Author: Robert Hyatt
Date: 21:06:22 11/08/00
Go up one level in this thread
On November 08, 2000 at 21:17:59, Christophe Theron wrote: >On November 08, 2000 at 15:01:23, Robert Hyatt wrote: > >>On November 08, 2000 at 00:41:05, Christophe Theron wrote: >> >>>On November 07, 2000 at 09:33:56, Joe Besogn wrote: >>> >>>>On November 06, 2000 at 10:41:01, Fernando Villegas wrote: >>>> >>>>>On November 06, 2000 at 09:58:56, Joe Besogn wrote: >>>>> >>>>>>You are, imo, running round in circles due to the usual reasons of vested >>>>>>interests, but also because you argue without even a basic understanding of the >>>>>>terminology. >>>>>> >>>>>>Try: http://cgi.student.nada.kth.se/cgi-bin/d95-aeh/get/kuhneng >>>>>> >>>>>>for a concise description of Kuhn's ideas in the "History of Scientific >>>>>>Revolutions". >>>>>> >>>>>>Then, perhaps, there might be some interest in reading what you have to say on >>>>>>the subject amongst the more enlightened. >>>>>> >>>>>>The descriptions in the article, imo, almost exactly mirror actions and progress >>>>>>within computer chess. That's imo. >>>>>> >>>>>>Although you are unlikely to reach agreement on new/old paradigms, existence of, >>>>>>or whatever, at least you'll have some new agreement on what words mean. That >>>>>>helps. >>>>>> >>>>>>Also useful for you will be the realisation that a paradigm is not a 'chess >>>>>>playing computer program', but a 'system of thought'. The fact that it is >>>>>>possible to take a conventional chess program and apply new ideas to it, does >>>>>>not mean that a paradigm shift has not taken place. The revolutionary shift is >>>>>>in 'ways of thinking' or in 'world view' - rather more difficult than changing >>>>>>code. The paradigm shift, therefore, is in you, in your own head. Some make this >>>>>>shift faster than others, one revolutionary starts it off, some see it soon, >>>>>>some see it later, some never see it at all. The ones that don't see it, deny it >>>>>>exists. The ones that do see it, say "you need to think different". The ones who >>>>>>see it late claim "it's evolutionary, I could do that". >>>>>> >>>>>>Why do I always try to help them ?! >>>>> >>>>>Hi. >>>>>You are right that sharper definitions of words helps, but not so much and not >>>>>always is neccesary, anyway. There is room, in a casual debate as those >>>>>performed here, to some fuzzy logic. I think every one here -or almost- knows >>>>>how radical a "paradigm" is, but nevertheless we all understand that, when >>>>>Thorsten uses it here, he is just referring to a more modest thing: a new way >>>>>to understand how to program certain functions of a chess program. In that sense >>>>>the word is useful and it would be a kind of pedantry to argue againts him on >>>>>the base of the exact definition of what a paradigm is. Besides, to define words >>>>>tends only to open another field of debate instead of solving it. >>>> >>>> >>>> >>>>>How much >>>>>radical a new way of thought has to be to be a paradigm? >>>> >>>>Powerful question. It has to be a new way of thought. In itself the new way of >>>>thought does not need to be so earth-shattering. Often it is a >>>>why-didn't-I-think-of-that or how-obvious or whatever. Einstein only questioned >>>>the idea that velocity vectors could be added together without limit. His >>>>theories flew from the simple-to-us-now idea that there was an upper limit on >>>>the speed of light and on information transfer. >>>> >>>>You use the word "radical". Isn't this the key? You imply the paradigm shift is >>>>political, and personal - which it is. Paradigms arise because humans do >>>>science. Science itself is an idea-theory-test-creation builder. Without >>>>personalities, empires, vested interest, humans - science could develop 'freely' >>>>- where every idea comes on its merits, where the how and why gets questioned, >>>>where no idea gains status attached to a book or a figurehead, where all ideas >>>>are equal, where teaching and student material is not restricted to any one >>>>dominant; except this is out of the question - we're humans. Humans make >>>>paradigms. How hard they fight for the old or the new, how much they invested, >>>>their personality, their politics, their assumptions, the things they didn't >>>>think of, how long they spent on it - these quantise your "radical". >>>> >>>>If I had to pick one technical assumption of the old paradigm which is turned on >>>>its head by the new - I'ld pick the concept of quiescence. The old paradigm says >>>>search to quiescence, then evaluate the 'simple' quiet position, don't stop the >>>>search in the unclear. The new paradigm says drive into the unclear and evaluate >>>>with knowledge. Unclear is a good place for the new paradigm, it's a killer for >>>>the old. >>>> >>>>Radical? It was said to be impossible. AFAIK the argument lasted (lasts) five >>>>years now, since 1995. First is was impossible as per Botwinnik. Not just >>>>impossible, he was smeared as a fraud. Later it was said, ok it's possible, but >>>>it doesn't work on a win/lose results basis - this was the anti-CSTal argument. >>>>Now it is said, maybe it's possible, but it's nothing new - the anti-Tiger >>>>argument. Or it still isn't possible - you lose in the endgame - another >>>>anti-Tiger argument. It could have been evolutionary, if different assumptions >>>>held. But they didn't. So it became political and personal. And radical. The >>>>chief proponents of the old paradigm tried to 'own' computer chess. They'ld been >>>>there, done that, could have done that, did that and proved it didn't work, worn >>>>the tee-shirt, read the book, bought the record. They knew it all. And stuck to >>>>it. For years. >>>> >>>>Hence the shift is radical, revolutionary, even. Because of the participants. >>> >>> >>> >>> >>>What you are actually saying here is that Bob is the guy who has helped us to >>>notice that Gambit Tiger a new paradigm? >>> >>>By his compulsory need to belittle the value of speculative evaluations and >>>categorize Gambit Tiger into the "old paradigm" by any way (when everybody with >>>a brain out there is able to recognize that GT plays both differently AND >>>strong)? >> >>That is simply 100% wrong. 