Author: Jari Huikari
Date: 04:20:40 10/27/98
Go up one level in this thread
On October 26, 1998 at 09:15:19, Robert Hyatt wrote: >As different moves become best, I sift >them to the top of the ply=1 move list, without disturbing the order of the >others, ... >Between iterations I keep the best move at the top, and sort the rest based on >the total nodes searched for each one, which will bring (generally) potential >new best moves to the front These sound good ideas. My temporary(?) solution was that I don't do cutoffs in ply=2 immediately when I find that some move is (slightly) worse than the best one. But I search until I find it's worse than the best minus SOMECONSTANT. So the values of the moves slightly worse than the best one will be safer to use, if needed. This means a longer time spent to an iteration. Perhaps move ordering is better and pays back some of the lost time. Now warn me if there are some dangers in this, I have not thought about. Also I'd like to hear what would be a good value for my SOMECONSTANT. Thanks for all the answers! Jari
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.