Author: blass uri
Date: 10:19:45 03/22/99
Go up one level in this thread
On March 22, 1999 at 00:18:12, Ed Schröder wrote: >>Posted by blass uri on March 21, 1999 at 12:58:11: > >>I think that we do not know if Rebel10c is weaker against humans because we >>have not data of many games of Rebel10c and Rebel10b against humans. > >>I think that the fact that Rebel10c is better in some tactics may help it also >>in games against humans and maybe adding the tactical knowledge is more >>imprtant then removing positional knowledge also against humans. > >This is very unlikely. Also my testing shows the opposite. I have posted some >of my testing on REBEL-BOARD. I did not see results of games against humans and you need many hundreds of games to prove difference of 20-30 elo against humans. It is clear that sometimes Rebel10c is better and sometimes Rebel10b is better I do not think that tactics is unimportant for computers against humans. Fritz and Junior cannot solve the nolot test positions in tournament time control and humans practically found the right moves so sometimes humans see tactics that computers do not see. 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.