Author: Bruce Moreland
Date: 17:08:04 08/17/99
Go up one level in this thread
On August 17, 1999 at 14:26:28, Jason wrote: >I'm kinda new to chess programming, but not programming in general. Anyway, I >found the below description of alpha-beta searching on Paul Verhulst's chess >page (http://www.xs4all.nl/~verhelst/chess/search.html). Abandoning a branch b/c >of a cutoff that's HIGHER in value, but somehow a worse move, doesn't seem to >make sense to me. You're in a position, searching moves for your opponent. You've made a move to get here, so it's your choice to be here. There is some other move you could have made instead, and you've searched that fully, meaning that you've looked at all of your opponent's replies, and you know exactly how badly he can bash you, which in this case isn't very bad. So here you are in this position. You search some moves for you opponent, and you find a move that's just tremendously awesome, it wins a rook for nothing. Well, you say to yourself, I know that I can do better than this. I don't have to make the moronic move that got me here and let my opponent win my rook, I can make some other move instead, which I have already searched fully, so I already understand it, and it leads to a way better situation than the loss of my rook. So what do you do, continue looking at the rest of your opponents moves here, to see if he can win your queen instead of a rook, or perhaps he can mate you? Or do you say, yuck, this situation sucks, I'll never allow this? Alpha-beta allows you to simply bag these bad positions, if you know you can do better some other way. bruce
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.