Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: which 6 man tablebases are the most important?

Author: Dann Corbit

Date: 16:32:33 04/05/04

Go up one level in this thread


On April 04, 2004 at 13:49:00, Uri Blass wrote:

>On April 04, 2004 at 12:48:00, Vincent Diepeveen wrote:
>
>>On April 02, 2004 at 23:17:58, Robert Hyatt wrote:
>>
>>DIEP raw compressed format :
>>
>>28-03-2004  02:15           218.108 kqqqkp_w.dtb.emd
>>28-03-2004  02:14           134.072 kqqqkp_b.dtb.emd
>>
>>Same tables in nalimov format:
>>-rw-r--r--    1 500      500      650212520 Sep 11  2003   kqqqkp.nbb.emd
>>-rw-r--r--    1 500      500      115294214 Sep 11  2003   kqqqkp.nbw.emd
>>
>>Factor 1000 compression or so here.
>
>
>I am not impressed.
>
>KqqqKp is almost always a win for the stronger side.
>
>I suspect that the only cases when it does not happen can happen only when there
>is a stalemate or when there is a pawn in the 7th and the drawing move is a
>promotion.

Which begs the question, if the excess information is excess baggage, then why
are the Nalimov tables so large.

>You simply do not need these tablebases because you probably can by static
>evaluation solve everything and in the rare cases when there is a pawn in the
>7th that can promote you can look in tables with no pawns.

Perhaps this idea can create a great compression for special tablebase files
like this one.  If we store ONLY moves that do not lead to a win for the
stronger side, then perhaps the whole thing can be stored in 10,000 entries or
so.

For really dominant tablebase files (lopsided 95% or more) it might be a good
idea.



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.