Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: "Accuracy" of SSDF list

Author: Dann Corbit

Date: 12:46:50 03/08/02

Go up one level in this thread


On March 08, 2002 at 08:07:24, Mike Hood wrote:

>As has already been discussed in other threads, ELO ratings only give the
>relative rankings of competitors in a limited pool,

This is true.

>although by extending the
>pool to encompass the whole world the ratings can be considered absolute.

This is utter nonesense.

>The SSDF rating lists can only be taken seriuously if the pool of computer
>programs is periodically calibrated by taking into account games between
>computers and humans.

Nonesense.  They can be taken seriously in any context in which they are
valuable.  A long time ago, the SSDF was calibrated using humans.  The problem
with keeping it current is that it would cost millions of dollars to do it.  Are
you willing to pony up the cash to have GM's play 1000's of serious games
against these computer systems?

>To what extent is this done?

It was attempted long ago.  I predict it will never happen again.

>Are only games between grandmasters and top programs considered? Or also
>verifiable tournament condition games between club players and computer
>programs?

The games in the list are computer verses computer.  No new human data is added.



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.