Author: Uri Blass
Date: 13:55:36 10/28/01
Go up one level in this thread
On October 28, 2001 at 16:32:49, Slater Wold wrote: >On October 28, 2001 at 16:14:31, Ron Langeveld wrote: > >>On October 28, 2001 at 14:28:12, Slater Wold wrote: >> >>>On October 28, 2001 at 13:48:01, John Merlino wrote: >>> >>>>On October 28, 2001 at 13:43:31, Slater Wold wrote: >>>> >>>>>On October 28, 2001 at 13:36:09, Chessfun wrote: >>>>> >>>>>>On October 28, 2001 at 13:28:37, Slater Wold wrote: >>>>>> >>>>>>>That *might* be the worst move ever. >>>>>>> >>>>>>>And they expect this thing to beat Kramnik? Geez.... >>>>>> >>>>>>If they, refer's to Chessbase, I don't think I've ever >>>>>>read they expect it to beat Kramnik. ;-) >>>>>> >>>>>>Sarah. >>>>> >>>>>Ha! Comparing it to Deeper Blue says otherwise. >>>>> >>>>>I don't think they have ever said, "..we will beat Kramnik..", but if they >>>>>didn't think they had a chance why would they play at all? >>>> >>>>Making it this far makes it easier for them to say they have the strongest chess >>>>engine available. This (I would assume) generates sales. >>>> >>>>I'm sure that's the main reasons for the whole thing -- advertising and >>>>exposure. >>>> >>>>jm >>> >>>I just don't think Fritz 7 (on an 8 way box) getting its butt whooped by Kramnik >>>is that impressive. It is exposure, but it's not going to get me to buy it. >>> >>>Seeing moves like this from Fritz 7 is what's going to make my mind up. >> >>Every engine has bugs, even your favourite one. Yet you still bought it. I'm not >>a big fan of Fritz, but your post simply begs for some relativity. >> >> >>Ron > >Let me give you some "relativity": > >Most programs, and yes, even my favorite, are going to make mistakes in their >eval. That's a given. No chess engine is immune to this. > >However, going from 0.00 to mate in x is *absolutly* unacceptable for a >commercial engine. I have yet to see my favorite engine do that. Note that Deep fritz does not see the 0.00 because it has no knowledge about the 50 move rule. I found other problems in the past about the 50 move rule with other programs. I remember that Junior5 lost a game from 0.00 evaluation because it assumed that 100 plies with no capture and no pawn moves is always a draw. It is not always truth and if the final position is checkmate then it is not a draw. Junior5 was not the only problem that suffered from this bug. It is better to be careful not to add knowledge about the 50 move rule unless you are sure that adding the knowledge is productive. It seems that chessbase did not do it and I guess the new knowledge that Fritz7 has about the 50 move rule is counter productive. 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.