Author: Robert Hyatt
Date: 17:23:57 05/23/02
Go up one level in this thread
On May 23, 2002 at 18:44:44, Rolf Tueschen wrote: >On May 23, 2002 at 17:02:49, Robert Hyatt wrote: > >>On May 23, 2002 at 13:41:05, Rolf Tueschen wrote: >> >>>On May 22, 2002 at 11:54:37, Robert Hyatt wrote: >>> >>>>On May 21, 2002 at 21:16:15, Rolf Tueschen wrote: >>>> >>>>>Did you know that you are one of very very few who are thinking about such >>>>>factors? Thinking and talking about. >>>> >>>> >>>>I don't think so. _several_ have mentioned "motivation" repeatedly. It takes >>>>a lot of motivation to go out and do something that is very difficult. Such >>>>as trying to push a computer off the board. You only have to look at the >>>>last Deep Blue vs Kasparov match to see just how difficult Kasparov found the >>>>match to be. If the absolute most weight you can lift is 500 pounds, you will >>>>find that going out many times and lifting 499 pounds is a _very_ difficult >>>>task to mentally prepare for, unless the reward is significant enough to make >>>>it worthwhile. >>> >>>That is a beautiful comparison. Let's add the social aspects if the crowd sees >>>the 499 or 450 already as a boring thing or loss. (Or and that might have been >>>the Kasparov Trap: he himself might have expected the 501 without preparation.) >>> >>>But honestly the question: Why, if all that is absolutely clear for you and >>>others as you say, why do you stress the meaning of contracts as fair which were >>>only accepted because all these aspects could not be foreseen by Kasparov? Why >>>do you put the factual over the human? Why do you pretend to be unaware of all >>>the social and psychological aspects, although you know them so well, as you >>>demonstrated here? >> >>I don't understand the question. The original goal of all computer chess >>projects was simply "beat the world champion in a match at 40 moves in 2 hours >>time control". For the longest that was totally impossible. So no questions >>were asked, no special issues about "fairness" were mentioned. > >Let's give it fair chance that we could understand the reasons for our different >opinions. Fact is I wrote "Why" because I couldn't understand you and you say >you can't understand the question. I think I have found the solution for our >different views. BTW thanks to Bruce too, while I had a debate with him about >the Jakarta 1996 topic (because the independance day of East Timor on May 20, >2002) in CTF included some of Chris Whittington's evidence. Bruce convinced me. >You in the US see the factual at first, then you decide on that data basis. If >there's a problem you search for a practical solution (JUNIOR going there as >tourists not official representatives of Israel). You do not ask yourself in >self-agonizing manner if perhaps the whole decision right from the beginning >(for going to Jakarta) could have been wrong! You just make decisions at every >needed moment. Hence if one of the participants later, after the event, begins >to ask questions which are moot for you, you say why the guy didn't ask that >before when the contract was signed. > >I, I don't speak for complete Europe, but I have always learned to proceed >differently. And therefore I asked now questions over questions, and could not >understand why you were so stubborn in my eyes and still so happy with the >presentation of your position. I always thought he should better hide that when >he has that opinion! So let's finally find some understanding for one another. > >You were never saying that 1997 event proved that DB2 was better than Kasparov >or stuff like this, but that in that event under the known unfavorable >conditions DB2 succeeded in winning in a match. The story with the games etc. >might be tragic but so was the deal. This is what you are saying. If it was fair >did never interest you because fairness should always judged in the light of the >contracts. Of course you say there could have been much more favorable >conditions for Kasparov in special, but he simply didn't 'ask' for in advance. >When I asked you, that IMO the scientists should have looked for fairness in >their own interest (because of the validity of the results) you answered that >they were all under contract with IBM. (etc.) I hope I recalled most of the >points correctly. All I ever said was that for a two-week period in 1997, DB proved that it played better than Kasparov for six consecutive games. Nothing more. Nothing less. Could it repeat that? Who knows? > >I wished only that you could also understand me. For me it is almost nonsense to >want someone whom I could present as the strongest human player (questions begin >here with "Is K. really the strongest posible against comps?")and then enjoying >his delusions, simply because the results are screwed, when the player reacts >more on the ambience than the pmachine itself. I would not let him in the >delusion that he must not prepare carefully and that Fritz could be enough if he >let run it a litle bit longer than normal. Simply because I wanted good data. Or >in sports the same. What is it worth if I win with doping? I could never accept >the position "No matter how, the main point is that I'm winning." In special if >I am sort of scientist. What is it worth if I win and people applaude me for >days or a week and later they begin to doubt the fairness. >I would even go so far that the mere factual itself can not be accepted as the >possible basis of my behaviour. If I accept or better I want a partner who is >unaware of the details and problems (all that could be said about Kasparov who >boasted himself with all his knowledge about comps) so that the results again >could not be taken for serious. > >But nevertheless I understand your position and opinions. I also understand now >why you always took it as offense if I simply asked questions. Then you thought >that I wanted to delete the factual results. On the contrary I wanted to have a >more sophisticated debae about the conditions without automatical consequence >that Kasparov therefore might have won the match. No! >But I thought that we could discuss in all freedom what led to what and why. >Just as it is in science. > > >> >>Then, one day in 1997, the unthinkable happened. The GM lost the match >>everyone assumed he would win (again) easily. And _only_ then did this >>fairness stuff come up. Not before. Not during. But _after_. > >Everyone except IBM/DBteam. And a few others. The IBM guys _knew_ he would have his hands full. But then they knew what they had been doing since the last match and how much better the thing was the second