Prospect choice is generally viewed as a game against nature. This article models prospect choice as an n-person game where each subject assumes that n-1 others will be exposed to the same decision problem (prospect choice set) as self; the goal is not to “beat nature” but to do relatively better than rivals exposed to the same problem. Preference becomes strategy choice in n-person Nash equilibrium. When symmetric pure strategy equilibria do not exist, choice is a symmetric randomized equilibrium; here, uncertainty (probabilistic response) becomes a method of dealing with uncertainty in nature. The approach produces, qualitatively, several empirical expected utility paradoxes (the certainty effect, intransitive cycles, and one form of reflection), and an evolutionary game-theoretic extension accounts for all the phenomena revealed by the research of Kahneman and Tversky.