http://deadspin.com/5926617/why-aaro...r-sabermetrics

It's an interesting article... in no way disproves Sabermetrics, but does show how in every statistical system there are outliers for whom the system doesn't work.