Author: Robert Hyatt
Date: 16:59:14 06/01/00
Go up one level in this thread
On June 01, 2000 at 14:47:31, blass uri wrote: >On June 01, 2000 at 14:33:47, Robert Hyatt wrote: > >>On June 01, 2000 at 14:13:10, Joshua Lee wrote: >> >>>Here's a game where today's programs would've smacked up one of the best of >>>15years ago.... >>> >>>1.e4 e5 2. Nf3 Nc6 3.Bb5 a6 4. Ba4 Nf6 >>>5. 0-0 b5 6. Bb3 Bb7 7.Re1 Bc5 8. c3 d6 >>>9. d4 Bb6 10.a4 h5 11. axb5 axb5 >>>12. Rxa8 Qxa8 13. Na3 exd4 >>>14. cxd4 Ba6 15. e5 dxe5 16. dxe5 Ng4 >>>17. Bxf7+ Ke7 18. Kf1 b4+ 19. Nc4 Rd8 >>>20. Qc2 Kxf7 21. Qf5+ Nf6 22.Qc2 b3 >>>23. Qe2 Nd4 24. Nxd4 Rxd4 25. Kg1 Bxc4 >>>26. Qf3 Qxf3 27. gxf3 0-1 >>> >>>22. exf6 hello isn't this the obvious move? yes as i see no way to lose after >>>this .... 22...Bxc4+ 23. Kg1 >>> >>>Can i make the conclusion that atleast for this game programs of today are >>>better than Hitech and Bebe of 1985? >> >> >>Of course not. Did you see the game where Fritz got slapped silly in the Dutch >>tournament? Does that mean it can't keep up with anybody today? No, it just >>meant it got into one bad game. Hitech can play with most any program of today. >>It would probably lose more than it would win since it has been passed by in >>terms of speed. But it would not be a pushover by _any_ stretch of the >>imagination, and would prove to be very dangerous. >> >>BeBe searched about 40K moves per second, and would be overwhelmed by most >>any program of today. But HiTech could hit over 160K nodes per sec, with a >>very good evaluation that didn't slow down as they added terms due to the >>hardware design they used. Hitech would be a very bad program to underestimate. > >I am not impressed by the nps because I believe that the search rules and >evaluation of today's program is better than hitech. > >If my memory is right >Ernst Heinz claimed here that hitech was not better than Fritz3(p90) > >Uri Hitech had a very good evaluation. There is plenty of literature from CMU covering the things they did, which in a way was a lot like the DB approach of doing all the evaluation in parallel so that new features didn't slow them down. Remember that HiTech produced ratings in the 2450+ range and was a real competitor to win the Fredkin prize stage II. Their search was as good as any of today, as it was a full singular extension implementation, with lots of other good ideas (mostly from Murray Campbell who was on the team for a long while.) Unfortunately, their speed is simply not great enough given today's hardware to have realistic chances against top programs. But I would _not_ bet on the outcome of a single game vs that machine. It was very strong. More than one IM/GM lost at long time controls to it...
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.