This article considers the problem of choosing between two treatments that have binary outcomes with unknown success probabilities p1 and p2. The choice is based upon the information provided by two observations X1 ∼ B(n1, ...