TJUSAMO Contest #3 Solutions

HMao
December 14th, 2006

1
Let O be the circumcenter of a convex quadrilateral ABCD. Let W, X, Y, Z be the foot of
the perpendicular from O to the segments AB, BC, CD, DA, respectively. If brackets denote
area, prove that [ABCD] = [ABZ] + [BCW ] + [CDX] + [DAY ].

1.1 Solution
W, X, Y, Z are just the midpoints of the sides of ABCD, so

[ABD] [BCA] [CDB] [DAC]

2
Prove that for every natural a, there exist infinitely many naturals n such that 10n a − 1 is
composite.

2.1 Solution
Let p be a prime that divides 10a − 1. We have 10a ≡ 1 (mod p), so p must be relatively
prime to 10. By Fermat’s Little Theorem, 10p−1 ≡ 1 (mod p), so 10k(p−1) 10a ≡ 1 (mod p)
for any natural k. Also, note that 10k(p−1) 10a is greater than 10a, so 10k(p−1) 10a−1 is greater
than p and also divides p, so it must be composite. Therefore any n of the form k(p − 1) + 1
makes 10n a − 1 composite, and we are done.

1
3
Find all finite sequences of nonnegative integers z0 , z1 , . . . , zn−1 , such that n is a natural,
and for any integer i such that 0 ≤ i < n, zi represents the number of integers j such that
0 ≤ j < n and zj = i. One such sequence is 3, 2, 1, 1, 0, 0, 0.

3.1 Answer
The only sequences that work are {1, 2, 1, 0}, {2, 0, 2, 0}, {2, 1, 2, 0, 0}, and for any integer
n > 6, z0 = n − 4, z1 = 2, z2 = zn−4 = 1, and zi = 0 for all integral i such that 2 < i < n − 4
or n − 4 < i < n.

3.2 Solution
Pn−1 Pn−1
First of all, it is obviousPthat n > 1 and z0 6= 0. Also note that i=0 zi = n =⇒ i=1 zi =
n−1
n − z0 . Additionally, i=1 zi is the number of indices i such that zi 6= 0, but there are
n − z0 − 1 such indices between 1 and n − 1, inclusive, so we have n − z0 − 1 natural summing
up to n − z0 . By the Pigeonhole Principle, one of them must equal 2, so the rest must be 1.
This means if i > 0, zi ≤ 2, so only one number in the sequence, namely z0 , can be greater
than 2. Therefore, if i > 2, zi ≤ 1. So either z1 = 2 or z2 = 2.
Case 1, z2 = 2, z1 6= 2: There are two twos in the sequence, and the only valid indices
for these twos are 0 and 2. Now we have the condition that all numbers in the sequence are
less than 2 except z0 and z2 , so if i > 2, zi = 0. The two possibilities remaining, z1 = 0 and
z1 = 1, give us the two answers {2, 0, 2, 0} and {2, 1, 2, 0, 0}.
Case 2, z1 = 2, z2 6= 2: There is at least one two in the sequence, so z2 ≥ 1, but
since z2 ≤ 2 and z2 6= 2, the only possibility is that z2 = 1, so there is only one two in the
sequence, so z0 6= 2. Thus, either z0 = 1 or z0 > 2. If z0 = 1, then if i > 2, zi = 0, so we
get the answer {1, 2, 1, 0}. If z0 > 2, we let n = z0 + 4. Now we have zz0 = 1 and zi = 0 for
all integers i such that i > 2 and i 6= z0 , so we get our remaining solutions, which then can
easily be verified to work.