[enriquto]

> Statisticians know that outliers need to be excluded from their results,

offtopic, but this is not true in general. This is one of my strongest pet peeves. For many distributions (e.g. Cauchy, power laws, Levy stable...) it turns out that all samples are "outliers" and you cannot really exclude them. It is only on the "degenerate" case of finite variance that you can exclude outliers. But some statisticians may still argue that this is still wrong: there are no outliers, only incomplete models.