Author: Robert Hyatt
Date: 10:02:11 08/30/00
Go up one level in this thread
On August 29, 2000 at 16:13:31, Severi Salminen wrote: >Hi! > >>I plan on implementing it to see what happens. This heuristic is related to >>killer moves and uses the from and to squares in a 64 x 64 array to maintain >>history information when moves are bestmoves or cutoffs. Each entry has 2 to >>the depth power added to it when a bestmove or cutoff is found. > > >Why do you add a value depending on depth (2^depth)? Why not just increment by >1? Just asking because I'm new to chess programming techniques and I'm starting >to program my own creature... > >Severi Think about this: which move would you trust as being "best"... a move near the root that failed high, or a root near the tip that failed hi? Using depth^2 (where depth is the number of plies remaining before quiescence) tends to favor moves near the root. This is good and bad, and probably needs thought. IE a move that fails high near the root may not fail high near the q-search because it isn't searched nearly as deeply and the underlying reasons why it failed high near the root might not be found with a shallow search. Similarly, moves that fail high near the tips may not fail high near the root as the 'near-the-root' search will be much deeper and see far more. A better approach might be to do this on a per-ply type of basis. So that at depth=N, you try the history moves found at other depth=N searches, before you consider history moves for non-depth=N positions. Would take some thought to make this work reasonably quickly. But it probably would be better. But as a sloppy first approximation, depth^2 works well, for the reason given.
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.