Author: Robert Hyatt
Date: 21:45:13 01/25/00
Go up one level in this thread
On January 25, 2000 at 13:07:38, Dave Gomboc wrote: >On January 24, 2000 at 09:09:50, Steve Coladonato wrote: > >>On January 23, 2000 at 17:06:14, José de Jesús García Ruvalcaba wrote: >> >>>On January 23, 2000 at 11:45:29, Steve Coladonato wrote: >>> >> >><snip> >> >>>> >>>>From this and Michel's response, is it not true then that the "best move" is >>>>stored in the tablebases and if so why have the program run thru the legal moves >>>>and probe the tablebases for each one? If you're concerned about legal moves, >>>>pick the best move as noted in the tablebase and then check to see if it's legal >>>>(in case of some error in the tablebase). I realize the CPU time here is >>>>minimal and I'm just trying to get the logic behind the use of tablebases. >>>> >>>>Thanks again. >>>> >>>>Steve >>> >>> Tablebases do not store moves (not even the "best" move). They only store >>>scores for positions. >>> (Assuming the root position is in tablebases and that there are no tablebases >>>missing) the program has to search all the legal ply-one moves, and probe for >>>every resulting position. This returns a score for each one, and the highest >>>scoring one is chosen. >>>José. >> >>Jose, >> >>After I sent the response I realized that the program would have to generate a >>move and then find that position in the tablebase. The table Michel included >>was really only a representation of the tablebase, not the structure of the >>tablebase. So the (init+move) would have to found again using the hash lookup >>and the scores for all (init+move)'s would be evaluated to determine the best >>one. >> >>Thanks. >> >>Steve > >Sounds like you've got it all. > >Dave Except there is no "hash lookup". It is an exact probe into the table made possible by computing a very precise table index that is unique for each possible configuration of pieces 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.