Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Accuracy of Selective Search Ratings

Author: Peter Kappler

Date: 20:52:25 03/27/99

Go up one level in this thread


On March 27, 1999 at 20:14:44, Leon Stancliff wrote:

>To any or all,
>
>
>
>  I have listed Fide ratings of the top eight humans in the U.S.A. and attempted
>to make a comparison by examination of Eric Hallsworth's Selective Search rating
>list for the highest eight Computer programs. When the computers are operating
>at 200-233 Mhz. there is a slight advantage in favor of the humans. If, however,
>the computer programs are allowed to use the 450 Mhz hardware now available it
>should be the reverse. Doubling the speed of the computer programs has resulted
>in about a 75 point rating increase in the past. If that is true, when we move
>from 233 Mhz. to 450 Mhz. the computers would seem to be capable of at least
>playing even at 40/2.
>

I suspect that we are entering the region of diminishing returns for 40/2
playing strength.  Blitz chess is where the effect of the new hardware seems to
still be dramatic.  Bottom line - you won't continue to see a 75pt jump every
time the hardware speed doubles.  (At least not for much longer!)

>
>
>  Is this pure bunko, or have we truly reached equality between the top human
>players in the U.S.A. and computer programs running at 450 Mhz? What would
>happen in a round robin tournament today with each of the top eight humans
>playing one game with white and one game with black against each of the
>following computers. All machines running at 450 Mhz.   Time Control 40/2.
>
>Hiarcs 7, Rebel 10, Fritz 5.32, Junior 5, MChessPro 7, Nimzo 98, Chessmaster 6K
>and Crafty 16.5  These may not be the best eight, but they are close!
>
>
>
>  Would anyone care to predict a total score in such a match?


At 40/2, I think the humans would still come out pretty clearly ahead.  I asked
a similar question earlier this week, only with Kasparov or Anand as the
participants.

The upcoming Yermolinksy-Hiarcs 7 match will be a very valuable benchmark.  I
predict a 4-2 victory for Yermo, but I made the same prediction for Kasparov in
his 2nd match against Deep Blue...

--Peter



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.