Author: Brian Richardson
Date: 07:00:17 04/08/00
Go up one level in this thread
On April 08, 2000 at 06:20:30, Severi Salminen wrote: >Hi! > >First of all I have to say what a wonderful feeling it is to see when your >program plays the first real chess game. You started from scratch and there it >is: a chess playing program! It almost won Rebel Decade (strenght=Novice, 2 >plys, no Opening book...)but it was a draw by repetition. But a question arose: > >I have now implemented only material and mobility to evaluation, but now when I >search 4 or 5 plies deep the program plays 1.g3? Why doesn't it play 1.e4? 1.e4 >would seem to give the best mobility after the first move. Is it because I >search ONLY 4 plies deep (so, no iterative deepening)? So, the program knows it >can make the move that increases mobility later? Problem is that it never does >that move. It moves the knight back and forth at corner (the Rebel endgame which >led to draw) thinking it can later move the rook to better position and so on. >Would the deepening solve the problem? When I do 1 ply search, it plays 1.e4. > >Severi Congratulations on getting your program playing. Try adding piece square tables to your evaluation function--see TSCP source for simple examples. This will make it much easier to direct opening play and piece development.
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.