Author: Osorio Meirelles
Date: 12:23:05 08/09/02
Go up one level in this thread
There are several types of math models we can build that would have a very good fit with rating as a variable to be predicted and ply as a predictor variable. The main chalenge is finding out of what a perfect rating would be. Anyway, I found a model with an R-square above 99% with the data you provided and the model is: Rating = 3735 - 3641/sqrt(ply) where sqrt(ply) is the quare root of ply According to this formula, with ply = 26 we would have a rating of 3022 ( which seems reasonable , not too high and not too low ) Unfortunately, in this situation we have to extrapolate and the error due to extrapolation could be greater than we assume initially.
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.