[ QUOTE ]
Coefficient of variation = standard deviation / mean.
In the 'real world' it's used mostly for variables which are always positive (dividing by zero is a bad thing), and mostly for things where the uncertainty is proportional to the value of the variable.
In poker, we usually standardize both the mean and the standard deviation by dividing by the big bet, instead of using the c.v. much. It does have a couple of theoretical uses: size of bankroll needed is proportional to CV, and expected time until showing a profit to CV^2.
[/ QUOTE ]
Bankroll needed is proportional to variance/mean, not SD/mean.
bankroll formulas and
derivation of bankroll formulas.