Author: Robert Hyatt
Date: 17:30:19 08/22/02
Go up one level in this thread
On August 22, 2002 at 03:42:13, Uri Blass wrote: >On August 21, 2002 at 23:06:47, Robert Hyatt wrote: > >>On August 21, 2002 at 18:26:43, Uri Blass wrote: >> >>>On August 21, 2002 at 17:57:51, Robert Hyatt wrote: >>> >>>>On August 21, 2002 at 14:49:00, Uri Blass wrote: >>>> >>>>>On August 21, 2002 at 13:33:55, Robert Hyatt wrote: >>>>> >>>>>>On August 21, 2002 at 07:49:10, Vincent Diepeveen wrote: >>>>>> >>>>>>>On August 20, 2002 at 20:27:18, martin fierz wrote: >>>>>>> >>>>>>>In these days all programs were so bad that games were decided by >>>>>>>who didn't give away most pieces *usually*. >>>>>>> >>>>>>>So in that respect all games from then are biased as the level of >>>>>>>*every* participant was 600 points lower than they are now. >>>>>> >>>>>>That's wrong. We had >2200 programs back then. Belle was > 2200 in 1983. >>>>>>In 1984 Cray Blitz was 2250 officially. Hitech was almost 2500 officially. >>>>>>Deep Thought was >2600 officially. All of those "official" ratings were earned >>>>>>by playing only humans, comp vs comp was never rated officially by any >>>>>>organization we dealt with... >>>>> >>>>>Fritz3(p90) had also IM norm against humans but if you >>>>>look at it's ssdf rating you can see more than 400 >>>>>elo difference relative to the top programs. >>>>> >>>>>If you remember that A1200 that is used by the ssdf is not >>>>>the best hardware you can say that Fritz3(p90) is almost 600 elo weaker than >>>>>the top programs of today in comp-comp games. >>>>> >>>>>Fritz3(p90) played in 1995. >>>>> >>>>>Most Programs that played against deep thought played >>>>>when p90 was not available. >>>>> >>>>>Uri >>>> >>>> >>>>The discussion was about deep thought/deep blue. It was not horribly better >>>>than _all_ its electronic opponents. Against the micros, yes it was pretty >>>>hopeless. But there were lots of opponents they played that were not on >>>>micros. Myself. Hitech. Lachex. Waycool. *socrates. You name it... >>> >>>What was the result of these players against the top micros? >>> >>>Uri >> >> >>At 40/2hr? Disasters. The micros played in things like the US Open, the >>World Open, and local events that had a GM here and there. No competition >>whatsoever at 40/2 back then from the micros... Which is what made Deep >>Thought's performance so startling to everyone. > >I know that the micro were weaker than deep thought but the question was about >comparison of the micro with other opponents except cray >blitz(Hitech,Lachex,Waycool,Socrates). > >I know for example that Socrates lost against Fritz3(p90) in 1995 in a game to >decide about the world champion. > >It's hardware was better than Fritz but I do not know if it was better and I >remember that it lost because of a tactical mistake. > >Socrates was probably better than other players in the list like Lachex and >Waycool. > >Uri I don't know much about waycool. It never seemed to be very strong. Lachex was another matter, however. It was very strong. It ran on the same sort of hardware we used and was reasonably close to us in strength. Major bits and pieces of the program came from us in fact, although the search and final evaluation were all done in assembler by Burton Wendroff and Tony Warnock... The problem with *Socrates was the same as the problem with Cray Blitz. Getting test time on these big machines is not easy. And without enough testing, disasters happen. Often... We had good results with little testing. And we had horrible results as well...
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.