Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Count of positions and encoding

Author: Dann Corbit

Date: 15:56:07 06/28/99

Go up one level in this thread


On June 28, 1999 at 18:39:34, Randall Shane wrote:

>I hope somebody knows the answer to this -- or can point me where to look.
>
>Like KarinsDad (and others), I'm trying to come up with an mimimal encoding
>method for chess positions. What's the best known upper limit on the number of
>unique positions (counting e.p., side-to-move, and castling status)?
>
>I figured out/proved for myself a upper bound of 2**151 -- and I think 2**149
>can be proved as well.  There's an enumeration/encoding method associated with
>it, but it's uglier than sin.  How far off of 'best known' is that?
The current CA format uses about 12.6 bytes per position, and includes both the
suggested move and current statistics!
I don't know if you can hold an arbitrary number of points in that format
without expanding the space per position.  If you could, that would mean that
there are about 100 bits of information in a chess position or less!

I will be interested to see any information on your encoding techniques, as
these are fascinating to me.



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.