Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Endgame questions-Tablebases and 50 move rule from actual game

Author: Uri Blass

Date: 07:18:39 06/12/01

Go up one level in this thread


On June 12, 2001 at 09:21:44, James T. Walker wrote:

>Here is an interesting game played by Fritz6 and an opponent (Beta) at
>Game/2hours.  At move 96 both programs called -#27 (Black mates in 27).  Notice
>also that Fritz had called mate in 35 on the previous move.  However the game
>ended in a draw.  Both programs were running in the Fritz6 GUI with two
>computers and auto232. Another problem is that the last pawn move/capture was at
>move 53 so they were running into the 50 move rule.  Maybe this explains the
>draw.  Is this a conflict between the engines considering the 50 move rule and
>the GUI using tablebases?  Or is the 50 move rule controlled by the GUI also?
>For the next few moves Fritz changed between mate in 29 & mate in 28 and then
>Fritz's score went down to around 1 pawn.  Then of course the 50 move rule ended
>the game in a draw at move 103.  Should programs continue to play for the mate
>even though the 50 move rule will cause a draw?  Did Fritz "give up" on the mate
>because it saw the 50 move rule being enforced?

1)I believe that Fritz simply ignores the 50 move rule in it's
calculations(otherwise it could see 0.00 evaluation).

2)I suspect that there was some null move problems because otherwise I cannot
understand getting down from mate to an advantage of 0.91 pawn.

I already saw a case when Deep Fritz got down from +5 or +6 to less than +1
against Junior.

In this case the +5 or +6 was correct but Deep Fritz blundered because of null
move pruning and did not put Junior into a zunzwang that it probably could see
some moves before.

Junior saw 11 pawns against itself and after the blunder was very happy with
less than 1 pawn against itself(unfortunately I cannot post the relevant game
because it is based on engine-engine game of one of my correspondence games)

3)In the game that you posted
It may be Interesting to know Deep Fritz's evaluations at the relevant positions
without null move pruning and you can change parameters to check it(My computer
is busy in another position so I cannot do it now)

Uri



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.