On this page you can find several studies dedicated to computer chess. The aim of those studies is to measure the level of play of chess players in circumstances where the rating is useless. It all began in 2007 when I performed a plain blundercheck analysis and posted results on chessgames.com. Since January 1st 2008 I have been active with analysis using the methods developed by Ivan Bratko and Matej Guid. In the course of it, I have been trying to enhance analysis methods and discover new ideas. The revolutionary article by Bratko ja Guid can be downloaded here: http://www.ailab.si/matej/doc/Computer_Analysis_of_World_Chess_Champions.pdf
Currently I have created the following studies:
- The latest study on the strength of play
- A study on playing strength of chess players with Rybka 3
- An earlier study on playing strength of chess players with Rybka 3
- My first study on playing strength of chess players with Rybka 2.3.1
- A comparison between Botvinnik and Keres in the candidates Tournament in 1948
What makes the work by Bratko and Guid notable?
Never before has anyone undertaken a scientifical research to find out the strength of play from the moves played. Previously rating systems have been used. But its disadvantage lies in the fact that the purpose of ratings is to compare the relative strength of chess players. In the course of time the average level of play changes and the number of players in the rating pool grows, on account of which the relation between the strength of play and the rating changes as well. The second fact is that they were the first to find how to measure the difficulty of positions and how it affects the accuracy of play. As far as I know, no one has ever attempted anything like this. Their research has influenced many subsequent authors, even if it can't be always said that they have correctly understood everything.