Author: Miguel A. Ballicora
Date: 10:09:36 01/15/01
Go up one level in this thread
On January 15, 2001 at 02:35:13, Jouni Uski wrote: >For FIDE GM title You need at least 2500 rating. Here are 3 latest >human-computer match TPRs: > >Rebel 11 ~2600 Wiel 2001 >Deep Junior 2703 Dortmund 2000 >Deep Fritz 2585 Dutch Championship 2000 > Being 2500 is only *one* of the requirements. Besides, it is a stable 2500 after being 2200, 2300, etc. That means a lot of games above 2500 strength. That is one thing. The other, to get the GM, traditionally you have to have peaks of 2600, which are the "norms". Unless FIDE changed things, those were 3 norms that should account for at least 24 games, in tournaments, in a period of less than few years (3? 5?). From the results you show above, Deep junior has a norm. There are many IMs in the world that stayed in that situation for a long time and never got the GM title. I personally believe that few programs would get a title if they are release to play like any player, but it has to be demonstrated... For now, it is not that clear. Anyway, like other poster said, regarding analysis they are still behind the GMs, because the understanding of the game is limited. I don't think that we can say "GM strength". Maybe we can talk about that computers might have a "GM performance" in practical play. That is, they can get good results but the strength is certainly not the same. At least qualitatively. Miguel
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.