Author: Uri Blass
Date: 11:42:41 05/23/02
Go up one level in this thread
On May 23, 2002 at 13:14:41, Rolf Tueschen wrote: >On May 23, 2002 at 03:02:12, José Carlos wrote: > >>On May 22, 2002 at 20:01:15, Rolf Tueschen wrote: >> >>>Do you want to participate against humans in human tournaments? If no, where is >>>the problem that you seem to think that the concept of playing against the FIDE >>>rules should not be discussed here in freedom but more be discussed with the >>>FIDE itself? I still think that it could be worth to search for a consense what >>>could be changed in programs to be more in accordance with the FIDE rules. Thank >>>you for your interesting comments so far. >>> >>>Rolf Tueschen >> >> I've participated in many tournaments, most of them with only humans, and only >>a few with computers allowed. >> I don't think the FIDE rules shouldn't be discussed here. > >May I ask you if you could read what I wrote about "playing against the FIDE >rules"? Please. > >That should be discussed. Only you had invited me to go and discuss the FIDE >rules with FIDE. Why should I do that? The players do not want to participate >against computers who play against the FIDE rules so that the FIDE excluded >computers. My proposition was to find a consense about a possible change in the >practice of computerchess. A computer should not be allowed (in FIDE tournaments >at least) to use opening books or endgame tables who enable him to >play way above his own strength. Now some here opposed such ideas with the >defense that all kind of code for evaluation and all must be forbidden in a >computer which is ridiculous of course. But some answer, no, what the computer >can achieve with calculating and the general chess content is his own result >comparable to the human players' but it's different when a machine simply >"reads" the pre-recorded content of the human GM book and plays it without a >single own calculation. > >Now, the overall mistake or fallacy or blind spot in all the negative answers to >my proposition is the popular logic in CC is the - ahem - reduction of the >thought process in little circles or single levels without considering the >context. For me from the outside it's strange to meet this thought process. >It was presented so many times from many different people from different >countries that it could help if we discussed it. > >I am unable to see the rational in the following: if someone says, it's against >the FIDE rules, which say "no use of pre-recorded stuff", meaning opening books, >and you in CC do that exactly, you let the machine read in books from human >history. You in CC reply this way: > >° books or anything in the code, if you forbid the books you must forbid all > >Who has demanded that at all? Again, I am unable to understand such a logic. Are >you worried about possible injustice among the different software codes in the >machine? (Or is it more some sort of hand waving because you know exactly what >that means in relation to questions about strength?) > >° human chessplayers do it all in the same manner - they play their pre-recorded >stuff, and yes, they do read it in their memory in their head, and that is just >the same, comparably, what the machine does, so no problem, period. > >But this is totally false. It's not even necessary to use the "understanding" >because it caused new confusion. It's simply a misunderstanding. GM players do >_not_ store anything without thinking and analysing, it's combined and connected >with the patterns so that it becomes a real weapon for the chess fights. But the >computer does something different. He gets to a move, which is not his, which he >wouldn't find through his own calculations included the evaluation stuff and >all, and most important, which is a move the machine would with certainty >replace by its own move - who would then lead to a catastrophe in the whole >game, probably a loss. This and only this is the cheating aspect so embarrassing >for human chessplayers. It is simply not fair. It's a kind of fraud. You should >also please bear in mind the financial aspects and consequences for the masters. > >Finally: the aspect that GM all the time respect the computers "as they are >actually". Because of the eidetics a real GM does not fear the computer. And for >his accepting the unfair time-relevant and physical fitness-relevant aspects he >gets paid. So that is how some masters see it. But if we enter FIDE tournaments >the whole question is different. And the participation of such a memory genius >could cause different influences on the human beings. > >I am absolutely sure, that if the computer would be really almost strong as >Kasparov (without the books and tables tricks) human masters would be very >interested in letting such a genius participate. But this (the fair genius) will >not happen in the next 50 years. IMO. I believe that you are wrong and the computer is already almost as strong as kasparov(there are positions when it is stupid but if humans do not special preparation before the game they have small chance to get the relevant positions and I expect computers to get GM results if you give GM's to play shuffle chess against them). I do not think that masters are not interested to play because of the opening book. This is only an excuse. The masters also memorize opening and learn them. The fact that they understand the reason changes nothing because it is possible to explain the computer the reasons and to teach it to find a lot of opening moves by itself based on the right extensions. You also have no way to know if the computer does not use a hidden book so a law that forbid books is a stupid rule. For the same reason computers are allowed in correspondence games in Israel and in a lot of countries. A more logical rule is to limit the hardware of the computer. In 50 years computers are going to be so fast that they will not need books to beat humans in chess. Even today I believe that Fritz7 with no book except first move has good chances to lead the ssdf list in the following conditions: 1)hardware that is 3 times faster than A1200. 2)Simple learning to repeat first moves that win and avoid first moves that lose. 3)forgetting all the learning after every match(ssdf do often matches of 40 games). It means that practically it is not going to fall into a trap twice in a match because in order to fall in a trap twice it needs to lose 20 games when it tries every possible legal move. I believe that Fritz7 with the fast hardware has good chances not to lose even with openings like 1.f3 or 1.e4 h6 or even 1.e4 f5 and Fritz is not going to choose that opening without losing many games with more logical moves in previous 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.