Author: Uri Blass
Date: 13:20:01 04/12/02
Go up one level in this thread
On April 12, 2002 at 11:50:03, David Dory wrote: >On April 12, 2002 at 03:51:14, Uri Blass wrote: > >>How can you know it when the thing never played on equal hardware against other >>programs? >> >>I believe that the algorithms of it were not world class and it was better than >>computers of that time thanks to better hardware inspite of inferior software. > >Uri, you're a chuckle! > >Hsu and team had ALREADY MADE A WORLD CHAMP - DEEP THOUGHT. Then they re-did it >with TWO more revisions, DB and Deeper Blue. Increasing it's hardware, which as >has been mentioned, is the biggest factor in chess computer improvement in the >past 20 years. (Try running Fritz on a real clunker CPU/system for a hoot!) I agree that they were better than programs of that time thanks to better hardware. I did not say that they were not the best at that time. I only said that I believe that they were the best thanks to better hardware inspite of inferior software and I believe that the top programs of today are better. > >Then they increased it's software side with lots of GM input. I do not believe that the GM's input is very important. programmers that are good players do not prove to be better than programmers than programmers with ratings of 1600-1800. > >But because of a move here, a move there, (and HSU has stated that DB had a bug >causing it to not always select the best move it found), some guys decide: > >** That's not good! ** This is a very significant bug not to play the best move that you find. > >When you read what Hsu wrote, you know right off - the guys a total bright bulb, >and he was totally into computer chess. Oh, and his team had several other >Ph.D's with lots of experience, also working on DB. Lots of backing from IBM. > >If a bunch of Ph.D's and/or World chess champions programmers (who had nothing >to gain from saying it, monetarily), were saying "DB is not as good as Fritz on >a good PC", I'd consider that. Amir ban already said it and I understood that Ed agrees with that opinion. > >But when I hear people who can NOT program a top chess player, go on and on >about "this move" or "that move" of DB being bad, well that just doesn't cut it >with me. Every program plays questionable moves now and then. We've seen >thousands of them in CCC/rgcc over the years. They're useless for any kind of >meaningful comparison of strength, and you know that better than most, I'm quite >sure! I know that every program can play bad moves but the fact that Deeper blue not only played bad moves but played no impressive moves give the impression that Deeper blue is worse than Fritz. Time will tell if I cannot program a top chess player. Certainly my program is not close today to the top programs. > >I'm not concerned about DB never playing Fritz on equal hardware. If you could >somehow compare just the software, because of the on-going work on Fritz, it >might be stronger than DB. But that's my point - you can't. Deep Blue was >hardware just as much as software, and that's why Deeper Blue still would be >world champ today if it was playing. I know that hardware is also important but I think that the software difference between deeper blue and the top programs is big enough. I also doubt if we know the real details of the hardware of deeper blue because IBM lost credibility by their actions. IBM says 200M nodes per second is not enough to convince me. > >Some day, a program (and system) will be stronger than DB, but that's not >happened yet. > >And it won't happen on a 600MHz laptop, no matter how much Kramnik may want it >to make us believe it, for his own financial gain! <chuckle, chuckle> > >Dave > >>I think that every result that is not convincing win for kramnik is a surprise >>under these conditions. > >I think you're putting your own spin on this, Uri. If Fritz is better than DB on >a little laptop, it should kick Kramnik's butt to the moon on a hot 8 CPU >monster, with all those "better" algorithms. Note that I did not say that Fritz on 600Mhz is better than Deeper blue but only that Fritz on the hardware that is going to be used against kramnik is better than deeper blue. Uri
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.