Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Mate solvers...

Author: Heiner Marxen

Date: 09:12:56 09/13/99

Go up one level in this thread


On September 13, 1999 at 03:43:34, Shep wrote:

>On September 12, 1999 at 10:45:48, Heiner Marxen wrote:
>
>
>>That depends upon the exact meaning of "solve":
>>(a) Find any forced mate, however deep it is.
>>    This is what normal playing programs are interested in,
>>    and what some of them are really good at.
>>(b) Find the absolutely shortest forced mate.  Be absolutely sure
>>    that there is no shorter forced mate.
>>    This is needed for chess problems (e.g. to check the correctness
>>    of a problem).
>>
>>(b) is much harder than (a) (i.e. takes more time).  And if you really
>>want (b) you must tell your normal playing program quite explicitly to
>>do so (if that is possible at all), or use a special mate solver program.
>
>And don't forget the most important one:
>
>  (c) Display a full variation tree for the mate.
>
>That's is what Mate 2.0 (Fritz package) does, and it's a great feature.

I don't know Mate 2.0, but yes, of course CHEST does that.  You can get
a full tree, optionally limited in depth (from the root), optionally
omitting duals (for a depth from the leaves), optionally reordered and
reduced by notations for "any move" and "all other moves".
Full solution trees can be *very* BIIIG.

Another related feature I found very important is a "refutation table",
listing all the non-solution key moves together with that answer move,
which CHEST found refutes.  This is important in order to spot problems,
when CHEST insists that something is not a mate in 8, while you are "sure",
it is a mate in 8 (CHEST is right most of the time :-).

Cheers, Heiner

>---
>Shep



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.