Computer Chess Club Archives


Search

Terms

Messages

Subject: Other thoughts

Author: Bob Durrett

Date: 07:05:19 02/22/04

Go up one level in this thread


On February 22, 2004 at 08:46:27, Uri Blass wrote:

>On February 22, 2004 at 08:27:55, Uri Blass wrote:
>
>>I am not talking about situation when we prove theretical result but about
>>situation when all comp-comp games at 120/40 time control between top programs
>>are drawn.

I guess you assume tournament books where the chess-playing program only plays
the best opening moves?  Otherwise, games might be won or lost just because the
opening book move was inferior.

There is also a phenomena seen in a recent computer chess tournament where the
programmer for one engine that lost said something like "the opponent engine
simply out-searched my engine" [paraphrased]. In other words, the idea is that
deeper searching provides better chess moves.  For the time to come when no
chess engine loses, it would be necessary that the moves found be "good enough"
so that a deeper-searcher could not get a decisive advantage.  The idea of
"accumulation of small advantages" is relevant here.  The deeper searcher may
consistently find moves which are slightly better so that there is, over the
whole game, an accumulation of small advantages sufficient to win.

Bob D.

>>
>>My guess is that we need more than 20 years but less than 50 years to achieve
>>that target.
>>
>>What is your opinion?
>
>I can add that I also mean that by practically solving chess I mean that it will
>be impossible practically for humans to win against chess programs even when
>they get computers to help them.
>
>It does not mean that computers will know to solve every position in chess and
>it is possible that they will not know to evaluate a lot of positions when one
>side is a pawn up but the point is that I believe that when computers search
>deep enough and their evaluation will become better they will not get into the
>situation when they have to go to inferior position when the opponent is a pawn
>up and it is not clear if it is a draw or a win.
>
>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.