Sample size of 1 (Iverson), or even 10 or 20 is not a valid argument against a statistical model. That is like saying you don't believe that cancer or heart disease (one of those two I think) is the number one killer of people because your Mom died in a car accident. It's super easy to find exceptions to models, the same way individuals don't always fit generalizations. That doesn't invalidate the generalization, rather it points out the faultiness of applying generalizations to individuals (like claiming that since one person scored higher in this model they are therefore a better player than someone that scored lower). Statistics are misapplied/misinterpreted constantly.