Probability dominance in random outcomes

Abstract

An intuitively appealing “probability dominance” relation for random outcomes is defined and investigated. Its properties, strengths, and weaknesses relative to stochastic dominance and meanvariance dominance are studied. It is shown that the proposed probability dominance complements existing solution concepts and strengthens one's confidence in decision making under uncertainty. Fundamental characteristics of nondominated random payoffs and methods for identifying them, for both general and specific classes, are reported. Application pitfalls and possible extensions are also discussed.