Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: give up

Author: Robert Hyatt

Date: 09:05:09 08/10/99

Go up one level in this thread


On August 10, 1999 at 02:29:51, blass uri wrote:

>
>On August 09, 1999 at 17:10:19, Robert Hyatt wrote:
>
>
>>I have yet to find a single game at 40/2 with no tactical errors.  Even by
>>Kasparov.  They might not be 'losing' tactical errors, but they are errors all
>>the same...
>>
>>That is usually the give-away, because someone using a computer might make a
>>weak move here and there, but they _never_ hang pawns or overlook winning them
>>when it is reasonable to do so...  just like a computer, and much unlike a
>>human, even a super-GM.
>
>1)What do you mean by tactical error?
>If you mean that the biggest change in the evaluation is not more than 0.5 pawn
>then I drew in the past a game at tournament tiem control of more than 40 moves
>with no tactical errors
>(I analyzed it with a computer and the biggest change in the evaluation was not
>more than 0.5 pawn)
>
>The opponent's rating was close to 1800 and my rating was 1985.
>
>2)It is also possible that a move that the computer evaluate as a tactical error
>is a good move and a human who use a computer  can play it (analyzing the
>position with a computer can help the human to be sure that the move is not a
>tactical blunder).
>
>Uri


I define tactical error as follows:

any move that changes the material balance by a pawn or more, whether or not
the move is really good or bad.  IE if a computer thinks that grabbing a pawn
is the right thing, and the human doesn't, I call that a tactical error.  Note
that this is from a computer perspective, because I have seen several positions
where I would _not_ grab the pawn, but any program I try would.

But please post the game if you'd like to have me check it by my 'definition'
to see what happened...

Note that my definition would _not_ be used to assign a ? to a move, in every
case, because often the computer is too materialistic and grabs pawns when it
really should not.  My definition simply means that someone's game agrees with
a 'materialistic' computer or not.  And if it _does_ agree 100% of the time,
it is very likely that the player is a computer or using a computer.  If this
is repeated over several games in succession, the probability goes _way_ up...



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.