Game Theory without Decision-Theoretic Paradoxes.pdfVIP

Game Theory without Decision-Theoretic Paradoxes Pierfrancesco La Mura Department of Economics and Information Systems HHL - Leipzig Graduate School of Management Jahnallee 59, 04109 Leipzig (Germany) plamura@hhl.de Abstract a game-theoretic context, linearity of expected utility en- sures that the best response correspondence is also linear, Most work in game theory is conducted un- and hence that any finite game has a Nash equilibrium. der the assumption that the players are expected In economics and the social sciences, the importance of ac- utility maximizers. Expected utility is a very counting for violations of the expected utility hypothesis tractable decision model, but is prone to well- has long been recognized (Tversky 1975), but so far none of known paradoxes and empirical violations (Al- its numerous alternatives (e.g., Machina 1982, Schmeidler lais 1953, Ellsberg 1961), which may induce 1989, to quote only two particularly influential papers in a biases in game-theoretic predictions. La Mura rich and constantly evolving literature) has fully succeeded (2009) introduced a projective generalization of in replacing expected utility as a standard foundation for expected utility (PEU) which avoids the domi- decisions under uncertainty, partly due to the great mathe- nant paradoxes, while remaining quite tractable. matical tractability of expec