DECISIONS TO BIAS CASE AND CASE-BASED DECISION Biased Choice/Case-Based Decision/Discrete Choice

The present model of individual decision-making is based on a bounded-rationality style assumption. The de-cision-maker's preferences over actions lead to choice probabilities that can be defined as the reflect of his be-havior when facing past cases of the same choice prob-lem, i.e. his choice memory. Actually, the actions utility is supposed to be cumulatively learnt, case after case. Satisfying a weak assumption of cognitive rationality, the decision-maker searches for relevant choices that in-crease the probability for the chosen action to be the best. Then, our main result establishes that he should design biased cases which are the only ones leading to relevant choices. Furthermore, we show that the bias introduced in the case must be in favor of the best past action and finally, we prove that each action that is chosen at the end of a biased case is helpful to bias next cases in order to make future choices relevant. And don't speak too son For the wheel's still in spin And there's no tellin ' who That is namin ' For the loser now Will be later to win For the times they're a-changin. Bob Dylan 1