Author: Dieter Buerssner
Date: 16:03:25 05/02/03
Go up one level in this thread
On May 02, 2003 at 18:48:03, Dann Corbit wrote: >>Mod 3) >>Added a save of 2 killer moves at each ply and re-sorted moves before searching You don't need to sort all moves allways. Selecting the currently best move is enough. Often the first or second selection will give a beta cutoff. How to select the best move? Well, it can be the inner loop of a selection search. While this is less effective for huge search problems, for a chess engine, it seems a good choice. Basically, you go over the move list, find the move which is most promising. You swap the place of this move in the move list with the first or last move, and by this make the move list smaller. So for the next selection you try, you only have to process a smaller move list to find the next best selection. >>This made hardly any difference in reducing the nodes searched to get to a given >>ply. It gives a significant difference for me. Perhaps you should carefully look over your code. You might have a bug here. >3. Null move. (make/unmake is the tricky part) Why is make/unmake the tricky part? To me, makeing a null move looks much easier, than making a normal move. All you have to do in a typical implementation, is delete the ep-target (perhaps reset 50 move counter), and increment some index (or pointer) to some move stack and update side to move (which may be implicit by the increment). Regards, Dieter
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.