Author: Earl Fuller
Date: 06:06:15 05/13/03
Go up one level in this thread
On May 13, 2003 at 00:30:53, margolies,marc wrote: >the basic problem with USChessFederation rating chess computers (as a means on >accurate understanding of the products abilities) is that the USChess also sold >computers ; their ratings were essentially a marketing tool and the marketing >executives who have run the USCF have looked upon their membership rolls as a >targeted marketing list who should be induced to buy things ( in contrast to >say, a form of political constituency which could advocate our sport and further >chess education in the public sector.) >(As an example chess is one of the cheapest subjects to teach hence its >popularity in the communist world but ask most educators and school system >administrators who will say that they can't afford frills like chess in an age >of budget cuts.) >While I am sure that many of the relative comparisons of machines which USCF >elected to sell (eg. recieved monies from the manufacturer) were pretty fair, we >should not doubt that the general abilities of these devices were exagerated to >generate excitement about them. >If a 1900 rated player does well against these machines, it still does not >account for ratings inflation like grade inflation as these machines were >competing against a pool of less sophisticated players in past times before good >chess knowledge was widespread. Yes i agree with what you have said! But i think the knowledge differance that players have today (say a uscf 1900 player) is in the opening! I don't think the knowledge base for human players has changed very much in the "middle game, or end game". So if one were to take an older computer, say like the Par Excellence (2100) as an example and upgrade the opening book to represent todays play, (not changing anything else) i think their is a good chance the computer would achieve a rating chose to what it had. Close would be like, maybe -100 pts. Hmmmm, lololo I don't know if people consider 100 pts close or not, but maybe something like that, because once the computer reached the middle game (from a new opening book)i think the computer and the human player would be on there own and the results would be close to the same as before. But maybe i'm wrong and the new test would show that they are a couple hundred points weaker, which would highlight your thoughts on the USCF ratings. Interesting! Earl
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.