Please stop punting

Supported by Greenhaven Road Capital, finding value off the beaten path.

Sloan Conference week continues! Monday it was how General Managers make decisions, today is all about the NFL.

It’s easy to say “coaches should stop punting!” yet they keep doing it because punting is a conventional failure. “No one got fired for buying IBM,” said Rory Sutherland. Along with career risk/career capital, there’s bias, inertia, and other factors. Here’s what these football guys said about analytics in sports.

…

Analytics isn’t a panacea. Nothing is. Numbers are merely tools that can lead to black and blue thumbs too. Mike Lombardi (more on him here and here) said that coaches need to figure out the figures and structure practices around them. It’s not about convincing players so much as, in Jocko Willink’s words, showing “commanders intent.”

This is when the buy-in happens. John Urschel played for the Baltimore Ravens and echoed what Lombardi said. If a coach found helpful statistics the players would readily adopt the strategy. However, said Tedy Bruschi, it had to work. He said there were times when Bill Belichick (more on him here) would stop following a model early in the season and never revisit it. Wrong priors were gone priors.

Actions in practice should reinforce the analytic indicators said, Lombardi. Should they punt? isn’t decided on Sunday. That decision is made on Tuesday. Teams will (should?) practice and prepare. The fourth down bot was fun but by then it’s too late.

There is in-game data that the players use. Former players Bruschi and Urshcel both said that it has to be simple. These aren’t statisticians, Urshel said, these are football players. Sometimes he and the other coaches had to help players figure out what this or that meant. Bruschi said that one or two best guesses are what the players need, especially during key moments of the game; third-downs, red zone, high confidence plays, etc.

There’s so much going on the players need to focus on the most important things. We saw this too with Eliud Kipchoge, the fastest marathoner in the world. It’s also something professional investors remind the amateurs; save enough, limit fees, diversify holdings, and be patient. Do the Most Important Things first.

Executives can roll around in the numbers a bit more. Lombardi said that individual player numbers can tell you when someone’s physical skills are declining or when a player needs more rest. Bruschi said that each team approaches numbers differently and that figures are all relative. Yards per game, for example, is a baloney number said, Bruschi.

The panel also discussed how difficult it was looking in from the outside. This is a common theme in sports from Andre Agassi in tennis to Ben Falk in basketball to Neal Huntington in baseball. People on the inside need thick skin for the criticism bombardment from people on the outside who don’t have the same information. It’s so nuanced, said Lombardi, that the same scheme like ‘Tampa Two’ is played differently in Pittsburgh than New England.

This internal/external dynamic also applies to trading players. Like used car sales, there’s an information asymmetry. How, asked moderator Bill Barnwell, do you evaluate Jimmy Garappolo? (This panel occurred before his trade).

Well, said Chase Stuart, look at your prior. What was his score when your team graded him for the draft? Then look at the evidence; completion percentage, time with the Patriots, etc, and update that view. The sample size is small enough that Stuart says to not update your prior. Bruschi said that the sample size is small but the comparables are good. Throwing forty passes in the NFL is a much lot better sample than throwing forty passes in college. But is it better than four years of passes?

There’s also the performance per dollar to consider said, Sandy Weil. Rookie deals are cheap, especially if a player is excellent like Russel Wilson during his rookie contract. Garappolo only has one year left on that type of deal so he’ll be more expensive.

This stuff is complicated – and that’s before the biases the other GM’s talked about. Weil said he was shocked to see that just a few years ago many NFL teams used the fastest forty-yard dash time in their player evaluations. That’s like an executive who picked his best quarter from the past three years.

You get this kind of thinking, said Lombardi, when you begin with the end in mind. If you know you need to solve for X – lineman, linebacker, kicker, etc – you’re more likely to justify why a player does that. Instead, outsource your scouting at the combine. Distance can create impartiality.

 

Thanks for reading.