The Deal-by-Deal Principle for Rational Choice on the Qui Vive

Based on the deal-by-deal principle, we propose preference indices as a starting point for the analysis of rational choice. They specify how good acts are as a deal. Thus, the recognition of reference points in Prospect Theory is interpreted as a crucial step towards modeling rationality, to be extended to a degree of goodness level for all acts, in order to reflect context dependency in a more refined way. We show how this leads to an uncomplicated syntax for updating, reconciling familiar concepts in dynamic choice theory. A representation theorem characterizes the weakly decomposable preferences in our framework.
We also describe a relaxation of gain-loss separability in Prospect Theory, by a refinement of comonotonicity, to bring it closer to our framework. Furthermore, we show that S-shaped probability weight- ing can be explained as an artefact under betweenness for binary lot- teries, and conclude that rationality comes nearer to behavioral evi- dence than generally believed.