Author: Dave Gomboc
Date: 21:12:36 10/15/00
Go up one level in this thread
On October 15, 2000 at 23:44:03, Christophe Theron wrote: >On October 15, 2000 at 23:25:46, Robert Hyatt wrote: > >>On October 15, 2000 at 22:41:31, Christophe Theron wrote: >> >>>On October 15, 2000 at 21:35:33, Robert Hyatt wrote: >>> >>>>On October 14, 2000 at 16:15:17, Thorsten Czub wrote: >>>> >>>>>[Event "Open Dutch CC 2000"] >>>>>[Site "Leiden NED"] >>>>>[Date "2000.10.14"] >>>>>[Round "02"] >>>>>[White "Tiger"] >>>>>[Black "Nimzo 8"] >>>>>[Result "1-0"] >>>>>[ECO "D20"] >>>>> >>>>>1.d4 d5 2.c4 dxc4 3.e4 Nf6 4.e5 Nd5 5.Bxc4 Nb6 6.Bb3 Nc6 7.Ne2 Bf5 >>>>>8.Nbc3 e6 9.a3 Qd7 10.O-O Be7 11.Be3 O-O-O 12.Rc1 f6 13.exf6 gxf6 >>>>>14.Na4 Nd5 15.Bc4 Na5 16.Ba2 Bg4 17.Nac3 Nxc3 18.Rxc3 Kb8 19.f3 Bh5 >>>>>20.b4 Nc6 21.b5 Na5 22.Qa4 b6 23.Nf4 Bf7 24.Rfc1 Bd6 25.Nd3 Rhg8 >>>>>26.Nc5 Bxc5 27.dxc5 e5 28.Bxf7 Qxf7 29.cxb6 cxb6 30.Qc2 Qg6 31.Qa2 f5 >>>>>32.Kh1 f4 33.Bg1 h5 34.Qe2 Qf6 35.a4 h4 36.h3 Qg5 37.R1c2 Rd7 >>>>>38.Qe1 Rdg7 39.Qe4 Rd7 40.Qe2 Rgd8 41.Qe1 Qe7 42.Qe4 Qg5 >>>>> >>>>>Here the thing begins ! >>>>>Rc6 is an easy move for Gambit-Tiger. It KNOWS that these kind >>>>>of moves are playable. and it gives the rook to make the bishop >>>>>on g1 active. >>>>>no search-based chess program can see this IMO. >>>>>this is the paradigm-shift. >>>>> >>>>>gambit-tiger believes that this move could make it. >>>>>it is not KNOWING it. it trusts the evaluations and the >>>>>knowledge it has. >>>>> >>>>>thats chris whittingtons way. here you see christophe >>>>>walking in the foot-steps of chris... >>>>> >>>>>leaving the hyatt-paradigm. >>>> >>>> >>>>I have to disagree. It is not hard to tune my (or any other) program to >>>>play this move. If you watch gambit tiger play, it has some _outrageous_ >>>>scores. In a game on ICC the other night, Crafty was at -.2, gambit tiger >>>>was at +3.2... would you care to guess _which_ was right? >>> >>> >>> >>> >>>Would you care to guess what is the current score of Gambit Tiger against Crafty >>>(including Crafties with an outrageous hardware advantage)? >>> >>>It's easy to laugh about a lost attack, but in this case it seems I'm going to >>>laugh more often than you will. >> >>That remains to be seen. I have watched my share of games where you were >>reporting +3, then 2, then suddenly +.5 and so forth. I Wasn't being >>negative about your program. I was being negative about this so-called >>"paradigm shift" that Thorsten likes to mention. Such a shift is _not_ >>hard to do. CSTal did it. But then making it _work_ is something else. >> >>If you do so, fine. But I have seen my share of evidence that says +3.0 >>scores when material is _even_ is _dangerous_. +1? Maybe. Even +2. But >>to give up a piece for an attack that might well fall flat is nice for wild >>chess, but I don't believe it will work for _consistent_ chess. > > > > >If by consistent you mean boring, then I have to agree. > > > > > >>Feel free to prove me wrong of course. I have been pretty speculative in my >>own way, contrary to what Thorsten might think. > > > > >So be a little bit more, the users of Crafty will like it. > >We are not going to spend the rest of our lifes looking at Genius-style chess >programs, are we? > > > > > >>>If it's not hard to "tune" your program to play moves like 43.Rc6 in >>>Nimzo8-Gambit, then why don't you do it right now? >>> >>>Do it, Bob. Do it right now. And let your overtuned version meet Gambit Tiger. >>> >> >> >>Why would I want to do this? Just because you say so? I don't like that >>type of chess from a machine. Because I _know_ that they don't "understand". > > > > >There are strong chess players, including Tal for example, who have said that >they have sometimes played sacrifices just because they felt it would work. > >Human players think this way, we could maybe learn the lesson. > >It is known since ages that a strong king attack can be worth more than a piece. > >If a chess program is not able to value a strong king attack this high, then it >is missing chess knowledge that even a 1700 elo