Author: Gregor Overney
Date: 01:41:40 10/10/02
Go up one level in this thread
This is also one of the misconceptions. I can build a chess engine that "nominally" out searches Deep Fritz by 4 ply running on my old AMD system. How do I do this? Excessive pruning! I am not going to search a huge search space. Just follow a few lines I think the opponent will play. And if he does not, my program will look like a fool, because my program's evaluation was based on the wrong assumptions. Somehow I get the impression that exactly this is happening to DF. Pruning, pruning, pruning..... (see 1949, Claude Shannon's two ways to suggest a chess playing program.) It always gets right back to basics. Too less mega nodes cannot easily be compensated by too "smart" evaluations and "clever" pruning. It, as we can see from Kramnik vs. DF, can easily lead to a rather silly play. Or how many times did you move a bishop and pull it back right after? Gregor On October 09, 2002 at 02:21:00, Gian-Carlo Pascutto wrote: >On October 08, 2002 at 20:59:39, Gregor Overney wrote: > >>It is easier to find flaws in certain evaluation functions than to outthink >>brute force searches with decent algorithms (such as implemented in Deep Blue). > >Depends on how low you lower your definition of decent algorithm. > >Deep Fritz at least outsearches Deep Blue nominally by 4 ply, >so your reasoning makes no sense. > >-- >GCP
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.