Author: Tord Romstad
Date: 09:25:30 08/19/03
Go up one level in this thread
On August 19, 2003 at 09:05:09, Omid David Tabibi wrote: >On August 19, 2003 at 07:20:27, Steve Maughan wrote: > >>Omid, >> >>>I was wondering, wouldn't it be better to write an SEE which neglects backed- >>>up attacks, and not use it for pruning? This way it yields a good move >>>ordering at no cost (e.g., using Ed's attack tables), and isn't liable to >>>risks of pruning. >> >>If you read Ed's page you'll see that Ed's method accounts for x-ray threats and >>is probably (almost certainly) more costly than SEE. The reason Ed makes it >>work is that he's using the information for other reasons as well. > >Well, I use them for king safety too. The problem is that it is too costly to >calculate the attack board at each node, and I still haven't found a good way to >update them dynamically. IIRC, Ed calculates them at every node, and Rebel is very fast, so clearly it is possible to do this efficiently. I wish I knew how he does it. :-) Phalanx is open-sourced and computes similar tables at all nodes, but is not as fast as Rebel. Tord >
This page took 0.01 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.