The theory of equal division payoff bounds in its modified form (SELTEN 1985) has proved to be more successfull in the prediction of 3-person characteristic function experiments than various versions of the bargaining set. In this paper the question is raised whether the hypotheses "order of strength" (applied to zero-normalizations) and "exhaustivity" introduced earlier as parts of equal share analysis (SELTEN 1972) lead to improvements of predictions, if they are added to other theories. Order of strength improves predictions for games where one-person coalitions receive zero-payoffs, but not for other games. Exhaustivity improves predictions for experimental procedures with free verbal communication but if communication is restricted to the exchange of formal offers, predictions are better without exhaustivity as an additional hypothesis. These conclusions are based on reevaluation of 11 data sets from various published and unpublished sources.
Order of Strength and Exhaustivity as Additional Hypotheses in Theories for Three-Person Characteristic Function Games, Reinhard Selten and Gerald R. Uhlich, in: Tietz, R., W. Albers, R. Selten (Eds.),Bounded Rational Behavior in Experimental Games and Markets,Lecture Notes in Economics and Mathematical Systems, Springer-Verlag, Berlin-Heidelberg-New York-Tokyo, 1988.