For problems involving choices over "certain x uncertain" consumption pairs, it is almost universally assumed that the decision maker's preferences can be represented by an expected TPC (two-period cardinal) utility function. In this paper, we present an alternative representation of preferences, referred to as the "ordinal certainty equivalent" hypothesis, which we argue (i) is at least as intuitive as the expected utility hypothesis, (ii) includes the corresponding TPC representation as a special case with the set of cases not expressible in the latter format being both large and important, and (iii) is based on a more sensible hypothesis concerning the connection between "risk" and "time" preferences.