Author: GuyHaworth
Date: 15:11:17 11/11/03
Go up one level in this thread
Russell Reagan seems to be doing some excellent research, flipping the roles, and seeing how often a chess-engine, e.g. CHESSENG, reproduces the moves of the GMs. This is most commendable. I don't know how this can be automated, but I'd certainly like to see this kind of comparison done across a range of chess engines and over more games. Let us suppose that in position P1, the probability of a GM and CHESSENG playing the same move independently is q1. Of course, the difficult is to assess what q1 is. Russell's work might evolve to indicate what this is. Anyway ... P1 is following by P2, P3 etc, and probability q1 is succeeded by probability q2, q3 etc. Thus, the probability of a GM and CHESSENG playing the same moves independently is p1*p2*...*pn. For a particular string of moves, this probability might be quite small. For another string, it might be quite near 1. For some string, it is a lot larger as there are many more strings on offer. Enough said - "a string of 8 moves" sounds impressive, but is it really? Note that enough events happen in our lives to make 'strange coincidences' occur and appear strange at the same time. There are afaik no canals on Mars, though many believed in them once [and I may be about to be proved wrong]. g
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.