Author: Howard Exner
Date: 23:23:54 11/26/97
Go up one level in this thread
On November 27, 1997 at 01:19:01, Jouni Uski wrote: >You are always praising Cray Blitz. But if I remember right it lost in >one ACM >tournament to Fidelity's 5 Mhz 8-bit 6502 program (probably with about >2000 >ELO) !?!? I think there are reasonable explanations to comp vs. comp "upsets". One weakness of computer programs is that the simply don't know who they are playing. They will play the board rather than realize they are playing someone rated 400 points below them. They would need some kind of "let's keep this position complicated" knowledge. A chess game may take a natural direction to simplify. Some games according to GM Kotov's writings require little calculation and to become overly aggresive can often lead to a loss. Examples are from openings that lead to early piece exchanges followed by the heavy pieces being exchanged along the one open file. The SSDF results are full of these "upsets". There could be one exception to computers not knowing there opponent and that might be Deeper Blue whose mission was to defeat one player.
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.