Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: New Site on my Homepage (Tests)

Author: Uri Blass

Date: 03:19:40 11/05/02

Go up one level in this thread


On November 05, 2002 at 05:34:41, Joachim Rang wrote:

>On November 05, 2002 at 02:37:11, Uri Blass wrote:
>
>>On November 04, 2002 at 21:33:36, Rob Shultz wrote:
>>
>>>Shredder 6.02 comes up with the following line:
>>>
>>>New game
>>>1k6/8/8/8/8/5p1p/PPP5/6K1 w - - 0 1
>>>
>>>Analysis by Shredder 6.02:
>>>1.b4 Kc7 2.a4 Kc6 3.a5 Kb5 4.c3 Ka6 5.c4 h2+ 6.Kxh2 f2 7.Kg2 f1Q+ 8.Kxf1 Kb7
>>>  +-  (15.67)   Depth: 30/42   00:14:47  844201kN, tb=4168129
>>>
>>>(Shultz, Delmont 04.11.2002)
>>>
>>>
>>>
>>>New game
>>>1k6/8/8/8/8/5p1p/PPP5/6K1 w - - 0 1
>>>
>>>Analysis by Hiarcs 8:
>>>  -+  (-1.70)   Depth: 20/44   00:10:16  151098kN, tb=296590
>>>1.a4 Kb7
>>>  -+  (-1.95)   Depth: 21/48   00:12:20  180842kN, tb=354248
>>>1.a4 Kb7 2.b4 Ka7 3.b5 Kb6 4.c4 Kb7 5.a5 Kc7 6.a6 Kb6 7.c5+ Ka7 8.c6 Kb6 9.c7
>>>Kxc7 10.a7 Kb7 11.b6 Ka8 12.Kf1 h2 13.Kf2 h1Q
>>>  -+  (-7.27)   Depth: 21/48   00:16:40  250066kN, tb=457173
>>>
>>>(Shultz, Delmont 04.11.2002)
>>
>>Can you play a game between shredder and Hiarcs8 and post moves+evaluations?
>>
>>Uri
>
>
>that's not necessary:

That is necessary to see when Shredder understands that white is lost.
>
>white is lost because he can't move his king, but the black king can stop all
>three connected passed pawns from white (!), because white neither can support
>its pawns with his king nor can put black into zugzwang (it is white who will be
>in zugzwang). White is forced to move its pawns forward until black will have
>all eaten them and then has to move its king which looses.
>
>This is another one of Eduards famous zugzwang-problems which are for most
>engines very hard. Congratulations!

zugzwang is usually no problem in pawn endgame because engines do not use null
move pruing in pawn endgames.

I was interested to know what wrong pruning shredder use in pawn endgame and I
am surprised if it is simple null move pruning.

I expected mistakes by some special pruning algorithm for pawn endgames so I
thought that we can discover details about it by finding when it fails.

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.