Author: Bob Durrett
Date: 12:10:10 10/23/02
Go up one level in this thread
On October 20, 2002 at 17:17:43, Jeremiah Penery wrote: >On October 20, 2002 at 14:37:09, Uri Blass wrote: > >>On October 20, 2002 at 14:05:48, Jeremiah Penery wrote: >> >>>Would you rather have a program with no king safety (no knowledge), or a program >>>that over-evaluates it (untuned knowledge)? >> >>It is not clear >> >>I think that a program with no king safety may be also better. >>The question is what is the size of the mistake in the evaluation. >> >>I agree that slightly untuned is better than nothing but the question is if >>deeper blue was only slightly untuned. > >I can promise you that if you have a GOOD king safety (maybe something like >CSTal 2 had), no matter how badly it's tuned, you will win way more than you >lose against an opponent with NO king safety, if search speed is roughly >equivalent. The reason CSTal lost so much against other computer opponents is >because it searched so much slower. Deep Blue supposedly evaluated the same >kinds of things as CSTal, but it did not have to be extremely slow because of >it. Hmmm. This suggests a possible new feature to add to current chess engines: Provide a way to disable the "king safety feature," whatever that is. Then the person playing with the computer could have fun attacking the program's king. Sounds like fun to me. It is always fun to mate. Bob D.
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.