Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Another User Desirement: An Emulator

Author: Uri Blass

Date: 10:51:25 10/31/02

Go up one level in this thread


On October 31, 2002 at 13:12:36, Bob Durrett wrote:

>
>The ideal sparring partner would be a chess engine which is perfectly human-like
>in all respects.  This is true if and only if the chessplayer is training for
>future human vs human matches.  I am not addressing sparring partners used to
>prepare for human vs computer matches.
>
>A sparring partner would be not useful in preparation for human vs human matches
>if the sparring partner engine could be defeated only by anti-computer
>strategies, because anti-computer strategies might not work well against a
>human.
>
>It should be possible to win against the sparring partner, but not by using
>anti-computer strategies.  Ideally, only anti-human strategies would be
>successful.
>
>This implies that the ideal chess sparring partner engine would have human
>susceptabilities and frailties.  Engines set to train weak players should have
>some serious human weaknesses!
>
>A chess engine programmer must have a keen understanding of human chess
>weaknesses if he/she is to design an ideal sparring partner engine which would
>EMULATE the play of humans.  [This partner should be useful for human players at
>all performance levels.]
>
>For some, this may be a problem.
>
>It may be easier to design a GM strength engine than to design an ideal sparring
>partner!  Of course, I wouldn't know since I am a chess engine USER myself and
>not a chess engine programmer.  But, as a USER, I know what I want.  I want a
>sparring partner which is as close to being human as a computer can get.
>[Olfactory and outgassing emulation not a requirement.]  I want a human
>emulator.
>
>Bob D.

I think that it is possible to do a tournament when the programs get pgn of
previous games of the players and try to guess moves in new games

I think that if we gives every program to try to guess 1000000 moves then we can
get information about the question which program is best in guessing human
moves.

regular programs when the programmer did no effort to guess human moves can also
participate in the tournament and it may be interesting to see what is the value
of the knowledge of previous games for guessing.

Uri



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.