Computer Chess Club Archives


Search

Terms

Messages

Subject: References for D = -400*log(1/y - 1) error margins

Author: Marc Boulé

Date: 13:11:44 05/21/02


References for "D=-400*log(1/y - 1) error margins"

A previous message (see below) showed such calculations but I would like to know
the source of these calculations so that I can reference them correctly. For
example, where does the std(y) formula and the 1.96 constant come from in what
was previously stated? If anyone has documents or links showing more explanation
on this, please let me know.

Marc

From a previous message:
-----------------------
std(y) = square_root( y*(1-y)/ n ) where n is the number of games.

A 95% confidence interval for y is:

[ y - 1.96*std(Y) , y + 1.96*std(Y)]

make h1 = y - 1.96*std(Y)   this is the y lower bound
make h2 = y + 1.96*std*(Y)  this is the y upper bound

Since x = -400*log((1-y)/y), the confidence interval for x would be:

-400*log( (1-h1)/h1) ) lower bound

-400*log( (1-h2)/h2) ) upper bound



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.