Analyzes statistical test tradeoffs to help compare power, sample size, and error risk.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.