Author: Robert Hyatt
Date: 19:50:07 08/14/05
Go up one level in this thread
On August 14, 2005 at 22:34:11, Uri Blass wrote: >On August 14, 2005 at 21:29:50, Robert Hyatt wrote: > >>On August 14, 2005 at 20:08:21, Uri Blass wrote: >> >>>On August 14, 2005 at 18:01:28, Robert Hyatt wrote: >>> >>>>On August 14, 2005 at 17:00:02, Bernhard Bauer wrote: >>>> >>>>>On August 14, 2005 at 16:47:43, Darrel Briley wrote: >>>>> >>>>>>On August 14, 2005 at 15:22:43, Graham Banks wrote: >>>>>> >>>>>>>Best wishes for the remaining games. It would be really good for this to be the >>>>>>>best Crafty ever! Fingers and toes crossed! >>>>>>> >>>>>>>Graham. >>>>>> >>>>>>1.5 with two blacks vs. Shredder and Junior. Impressive indeed. >>>>>>Congratulations Bob! >>>>>> >>>>>> DB >>>>> >>>>>As Bob has pointed out earlier this is pure luck. >>>>>Crafty has serios tactical weaknesses as all the >>>>>computer chess specialists know. Crafty is not even commercial, >>>>>so how can this be explainded in any other way, but luck? >>>>>Perhaps Bob has found a very old version and is running this now, 6.4? >>>>> >>>>>Anyway congratulations and >>>>>Kind regards >>>>>Bernhard >>>> >>>> >>>>I'd bet that those "serious tactical weaknesses" are not to be found on this >>>>hardware, at least until someone is searching far faster than I am... >>> >>>Tactical weaknesses can be found also in software. >> >> >>And they can be compensated for by fast hardware. >> >>> >>>> >>>>Crafty's "serious tactical weaknesses" are more urban legend than fact... >>> >>>Comparing Crafty with fruit Crafty does not do checks in the first ply of the >>>qsearch when fruit does it and Crafty does not use history based pruning when >>>fruit does it. >> >>So? What says checks in the first ply of q-search is better? Crafty early >>versions prior to the Jakarta version did checks in the q-search. Any idea why >>I might have stopped? Both Bruce and I (Ferret) decided after much testing that >>it was no worse at all, and the extra search speed was actually beneficial... > >Interesting that other programmers found different things so maybe something was >wrong in the implementation. Or maybe the programs are different enough? I go for extending in the basic search which is more accurate. I avoid extending in the q-search which is less accurate by definition. I just happen to believe that idea is right. I can think of one commercial program that has taken this quite a bit farther than I have, if you know who I mean. :) And it seems to be working for him perfectly well also... I don't participate in the "herd mentality". Just because everyone else is doing it doesn't make it the right thing for me to do... > >I know that you used checks in the qsearch and stopped but the question is if >you used it in the right conditions. I tried _dozens_ of conditions, all the way back to Cray Blitz which also did them until the last two years of its 'life'... > >> >>forward pruning certainly introduces tactical problems more than it cures them. >>Look at the Rxh6 game vs shredder the other night. Whredder somehow just >>overlooked the hung pawn... > >I am not sure about what game you talk >I see no possible Rxh5 in the games of shredder in this tournament. No, this was a game DB posted from Crafty vs Shredded when I was testing the opteron system on ICC a couple of nights ago... > >Note that in the case of fruit to be technically correct it is not exactly >pruning but reduction and the idea is that moves that almost never failed high >are searched with reduced depth if they are not in the top of the list and only >if they fail high a research is done with the original depth. > >Uri I understand that idea. It isn't new. You can find a discussion about it between Bruce Moreland and myself (among others) probably around 1996-1997 or so...
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.