Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Tablebase question/proposal (reasonable subset of 6-man)

Author: Anthony Cozzie

Date: 05:17:01 12/06/03

Go up one level in this thread


On December 05, 2003 at 21:26:49, Robert Hyatt wrote:

>On December 05, 2003 at 12:53:17, Anthony Cozzie wrote:
>
>>First, how far along are we (we being Nalimov) in generating the complete set of
>>6-man TBs? I was under the impression that the whole set is 1TB, so we were
>>about 1/2 done.
>
>Roughly.  I don't think the _real_ value will be known until they are
>finished and compressed...
>
>>
>>I would be really interested in compiling some statistics about the relative
>>usefulness of some of the tables.  There are a lot of 6-man tables that are
>>completely useless [e.g. KQQQKP] that still take up quite a bit of space.  What
>>I propose is that someone (someone being Bob & Nalimov) runs a program that
>>simply checks the win rate for side-to-move for each table.  We then order all
>>the tables by "difficulty to win", so the list looks like:
>>
>>krpkrp 5GB 5GB    50%
>>kbqkbq 5GB 10GB   50%
>>.....
>>krrknn 2GB 250GB  70%
>>.....
>>kqqqkp 3GB 1003GB 99%
>>
>>etc.
>>
>>This way those of us with money/noise constraints can pick out a reasonable
>>subset of the 6 man tables.
>>
>>anthony
>
>That might be reasonable.  the .tbs files contain the data and it should not
>be hard to read them in and summarize each one.  (hint hint).
>:)

I thought I was going to have to run through the entire TB.  Your site seems to
only have tbs files for the 5 man set though.

anthony



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.