Author: Robert Hyatt
Date: 07:49:39 10/16/00
Go up one level in this thread
On October 16, 2000 at 05:53:47, Thorsten Czub wrote: >On October 15, 2000 at 23:25:46, Robert Hyatt wrote: >>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. > >It works Bob. the main problem with CSTal (that CSTal was not that strong) >wasn`t that Chris was a lousy programmer. He knew about chess and imlemented new >ideas, he was also risky enough to do such a weird program, BUT his programming >abilities were limited. Here is my take: a lot of hyperbole, cast around some pretty neat evaluation ideas, in an effort to make something into a legend before its time. The average consumer might buy into the "search toward unclarity" stuff, but not "this consumer". I don't see any magic in this approach. Nor do I happen to believe that searching into positions that are unclear is a smart thing to do. As far as Tiger goes, I have not seen Christophe claim that he searches toward positions that are vague. Rather, it is a combination of evaluation tuning to exaggerate king safety terms, plus (maybe) some specific search extensions for positions where certain evaluation characteristics are noted. The first approach is more hyperbole than anything else. Because it doesn't matter if you search into positions that are unclear, and then lose them. Sure, more speed or better search would help. But that is true for _all_ of us. Yet we have to make our compromises as to what is included and what is excluded, to maintain what we consider to be a search fast enough to avoid tactical pitfalls. As far as wild play, SJLim on ICC played around with the tuning options in Crafty, which let him turn it into a quite wild player. Or should I say wild and unsound player. The only thing I was pointing out is that if a program evaluates something as +3, when material is even, then it will evaluate that same position as even, when it is a piece down. It had _better_ be right. And from what I have seen, most are not. Tiger might well play quite aggressively without making this mistake. I only noted that in some games I watched (long games, not blitz) its play would be called "risky". Kasparov has shown that risky chess can work. Karpov showed that safe/sound chess could _also_ work. There is certainly room for both. Particularly until it has been shown that "risky" can beat "sound" more than it loses. >CSTal works. it would kill its enemies like hell when it would make more >NPS and when the bugs would be out of the program. But the idea works. >Its not important that the score changes. the score MUST change. its only >important to win. cstal was not always able to win, but not because it played >weak, it lost because it was outsearched. >Gambit-Tiger is a program you cannot OUTSEARCH that easy. And it makes 8 x more >NPS. And you _really_ think that the algorithms used in Tiger are related to the algorithms in CSTal? I would make a wager that they are _not_ similar in any way. Tiger is still a fast searcher, with a relatively simple evaluation (relatively simple -> similar to the rest of the best). Just because you happen to like its style of play doesn't mean it is something remotely related to CSTal, IMHO. > >The paradigm is shiftet Bob, in a few years ALL chess programs will have to >play like gambit-tiger and cstal. if they don´t do, they can never catch >gambit-tiger. This is an easy hypothesis to test. We only have to watch and wait. As time goes forward, it favors the more accurate searcher, due to hardware speed improvements. > >Its over IMO. the days of accurate computerchess have are over because they >don´t get the most points anymore. you get more points playing >unsound moves. chris was right. Kust my opinion. Nothing personal Bob. >why not a paradigm change ? a new century has just began a few months before. >why not change old ideas ? when it works. Show me the solid evidence that shows it works. I played a zillion games vs CSTal (on ICC) using equal hardware. (P6/200 single cpu at the time). It didn't work then. 2 of every 3 games ended in an endgame. and 9 of every 10 of those ended in a loss for the speculative program. In the games that didn't reach an endgame, CSTal wasn't winning more than it lost. We even played games with my king safety totally disabled. Which made Crafty also play quite aggressively. And the search speed _still_ decided the game, even though one program claimed to understand king attacks and the other had _nothing_ in the evaluation about king safety at all. I've always been good enough to recognize things that work and try to incorporate them into my program when needed, but I don't pay a lot of attention to results obtained by beta testers, since _every_ new program always produces significantly better results against the rest, until the program becomes available where it can be studied a bit. > >>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. > >how can gambit-tiger win dutch-championship and wipe out shredder4 with 11.5-2.5 >(only 1 loss !! rest wins and 3 draws) when it WILL NOT WORK FOR CONSISTENT >CHESS ???? can you beat a world-computer-chess-champion (paderborn) 11.5-2.5 >by playing Inaccurate ?? you can ! >therefore i do say: the paradigm has shifted. chris w.