Author: Robert Hyatt
Date: 20:11:33 07/29/99
Go up one level in this thread
On July 29, 1999 at 16:48:54, José de Jesús García Ruvalcaba wrote: >On July 29, 1999 at 15:08:36, Marc van Hal wrote: > >>Table base actualy do get importand if in the table base would come more pawns >>fot instance a king rook pawn versus king rook 2 pawns > > This is a seven men ending. > >>and this alo for instance >>king 2 pawns bischop versus king rook > > A six men ending. > >>because this kind of positions acuare a >>lot of times in many games it even would be more intresting to have a table base >>of king rook and 5 pawns versus king rook 4pawns > > Wow! This is a 13 men ending! We are very far from this one. I do not know of >any complete five men tablebase set, but definitely it is posible with current >technology. The six men ending you want is posible too, but I have never seen it >(I have seen only pawnless six men tablebases). The seven men one might be >posible during our lifetime (if we have a long life), but I do not know. And the >13 men, well, maybe for year 3000. >José (: Two things: (1) we already have the complete set of 3 vs 2 5 piece files. We didn't do the 4 vs 1 although Eugene's code will make them. But I didn't see much point in such since they should be easily winnable without tables. (2) KRP vs KR is used a _lot_. In KRPPP vs KRPP, deep in the search, many positions simplify into KRP vs KR, and the table score is perfect. Remember that _some_ programs only probe at the root. Others have 'woken up' and are now probing deep in the search. And by doing so, the utility of the tables goes way up. I see lots of tablebase hits when the root position has 12+ pieces still on the board...
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.