Author: blass uri
Date: 07:46:22 05/26/00
Go up one level in this thread
On May 26, 2000 at 10:37:54, Albert Silver wrote: >On May 26, 2000 at 09:48:19, Frank Wolf wrote: > >>I've been playing around with Hiarcs 7.32, Fritz 6 and Junior 6 to see which >>program gives the best analysis. I use Blundercheck on long time controls >>overnight. Hiarcs 7.32 and Fritz 6 are close, the lines and evaluations may >>differ a little but they basicly say the same thing. Junior 6 however will >>sometimes miss major mistakes. Lines rated minus 10 or worse by 7.32 and Fritz >>will be rated zero on Junior. The bottem line is I cannot trust the analysis of >>Junior 6 (with upgrade). > >Since Junior is probably stronger, I would have thought the bottom line was the >other way around. If it beats those programs, then it's eval is superior, >particularly since it is known that it isn't winning due to better tactics. >Besides, minus 10? That's a tenth of a pawn according to Fritz or Hiarcs. Not >much. > > Albert Silver The post was about analyzing and not about playing. The main difference is that in analysing the program goes backward and in playing the program goes forward. It is possible that Junior6a has some bug in analyzing when it goes forward but I do not know if this is the case. I have not Junior6a but I suspect that programmers tested their programs in game and not in cases when you need to go backward. I found that hiarcs7.32 had a bug in going backward and in one case it did a mistake in blunder check only(I found that when I went backward and gave it to watch the game it also did the same mistake unless I give it enough time when it does not do the same mistake if I give it to go forward in the game) It is possible that Junior6a has more bugs in going backward but I do not know. I trust no program in blundercheck only. 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.