Amarin Corp Plc (AMRN)

[jfmcrr]

If it is perfect Gaussian distribution, a perfect bell, symmetric around a center line, mean is same as median

But often time it is tilted, not perfect

The issue is that in large part, along with the diverse population, the differential etiology of the risks, and the tendency toward "fat tails" due to real world constraints, you end up with something you can work with in practical terms of applying statistics, (think "T" vs "Z" tests), but overlaid with a Gaussian or binomial distribution, you find imperfections.