Arrest That Man; He Speaks in Maths

One of the numerous axes that stats-minded sports fans (like the guys at FJM, who I can't recommend enough) like to grind is that coaching almost certainly matters a lot less than you think. "Coaching," as a variable, is of course nearly impossible to measure in the first place. You might know it when you see it, with broad strokes; we can probably agree that football coaches who don't know when to use timeouts (Herm Edwards, Art Shell) or baseball managers who don't understand the relationship between fatigue and injury (Dusty Baker) suck, but beyond that it's really hard to say anything meaningful about whether one coach is "better" than another.

The effect of coaching is similarly murky. Gregg Easterbrook's TMQ column this week argues that the perceived impact from preparation and planning is overstated because of an observer bias that likes to think that someboldy's at the controls. I tend to agree with that, but would also add that there's quite a bit of rent-seeking among commentators and pundits in the sports media. Many of them are former players whose endorsement deals depend on the perception that there's more going on than a child's game in which the better players are usually going to win; many of them are former coaches who want to get back into the game, and their career prospects depend on a wealthy owner believing that they have an Answer that no one else posesses.

Or more simply, the whole brutal charade of the post-game show, pre-game show and highlights show can only exist so long as enough viewers think that there's something that Howie Long knows about NFL football that they don't (even when it's abundantly, painfully clear that it's not true).

Which brings me to my original motivation for writing this post. The only franchise that seems to have any grasp of this is the NFL's Pittsburgh Steelers. The Steelers have had exactly two (2) coaches in the last 37 years. Bill Cowher, who resigned the other day, has had some very good seasons (2 Super Bowl appearances; one win) and a couple of poor seasons. In particular, he's done well with good rosters and had losing records with shitty ones. He won a Super Bowl the year before last with a 2nd year QB playing out of his gourd and an aging star RB that the rest of the team rallied around; this season he went 8-8 after that QB almost killed himself in a motorcycle accident and the RB retired to a career of laughing at awkward jokes from Sterling Sharpe.

And since those are the things that determine how many games you win, there's precious little point in throwing good money after bad to hire, say, Nick Saban.