Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fritz analysis at 21/21

Author: Robert Hyatt

Date: 07:30:17 11/30/98

Go up one level in this thread


On November 30, 1998 at 09:27:54, Dave Gomboc wrote:

>On November 29, 1998 at 21:39:56, Geo Disher wrote:
>
>>11/25/98 5:40am 19/19 Qb6 eval=1.06
>>11/25/98 9:30pm 19/49 Qb6 eval=1.00
>>11/29/98 6:57pm 21/21 Qb6 eval=.88
>>
>>Best line is 36)Qb6 Qe7 37)axb5 Rab8 38)Qxa6 e4 39)Bxe4 Re8 40)Kf1 Qxe4
>>
>>It is interesting that the longer Fritz thinks the lower the eval for this pos.
>
>This is pretty interesting.  Fritz is sort of noted for its pawn-grubbing, isn't
>it?
>
>Dave Gomboc


yes... but it is going pretty deep.  I watched Crafty play a game last night
against a GM... it was a pawn up and grabbed a second, but let the position
become *totally* locked up... no way for either side to break anything open
as there were no pawn breaks available.  No way to do anything (crafty had a B,
the GM a N)  Crafty's eval steadily went down, because to prevent repeating,
it had to back its king further and further away from where it wanted to be
(in the center).  I suspect this is what Fritz is seeing...  It grabs the pawn,
but then has to slowly wreck what it thinks is a good position to avoid the
series of checks.  And it might not *ever* get down to zero either.  In this
game last night, I restarted later and let it search for an hour, and it
never got down to even +1, although it got close..  And it would have eventually
pushed the only pawn it could, and lost that pawn, to avoid the repetition,
which it did.  And then it kept dropping until it finally offered the draw.



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.