F. J. Anscombe and R. J. Aumann (1963) showed that, if one accepts the existence of a physical randomizing device such as a roulette wheel, then L. J. Savage's derivation of subjective expected utility can be considerably simplified. They, however, invoked compound gambles to define their...