Please use this identifier to cite or link to this item: http://hdl.handle.net/10419/78407
Authors: 
Selten, Reinhard
Abbink, Klaus
Buchta, Joachim
Sadrieh, Abdolkarim
Year of Publication: 
2000
Series/Report no.: 
Bonn Econ Discussion Papers 3/2001
Abstract: 
We report an experiment that uses the strategy method (Selten 1967) to elicit subjects' general strategy for playing any 2-person 3x3-game with integer payoffs between 0 and 99. Each two subjects' strategies play 500000 games in each of the 5 tournaments. For games with pure strategy equilibria (ca. 80%), the frequency of pure strategy equilibrium play increases from 51% in the first to 74% in the last tournament, in which there is equilibrium play in 98% of all games with only one pure equilibrium. In games with more than one pure equilibrium, a tendency towards the selection of the one with the maximum joint payoff is observed. For games without pure equilibria, subjects’ strategies do not search for mixed equilibria. The strategy programs are based on much simpler strategic concepts combined in various ways. The simplest one is MAP, maximal average payoff, the strategy which maximizes the sum of the three payoffs obtainable against the possible choices of the other player. BR-MAP, the best reply to MAP, and BR-BR-MAP, the best reply to BR-MAP, are also important ingredients of the strategy programs. Together these three form a hierarchy to which we refer to as the best-reply cascade.
Subjects: 
2-person games
experimental economics
JEL: 
C78
C91
C92
D82
Document Type: 
Working Paper

Files in This Item:
File
Size
247.99 kB





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.