Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Khalifman and Gelfand on computer

Author: Alvaro Rodriguez

Date: 02:02:07 05/21/00

Go up one level in this thread


On May 20, 2000 at 17:25:28, Mogens Larsen wrote:

>On May 20, 2000 at 16:08:00, Alvaro Rodriguez wrote:
>
>>You mean like a learning function? Learning from games they have played
>>themselves without opening book? I think that some programs plays pretty strange
>>moves and completly dislike some known theory without an opening book..
>>I´m not really sure, but I´ve heard this many times..
>>Maybe someone would like to do some testing, playing different programs against
>>each other without opening book, and then compare them to the games with opening
>>book.
>
>Yes, that was the idea. To make the program "discover" by experience what moves
>are correct and what are not. It might even extend known opening theory.
>
>>> Gelfand believes that exterior means
>>>improve the chess programs, which I think is correct.
>>
>>Improve what?
>
>Opening and endgame play to mention a couple of examples.
>
>Sincerely,
>Mogens

Yes, it is a good idea. Then, the GM´s can´t complain about opening books or
tablebases.. The programmers need to agree of course :-)



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.