time around. The rest of us assumed it was just faster. It was much more than just faster... we just didn't know it at the time. > >> >>So, let's ask a new question now: >> >>"When will a computer be able to beat the world champion after it has played >>hundreds of public games so that its weaknesses and strengths are less of a >>mystery?" >> >> >>Perfectly reasonable question. Just not the question asked in 1997. >>Nor was it the question answered in 1997. That was the _original_ question >>that was being asked and answered back then. >> >> >>> >>>What was it after all, the strength of DB or the confusion of Kasparov in the >>>social situation? Don't you see that artefacts might destroy the sense of >>>results? And was that the intention of the DB team? I still can't believe it. >> >>None of that had anything to do with anything. DB2 was completed immediately >>prior to the 1997 match. There was _no_ way Kasparov could have seen games >>played by it to study them. They didn't exist. That wasn't part of the >>question in 1997. Until after the match ended in the surprise fashion we all >>watched. > >Here I dare to ask the question from long ago. Why then - if no games were >existent - Joe Benjamin told the crowd during on of the games, answering a >question from the stage, that he would present 'them' after the match??? I have no idea what he was referring to. Hsu, in his book, _clearly_ explains exactly when the new chips arrived and how long it took to get things working properly with a few hardware glitches that were unexpected and needed some clever work-arounds to solve. Perhaps Joel was talking about "deep blue junior" which was the machine he did almost all of his playing against... > >Of course he had games. At the time you or someone answered, that Benjamin was >forbidden by IBM to present them. Something Benjamin didn't know during the >match. Although DB2 was dismantled the games were never published as promissed. > >Isn't that a situation that could cause you disturbances? > > No. written contracts are binding. I have seen lots of cool things over the years but been unable to comment on them due to signed non-disclosure agreements between me and the vendor. Contracts are a part of the world, even when we don't like some of the conditions they impose on us... > >> >>Prior to the match Kasparov was not complaining about the lack of games to >>study. He was busy predicting his victory based on practice games against >>Fritz and the games against the previous version of the hardware. It was only >>after he lost that the issue of "fairness" was raised. Notice that in 1996 he >>_won_ the match, under _identical_ rules as those used in 1997, yet not a >>single person mentioned the fairness issue. Wonder why that is??? I don't. >> > >I understand what you want to say. Let me ask you a real expert question: >°°° What do you think how much the whole preparation with Benjamin was worth? Which kind of preparation? Opening book? probably worthless. kasparov tried early on to bypass book preparation and did so effectively. Analysis of weaknesses? Probably invaluable as Benjamin is a strong player, with lots of computer savvy (he played Cray Blitz 20 years ago for example, so computers were not mysterious to him at all), and his analysis of weaknesses in DB could certainly aid the team in plugging the holes. That led to the re-design of the DB2 chips to incorporate more evaluation knowledge, for example.. And I think that Joel (and the other GMs involved including Byrne, Kaplan and others) was essential. > >°°° Would you say that no matter of the details with Kasparov DB2 really was >areally strong chessplayer. Or do you see a correct in my assumption that >special preparation by human GM in a group would create a "new" chess against >computers that could cause many difficulties for the machines? Or is this >against DB2 from your view and knowledge almost impossible to achieve? It depends. Did it have weaknesses that might be found and exploited given enough time? Probably. Was it incredibly strong at that point in time? definitely. > >°°° Why Kasparov was unable to repeat the play of the first game in '97? When >DB2 played very weak moves? I don't understand the question. None of the other games were attempts at repeating this game so I don't follow... > >(Excuse the whole topic with DB2 but this here could be appropiate in my >question about the strength of "engines" in a program. If we understand the >engine as the place where the strength of a machine is. Books and tables make it >even better, but IMO in a forbidden way. Please allow these thought to be >mentioned, because you yourself always state that Kasparov could have requested >better conditions. So I think FIDE could request no or strictly defined books >and no tables.) FIDE could make any rule they want. Including that the computer has to run only on AA batteries and no more than 4 of them. If they want. What the point would be however, is a good question with no good answer. Why restrict the computer, period? Or why not restrict the human as well? > > >> >> >> >> >> >>> >>>But the consequences are visible: FIDE banned computers in tournament play. >>>Kramnik wants FRITZ (sic!) 6 months in advance and some more rules. The FRITZ >>>even Eduard Nemeth with Elo 2100 can beat almost at will? Strange. >>> >> >>First, I don't believe Eduard can beat it "almost at will". I'll bet he can't >>beat it one game out of every 10, played on a real board with no take-back, no >>looking at scores and analysis. And I'd take that bet with him playing against >>Crafty on ICC if he (or anybody else) would like to try. It simply isn't _that_ >>easy to beat any computer in a _real_ game... > >I will tell him of your proposal. Would you allow all time controls or is there >a limit? And could he train a little bit before the match? I think that he's not >able to play Blitz against Crafty online. What is a practical time for you on >ICC? Crafty is on 24 hours a day, 7 days a week, so any time is fine. And any time control is fine, within reason. IE no game/24hour type stuff. :) > >Rolf Tueschen > > >> >>Kramnik made the rules as favorable to himself as possible, just as Kasparov >>_could_ have done. IBM wanted _him_. He held all the cards. Kramnik just >>managed his hand a bit better and made a better deal. Kasparov _could_ have >>had a better deal (he couldn't have had access to the real DB2 of course as >>6 months prior to the match it was still being designed.) >> >> >> >> >> >>>Rolf Tueschen
This page took 0.01 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.