Author: Robert Hyatt
Date: 05:40:16 11/10/98
Go up one level in this thread
On November 10, 1998 at 04:08:17, Amir Ban wrote: >On November 09, 1998 at 23:14:16, Robert Hyatt wrote: > >>On November 09, 1998 at 19:50:27, jonathan Baxter wrote: >> >>>On November 09, 1998 at 17:11:48, Robert Hyatt wrote: >>> >>>>>I am not accusing anyone of lying, I was not aware that feng hu personally told >>>>>bob hyatt about the match results. Still the question has yet to be answered why >>>>>haven't the games been revealed? This is a very simple question, If such a match >>>>>between deepblue and the micro's did in fact occur it is a tremendously >>>>>important event and should be published! >>>>> >>>> >>>> >>>>Hsu told me about this privately. He and Murray both have discussed this at >>>>a couple of presentations they have given since the DB match. however I doubt >>>>they will say much here as it would only take a word by one of them to get the >>>>personal attacks going like crazy, by those that "hate big iron." >>>> >>> >>>I don't hate "big iron" and I am not interested in attacking these guys >>>personally. But I do find it terribly disappointing and very surprising that >>>DB---the most exciting thing ever in computer chess and possibly AI as >>>well---only played one, inconclusive match against the human world champion and >>>then quit forever. >>> >>>It would have cost IBM almost nothing to put DB on ICC and have it play top >>>GM's, and the results would have been incredibly interesting for both human and >>>computer chess. >>> >>>Given the extent to which DB built upon freely published research, and the >>>amount of publicity IBM gained from the last Kasparov match, it seems to me that >>>letting DB play in open competition is the least IBM could have done to repay >>>their debt to the research community. >>> >>>Jon >> >> >>You think way too small, as do most of us. When you look at the value of >>another Deep Blue vs someone match in a year or two, then you begin to >>understand why it is kept under wraps. Interest is high. A Karpov match >>would draw another crowd, as would a match against maybe Anand... >> >>This is all about money... and once it becomes about money, the guys doing >>the research lose control to the guys counting beans... > > >You are both wrong. Jon expresses a naive wish, but your explanation of why this >won't happen is in the wrong direction. Deep Blue will not play again, and has >already retired from computer events several years ago. I'm not sure how you, Shay, or anyone else can make this statement. DB will play again if IBM thinks that it will be worthwhile.. and probably not before. But, as I was taught in elementary school, "never is a long, long time. So long that it should not ever be used, in fact." > >It so happens I have a draft Shay sent me of an article he's writing for some >purpose, in which, in passing, he reviews the Deep-Blue affair. Here's the >quote: > > ... > Indeed, in May 1997, an IBM "monster" named Deep-Blue managed to >beat Gary Kasparov, the reigning human world champion, in a six-game match >3.5-2.5. IBM cleaned up, achieving an amazing financial gain: > 1. Its stock jumped up 20% > 2. It had planted the notion of supremacy in computing > 3. Its sales of super computers grew as it branded the name "Deep Blue" >in its product line. > By refusing to play a revenge match and dismantling the Deep Blue >project, IBM ensured that its result would be perceived as an ultimate proof >of supremacy and prolonged the public relations effect to the maximum. >Kasparov, shattered by his loss, has never played a computer since then. Shay's choice of words is unfortunate and inaccurate. "Dismantle" means to permanently disassemble. This has *not* been done to DB. The chess processors were removed from the SP and put on a shelf. It could be ready to play in a few hours if they wanted. The software is still there. The databases and so forth are still there. The chess processors are still functional and ready to go. SP's are still shipping so they could grab one and be ready to rock in a short amount of time. Whether it *will* play again is unknown and subject to his (or anyone else's) speculation. But it *can* play again easily. DB Junior is alive and well last time I heard... > A few facts regarding computers and the game of chess can cast some >doubt about Deep-Blue's true strength. Few people know, for instance, >that prior to its game against Kasparov, Deep-Blue with all its calculative >might, was not even the official world computer champion. > It lost this title, lo and behold, to a PC program named 'Fritz', from >Germany, that managed to defeat it at the official world computer >championships held in Hong-Kong 1995. This is one of those true statements that is dead wrong. "Deep Blue" did *not* lose its title to Fritz. It *never* held any title. The old deep thought program, running the "deep blue prototype" software lost to Fritz in Hong Kong... but that machine was based on hardware chips two versions behind the DB-2 chips that played against Kasparov. To even mention such a statement shows both a lack of regard for "the truth, the whole truth, and nothing but the truth" as well as a lack of understanding about the Deep Blue project and the various chip re-designs that were undertaken. DT/95 was doing 2-3M nodes per second. DB-2 was doing 100 times that with an evaluation that was over 10X more complex and sophisticated. So the statement is the truth, only when stretched to the max, since DT was the world champion in 1989, any lost to Fritz using the same hardware. But this has *nothing* to do with the DB that beat Kasparov any more than Cray Blitz had anything to do with the chess program that played in human tournaments in 1972... > Since then, the IBM team carefully avoided any more "embarrassments" >and did not confront any other computer programs officially. In fact, it >reduced its opponent field to just one person - Gary Kasparov. I doubt they worried about "embarassments" if this word is a substitute for "losing to another program." It only happened twice in 15 years or so. I'd say they made their point during that time-frame... > ... > >Amir
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.