Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Mate-in-18 revisited

Author: Tord Romstad

Date: 08:14:44 04/15/05

Go up one level in this thread


On April 15, 2005 at 08:52:26, Kolss wrote:

>On April 15, 2005 at 06:01:05, Tord Romstad wrote:
>
>>I wonder why Crafty, Rebel and The King have problems with this position.
>>Glaurung does not use any special tricks, except the usual extensions for
>>checks, single-reply-to-checks, mate threats and checks at the first ply of
>>the qsearch.  I would expect most programs which use all of these common
>>extensions to find this mate in just a few seconds.
>>
>>Tord
>
>
>Hi Tord,
>
>In my experience, times to find these mates highly depend on the value of the
>single-reply-to-check extension. If you do not use fractional plies, this will
>explain why you find the mate amazingly fast. If I recall right, changing this
>extension by e.g. 1/8 ply can drastically affect solution rates for such
>problems (=> easy experiment to do...).

I do use fractional extensions.  The only full-ply extension I use is the
check extension.   All other extensions are 3/4 of a ply or less.

I do not think my program extends more heavily than most others.  If
anything, the opposite is true (for instance, forced recaptures are not
extended).  Glaurung is usually not particularly fast at finding forced
mates, and I doubt that it is faster than most other engines in the
position discussed in the current thread.  I think we have just stumbled
upon a position where Crafty, Rebel and The King simply happen to have
unusually big difficulties.

Tord



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.