The winning t-set problem is the problem of choosing nonnegative real numbers , subject to and , so as to maximize the number of t-sets I such that . For t = 2, the optimal nontrivial solutions use at most two distinct nonzero values, and the optimal values and multiplicities can be computed quickly from n, a, b. It is conjectured that the phenomenon of using at most two distinct nonzero values persists for .
