ABSTRACT
This chapter discusses the under-weighting, over-weighting and the reasons for the 'description-experience gap', why do preferences reverse and the four versions of the 'computerized money-machine' task. It also covers Erev and Haruvy emphasize, the study of decisions from experience is not designed purely to test or refine theories based on modifications of expected utility theory. Although much of the initial interest came from the call to arms for a new theory to explain surprising patterns of behaviour, under-weighting where over-weighting was expected. The reversal of the certainty effect subsequent research has used the money-machine task to achieve what Erev and Haruvy argue is the more important goal, namely to expand the set of situations that can be addressed with economic models that provide clear and useful predictions. Successful class of models is based on the Instance-Based-Learning architecture, a memory-based decision model uses a combination of exemplar memory and blended calculation of the expected value of different options to predict choice.
