In three experiments, we studied the extent to which theories of decision making and memory can predict people’s preferences. Studying risky decisions, we aimed to answer questions about human preferences, prompted by similarities between the leading economic theory, expected utility theory(EUT), and the leading psychological theory of human choice under risk, prospect theory (PT). Accordingly, we studied (1) the lability of human preferences and their relation to choice justifications given in risky decision-making scenarios, (2) the dynamics of preference formation for choice with monetary gambles, and (3) the limits of existing theoretical accounts (EUT, PT, experience-based decision research) by contrasting them with a new theory of risky choice based on the impact of context, complexity, and prior choices. We found evidence that people do not have underlying preferences for risk; instead, context, skills, and memory determine preferences even when the utilities (risk and reward) of alternative options are known.