I had a little insight, and I'm curious whether or not it's true.
What I conjecture:
Say we have a set of functions S = {f(a, n) = b_n * a^n + b_n-1 * a^n-1 + ... + b_0} where a and n must be both natural numbers, and b must belong to {-1, 0, 1}. Then for fixed a = a_0 and n = n_0, we...