The Law of Genius and Home Runs Refuted

In a lively, provocative article, DeVany claims inter alia that the size distribution of home runs follows a continuous "power law" distribution which is nested in a larger class of "stable" statistical distributions characterized by an infinite variance. He uses this putative fact about the size distribution of home runs to argue that concern about the use of steroids to enhance home run ability is necessarily misplaced. In this article, we show that the initial claim is false and argue that the subsequent claim about the potential importance of steroid use does not follow from the first. We also show that the method used to establish that the size distribution of home runs is characterized by an infinite variance is unreliable and will find evidence "consistent" with infinite variance in all but the most trivial of data sets generated by processes with finite variance. Despite a large and growing literature that spans several fields and uses methods and arguments similar to DeVany's, we argue that mere inspection of the unconditional distribution of some human phenomenon is unlikely to yield much insight.