Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Junior 9 Test (40'/40) after 300 games

Author: Sandro Necchi

Date: 07:56:02 01/01/05

Go up one level in this thread


On January 01, 2005 at 10:44:43, Uri Blass wrote:

>On January 01, 2005 at 07:46:46, Sandro Necchi wrote:
>
>>On January 01, 2005 at 07:40:58, Harald Faber wrote:
>>
>>>On December 31, 2004 at 15:21:23, Sandro Necchi wrote:
>>>
>>>>Again I agree. Some years ago there was no learning/book learning at all and
>>>>that was a big missing...which made a big difference with the human players.
>>>>Now that we have them we need to improve the learning features and not to turn
>>>>them off!
>>>
>>>
>>>There is danger that one might test the best learner.
>>
>>Yes, but I think a chess program is a chess player and so it should be tested in
>>the best mode, suggested by the programmer, and without removing anything.
>>The learning feature is part of the program.
>
>I think that it is different.
>The main difference relative to humans is that humans cannot buy a copy of their
>human opponents.

Yes, this is true, but humans can study their opponents with the help of
databases and prepare "something" on the variations they normally play.
This is called "home preparation".

Home preparation is becoming more and more important in modern chess both to
human chess players and computer programs.

I mean that one engine can be tuned to score better against another one as a
main purpose to get a higher score, If this happens the opponent may needs some
games with learning feature to improve its games.
To remove such a feature damages the old program versus the new one.
= less accurate strenght value.

>
>If you take some old programs and prepare a killer book against them you can
>earn ssdf rating and even if the old programs have learning it cannot help >them because they cannot learn from games that they never saw.

Yes, this is possible and the only way to help the programs that are facing that
situation is to let them use the learning feature to avoid losses as much as
possible.
Of course this is allowed and cannot be avoided.
I am saying that in order to get estimate of the strenght of a program the
learning feature helps so it should not be turned off.

>
>Not that I think that you do it but the possibility to do it is unfair and it >is possible that some weaker program than shredder can top the ssdf because of
>these tricks.

Yes, this is unfair, but allowed as this would be a sort of home preparation, so
only by testing the programs with all their features we can reduce the impact of
this as much as possible.
>
>Uri

Sandro




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.