In this paper, we introduce a chess program able to adapt its game strategy to its opponent, as well as to adapt the evaluation function that guides the search process according to its playing ...experience. The adaptive and learning abilities have been implemented through Bayesian networks. We show how the program learns through an experiment consisting on a series of games that point out that the results improve after the learning stage.
•We use move similarity techniques to explore creativity and originality in computer chess programs.•We apply the results to the specific case of Rybka and Fruit.•The move similarity data is also of ...interest to questions of “style” of engines.
In June 2011, the International Computer Games Association (ICGA) disqualified Vasik Rajlich and his Rybka chess program for plagiarism and breaking their rules on originality in their events from 2006 to 2010. One primary basis for this came from a painstaking code comparison, using the source code of Fruit and the object code of Rybka, which found the selection of evaluation features in the programs to be almost the same, much more than expected by chance.
In his brief defense, Rajlich indicated his opinion that move similarity testing was a superior method of detecting misappropriated entries. Later commentary by both Rajlich and his defenders reiterated the same, and indeed the ICGA Rules themselves specify move similarity as an example reason for why the tournament director would have warrant to request a source code examination.
We report on data obtained from move-similarity testing. The principal dataset here consists of over 8000 positions and nearly 100 independent engines. We comment on such issues as: the robustness of the methods (upon modifying the experimental conditions), whether strong engines tend to play more similarly than weak ones, and the observed Fruit/Rybka move-similarity data.
In this paper we demonstrate how genetic algorithms can be used to reverse engineer an evaluation function’s parameters for computer chess. Our results show that using an appropriate expert (or ...mentor), we can evolve a program that is on par with top tournament-playing chess programs, outperforming a two-time World Computer Chess Champion. This performance gain is achieved by evolving a program that mimics the behavior of a superior expert. The resulting evaluation function of the evolved program consists of a much smaller number of parameters than the expert’s. The extended experimental results provided in this paper include a report on our successful participation in the 2008 World Computer Chess Championship. In principle, our expert-driven approach could be used in a wide range of problems for which appropriate experts are available.
PDF
