In this paper a new mathematical model of human behavior in positional combinatorial games is stated. The considered in the paper version of the model describes two basic factors which have influence on the efficacy of searching procedure - that are differences in structure between model tree and exact game tree and human's calculations upon sets of different positional factors. The efficacy of searching procedure is estimated by using dispersion of comparative deviation in model evaluation and also by calculation of frequency for best (or worst) moves selection. The main object of this paper is to set the fundament for the common theory of human's errors in enumeration of variants in combinatorial problems (such as different games). From this point of view some qualitative consequences and recommendations are presented. These results can be applied to the vide class of game-like problems for diminishing errors and their effect on the final game result.

Mathematical modelling in actual problems of science and technics

Medvedev V.R.