Game theory and sociology have nearly sixty years of relations. Published 54 years ago in the U.S., and this year in Poland, Thomas C. Schelling Strategies of conflict was an important step in the perception of game theory by sociologists. Despite initial signs of enthusiasm, especially in the 60s, and the publication of a number of major sociological works referring to the theory of games, in the end position of game-theoretical approach in sociology turns out to be marginal, even within rational choice sociology. Using the background of the history of the relationship between sociology and game theory, this paper analyzes factors that make game theory a potentially attractive tool for a sociologist, and the reasons why sociologists rarely use it.

This study presents the results of an experiment on determinants of the strategy used by the participants and their profi tability. The study was based on N-person games in normal form 3x3. All games have clearly defined theoretical solution (the unique Nash equilibrium, which can be reached using the iterated elimination of strictly dominated strategy).The most important conclusion that can be drawn from this analysis is the failure of the theoretical model predictions based on Nash equilibrium. The use of non-equilibrium strategy for people with relatively little theoretical knowledge was due to the inability to find such a strategy. However, participants with a relatively high theoretical knowledge use non-equilibrium strategies deliberately expecting higher profi t. That is why theoretical knowledge is useful, but its use should be adapted to the situation in the real world. This enables players to achieve extra profits.

This paper presents the results of study on the probability distortion. The distortion is studied at two levels: (i) in the decision maker's perception (expressed as subjective judgements), and (ii) in the decision maker's choices (expressed in decision weights). The study examined how the two types of distortion – concerning perception and weights – depend on the way information on probability is presented. Two formats of probability information were distinguished: descriptive (expressed in numbers) and experience.The results confirmed the hypothesis of lower perceptual distortion of objective probability when the information on payout distribution was presented using the experience-based format than when the information was given as a description. The hypothesis on the lower distortion of probability in decision weights with the experience probability format was not confirmed. An unexpected result was a lower risk seeking with the experience format. This effect seems to be a result of attributing lower weights than with the descriptive format. Furthermore, it appears that the probability distortion in decision weights at least partly results from the distortion in perception.