Author: Severi Salminen
Date: 23:54:57 02/03/01
Go up one level in this thread
>It sounds like I can have several null moves going at one time. >Lets say that its whites move. >White does a null move, black cannot do a null move because white is doing one. >White gets to move again, but since black did not do a null move, white can do >another null move. >The search goes back to the first ply and Whites score is not >= beta, so it >continues and makes a move. Black sees that white is no longer doing a null >move, so black makes a null move. At the third ply, white cannot make a null >move because black is making a null move at the 2nd ply. > >Am I way off base here ? No, you got it totally right. I was also confused at the first time because I thought that you try only _only_ nullmove at a time - (I wonder how this would perform?). But you can have _many_ null moves going at the same time. Just find a way to inform the next ply whether you tried a null move or not. I first made the mistake _not_ to clear en passant information. This resulted some interesting positions: 1.e4 NULL 2.fxe3ep. After these moves I had a white pawn at e3 and e4 and a _black_ pawn at e2! Funny - all because of my makemove(). And BTW you can search using (-beta,-beta+1) instead of (-beta, -alpha) - we only want to know if the score is higher than beta our beta. If it is below beta or alpha we don't care (unless for threat extensions...). Severi
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.