Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: likelihood instead of pawnunits? + chess knowledge

Author: Ingo Lindam

Date: 10:49:07 10/26/02

Go up one level in this thread


On October 26, 2002 at 07:40:16, Josť Carlos wrote:

> Thanks for the example. Now I understand.
> I have thought about patters sometimes. The problem I've always found is: a
> pattern (in this context we're talking about) is a subset of the board. In an
> endgame position with few pieces you could probably use that scheme, but what > do you do in a midgame position full of pieces?
> If you use locality to determine the patterns (in your examples, all pieces
> are within the same area) you miss the action of sliding pieces from the
> distance. It could be possible to find the distant influencing pieces, and
> include them into the pattern... right?
>
> Josť C.

Oh well , I am particulary thinking of midgame positions. Not only subsets of
the board are possible pattern. A pattern of a position might be ofcourse a Kg1
or a Ph3 but also a R somewhere on the h file or a rook on any open file, the
number of pawns, Queens on board, having the pair of bishops, ... and ofcourse
also distances between a piece and an subset of the board or a particular
square. I will not stop the computer to find something important I did't think
of.

But I am not very interested in patterns I can't find in my huge amount of games
or in patterns that don't tell me something about the probability whether game
will end in a win, draw or loss.

Ingo


And with a, b being patterns also a||b and a&&b are pattern candidates.



This page took 0.01 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.