2 ms·

I think this is an overzealous statement of the central limit theorem. A log-normal distribution would probably be a better model in this case.