Author: José Carlos
Date: 08:20:11 11/16/99
Go up one level in this thread
On November 15, 1999 at 15:28:21, Sven Reichard wrote: >On November 15, 1999 at 03:43:15, Shep wrote: > >>As a mathematician, I have to jump in here and correct you. :) >>You probably mean "statistically improbable". >>It is not _impossible_ to find the optimal CM settings (or to win the lottery a >>million successive times) because the number of settings is finite, so you could >>(given a large, but finite amount of time) try them all out and come up with the >>best one. >> >>Sorry to be nit-picking, but I felt that had to be said. :)) >> >>--- >>Shep > >As another mathematician, I have to jump in here and correct you :)) >If there were an algorithm that would give us a numerical playing strength of >any given engine, your approach would work (in a finite number of steps). >However, lacking such an algorithm, your optimization problem (on a discrete >space, to make things worse) is not well-defined. > >Sven. I'm not a mathematician :( but I think we could define an algorithm to evaluate who's best from two players, for example, best score of 1 million games match :) Then, we'd only have to make those matches and decide wich is the optimal settings by counting victories, or something like that. I would not be a "real" measure of optimal settings, but a well defined one, I think... José C. PS.: Sorry for not being a mathematician :)
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.