Author: Matthew Hull
Date: 20:28:18 07/21/03
Go up one level in this thread
On July 21, 2003 at 21:30:10, Sherry Washington wrote: >On July 21, 2003 at 16:26:38, Matthew Hull wrote: > >>On July 21, 2003 at 16:02:23, Sherry Washington wrote: >> >>>I'm amazed that everyone here seems to take the word of Edward nemeeth and kurt >>>U for gospel, I mean does anyone know these people personaly? Why do they >>>deserve so much credibility? If I post a game and say that I won, without any >>>eyewitness to say that i didn't take any moves back will the game be taken for >>>fact? I am not saying these gentlemen are lying, but it does seem strange that >>>these two can accomplish better results then the hundreds of public verified >>>games we have with grandmasters and masters. I Have several master rated friends >>>and they get slaughtered everygame, one is rated 2319 and hasn't even drew out >>>of seven games at 60/ 0 and that was when i had only a 1 gig processor. All i >>>can say is that I am very sceptical of these gentelman and think their games >>>should be taken with a grain of salt. Kasparov himself recently said that these >>>people who are claiming wins against the top programs are full of crap. >> >> >>I don't think so. The procedure for obtaining a win or draw against a very >>strong program is one of trial and error. If you look at E. Nemeth's match >>requirements, he states that he needs a copy of the program a few months in >>advance. This permits experimentation in probing the program's weaknesses (and >>they have weaknesses!). Then, he wants a freshly installed copy of the program >>for the match which is identical to the copy he was given earlier. This ensures >>that no "learning" files from previous match-play or experiments get included. >>Learning files are position histories that some programs keep for games they >>have lost so that they don't repeat the same mistakes. >> >>This means that E. Nemeth can essentially re-play already won games from his >>trial and error experiments. >> >>So he is not really full of something. The programs have weaknesses that can be >>demonstrated under certain conditions. He is simply taking advantage of this >>procedure to "show-up" stronger players than himself and to tweak the egos of >>weak chessplayers who can only participate in high level chess vicariously >>through the ownership of a GM-strength program. >> >>Regards, >>Matt > > >The process you describe has a flavor of dishonesty in my opinion. I see no dishonesty. His match conditions are pretty transparent, and so no one has been willing to take up his offer because of that. However, I think vendors have nothing to fear. It's pretty hard to win a 6 game match against one of these things, especially if you aren't a GM. You have to really believe you can steer better than three games into one of these positions and pray that time variability doesn't cause the program to search that one node he did not search in the experiments and bust you. I think he'd get waxed. Matt
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.