Author: Uri Blass
Date: 19:41:30 12/02/01
Go up one level in this thread
On December 02, 2001 at 18:54:27, Fernando Villegas wrote: >No, no no. >I do not underestimate nothing. Less tactics, core of the game. What I say is >that, no matter how much you search, if you does not know how to evaluate those >things you see in the search, then you get just the most elemental. So, a good >evaluation function such as that of rebel - call it positional understanding or >whatever, I do not mind- is neccesary with deeper search, not the opossite or >something like a different approach. Both goes and must go toguether. I think >all this is obvious. >My best >Fernando 1)I also do not think that Rebel or shredder has better evaluation function than other top programs. There are positions when they are better but there are also positions when the opposite happens 2)Chess is a complicated game and it is possible to find tactical blunders of programs in games that are considered to be positinal games. If you take the losses of Junior against kramnik or piket and give only material evaluation to analyze positions for 24 hours per move then it is going to find part of the mistakes of Junior that without them the opponent could have more practical problems to win. When humans have to play against good defence they may have problems and they may get a good position only to fail to win it(kasparov failed to win deeper blue in most of the games inspite of getting better positions and I believe that top programs at 24 hours per move are better). Uri
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.