[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Re: [computer-go] Statistical Significance



One difference between Go and chess rating is that Go has a handicap systems. Go is more fine grained than chess. Assume a rating of 30k for a new player and 1 kyu separation means 1 stone handicap. On IGS it use to be one win against an equal ranking or higher ranking player = 0.2 point and 1.0 point is 1 kyu. But now it seems one has to win several games against higher ranking players to advance. It's impossible to advance if only play against equally ranking opponents. This system has a draw back. Because most people don't like to play with lower ranking players. This situation hinders the improvement of players' playing skill. In other word it lacks the dynamics. A player's ranking should be allowed to be like a Browning motion. He will find the right place after a large number of game play. This way players between different ranks can exchange knowledge, mostly from the higher ranking one to the lower ranking one. There is a very distinct difference in playing knowledge between players seperated by even only one 1kyu. I can almost list the knowledges each rank of players have up to 3 kyu. By the way the statistical variations in the game results is largest when two players have the same level of skills.

Daniel Liu
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/