This is about baseball, but if you care about analytics you should read it. It’s about the lack of Sabermetric-minded managers making on-field decisions. The money quote:
There are at least a dozen major league teams that gobble up sophisticated analysts from all corners and use them in any way they can conceive of to improve their valuation of personnel. For years, I wished that the Cubs would be one of those teams, and now that they are, I couldn’t be happier about it. But it’s not 2003 anymore. What kind of gains can be expected from an analytical front office when there are at least a handful of others who are equally adept at analyzing the numbers?
That’s why I’m flummoxed as to why not a single team has placed an analyst in charge of on-field strategy. At least once every other game, I see a manager make a decision that seems obviously wrong, and I don’t usually pay attention.
We’re nowhere near the point where anyone knows anything concrete enough about on-field analytics in football to need a guy making on-the-fly decisions based on numbers. But I’ve been musing recently on some problems with the role of statistics and analytics experts like performance analysts and technical scouts at football clubs under a coach or manager who doesn’t respect their work, or does but doesn’t understand it.
For example, while we don’t know much about technical scouts, we do know that some look for specific attributes favoured by the manager/coach. But what if those attributes in isolation aren’t that important?
One thinks here of the tale of Sir Alex Ferguson selling Jaap Stam because he didn’t tackle enough. Simon Kuper wrote of the sale, “As Ferguson later admitted, this was a mistake. Like many football men in the early days of match data, the manager had studied the wrong numbers.”
Stam was defending brilliantly, so much so he didn’t need to make desperate last-man tackles. To Sir Alex, it simply looked like Stam wasn’t doing his job. We might see the same issue arise when a manager asks for players with a trait beyond the obvious (they score lots of goals), but one that might not be the best fit for the existing team—or worse, a statistical chimera that might seem impressive on paper but doesn’t translate to a better team performance.
For example, a technical director, a chief of scouting, or a performance analyst may come to a conclusion that counters their manager’s strategy, or at least changes it subtly. A manager more amenable to accepting their conclusions may make changes accordingly, but a far more conservative leader may restrict the role of their staff in making on-field decisions.
As football front offices rely more and more on data-gathering and—far more important—the intelligent use of that data, there could be more pressure from below on managers to use it properly.