Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: WAC 141 blowup

Author: Stuart Cracraft

Date: 14:42:55 09/01/04

Go up one level in this thread


On September 01, 2004 at 05:35:08, Uri Blass wrote:

>On September 01, 2004 at 05:09:49, Chris Welty wrote:
>
>>On September 01, 2004 at 03:58:46, Uri Blass wrote:
>>
>>>On September 01, 2004 at 03:30:42, Chris Welty wrote:
>>>
>>>>On September 01, 2004 at 00:51:47, Stuart Cracraft wrote:
>>>>
>>>>>I know this is a long shot throwing this out there without much more
>>>>>info, but perhaps I'll get lucky and someone has seen this before.
>>>>>It's appeared on this bulletin board about 6 years ago.
>>>>>
>>>>>[D]4r1k1/p1qr1p2/2pb1Bp1/1p5p/3P1n1R/1B3P2/PP3PK1/2Q4R w - - bm Qxf4; id
>>>>>"WAC.141";
>>>>>
>>>>>without mate threat code
>>>>
>>>>>10/46  g2f1 471.92 -825 92873175 g2f1 e8b8 h1g1 f4e6 h4h5
>>>>>11/46> g2f1 666.85 -575 133140320 g2f1 e8b8 h1g1 f4e6 h4h5
>>>>
>>>>Perhaps the problem's not in the mate threat extension but somewhere else? When
>>>>Altamax sees Qxf4 the PV is only 10 ply long:
>>>>6 347 0 196349 Qxf4 Bxf4 Rxh5 gxh5 Rxh5 Bh6 Rxh6 Qg3+ Kxg3 Kf8  196kn /2.80s =
>>>>70.2kn/s
>>>
>>>Does it use null move pruning?
>>
>>Altamax does. In this line giving black 2 moves in a row would either lead to a
>>free piece (after Qxf4) or mate (after all other moves) so null-move pruning
>>should not happen.
>
>I guess that null move pruning is not the reason that movei needs bigger depth
>but evaluation based pruning because movei as far as I know detect mate in 1 in
>the qsearch(I will check if I have no bugs there) but it simply skip the qsearch
>when it is using evaluation based pruning(I may try to improve my evaluation
>based pruning in order to smell mate danger and not to skip the qsearch in
>minority of the cases).
>
>Stuart's program does not detect mate in the qsearch so for him it is clear that
>null move pruning is a problem.
>
>Uri

I suspect I have a problem in that my null move returns beta not
value and at various other places in my main search (and in the
quiescence), I return mostly alpha or beta, rather than value.

It sounds like this is an issue for solving WAC 141 since the
null move never would return a mate.

I've tried adding Botvinnik-Markoff extension, searching all
initial quiescence ply checking moves (in addition to all
check-evasion moves throughout quiescence), using return
of value instead of beta in null move, and a combination of
all and none of these, but still WAC 141 eludes it, usually
blowing up at ply 8.

Not sure how to proceed next on WAC 141. It's got me for sure.

Stuart



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.