Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: DJ - GM or Super GM ??

Author: Robert Hyatt

Date: 08:53:21 02/06/03

Go up one level in this thread


On February 06, 2003 at 05:42:29, Michael Vox wrote:

>
>Still the question remains unanswered is DJ - GM or Super GM ??
>
>It seems based on the F7 vs Kramnik and now DJ, that the SuperGMs are about even
>with the bots... Just one man's opinion!


I have said for a couple of years now that the comps are probably in the low-GM
range for playing skill.  That hasn't changed a lot in my opinion.

I would assess things as follows, after having watched the DB vs Kasp match,
then DF vs Kramnik, Hiarcs vs Bareev, DJ vs Kasparov, etc.

In a single game, Kasparov is probably 200 rating points _better_ than any
computer
currently playing chess.  For example, can you think of a _single_ 2500 player
that
would have played the _first_ game from the DJ side?  I can't.

In a match, this changes.  It would seem that the fatigue factor rates far
higher than I
would have imagined, bringing the humans _down_ a couple of hundred rating
points,
rather than the computers being that much stronger.  Why this is, I don't really
understand.
I have, on occasion, played a serious game (or two) every day for two weeks,
tuning Cray
Blitz for a major event.  Granted, there was not any money at stake, but you can
bet I was
not interested in "rolling over" as I had a high level of motivation to repeat
as WCCC in
1986 for example.

It is going to take some thought to understand what must be called the "fatigue
factor".
But it is obviously there.  15 years ago we referred to the "computer shock
factor" where
players met a machine for the first time and didn't know what to expect.  They
would
often fall into a pit and lose, then come back and win much easier the second
time around.
Now we have a completely different "issue"...



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.