Two Strongly Truthful Mechanisms for Three Heterogeneous Agents Answering One Question

Peer prediction mechanisms incentivize self-interested agents to truthfully report their signals even in the absence of verification by comparing agents’ reports with peers. We propose two new mechanisms, Source and Target Differential Prediction, prove very strong guarantees for a general setting. Our Prediction are strongly truthful : Truth-telling is strict Bayesian Nash equilibrium. Also, truth-telling pays strictly higher than any other equilibria, excluding permutation which same amount as truth-telling. The hold asymmetric priors among agents, need not know ( detail-free ) single question setting . Moreover, they only require three , each submits item (answers), her forecast (prediction one agent’s reports). proof technique straightforward, conceptually motivated, turns on logarithmic scoring rule’s special properties. we can recast Truth Serum mechanism [ 20 ] into our framework. also extend results continuous slightly weaker guarantee optimality