Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: mate threat extension/null move

Author: Stuart Cracraft

Date: 16:37:34 10/01/04

Go up one level in this thread


On October 01, 2004 at 17:09:08, Uri Blass wrote:

>On October 01, 2004 at 13:32:27, Stuart Cracraft wrote:
>
>>On October 01, 2004 at 13:17:27, Uri Blass wrote:
>>
>>>On October 01, 2004 at 12:43:48, Stuart Cracraft wrote:
>>>
>>>>On September 29, 2004 at 22:04:41, Don Beal wrote:
>>>>
>>>>>On September 29, 2004 at 10:23:20, Stuart Cracraft wrote:
>>>>>
>>>>>>For this one, I am *paying* $50 to the first person who can point out
>>>>>>the solution to make my mate threat work and make the Win at Chess #141
>>>>>>problem time drop from 95 seconds on a 1ghz p3 to ~30 seconds or less
>>>>>>on the same box.
>>>>>
>>>>>$50!  :-)  I'll have a go.
>>>>>
>>>>>In the code you posted,
>>>>>instead of:
>>>>>  if (!threat && donull && (material[stm^1]>weights[0][stm^1][rook])
>>>>>      && !checked && !inpv) {
>>>>>use:
>>>>>  if(!threat && donull && !checked) {
>>>>>and instead of:
>>>>>  if (value >= beta) {
>>>>>use:
>>>>>  if (value >= beta
>>>>>  && !inpv && (material[stm^1]>weights[0][stm^1][rook]) ) {
>>>>>
>>>>>
>>>>>The idea of these changes is to apply the null move search
>>>>>(and hence get the mate threat extension working) in the
>>>>>pv as well as other variations, while still preventing
>>>>>you pruning in the pv or believing in null move betas
>>>>>when material is too low.
>>>>>
>>>>>
>>>>>In the unlikely event that:
>>>>>(a) I have understood your code
>>>>>(b) I have correctly diagnosed a problem
>>>>>(c) my fix does not conflict with other things you do
>>>>>(d) it achieves your desired 3-fold speedup on 141
>>>>>then the $50 will be gratefully received, but don't
>>>>>send it to me - send it to the Steve Schwarz CCC
>>>>>donation fund!
>>>>
>>>>Hi Don -- I remember reading your articles through the years
>>>>especially that one on null move. Thanks for your interest.
>>>>
>>>>I did take your code and put it into my program and removed
>>>>both Ross Boyd's and Uri Blass's ideas to test. Since they had
>>>>already solved the problem of my slow WAC 141 from 95 down to
>>>>5 seconds but Bob now tells me it's no good since it throws
>>>>out the null move by disabling it for depth > 1 which is a
>>>>very bad idea. So I can't name Ross nor Uri as receivers of
>>>>the bet after all.
>>>
>>>I only suggested to limit the recapture extension because the way that you use
>>>it is clearly wrong.
>>>
>>>correct recapture extension should not change much in 141.
>>>Of course if you limit it to have recapture extension only in the first 5 plies
>>>then it can help much and it is still better than do it everywhere but generally
>>>it is not a good idea.
>>>
>>>If the target is solving 141 and not productive recapture extension then you
>>>have 2 ways:
>>>
>>>1)having counter productive recapture extension.
>>>2)having mate threats extensions and checks in the qsearch
>>>
>>>Uri
>>
>>Without all that, do you think that the mere inclusion of
>>checking moves at first ply of quiescence would solve WAC 141?
>
>I think it can clearly help but you probably also need mate threat extension to
>solve it.
>
>Movei of today can detect mate threats by null move search but it does not solve
>141 very fast because of not having mate threat extensions and the reason that I
>have not mate threat extension is simply that I do no t store this information
>in the hash.
>
>I plan to have mate threat extension when I may limit them by not extending mate
>threat that is a reply to extended move.
>
>Uri
>>
>>What are ALL the requirement that you see for a program to be
>>able to solve WAC 141?
>>
>>Stuart

It sure seems like knowing all the extensions along the entire variation
to the current node, each extension count and the type of extension
for all previous 0...ply-1 plies is important in deciding whether or
not to extend at this ply. What I do is use only information at the
current node and in the game history list of moves so far to decide
but I don't keep track of say the number of extensions done at the
last ply or what type of extensions they are. I will begin to do that
as it seems that fractional is not enough. One also has to keep historical
extensions.

Stuart



This page took 0.06 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.