Bob: > By taking the next > step, using Expected Utility Maximization (EUM) to solve the problem of > decision making under constrained cognitive and computational resources > you have increased the level of complexity by orders of magnitude. So > how is our poor cognitively deprived human to master that one?
Nice one :-) But I am not claiming that humans consciously apply EUM in these circumstances - we agree on this. The issue I'm raising is your claim that such decisions have been shown to produce cases that are not representable by EUM. In order to convince me of this you would need to demonstrate that the decision made in the decision-under-uncertainty problem violates EUM. > What Kahneman and Tversky and others discovered is that those > short cuts learned from training don't always work, especially when > taken out of context in which they were learned or in otherwise strange > situations. Nevertheless, they may give the best decisions available for our "cognitively deprived human". And these may turn out to be the EUM decisions for the decision-under-uncertainty problem. > But I suspect we have no disagreement On the prescriptive side we agree. On the descriptive side we agree that, under assumptions such as perfect cognitive function and sufficient time and motivation, EUM does not work. My point is just that we should not assume that EUM would continue to be a bad descriptive tool if we were able to evaluate it after relaxing those assumptions. Konrad
