Go take a look at some solutions to the multi-armed bandit problem. A common trait is assigning value (ie. monetary) to each hypothesis. A weaker positive-outcome hypothesis is worth less. The method takes that into account.