Author: Bob Durrett
Date: 18:41:50 12/10/02
We all know that Deep Fritz was designed solely for the purpose of "killing" Gary Kasparov. It seems likely that the Fritz people had the same fate in mind for Kramnik. No doubt, the Junior programmers are also out to "kill" Kasparov in January. You don't have to be on a computer chess bulletin board to know that. _ _ _ _ _ _ _ _ _ _ But the real, highly technical, question is: HOW? What changes must be made in an existing chess engine to optimize the engine's chances against a top GM? Or, is it necessary to start from scratch and reexamine the method of playing? The Deep Blue people didn't have a top-ranked engine to start with, I presume. So they had the luxury of optimizing the design from the ground floor up. The Fritz people did not have much time, and probably had to modify their existing engine. Obviously, the Fritz people examined Kramnik's openings to create their anti-Kramnik opening book, just as the Deep Blue people optimized their opening book as an anti-Kasparov opening book. But what about the middlegame and endgame? _ _ _ _ _ _ _ _ _ _ If a chess programmer has no knowledge at all about his/her chess engine's opponent, not even knowing whether the opponent is human or machine, then the programmer must make an engine which will do reasonably well against all possible opponents. That might require some tough decisions and compromises. But if the programmers knew exactly who or what the engine would have to play against, it would be at least theoretically possible for the programmer to utilize all available information about the opponent to make the very best engine. What if a programmer wishes to make a chess engine that would do well against the top twenty [or 50] human GMs? The first step would be to obtain information about the GMs and characterize these top GMs in some useful manner. Next, the programmer would have to strive to find innovative and effective ways to take advantage of the human weaknesses. Hopefully, the engine would also steer the game away from the human strengths. So far, obvious. _ _ _ _ _ _ _ _ _ _ But HOW? Bob D.
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.