1. I didn't "belittle" speculative evaluations. > > > >You have been jumping systematically into every thread where a successful >speculative move of Gambit Tiger has been shown, and every time your message has >been "I have seen many cases where it did not work, speculation is dangerous". > > > > Saying "something can be dangerous" and "belittling something" are two totally different things. I haven't jumped into "every" thread, either. I am obviously interested in the subject of speculation, as I have also been working on this for a long while. But when someone says "new paradigm" then I certainly take issue with _that_ in this context. To put this in simple terms, I think your program plays great chess against computers. I don't know that it plays great chess against humans for reasons we can talk about if you want. I am not interested in putting your program down, nor in putting any other program down. However, I would like to see discussions kept inside some reality bound that is reasonable. CSTal "drove into the fog" and it died in horrible wrecks caused by doing so. I don't sense the same sort of play at all from Gambit Tiger. Just play that is very speculative and might (or might not) be optimal as other programs adjust to a slightly more aggressive opponent. It might be that your search is good enough that you can make speculative mistakes and still escape with your life. Or maybe not if your opponent is tactically as strong. Or if your opponent has superior hardware to help with tactics. Speculation that is too big is just as bad (or even worse) than no speculation at all. And from the games I watched carefully, my first impression is that your program might be a bit "over the top". Only my impression of course. And it doesn't mean I don't think a speculative program can compete. But speculation done right can compete even better, IMHO... > >> I >>have more than my fair share of speculative terms, from the two pawns on the 6th >>discussion that was held here a couple of years ago, to king safety and other >>things. I simply pointed out that in some cases _I_ had personally seen, your >>scores looked suspiciously high to me. Others chimed in and agreed in places. >>That isn't to say what you do is wrong, nor bad. And if you don't want any >>"pretty accurate' critiques from me, I don't mind "going away" and letting you >>do your own thing without any further comments from me. > > > > >If your "pretty accurate" critiques are limited to parroting "I have seen many >cases where it did not work, speculation is dangerous", then I can certainly do >without them, indeed. I have given more detail than that. I even gave a game at one point. I gave another example a day or two back. If you don't want any input from me, that's ok... I try to accept comments and suggestions from everybody. Some turn out to be quite good. > > > > > >> I pointed out some >>horrible king safety issues you had a year or two ago. You apparently fixed >>them. I just recently pointed out some severe endgame misevaluation in a game >>played vs my program. If you think that is unfair criticism, I suppose I would >>be quite happy to post a screw-up by my program everytime I post a screw-up by >>yours. I don't post mine very often because it is my task to fix them. > > > > >The post you are talking about was fair and accurate, I have no problem with >this one. > > > > > >>2. I don't believe I have _ever_ said that GT didn't play "strong". I did say >>that in several games I had seen, it thought it was winning big, and it >>eventually drew or lost enough of the games to make it suspect in those cases. >>I generally don't criticize programs just for the sake of criticizing them. > > > > >In the case of Gambit Tiger, let me have a doubt about it. > > That is what you should be doing of course. But outside impressions are also valuable. You often see what you want, rather than what is really there. I have done the same _many_ times when looking at my program play games. > > > >>When someone posts a criticism about Crafty, giving a position where it really >>seems clueless, that gives me something to fix. > > > > >Of course. When critisism is not as general and useless as "speculation is >dangerous, I have tried before". > > > > > >>3. If you want to say that your new evaluation ideas are "revolutionary" and >>belong in a "new paradigm" bucket, feel free. If that is the case, however, >>then there have been so many "new paradigms" over the past 20 years that they >>are beyond counting. From watching, your search hasn't changed. Unlike the >>search in CSTal. Only your scores have changed. Anybody can do that. Maybe >>you have done it better than anybody else. Maybe not. But "new paradigm"??? >> >>> >>>That's maybe indeed a way to detect new things. The establishement is indeed >>>always very afraid of new ideas, and especially when they are strong. >>> >> >>It also seems that those younger seem to write off the comments of those >>that are older as "they are over the hill and don't know what they are talking >>about." You might discover that this is wrong, one day. > > > >If by younger you mean that you are in computer chess since 30 years and I'm >into it since "only" 20 years, then I'm indeed "younger". I mean younger as in when you were born. 3/18/48 for me. > >On the other hand, I don't believe anybody expects the next revolution in >computer chess to come from you. You never know... I have my days... > >Unless the idea to try to beat a one-processor program by stacking 8 processors >or more on top of each other is what you call "revolutionnary". > > > > Christophe
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.