Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: how do programmers handle this?

Author: Frank Phillips

Date: 03:23:29 04/29/01

Go up one level in this thread


On April 29, 2001 at 06:10:13, guy haworth wrote:

>
>See ICGA_J (2000) for Nalimov, Heinz, Haworth article on
>
>"Space-efficient Indexes for Endgame Tables".
>
>There's a factor of 8 in play because there aren't any Pawns.  But with Pawns,
>after you have computed the potential successor-endgames, you can slice up this
>endgame according to the position-profile of the Pawns on the board.
>
>Thus one Pawn gives you a 4*6 = 24-way slice, 2 Pawns gives you a 24*47-way
>slice.  Potential for parallelising EGT-generation across a syndicate of
>PC-owners.
>
>G
Thanks.  My question was different and somewhat simpler.  I meant to say: the
material count for R+B is 800 (say) compared to 700 for R+2P.  Without six man
EGTB, the latter would presumably give more practical winning chances despite
seeming less advantageous froma purely material count prespective.  (As a
complete patzer I would prefer the 2P to the bishop in this position.).

Frank



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.