Author: Sune Fischer
Date: 07:10:24 08/17/03
Go up one level in this thread
On August 16, 2003 at 02:28:30, Johan de Koning wrote: >>Clearing TTs makes the engine, IMO, significantly weaker at short time controls. >>It is simply too expensive to throw away the little information the engine has >>collected, and the fraction of a second the clearing itself takes is no small >>handicap (guess there are tricks to speed this up? :). > >Time control is not an issue here since. How so? >>It also makes backwards (=retograde?) analysis impossible, and as Gian-Carlo >>mentioned smp is not possible to do in a determanistic manner. > >Valid points. >But IMHO not a good reason to always behave randomly. I think "randomly" in the sense "use past experiences" is a positive thing. As we've been over certain things are just impossible with the pridictability straight jacket on. Now I can see why you would want pridictability during testing or bug fixing phases, but when playing for real all you want is maximum strength. I have come to learn, that chess is really a game of chance, everything being horizon limited makes it all pretty random anyhow. Maybe pridictability is even a bad thing here, it disrupts the natural laws of statistics, making it possibly even harder to test. >Bells and whistles are fine, as long as you can control them. No, bells and whistles _rock_ when you can control them :) >That's what 1950s AI (p)research was about. :-) Heh, given that some of the top engines of today can't even mate with a rook without EGTBs, maybe it's time to reconsider what it is we want in an engine. -S. >... Johan
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.