Computer Chess Club Archives




Subject: Re: Singular extensions

Author: Amir Ban

Date: 09:26:47 01/09/98

Go up one level in this thread

On January 09, 1998 at 11:45:51, Bruce Moreland wrote:

>On January 09, 1998 at 08:51:07, Amir Ban wrote:
>>On January 08, 1998 at 20:00:31, Robert Hyatt wrote:
>>>Does anyone else get a real mate here in reasonable time?  I've been
>>>extensions and none seem to be "out of control"..  However, this has
>>>been *the*
>>>hard problem for Crafty for a couple of years, so at least it now finds
>>>right move fairly quickly...
>>On P5/133 I see Qxf4 after 28 sec. and mate in 6 after 84 sec.
>>Why are singular extensions supposed to be important here ?
>This is a position where white goes down big material then wins without
>any checking moves.  He makes a series of moves that threaten mate, and
>black can do something about all of them except the last one, although
>at the end he's having to throw away material.
>A critical line for computers would be Qxf4 Bxf4 Rxh5 gxh5 Rxh5 Bh6 Rxh6
>Qg3+ Kxg3 and white is ahead material.  I might be slightly wrong about
>this line, I'm doing it from memory, but you get the idea.
>This problem has always been hard for my program for some reason, and
>Bob has the same problem.
>Interestingly, Fritz gets it pretty quickly.  You'd think that Fritz
>would have bad null move problems here, but it doesn't.
>To some other programs this position is no problem at all.  Vincent's
>program gets this in 3 plies, for instance.
>I had always assumed that other programs solved it with a mate threat
>extension, but perhaps there are other ways.
>Genius solves this problem instantaneously, and I was wondering if it
>might have something to do with this rumored singular extension thing.
>It takes mine 14 seconds on a P6/200 to find Qxf4, and I do it with a
>mate extension.  How do you do it?

I do mate extensions, yes. I guess it helps a lot here.

I think when someone says that singular extensions help here, they are
thinking how black's desperate delays can be handled, but there's no
reason to assume that there will be only one such move every time.

I don't believe in singular extensions because the "singular" part of it
is completely arbitrary. What if there are two possibilities ? You can't
base your search on it unless you have a whole bag of additional tricks
to plug the holes in it.


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.