[m3kw9]

There seem to be survivorship bias and the examples he demonstrated are probably at the tail end of a distribution