Economists used to assume that people make optimal decisions based on perfect information about their options and about the consequences of their decisions.
Joe showed over and over the profound implications of relaxing the perfect-information assumption, while maintaining the optimal-decision assumption in order to make his conclusions more palatable to the economics profession.
George Akerlof, another pioneer in the economics of imperfect information gave a talk (indeed the first talk) at Joe's Festschrift noting that psychologists (such as Kahnemann and Tversky) have convincingly demonstrated that humans are imperfect decision makers.
In my talk, I noted that regression to the mean is a nice example of the combination of imperfect information with imperfect humans.
I showed how a recognition of regression can improve our estimates of a player's ability and can improve our forecasts of company’s earnings.
Wherever there is imperfect information, there is the likelihood of regression, the possibility of being fooled by regression, and the potential for improved estimates and forecasts by taking regression into account.
That is the theme of my next book. Stay tuned....