The similarity of certain games and certain social situations is, of course, not accidental. A considerable body of psychological research deals with the transfer of patterns of social behaviour into the play activity of children and back from the play of children into social life. It seems plausible that adults, as well as children, may tend to find those types of games more interesting which would permit them to adopt patterns of behaviour which they could also apply to some social situation, or which would permit them to act out, as games, those patterns of behaviour which were initiated in some experience of social life but which only in the innocuous form of games can be carried to completion. Though we cannot wage private war and kill our opponent, we can play chess and checkmate his king; and the art of deceiving others profitably is more safely practised first in the game of poker than in politics or economic life. Granting the potential relevance of games to the analysis of political behaviour, the approach of the theory then consists first of all in analysing simplified prototypes of such games as chess, poker, and others; to calculate as accurately as possible, the winning chances for each player and each hand; and to determine the conditions under which advantageous coalitions can be made, or alternative strategies can be evaluated for their chances of success. *This paper was read in draft form to the Center of International Studies at Princeton University on October 8, 1952. In writing it, I have had the benefit of discussions with Professor Norbert Wiener of the Massachusetts Institute of Technology and Dr. L. C. Haimson of the American Museum of Natural History; with the members of the Research Seminar on Comparative Politics of the Social Science Research Council, under the chairmanship of Professor Roy C. Macridis; and with Dr. Martin Shubik, Professors Oskar Morgenstern and Richard C. Snyder, and Mr. Harold W. Chase, all of Princeton University. While all these have been helpful, the responsibility for the present paper remains, of course, my own.