Author: Uri Blass
Date: 22:14:21 05/11/01
Go up one level in this thread
On May 11, 2001 at 19:36:46, Bertil Eklund wrote: >On May 11, 2001 at 13:16:43, Larry Proffer wrote: > >>On May 11, 2001 at 12:36:50, Bruce Moreland wrote: >> >>>On May 11, 2001 at 10:47:49, Robert Hyatt wrote: >>> >>>>It really doesn't show anything. Ferret at that time was running on a >>>>single cpu pentium pro 200 I believe. At least that is what I was using >>>>at that time and Bruce and I were using identical hardware. >>>> >>>>It was a reasonable anti-computer game. Whether he could do that with today's >>>>ferret and today's hardware is a totally different question. >>> >>>That game was played on a Pentium 133 that had two copies of the program running >>>simultaneously. I was at work, and for such an important game I didn't want to >>>have someone burst into my office and demand I do something intensive on my >>>computer, so I asked my wife to start it for me at home, and she got two copies >>>going at once by mistake. I did not announce this at the time, because I didn't >>>want to be accused of whining about the losses. He won against the best I could >>>do that day, and I accepted that. >>> >>>Those two games (Shirov won both) have become signature games for my program for >>>some reason. I still hear about them despite their being played approximately >>>five years ago. >>> >>>But that *was* old hardware, and the version I was running was approximately >>>*800* distinct versions ago. >>> >>>I was honored that my program was able to play some 30 0 versus Shirov, and he >>>did a great job bashing in the program. Shirov is *in my own limited >>>experience* the strongest anti-computer GM, and he could have probably totalled >>>the program on any hardware. >>> >>>I will say that I spoke with Shirov after he played some 5 0 against Ferret on >>>my 533 mhz Alpha in approximately 1998. He told me that he thought it seemed as >>>if it was seeing as much in 5 0 as it did in those 30 0 games. It's very >>>interesting that 500/133 or 500/66, whichever you wish to use, is approximately >>>6, so he is pretty close to right. Of course now it would be seeing the same >>>stuff in a 2 0 or a 3 0. >>> >>>I'm surprised that Keene couldn't find something more modern. If he's looking >>>for bad games against humans, there was the game against Vaganian at the 1999 >>>WCCC post-tournament exhibition, where my program moved its QB 4 times in a row, >>>in order to get it from c1 to d2. If he wants to see a bad game against a >>>computer, he can pick the one against Hiarcs from the same event. If he wants >>>to see a good game he can examine the one versus Fritz at that event. >>> >> >>It is deeply unlikely that Keene spent his time running around trying to dig up >>games played by Ferret. >> >>The game data will have been presented to him by somebody else. >> >>The game was a 'typical computer game' - meaning crappy old material grabber. As >>you point out they could have found good games to illustrate Ferret - but it >>appears they illustrated Ferret with a bad game. They could have illustrated >>Ferret with ICC grade information - with game results, instead of game scores. >> >>But game results look good. This game score makes Ferret look bad (especially if >>you conveniently forget 1996 and P133's). >> >>We know Enrique and Bertil don't usually deal in game scores. They deal in game >>results. Enrique and Bertil are not specially known for their talents in >>persuading grandmasters of the merits of programs based apon their chess >>knowledge. > >You know very well that the comments from the grandmasters and IMs are just a >joke, just look at the comments of the Odyssey tournament ( no harm intended, it >is interesting anyway) they look at one game and understand nothing but people >that follows a 40 game match has a much better feeling of the abilities of a >program. I think Thorstens comments in general are much better then those >experts that looks on one or to games from a particular program. I agree with >Thorstens comments in general about the playing-style of the programs except for >Nimzo8, that has fooled him a lot (and he know it now I think) Or just take the >revue of Tiger12 on the Rebel-site from Elvis, the shepherd of the >Chessbits-Schroeder-Forum (a strong chess-player and the excellent book-maker of >Shredder4), can anyone take this well-written revue seriously when he describes >the Tiger12 program as a very active and aggressive program when everyone that >has played some games with the program,compare it with the style of Genius, >Andersson or Petrosjan. I think even mr Theron agrees that Tiger12 played a bit >boring chess. Yes today Gambit2 rocks, I fully agree!! The Tiger(s) are >fantastic, the positional play are still weak sometimes, so it is good to see >that mr Theron still can improve his program! > >In example there is some terrible weaknesses in the Rebel-programs, I guess you >know them very well as a "fair" chess-player, and if Ed worked a bit with them, >his program could compete with the best programs of today but instead he is >blaming me or anyone that critizise his program for being antagonistic or being >bought or whatever. 1)I guess that Ed knows about weaknesses of Rebel. I already exposed one weakness of Rebel based on the ssdf games in order to prove that the ssdf games can be productive to learn about Rebel's weaknesses inspite of the problems with the autoplayer. 2)I do not think that it is easy to fix the weaknesses I guess that if Ed did not do it then it is not easy and you may fix one weakness and create another weakness. 3)I did not read that Ed blamed you for Rebel's weaknesses. I remember that he admitted that Tiger is better than Rebel in comp-comp games. 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.