Author: Uri Blass
Date: 04:20:36 01/21/02
Go up one level in this thread
On January 21, 2002 at 06:42:44, David Rasmussen wrote: >On January 21, 2002 at 05:15:45, Uri Blass wrote: > >> >>You assume that material is evaluated as a constant and positional evaluation is >>added later >> >>It is not the case in my program >> >>I have a constant array for evaluating white pieces >> >>int pcsq[6][64] >> >>When I need to evaluate a knight at b2 I simply translate knight to 1 >>I translate b2 to 9 and I add pcsq[1][9] to my evaluation. >> >>Can you tell me based on the array pscq[6][64] what is the value of the knight >>in my program? >> >>Uri > >No matter how you do it, you still have some sort of base value that will tell >you that a knight is not as good as a queen. The average value of a knight in >your case is probably around 3, I would guess. It doesn't make a difference >whether you use piece/square tables. You still have to decide. Piece/square >tables can't solve all positional problems either. A knight on h4 isn't always >bad. Of course, if you use pre-evaluation piece/square tables, it can get pretty >strong. But you still have base values that are supposed to take care of >material inbalances. > >/David The problem is how to decide what is the base value. I can do average of the piece square table but I think that in this case the number that I get is too small because a knight has not the same probability to be at every square and in most practical cases it is in relatively good squares. Uri
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.