Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Endgame code instead of Tablebases

Author: Dave Gomboc

Date: 17:10:07 04/16/99

Go up one level in this thread


On April 16, 1999 at 19:53:24, Paul Massie wrote:

>For what it's worth, I think if you could codify in some manner algorithms for
>playing nearly perfect endgames there'd be a lot of chess players eager to
>understand those algorithms and try to apply them.  Although I may be in a
>minority here, I do not believe there is anyone in the world today, Kasparov
>included, who truly understands this issue.  That's not to say you shouldn't
>attempt it, because attempting what has never been done is how fundamental
>progress is achieved, but I do believe the size of the task should be clearly
>understood.
>Paul

We have an example of this in the case of KBB vs KN.  The bottom line for
humans?  "The idea is to use mating threats and zugzwang to force the win of the
knight.  A good first step is forcing the knight and the king to separate.  This
is always possible: even the (defender-desirable) fortress demonstrated by
Horwitz and Kling in 1851 can be broken (black king on g2, knight on f2/g3,
white king and bishops somewhere).  However, it takes very accurate calculation
to do this properly against best play."

Did I miss anything really helpful that humans can use OTB?

Dave



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.