Author: Mark Young
Date: 12:03:45 06/25/01
Go up one level in this thread
On June 25, 2001 at 14:39:30, Robert Hyatt wrote: >On June 24, 2001 at 13:16:56, Chris Carson wrote: > >>On June 24, 2001 at 13:08:35, Mark Young wrote: >> >>>On June 24, 2001 at 03:25:13, Uri Blass wrote: >>> >>>>Fritz is better than 2600 in open positions and 2200 in closed position. >>>> >>>>2900 in open positions and 2500 in closed position seems to be more realistic. >>> >>>Interesting, I don't know if you are right are wrong with these numbers. Could >>>you explain your reasoning behind them in more detail? >>> >> >>Uri and Mark, >> >>I find this interesting as well. I have not looked at all the games (232) that >>I have to check this. If either one or you wish to do this, let me know and I >>will send you the latest set. I am inclined to think that Uri is correct when >>you actually look at the games. >> >>Best Regards, >>Chris Carson > > >I don't think computers are 2500 in closed positions. If they were, they would >be unbeatable. Neither are they 2900 in "open positions". I don't think they are either, but I just don't know. That is why I asked Uri for details....Uri did not provide any...so I assume this was just a gut feeling. So I will doubt his numbers until I see data. It is much easier to find out overall strength, when you start trying to make subratings for certain kinds of positions you are asking for trouble without data to back it up. One flaw it this approch is you must do find the ratings for the human players also for open and closed positions, and you must define exactly what a open and closed positions are, as there is much grey area known as semiopen or semiclosed positions. Just find the Shirov >game where he played Bh3 which was a forced win. See how long it takes a >program to become convinced of that. Then think about what the human had to do >as he calculated this to a forced win from the start. > >These kinds of positions show that there are some tactical positions where the >computer is simply "out of it totally." In wild positions, with pieces hanging >everywhere, yes they are tough. In positions with a deep forcing variation, >they are like a duck out of water... very slow and clumsy, and easy pickings >for "the fox". > >We are doing pretty good at the "mid-range tactics" because we are not searching >like a human at all. But the "long-range tactics" are totally beyond the >programs of today. And will be for many many years.
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.