It does follow from repeatability -- Warren Buffet and Goldman Sachs are not explained by the lottery analogy.

It doesn't follow from repeatability. Or rather it doesn't follow from the existence of examples of repeated success. A normal distribution of returns with repeated play will still lead us to expect at least one investor with Buffet's track record (if my econ. prof's math is to believed----I haven't have cause to look closely).

It would follow from "significantly more repeated success than a normal distribution would predict".

I do believe that Buffet's success is more than luck, just not that this particular argument follows.