Author: James Long
Date: 17:39:39 12/08/98
Go up one level in this thread
On December 08, 1998 at 00:55:44, Will Singleton wrote: >On December 08, 1998 at 00:12:37, James Long wrote: > >>I'm in. :-) >>Include "Galahad." On ICC, it plays as "jlong." >>The program is brand new and still lacking any >>sophisticated heuristics, but it should improve >>rapidly. When it surpasses Tristram, I'll probably >>have you remove Tris from the list. >> >>-- >>James >> > >Aha, I sense a program redesign. What, bitboards? > >OK, Galahad is on the list (no blitz rating as yet.) > >Will Great job Captain Obvious. :-) Tristram uses bitboards, but it has some serious shortcomings. I finally got tired of patching the same old code. There were a few structural issues that would've been a nightmare to fix. There are a couple of new things I'd like to try, too. I'm wondering just how selective I can get with move generation before it "backfires." For example, is it plausible to develop some algorithm to determine "in this position, a knight move would probably be best," and then gen only knight moves? If so, would it be worth it? Would the time it takes to determine this be as fast as generating the rest of the moves? It could come down to how accurate this algorithm is... Anyway, this is a topic for another discussion, and another time. I have a long way to go before I reach that stage. --- James
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.