Analytics have an important role in football, but certain factors complicate things
Tuesday’s rare public appearance from Browns chiefs strategy office Paul DePodesta has sparked renewed discussion about the role of analytics in football. The Browns claim that they use analytics no more than any other team; people with other teams who have gotten some exposure to how the Browns use analytics through the recently-completed interview process, as Simms said on Wednesday’s PFT Live, disagree.
Regardless, every team uses analytics in some way or other. It’s important to know their limitations in football, however.
“I think people have a really warped view of what analytics is,” DePodesta said at the Tuesday press conference introducing new Browns coach Kevin Stefanski. “I think I may have a very different conception of what everybody else has. When I think of analytics, I just think of having certain frameworks to make decisions under uncertainty. I mean, look, everything we do in these jobs, is really built around uncertainty. What players are we going to take in the draft. What we are going to call on third-and-eight. It is all about uncertainty. So what frameworks can you create that at least stack the odds in your favor? Give you a better chance of being successful and whether that is drafting a player, hiring a coach or calling a play? It is not necessarily about numbers and spreadsheets.”
But it is an objective process, aimed at trying to minimize uncertainty through the manner in which similar decisions made under similar circumstances have worked out in the past. That’s really what analytics is -- an assessment of how past events predict future outcomes.
The problem comes from the potential dissimilarities between past situations and current ones.
“I think what’s most important [is] what ones you don’t want to pay attention to,” Saints coach Sean Payton said during a November appearance on PFT Live. “In other words, you have a lot of information, but what is it that you look at that means something substantial? Because different than baseball, in our game, you might have an injury or your left tackle is having trouble blocking the defensive end so the numbers compute to say, ‘Hey you need to go for it here,’ but those aren’t accounting for the injury at left tackle, or the pressure your quarterback is having.
“There’s so many other factors other than an at-bat versus a pitcher, left handed, right handed. However, there are some things that are extremely important and Ryan Herman in our [football research] department here each week will do several reports. We’ll do a ‘keys to victory’ analytics report, separate from our regular keys to victory. In other words, outside the box thinking, here’s what he would say relative to what he’s seen from the opponent. So I think more importantly than anything, what specifically do we want to pay attention to and which ones do we want to ignore?”
That’s the key. On Sunday, for example, the odds were low that the Seahawks would convert on fourth and 11 from their own 36 while down by five points with roughly three minutes left. The analytics surely said the chances of victory were greater if the Seahawks punted and tried to stop the Packers and get the ball back than if the Seahawks failed to convert and gave the Packers a short field with a chance to extend their lead to eight or, possibly, to 13 or 14.
But any numbers relevant to those circumstances were generated from scenarios that didn’t include: (1) Russell Wilson having a vintage day for Seattle; and (2) Aaron Rodgers having a vintage day for Green Bay. That’s where a football team needs to be willing to ignore the analytics and say, “We’d rather trust our guy versus their defense than our defense versus their guy.” Because Seattle’s guy and Green Bay’s guy were both playing great.
And that’s why the head coach gets paid so much money, and why he’s always a bad season or two away from being fired. He’s the one who should be charged with knowing when to embrace the analytics and when to disregard them.
It’s also why Bill Belichick has six Super Bowl victories in 20 years with the Patriots. He processes all relevant information, both objective and subjective, and he makes the best decision under each situation he faces, factoring in the odds of a certain approach working, the specific realities of the game unfolding in front of him, and his own sixth sense from a lifetime spent around football as to how he can best bend the outcome in his favor.
That’s how Josh McDaniels has learned the game during his time with Belichick, and it’s probably one of the reasons why McDaniels didn’t get the job. He has a handful of rings that prove the very real limitations of relying too much on formulas and equations generated by past events.