Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Congratulations to Rebel Century

Author: Stephen A. Boak

Date: 19:14:01 10/04/99

Go up one level in this thread


On October 04, 1999 at 20:24:14, blass uri wrote:

>On October 04, 1999 at 15:11:45, Dave Gomboc wrote:
>
>>This might be a really interesting position to check out.  Did you record it?
>
>Here is the relevant position:
>7k/4K2p/7P/3p4/8/4Q3/1q6/8 w - - 0 1
>
>Fritz5 find the right move Ke6(leading to mate) but the evaluation is strange
>evaluations of Fritz5:
>
>depth 8  0.00
>depth 9  5.16
>depth 10 4.81
>depth 11 4.50
>depth 12 4.00
>depth 13 3.69
>depth 14 3.38
>depth 15 3.06
>depth 16 2.75
>depth 17 2.44
>depth 18 2.13
>depth 19 1.81
>
>Uri

Seems like a bug (unintentional and bad) to me, not a feature.  :)
Rebel10c finds White Ke6 in a split second, and after 13 seconds, on ply 9.00 it
finds and announces Mate in 8 for White.  During the earlier ply searches, and
until the mate is discovered, it shows approx 1/3 pawn advantage for White.

I wouldn't extrapolate what intentionally programmed features of the program
might cause this strange behavior of Fritz (a mix of Brute Force with Selective
Search, for one example) but instead would chalk it up to a bug in Fritz's
programming.  If Fritz has reached 19 plies without discovering the mate (15
plies only), there is obviously something wrong (my opinion only) with the
working of Fritz (whether intentional or unintentional).  Maybe there is more
than one (or more than one successive) quiet move by white in the mate, which
cause Fritz to miss it (null-mover problem?).



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.