Author: David Blackman
Date: 02:32:32 07/29/99
Go up one level in this thread
On July 27, 1999 at 20:43:32, syed wrote: >How necessary is it to be a good chess player in order to be a good chess >programmer? >Arshad It doesn't seem to be essential. Ken Thompson (Belle) and the guy who did Fritz (Franz Morsch ?) are supposed to be very weak players but both produced programs that were very tough. I suspect most of the successful programmers are around average club strength, maybe 1300 to 1900. A few are master strength. For a weak player, it probably helps to read a few books about positional play. My System by Nimzowitsch is a bit obscure, but seems to be very good. Silman's Reassess Your Chess is supposed to be much easier and also good but i haven't got hold of a copy yet. Modern Chess Strategy by Pachman and Pawn Power in Chess by Kmoch are other old favourites. Probably it also helps if you can talk to a stronger player, and especially if you can get them to analyse a few of your program's games. Not too heavily, just to look for obvious positional mistakes you might not be aware of. And to tell you that some of the obvious positional mistakes you did see are probably good moves after all. Of course tactics are the single most important thing in chess. To write a good tactical program, you don't have to be good at tactics yourself. The well-known methods work fairly well. You can sniff out bugs and look for improvements by playing games against other programs. Crafty or most of the commercial programs will punish tactical mistakes most of the time. Tactical problem sets can be useful in debugging and tuning because they provide repeatable positions so you can compare with your last version. But don't believe in them too much. A lot of the time, tactics from real games are both simpler (not as deep to see the main idea) and more complex (more counterplay).
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.