Small sample sizes

Extrapolating wildly from a small subset of performance is so common to sports analysis as to seem intrinsic to the enterprise. The problem is especially acute in football. The season is dramatically shorter than are the other major pro leagues, the sport receives more media attention than any other, and the physical realities of football result in more year to year variation in performance.

This past week provided a particularly good illustration. Example #1 is Tom Brady and the New England Patriots. The Pats have put together eleven straight seasons of ten or more wins. In eight of those years, including each of the past four, they’ve won 12 or more games. And yet, after beginning this season 2-2, the near consensus among sports media was to bury them. Brady is old. Belichick’s years of trading away talent has finally caught up to him. Their offensive line stinks. Brady has no weapons with which to work.

This is based on four – count them four – games. Is Brady the player he was in 2007? No. Nor is the offense what it once was. But as of this morning, the Patriots have scored the third most points in the AFC (in fairness, Denver’s played one fewer game). Is that a misleading data point, since New England’s only played five games and put up a large chunk of that total in one game? Well, that’s the point. How much can you really say about a team based on so little time? Just as commentators were getting ready to declare the Brady/Belichick era over, many of the same folks are now gushing about the Pats after their 43-17 pounding of the previously undefeated Bengals last night. They’ve “figured it out,” in this new telling of the Pats’ season.

I am not a Pats’ fan, so the thought of their finally entering a decline phase would be a pleasing one to me. But if you’ve got thirteen years of data on one side, and one or two bad games on the other, you might consider not giving undue weight to the latter.

Having said all that, I am just full of sour grapes for having fallen into the same analytical black hole I’ve just complained about. Last week, i plucked the Green Bay Packers defense/special teams off the waiver wire in my fantasy football league. I reasoned that, despite their poor play so,  they were going to be playing the Viking, a team that would be starting a third string quarterback (and it would be Christian Ponder, no less). I saw last year how a string of third string quarterbacks helped the Giants (sort of) salvage what was shaping up to be an historically bad season.  Since I didn’t draft one of the league’s really good defenses, I am patching it together week-to-week, and I felt OK about patching in the Packers last Thursday night. I had the Steelers’ in my starting lineup, facing another bad offense, the Jacksonville Jaguars. But I’d told myself that I’d switch out the Steelers for Green Bay before game time.

Then I started listening to the analysts. The Packers were giving up six hundred million yards per game on defense through four games. The Vikings surprising 41-point showing the previous week against the Falcons was, in fact, no fluke. They had a great running back tandem. By the time I arrived home about a half hour before kickoff last Thursday night, I was more or less convinced that Jerick McKinnon and Matt Asiata were Mercury Morris and Larry Csonka incarnate. Adrian who? That the quarterback who led the Vikes against the Falcons, Teddy Bridgewater was out of the game? Please. An afterthought.

A few schemes this, schemes that and another sprinkling of blah, blah, blahs later and I was mortified of inserting the Packers into my starting roster. So I didn’t.

And it took roughly 12 seconds of game time to realize I’d made a mistake. Ponder was battered. Once the Vikings fell behind, they had to throw to play catch up and Ponder was a sitting duck. Naturally, the Pack ran back an interception for a touchdown. Six sacks, three turnovers, a defensive touchdown and ten points allowed, for those of you who don’t play fantasy football, adds up to a huge night for the defense – a whopping 23 points in standard scoring (you’re doing well most weeks if your defense gets you ten). And I missed out because I listened to the freakin’ ESPN experts tell me how good the Vikings running game is, based on a couple of good games.

The moral of all this:

I am an idiot.



  1. I believe if you put bad teams in the playoffs, some of those bad teams might win more than people expect. Especially in hockey, baseball, futbol, and american football(possibly basketball). There isn’t absolute truth in sports. People don’t want to believe it though until they get an undesired result. Same people who talk about Goddell being reactionary. Oh the hypocrisy.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s