Author: blass uri
Date: 04:51:03 08/02/98
Go up one level in this thread
On August 02, 1998 at 03:33:52, Thorsten Czub wrote: >On August 01, 1998 at 08:34:15, blass uri wrote: >>The reason I believe Fritz5 has better general position understanding >>than Junior4.6 is that Fritz5 had better results when mark young gave it more >>time. > >Junior plays very okish games. This remembers me on how rebel plays. Many moves >Junior or Rebel play, I would have played too. The moves they play look NORMAL. >Human like. They play very good. It can be that Fritz kills Junior in longer >time controls. I am sure fritz-tree is more effective than junior's tree is. >Juniors biggest problem is in my opinion the horizont thing. >It generates very likely horizont ideas into its main-lines. When analyzing with >junior on a position, you often see that it moves a problem over the horizont by >creating an interesting , very artistique bluff mainline, that does work within >the range of the search, but does not work out of it. >I guess the reaso for this is the way junior searcher, in opposite to the way >fritz searches. Fritz outsearches junior at longer time controls. Yes. >But i like the way junior plays. And I know other programmers have found the >same. Junior plays homogen, as if the whole game is ONE BODY, not this up and >down chess, it plays elegant. > >>I am not sure about Junior5 and I should do another test >>I think to let Junior5 play against fritz5 all the positions after the 10th move >>of black in the match rebel-anand and to see what is the result. > >I guess you should try to play fritz from the 10th moves, against junior, but >also against Hiarcs. hiarcs would not allow fritz to play POSITIONAL ! >maybe fritz outsearches hiarcs, but fritz will not impress hiarcs with >positional stuff... > >Yes, the anand matches , especially the 2 tournament games could be interesting >to see played by two computers. Why not use the 2 40/120 games and replay them >with computers ?! I use one computer for Fritz-Junior game and the game is automatic. I think to test the 2 40/120 games and the games morovich-ferret in 30 minutes per game+30 seconds per move and in 10 minutes per game+ 10 seconds per move. I will use the last version of Junior Amir Ban sent me(he improved the last version he sent me). Uri > >>This will give an indication which program is better in general position >>understanding. If Junior's result improves with time than Junior has better >>positional understanding. >>I hope Junior will win this match > >I like that you have not seen my statements as a personal attack !! > >I see your posts very constructive. I hope we can discuss things later... > >Try the anand-match games instead of the nunn-stuff. > >>Uri >>understanding > >Thorsten >apologizing for his overreaction !
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.