Author: Robert Hyatt
Date: 18:31:01 09/29/98
Go up one level in this thread
On September 29, 1998 at 16:38:45, John Coffey wrote: >On September 29, 1998 at 16:32:14, Robert Hyatt wrote: > >>On September 29, 1998 at 16:04:45, vincent dichiacchio wrote: >> >>>Can anyone explain how opening book transpositions are identified? My guess is >>>that opening book files are just permanent hash coded transposition tables, but >>>I am not sure. >>>Thanks, >>>Vince >> >>This is the way I do it, then "transpositions" are handled without trying. > > >Dr. Hyatt, > >Didn't you say that you use the parent position as part of the hash key? >(this might have been someone else?) Anyhow, that would prevent transpositions >would it not? > >John Coffey yes and no... there are two types of transpositions... here is one: e4 e5 Nf3 Nc6 Nc3 Nf6 e4 e5 Nc3 Nf6 Nf3 Nc6 my approach will notice that the position after the above 6 moves are identical, because I don't store paths at all. But it *intentionally* doesn't do the following: e4 e5 Nf3 Nc6 Bb5 a6 e4 e5 Bb5 a6 Nf3 Because white plays Nf3 to reach a known bok position, and then black plays axb5! rather than Nc6, following book. Cray Blitz didn't do this and would fall for the above silly problem... Crafty won't... but once *all* the moves in a book line have been played it will instantly recognize that this is a known book position and start using the book again...
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.