Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: A good idea for saving space in tablebases

Author: blass uri

Date: 11:47:51 10/10/98

Go up one level in this thread



On October 10, 1998 at 14:11:47, José de Jesús García Ruvalcaba wrote:

>On October 10, 1998 at 14:03:26, blass uri wrote:
>
>>
>>On October 10, 1998 at 13:55:53, José de Jesús García Ruvalcaba wrote:
>>
>>>On October 10, 1998 at 11:33:14, blass uri wrote:
>>>
>>>>I think that in some endgames like KRN vs KR you can save space
>>>>by defining rules to do the right moves:
>>>>
>>>>If there is conversion in a small number of plies you can find it by search and
>>>>not use the tablebases to tell you the right move
>>>>
>>>
>>>You can do the search if the position is in the root.
>>>But what if it appears after a 18 ply search?
>>you do not define 18 as small
>>you define small only if you are sure that you can do practically the search
>>let say less than 8 plies
>>
>>If the rules for finding the right move does not tell you the right move and it
>>is at least 8 plies to conversion you save it in the exceptions.
>>I believe that there are not many positions like this.
>>
>>Uri
>
>Perhaps I did not state my point clearly.
>	In the search tree, do appear many weird and ridiculous positions. Imagine that
>after a 18 ply search in one leaf node you have a position like KRNKR. It is
>unlikely it is a "sane" position. But you need the score, and if it is an exact
>one it is best.

I need the score for win,draw ,loss if my rules for win,draw ,loss that are not
based on search are wrong but I do not need the exact number of moves before I
go to KRNKR
and when I am in KRNKR I can find the right move by search.

I have the following steps when I am in KRNKR:
step 1)If position is in the exceptions look in the tablebases for the right
move otherwise continue to step 2
step 2)If I find conversion in a small search I play the move that is leading to
conversion as soon as possible otherwise I continue to step 3
step 3)I play the move by the rules I defined

The problem is to define rules such that the set of the exceptions will be
small.

I store in the tablebases the set of positions when my score for result by the
rules I define is wrong with the right score,
and the set of positions when my move is wrong by the rules I define with the
right move.

Uri

>	Is it affordable to extend eight more plies on that position? It is very likely
>that many more leaf nodes with the same reduced material appear. Is it
>affordable to extend eight more plies on an unspecied number of them?



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.