The confusing part, as Bret mentions, is the lack of context with those stats: Rank amongst how many?; how large is the gap between 1st place and say 10th, or 10th and 20th - and are those gaps meaningful or trivial?; what kind of sample sizes?; etc. Too often there's some stuff that seems so counter to observations that I'm reluctant to trust the remainder. for ex: I remember Rob Neyer examining the Hampton/Bell-Cedeno/Dotel trade all those years ago and concluding (among other things) that the Mets would be hurt by it becuase Roger Cedeno had been a "dominant defensive RFer" - and of course he had some kind of stats to back up his claim. Now Roger verison 19.99 certainly wasn't as bad as the version we got back a few years later but, even then, anyone who had watched him play that season would know that the notion of calling him "dominant" defensively was simply laughable.