Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: a bug in yace?

Author: Dieter Buerssner

Date: 18:03:47 01/19/03

Go up one level in this thread


On January 19, 2003 at 18:42:08, Uri Blass wrote:

>New game - Yace 0.99.56
>8/8/4r2p/5R2/2K3P1/7P/4k3/8 b - - 0 1
>
>Analysis by Yace 0.99.56:
>75...Re6-e3 76.Rf5-h5 Re3-e6 77.Kc4-d5 Re6-b6 78.Rh5-f5 Rb6-b3 79.Rf5-h5 Rb3-b6
>80.Rh5-f5 Rb6-b3 81.Rf5-e5+ Ke2-f2 82.h3-h4 Kf2-g3 83.Kd5-e6
>  ±  (0.92)   Depth: 19   00:06:53  57404kN
>
>(Blass, Tel-aviv 20.01.2003)
>
>
>This analysis is by yace0.99.56 with part of the tablebases
>pay attention to the line
>78...Rb6-b3 79.Rf5-h5 Rb3-b6 80.Rh5-f5
>
>This line is a repetition in the search so I wonder what is the reason that yace
>searches after it.
>
>I can understand not evaluating repetition of previous games positions as a draw
>but I do not understand evaluating repetition of positions in the search
>not as a draw.

This indeed looks strange. I cannot reproduce it at the moment. With other hash
size or another set of TBs, it is not too surprising, that I cannot reproduce
the exact line. The only explanation I have is, that part of the line was from
the search, and part of the line was reconstructed from the hash tables when
displaying the line. When using the console mode of Yace, it will try to warn
you about this, by attaching an "H" to the moves from the hash table. Under GUIs
you won't see it. The search logic should not allow a repetition inside the PV.

Regards,
Dieter




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.