Proper Scoring Rules for General Decision Models

Abstract

On the domain of Choquet expected utility preferences with risk neutral lottery evaluation and totally monotone capacities, we demonstrate that proper scoring rules do not exist. This implies the non-existence of proper scoring rules for any larger class of preferences (CEU with convex capacities, multiple priors). We also show that if an agent whose behavior conforms to the multiple priors model is faced with a scoring rule for a subjective expected utility agent, she will always announce a probability belonging to her set of priors; moreover, for any prior in the set, there exists such a scoring rule inducing the agent to announce that prior.

Files

Additional details