Strategy-Proof Incentives for Predictions

Our aim is to design mechanisms that motivate all agents to reveal their predictions truthfully and promptly. For myopic agents, proper scoring rules induce truthfulness. However, when agents have multiple opportunities for revealing information, and take into account long-term effects of their actions, deception and reticence may appear. Such situations have been described in the literature. No simple rules exist to distinguish between the truthful and the untruthful situations, and a determination has been done in isolated cases only. This is of relevance to prediction markets, where the market value is a common prediction, and more generally in informal public prediction forums, such as stock-market estimates by analysts. We describe three different mechanisms that are strategy-proof with non-myopic considerations, and show that one of them, a discounted market scoring rule, meets all our requirements from a mechanism in almost all prediction settings. To illustrate, we extensively analyze a prediction setting with continuous outcomes, and show how our suggested mechanism restores prompt truthfulness where incumbent mechanisms fail.