´s crazy ideas >have been proved right. First I don't think Tiger has _anything_ to do with "Chris w's crazy ideas". Nothing at all. Other than an evaluation that is tuned to be speculative when dealing with king safety terms. Second, let's see what happens when the program 'goes public' where everyone can look at the positions and evaluations to see what is going on. I remember a prior Nimzo version that was beating everyone in endgames. It later turned out that it was doing so by using larger than usual scores for passed pawns. Most adjusted and this advantage is now _zero_. In my case, I will believe that kingside attacks work when I see them work. So far I have not seen them work. Tiger is simply a good program. I am not yet convinced that the speculation is needed, nor that it makes it better. > >what a surprise, isn´t it :-)) > >>Feel free to prove me wrong of course. I have been pretty speculative in my >>own way, contrary to what Thorsten might think. > > > >>>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". >>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. > >no. relying on I HOPE is pretty normal. your relying on EXACT is the exception. >you lived 40 years in the paradigm of EXACT computer-chess, like newton lived >in his flat-2-dimensional world of forces. Then came einstein and >showed that the forces are just part of the space not beeing flat. >Believing is what makes a human. Hoping. and you cannot measure hope or >believing. how do you want to measure Rc6 ? >is it working ? is it not working ? is it correct or incorrect ? I rely on exact calculation whenever I can, when I play chess. I am _certain_ that Kasparov does the same. I doubt that against other GM players he casually says "I play this, it seems very unclear". Rather, I have watched him calculate for many minutes to be sure that he likes the position he is playing towards. > >you let your program NOT play those unsound moves, therefore they cannot >defend against christophes gambit-tiger, because they cannot SEEEEEEE the things >come. What if there is nothing to see? That has happened. >And if they eat anything he gives, they are only PASSIVE followers. they have to >do what gambit-tiger gives them. if he sacs, they have to eat. they lost >completely the control about the game. maybe or maybe not. If he offers pieces, and the attack fails, they are not really "following" but leading, themselves. > >Or what do you want to do with moves like Rc6 ? >You cannot defend against those moves. not with a hyatt-paradigm- >program. > You are certain that wins? I am not. That is what I do with those moves. >You would need what you don´t have. > >>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... > >?!? depends. i don´t find it boring. and i don´t find gambit-tiger boring. > >>>> 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. > >of course !! because GM´s have THIS kind of knowledge much better in their >brains. >GM´s and any human see that Rc6 is the right move to play. >without thinking they play moves like Rc6. >This is the difference between quantity and quality. >but christophe will teach gambit-tiger about this knowledge, as >chris w. has done, and then the GM´s will NOT beat gambit-tiger >more easily. >don´t you understand ? >the reason they beat a mechanistic-chess-program less good is that >they know about Rc6 believers moves, and have problems (kramnik game4) >to play accurate. therefore they win maybe better against gambit-tiger >(if you are true) and this will be over when christophe (or chris) >have tuned and programmed on this. >meanwhile it is good enough to beat the bean-counter programs not recognizing >stuff like Rc6. otherwise a 11.5-2.5 against shredder4 (that is not a weak >program at all) would not be possible. > > >>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. > >and in future they will have more problems. but this is the only >way to beat both. humans AND dump-chess programs. to shift the paradigm. > I disagree. obviously. There are other ways to accomplish the same thing. > >> 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. > >?!? it is better. >it can play chess, the others simulate it. > > >>We will see, I suppose... > >yes. time will tell it. > > >>>That's where computer chess finally meets human chess. > >exactly. this is why the old paradigm is dead. A quote of "the old paradigm is dead, but reports of its death have been greatly exaggerated. > > >>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. :) > >:-))) > >learning swimming is enough to get to the girls.
This page took 0.02 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.