/cdn.vox-cdn.com/photo_images/3675708/121092608.jpg)
Yesterday, I devoted 1,400 words or so in response to a claim that St. Louis Rams' QB Sam Bradford's rookie season was vastly overrated. In short, yes, it was overrated, but author making that contention overlooked several important bits of context in a rush to throw cold water on fan and media enthusiasm.
Background makes all the difference in judging a picture. I'm running too close to empty to delve into critical theory, so I'll spare you the academic lessons...consider yourself lucky. It was kind of ironic, I suppose, that the author of that piece focused on the tendency of people to look mostly at yardage and completions when judging Bradford's rookie season, then ignoring defense-adjusted stats (from an article he cites) showing just how blah the receivers were. And of course there were the coaching issues and the fact that Bradford hadn't really played football the year before being drafted (sort of a them among St. Louis' first round picks the last two years).
There's another stat comparison the Pro Football Reference author overlooked, a measurement from his own site. In a great post from Smart Football (one we'll talk about again soon), the author makes this catch.
Moreover, the statistics are not all bad. Bradford's 5.4 Adjusted Yards Per Pass Attempt (Pro Football Reference's vaunted quarterback stat), although not great, was better than the rookie number for another highly touted rookie: Peyton Manning only had a 5.2 AY/A in 1998, his rookie season. My point is not that Bradford was 0.2 better than Manning, but instead simply that with young quarterback's it's a guessing game. Remember too that Bradford was coming off a college season where he barely registered any snaps due to injuries, and logic indicates that he's at least on the right direction.
Some people tried to turn this whole overrated argument into the old chestnut of stats versus observation, or whatever the hell you want to call it. That was not at all the point. Context is key to understanding things, stats require context just like any other data, quantified or otherwise.