Author: Hans Gerber
Date: 04:56:09 06/28/00
Go up one level in this thread
On June 27, 2000 at 16:47:53, Robert Hyatt wrote: >On June 27, 2000 at 13:53:37, Hans Gerber wrote: > >>It was shown by you yourself that the outlogs printed one hour or one week after >>the initial event could very easily be changed and nobody would discover it. >>Now, I do not say that because this is possible that IBM did it. I am more so >>convinced that the DB team, most of them scientists would never cheat this way. >>This is the one side. Now let's talk about the other side of the medal. If it >>can not be proven beyond any reasonable doubt that DB and IBM "could" not cheat >>the logfiles even if they wanted, the logfiles as they appear on the IBM site is >>not proving that no cheating had happened. >> >>Could we find aggreement on this? > >That was the point. There is _no_ way to prove they "didn't cheat". But they >have provided everything they have. To some, it is enough. To others, it >is not. To some nothing would be enough. Could you say something about the history of such tries, failures and the actual knowledge that no proof is possible? As you know in science it is not the most common case that you can't show why that what you are searching for is exactly what you have found. >>Is this your opinion or is it hard evidence? How about your own declaration that >>unallowed influence from the outside could well happen via all kind of wave >>transmissions. At least I had understood you this way. > > >Then you misunderstood. There are _many_ ways to influence the game, from >lasers, to magnetic radiation, to high-frequency audio, to you name it. But >these would _require_ some hardware in the DB box, and hooks into the DB >program. That would be nearly impossible to do without 'inside help'. DB is >a very complex program. The SP is a very complex piece of hardware. wiring >it up for cheating would be quite a trick. Doing so in total secrecy without >any of the DB team or IBM knowing would be a miracle or better... You would say that if something irregular had happened the DB team must have realised it? Absolutely in accordance with your verdict that they had total control? >>Look, now you say that "the DB team had t-o-t-a-l c-o-n-t-r-o-l of the >>hardware/software". Then we have a new problem. Because if this would be true, >>then why the DB team can not prove beyond all reasonable doubt that at least not >>from the outside no cheating could exist. Next problem, why the DB team can not >>prove the authenticity of the logfiles? Do you not remember that you declared a >>couple of weeks ago exactly that they simply could _not_ prove this! Out of >>principle reasons. > >How can you _prove_ the authenticity of something that can be modified in >a few seconds? IE I once saw a child's head on an alligator's body, on the >cover of the National Inquirer. I _know_ that photograph was legitimate, >because I saw it myself, right? > >Same problem with logs. what you want is impossible to provide. > It is not that I wanted it, it is a question of how you could exclude doubts about the output of the machine. A basic question for computerchess and science. Could you explain what you meant? What is the important aspect? Why should it be impossible to make valide and reliable data recording? >>Please make a clear decision about what is the case. >> >>Hopefully you are aware of the consequences if they had "total control". Because >>we have the old and still unanswered question why they (the scientists!) did not >>avoid the development of a psycho war against their own testing person? Some >>weeks ago you said that the logs would not have helped Kasparov for his >>questioning the authenticity of the moves. So we have the next problem, why did >>the DB team not provide a convincing method to prove that? As I said, this is >>not about the question of trust, this is about the standards of science. Since >>the times of the historical Turk the question of authenticity is well known in >>chess and computerchess. Could we find agreement that the team should have >>guaranteed a simple method to prove the authenticity? >> > > >pure rolfish rambling now. There was no psycho-war. Only a psycho-collapse >after 5 tough games. How do you interprete the relation to their testing person when they at first agreed and then suddenly denied to give him the logfiles? What is your term if it was not a psycho war? Remember please that it all happened after game two, not just after game five or six. >>In general this is possible in science. Simply with the exact documentation that >>allowed others to receive the same results for a comparable setting. Why the DB >>team and Hsu in special did never care of? They had plenty of time, whole years, >>to think about. >> > > > >again, it is _impossible_ to prove they didn't cheat. Can you prove you >didn't steal a dollar at the last restaurant you ate at? How would you >prove it? I might prove that you did, with a serial number record, or a >video of your doing so, or whatever. But I certainly can't prove that you >_didn't_. > Why should it be possible to observe a cheating by video, but it should not be possible to observe if no cheating had happened? If the video would be recording the relevant process? I do not see the principal difficulty. >>If Kasparov has read what you have written about possibillities of cheating he >>will not try this. Didn't you explain that such cheating could not be proven nor >>prevented? > > >yes. > > >> >>However it is well known in science that scientists should take care of that >>their setting is controllable. If it is true that they had _total control_, then >>they should be able to prove what Kasparov is demanding -- for a couple of years >>by now. > > > >If the setting had to be controllable, then the match simply could not have >been played. That angle of argument is pointless... Please give more details about the astonishing fact that you have _total control_ over your setting, but still you can't make the setting "controllable". Perhaps I miss a lot because I am not familiar with the English? Hans Gerber
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.