Savage revisited (with discussion) (Q1819479)

In his ''The foundations of statistics.'' Wiley (1954; Zbl 0055.12604) \textit{L. J. Savage} showed that, under a certain set of postulates for rational behavior, one should make choices that maximize subjective expected utility. However, there is some empirical evidence that people often violate these postulates. According to the present revision of Savage's arguments, the analysis of a decision problem by subjective expected utility is merely an argument by analogy. It draws an analogy between that decision problem and the problem of a gambler who must decide how to bet in a pure game of chance. Sometimes such arguments are cogent; sometimes they are not. Sometimes other kind of arguments provide a better basis for choosing among acts. Thus, subjective expected utility is just one of several possible tools for constructing a decision.