. Observe that 95 = 3 × 25 − 1. Therefore, by the binomial theorem,

952000 = 1024 × 188 − 191999 + 210 (m − 188) = 513 + 210 (m − 188).

7. Since 6492 − 13 × 1802 = 1, we see that

Squaring both sides, we obtain

1. What is the maximum value of the function

f (1) = 1, f (2) = 3, f (3) = 5, f (4) = 7, and f (5) = 9.

What is the value of f (6)?

T4. For any two real numbers x and y, dene

p of C . The area of C is 9. Since

T7. For k 1, dene ak by ak = 2k + 3. Thus, a1 = 5; a2 = 7; a3 = 11; : : : . It is not

1 = cos(60 ) = cos(6) = g (cos()) = 32u , 48u + 18u , 1:

Thus, for u = cos() with = 10 , we have

It follows that 2 cos(10 ) = 2u is a root of x , 6x + 9x , 3 . (The answer is unique.)

is 0. The Binomial Theorem implies, however, that the coecient of x in u(x + 1) =

On the other hand, for n 4, the value of 10 =2n is an integer, namely 2 ,n 5 .

Observe that for n 3, the number 10 =2n = 2 ,n 5 is divisible by 10 , = 10 so

Since k is an integer, e2ik 1 = cos(2k) + i sin(2k) 1 = 0. Hence, the numerator

ANSWERS: T1: 11=32 T5: (p+ 8)=4

Taking t = , we see that p p

Since an 6 4 44 5 41 = 29, a negative number an 6 will make the coeÆcient of

Combining the above, we deduce T = 16 + 3ε1 − 4ε3 + ε2 + ε5 − 2ε6 + 4 . Since each