Author: Don Dailey
Date: 14:07:25 12/28/97
Go up one level in this thread
H Fernando, I read your message and understand your point. From a human competitors point of view I agree completely. I am not against any technique that may strengthen the computers game against me unless it involves lasers or radiation of some kind. I think what everyone is groaning about is the fact that we think of chess strength as being absolute (but it's not quite) and expect that the program we buy will perform "about" the elo strength advertised or seen on the SSDF rating list. This is not an unreasonable expectation is it? But what if you bought a program and it said this on the box: This chess program, although earning an SSDF rating of 2700, only plays at the 2400 level unless it's playing against Rebel or Mchess. I'm exaggerating the numbers of course but this is what a lot of people on this group are having a problem with. By the way, I'm not blasting Mchess or anyone in particular. I have no idea who may be involved or what they are specifically doing. Mchess is a great program to own and quite strong. -- Don On December 28, 1997 at 14:54:36, Fernando Villegas wrote: >Hi Chris, Thorsten, Robert and the rest of people discussing this: >There is, maybe, a different and complementary side to judge all this >matter about book cooking, auto learn ad anything that means an enormous >amount of opening knowledge and very long lines of instant moves before >the engine begin to compute. Before, let me tell you that nothing of the >stuff I am going to say means that other visions are wrong. I don't want >to polemize with nobody, not to say something is unmoral or not, fair or >not. I suppose the commercial problem that could arise from deceiving >ratings can be coped with more information for the customers, and even I >presume the market of this kind of chess software is already well >acquainted with the fact of the relative value of Elo list made out by >SSDF people. What I want to put here is how this is or could bee seen >from the side of us, human beings, and specially professional customer, >that is to say, people that use chess programs not only to get fun, but >to learn and improve his game. And as much as M7 is precisely a kind of >software that is very fitted to that people, I think that the following >remarks are appropriate. >a) You don t play a computer, you play against moves made by a computer. >This is not a triffling, silly distinction: if you focus attention in >the board and not in the computer at the other side of it, then is not >important if the moves were stocked or have been calculated just now: >the point is to get moves to which you must respond with the better move >you can get from you, calculated or also remembered from a previous >theoretical knowledge. >b) If the lines the computer play are sound and you are defeated because >of them, it does not matter if the computer just remembered them or >calculated them; in any case it has been a tool for you to see your >mistakes and learn a new, better way to play that line. >c) If the line played by the computer is wrong and you defeat it because >of that, the computer will not repeat it again and so you and your >teacher have learned something for a better and more challenging game, >later. >d) If you get out of the theoretical or cooked line very soon because >you are creative or just because you played badly or/and you did not >remember the best line, then you will begin to play at once against the >engine and so the problem , if that s the problem, disappear at once. >e) If in any case you just don t like to play against a computer that >does not appear computing very soon, you are always with the right and >chance to challenge the program with something new and this is, also, a >great way to improve. >So, ¿where is the problem FROM the point of view of the customer and >user? >Non... >For us the program is a tool because he calculates very well or because >he has a perfect memory, not because he seems to be a partner on the >club, thinking and perspiring. In any case, the computer thinking or >just remembering, you play against good or bad moves to refute, in any >case you learn, in any case you get a challenge. What a hell matter to >you if the computer proceedings are thorough calculation or just >reproduction? For the same reason I think nobody will be upset because >of the perfect ending database, where all is pre calculated. In fact, >for me, a lazy player with very little knowledge of openings and >endings, to be defeated in that stages of the game for a computer that >does not think at all because he "knows", is a lot better than to play >against a 1980 vintage computer that does not know how to punish my >mistakes in the technical stages of the game and so give me a break. >Even more, if I can win my shortcomings will not appear, giving to me a >deceiving high impression of my play. In fact, many of us were deceived >in that way with the first and second generation of chess programs. It >has been only with programs gifted with very big books -and now endings >tables- that I have realized the great gaps of my game. And of course >how the computer got his knowledge, if thorough programming, autolearn >or cooked lines against other programs, I don t care a shit.
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.