> I like the idea, but how do you apply this to power law distribution outcomes and get any statistical significance? I don't know the answer. See my post htt

> I like the idea, but how do you apply this to power law distribution outcomes and get any statistical significance? I don't know the answer. See my post https://news.ycombinator.com/item?id=10484602 https://news.ycombinator.com/item?id=10484602 where are distribution-free. So "power law", Gaussian, anything else, doesn't matter.