player has. > > > > > >>And relying on 'I hope" is dangerous. Perhaps your search is good enough to >>weed out some of the failures, I don't know. But it isn't weeding out anywhere >>near all of this stuff. >> >>If you can beat my ears off, so be it. But don't be surprised when some start >>taking advantage of such 'speculation'. It isn't difficult... CSTal was >>sort of scarey at first. Then it because sort of boring... > > > >I'm waiting for others to adapt. > > > > > >>>> I will be more >>>>impressed when I see lots of such moves where _most_ are right... >>> >>> >>> >>>I will be impressed when I see Crafty playing just ONE move of thing kind. >>> >>> >> >>I will be more impressed by other things. IE I have seen GM's have it >>easier with gambit tiger than with regular tiger. based on some games >>on ICC. They don't seem to be afraid of what seem to be the typical types >>of king safety analysis dealing with open files and pieces close to the >>king... If they are convinced the attack is going nowhere, they will snap >>up the piece and take it home... And they have. No, I don't know how the >>new program is doing vs all players. But it doesn't seem to be significantly >>better or worse than any other new program, IMHO. >> >>We will see, I suppose... > > > > >Hey, this is Gambit Tiger 1.0. > >This engine is almost not tuned at all. It was supposed to be just a bonus for >the Rebel 11 CD! > > > > > > >>>> IE CSTal >>>>finds great looking moves that often backfire. Too often. >>>> >>>>Speculation is one thing. Accuracy is another. There _is_ a middle >>>>ground... >>>> >>>> >>>> >>>> >>>>> >>>>>is there any chess program playing Rc6 ? >>>>> >>>>>brilliant. i hope more will follow. >>>>>because this steps into a new quality in computerchess. >>>>>the moment when chess programs BELIEVE !!!! instead of knowing. >>>>>or seeing in the tree. >>>> >>>> >>>>This is so inherently dangerous it is obvious. Just because you _believe_ >>>>you can walk on water doesn't make it so. You had better _know_... If you >>>>are only 10 feet from shore, you might not drown. In the middle of the Pacific, >>>>you had better be _certain_ you can walk... >>> >>> >>> >>> >>>Now I understand better what Chris Whittington calls "the Hyatt paradigm". :) >>> >>>I'm sorry, I never thought I would one day share CW's point of view, but you are >>>pushing me in that direction... >>> >>>Gambit Tiger was not certain that 43.Rc6 was winning. Actually it might even >>>turn out that this move is incorrect. I don't know. >>> >>>And you know what? I DON'T CARE. >>> >>>The only thing that matters is that the opponent has been UNABLE to show it was >>>incorrect. >>> >>>That's where computer chess finally meets human chess. >>> >>>If your program can't swim, then I guess it's better for it to stay close to the >>>shore. >>> >>>Mine is learning how to swim. There are sexy girls on the islands, I guess he >>>wants to go there. ;) >>> >>> >>> >>> Christophe >> >> >>There is also deep water. I'm only aware of _one_ person that was able >>to successfully walk on water. Some 2,000 years ago. :) > > > > >Oh yes. His name began with the letters "C", "H", "R", "I", "S"... > >Mmh... Now I'm in deep water. :) > > > > > Christophe How does that salt-water taste? :-) I think that if a reasonably deep search isn't refuting the attack outright, that itself shows that it's reasonably plausible. There's a lot of room in computer chess for understanding high-initiative positions better. Nonetheless, I don't view this as a "paradigm shift". The function being maximized by the search has been substituted by another one that gives weights much larger than zero to some factors that probably weren't assessed highly enough traditionally, and in the long run (that is, after a certain amount of experimentation) it will likely converge to settings that more accurately capture what's really going on in a position than the function being replaced did. Sorry to seem as if I'm taking all of the fun out of it. <grin> Dave
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.