Well, it is pure lua (<100 loc) so probably it is not
'high-performance' enough. However it works on running datas and it
keeps the sample count, min, max , mean, standard deviation, (a
definition of) Skewness and Kurtosis. I do not implemented a generic
formula so for the other distribution momentum you have to make some
change (there is a reference to a general formula in the source, if
you want to give it a try).
It also let you to keep two different indices of the same
distribution, and then merge them in one without losing information.
 I just noticed that the documentation do not cites count/min/.max.
I will fix it soon.