Author: Robert Hyatt
Date: 18:52:38 03/23/98
Go up one level in this thread
On March 23, 1998 at 08:12:53, Simon Read wrote: >On March 22, 2098 at 16:49:36:93:54:124:79:2:11347:55a, Robert Hyatt >will >write, among other things: > >> First, you can cull >> moves that were only played once or twice, as those are nothing but >> moves from a single game and are not worth considering without a search >> to verify them. I use the "minplay" option set to 3 when I build a book >> for Crafty, meaning that in the 330,000 game file I use, if a move was >> not played at least 3 times, it is tossed out. > >What about positions? If there's a transposition, a position may be >reached by different routes, so the move immediately before it might >only have been made once, but the resulting position may have occurred >several times. > >This might mean that you throw away a position which has occurred >more than three times because there are 2 ways of getting there, >each of which has been used twice. > >Regards >Simon Crafty doesn't work like that. As I read in PGN for the book, I store the hash signature for *every move* I read. I then sort these so that identical positions group together, then the final phase of the book create squashes duplicates out and counts them. So your example can't happen. the move that causes the transposition might be a singular move and get eliminated, but the position below that position would be kept if it was seen more than twice...
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.