Author: Dann Corbit
Date: 15:21:45 09/03/99
Go up one level in this thread
On September 03, 1999 at 17:59:46, Heiko Mikala wrote: >On September 03, 1999 at 02:59:26, Dave Gomboc wrote: >>Crack open an intro to statistics book. >Dear Dave! >Thank you very much for this tip and your always friendly replies to my posts. > >If you should ever feel the desire to find out more about me, a good starting >point might be to ask me about my education and my profession. What is your education and profession? You have made me curious because my degree is in Numerical Analysis. >Hint: I may have "opened" more mathematical books than some other people have >ever dreamt of. I have at least a dozen books on special functions. How's that for perverse? >Respectfully, > >Heiko =:-( I think you want the mathematical answer, then since you have hinted that you are a mathematician. I am guessing that your degree is in pure math, so the answer is: An infinite number of games is needed to be absolutely certain that program A is better (or equal or worse) than program B. A lesser number of games will determine the same outcomes with less and less degrees of certainty. I think what Dave was hinting at is that it is a fairly standard statistical problem. Probably most statistical texts will have some sort of discussion on this sort of test. E.g. : Hypothesis: "The Douglass Fir's average height is greater than that of the Ponderosa Pine" How would you attempt to prove/disprove this hypothesis? The best description is probably ELO's book itself (as a wild guess, since I don't have a copy, but I do know that he talks about games and probability and derives a boatload of formulae). HTH P.S. Dave wasn't grumpy. He's a grad student. ;-)
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.