Author: Bruce Moreland
Date: 01:27:01 08/02/01
Go up one level in this thread
On August 02, 2001 at 03:54:39, Dann Corbit wrote: >On August 02, 2001 at 03:44:01, Janosch Zwerensky wrote: > >>Hi all, >> >>I read some time ago that Deep Blue wasn't using heuristic game tree pruning >>methods (like, for example, the null-move technique). >>Since null-move was known when DB was around, can anyone here tell why the DB >>team decided not to use it (or wasn't able to do so)? > >Safety. >But don't imagine brute force mini-max. Not like that at all. As a matter of a >fact, beyond 30 seconds, wonderful things might happen. When I started doing computer chess there were people searching (gasp!) nine plies, and some of these people were talking about how the tree is way different when you search that deep. There was talk of tactical sufficiency and lots of other craziness. Now that we can do nine-ply searches in blitz games, a lot of that talk drops by the wayside. I've always wondered about DT/DB and null move. It may be that they had everyone so incredibly supersetted that they didn't need to mess with stuff like this, but I'd think it was pretty incredible if they were still not using null move. Null move is great, and as far as I can tell it works at any tree depth. This is one of the reasons I don't just keel over and die when Bob argues that they are so bloody fast, and therefore they must be godlike. Yes, they are fast, but they don't use the same kind of search. Maybe 98% of that tree is crap, because against a human or a micro program running on a 286, it makes sense to not make pruning mistakes rather than search an extra five plies (all numbers approximate and probably way off). I don't know. There are ways to find out, but they involve being able to test the thing. 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.