Author: José Carlos
Date: 01:02:06 01/12/01
Go up one level in this thread
Not responding particularly to the post this one is below, but mostly to the
whole thread, I have to comments:
- Definition of "GM strength". I think you're trying a mathematical approach
without having a precise definition of the terms (correct me if I missed any of
the post where this was stated). So, to _start_ arguing, we need a definition of
what you consider GM strength.
- Versions of programs. This is much more difficult, IMO. Suppose I test a
version of Averno (say 0.31) and I notice it lacks of endgame knowledge (among
zillions of other things :)). So I write some code to handle this and I release
0.32, which rating goes up 50 ELO due to better endgame playing.
Now suppose GM A notices he has a poor understanding in a certain kind of
position. He studies Capablanca's games (once again) and get a better
understanding on it, so his ELO increases in the next list.
Now suppose the GM lack of knowledge was not about a certain position, but
in playing computers.
Should we consider both versions of Averno as _the same player_?
Should we consider both "versions" of GM A as _the same player_?
How can we calculate (in a mathematical approach) the strength of the meant
players?
José C.
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.