Author: Antonio Dieguez
Date: 12:39:00 01/04/03
First, thanks for documenting your (special) program. About Lazy Eval, I had a hard time figuring exactly what you mean. I see that to use it the following condition must be true: I_SCORE[ply] + (I_SCORE[ply-1] - SCORE[ply-1]) + MARGIN <= ALPHA I dont know if the expression in parenthesis must have the other sign or if you use an abs() but that's not very important. After a while I understood it, and it is very similar to what I actually use, except that instead of I_SCORE I use only material, and I do it at the next ply, so checking if something is >=BETA, etc. Well, I think you can explain it better man!. Initially a confused pal (like me) could argue you this: "I think if I_SCORE and SCORE differs too much, then it looks to me ok to increase the margin, but, if I_SCORE and SCORE are pretty much the same I don't think the position is necessarily quiet, for example let imagine a position with oposite castles and both sides attacking, where anything can happen, the difference there can be very small but leting the margin be 0.5 or close to 0.5 is not a good idea." What is natural for me is the formulation as below: (note that scores and i_scores are respect to the side to move in each ply) -SCORE[ply-1] + (I_SCORE[ply] - -I_SCORE[ply-1]) + MARGIN <= ALPHA But it is exactly the same as above! wow! a miracle... PD: I am still a bit confused :) I have to clear my mind...
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.