Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: I think it's pretty Common Knowledge now

Author: Uri Blass

Date: 13:33:12 03/26/02

Go up one level in this thread


On March 26, 2002 at 16:18:35, Miguel A. Ballicora wrote:

>On March 26, 2002 at 16:01:47, Uri Blass wrote:
>
>>On March 26, 2002 at 13:56:41, Andrew R. Case wrote:
>>
>>>Exactly the point I wanted to make, if the programs played numerous games
>>>against GMs, the GMs would eventualy find a way to beat them on a regular basis.
>>>Also I would be willing to bet that the programs used in the Gulko match are not
>>>the same versions(with the exception of Shredder) that most people have, and
>>>certainly not the same opening books.
>>
>>I see no problem with it.
>>
>>The program and the book should not be available in order to prevent unfair
>>advantage.
>>
>>I am sure that if Gulko becomes commercial and everybody can buy it and play
>>against it when the real gulko knows nothing about the games then Gulko's elo is
>>going to be smaller but it is not fair.
>>
>>If we want to know the real elo of programs then the programs should not be
>>commercial and the program should have the right to learn from the programmer
>
>Then you know the elo of the programmer+program, not the elo of the program.

No
The programmer has the right to do nothing during the games but only between the
games.

By the same logic if Gulko has a human coach who help him (between games) you
know only the elo of Gulko and his coach and you will never know the elo of
Gulko.

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.