Author: Robert Hyatt
Date: 18:40:37 06/21/00
Go up one level in this thread
On June 21, 2000 at 19:00:45, Mogens Larsen wrote: >On June 21, 2000 at 17:06:29, Robert Hyatt wrote: > >>I then backtrack thru the book line averaging in this learn value until I reach >>a point (going backward) where the program had a choice. I reduce the learn >>value by dividing by the number of choices and continue working back toward the >>root, dividing whenever there are choices. > >How do you do the averaging with the learn value when backtracking through a >certain line? Are you reducing the number of evaluations that were used to >calculate the learn value and recalculate as you go back? > >Best wishes... >Mogens First, the 'learn value' is computed once and remains the same as I start backing up thru the book line. If there is only one choice at a node in the book tree, the learn value is just "averaged" in with whatever is already in the learned value for that book move. Whenever there is more than one playable alternative, the learn value is divided by N (N=number of playable moves only at this node, does not include moves that are marked as never play, or that have been learned to be never played). This decreased value is used above this point in the tree.
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.