There was a time when NFL film study was perhaps more valuable than statistical analysis just because it was really difficult, or impossible, to obtain reliable data. The 1970s Dallas Cowboys had no exportable box scores or play-by-play data, so all of that information would need to be collected by hand.
Even when I began analyzing the Cowboys and NFL just a few years ago while in college, certain types of analysis would take me hours and hours to complete. I tracked and categorized every Cowboys play, but I didn’t have any unique data on other teams. If I wanted to know how often the Redskins pass on first down, for example, it might take me quite a bit of time to uncover that data.
Nowadays, there are all kinds of awesome analytics tools available for every fan—and team. I say team because I think organizations could really benefit from using tools like Pro Football Reference’s Game Play Finder to scout upcoming opponents. Using the tool, you can uncover extremely in-depth information in seconds. I just got all kinds of data on the Redskins’ red zone rushing rate in games they’re trailing while on the road in less than five seconds. Nbd.
One reason that using analytics is better than traditional film scouting is that the differences in efficiency and accuracy aren’t even comparable. It takes countless hours for a scout to study film and come away with a single (informed) opinion about a team. Worse, that opinion might be the result of specific plays that stick out in his mind; he might have just spent a day studying film to formulate an inaccurate conclusion.
Another reason that stat analysis is better than film study is that it is evolutionary. When I create a model to predict a specific outcome, I can and do alter that model as new evidence comes in. I tweak the algorithm to make it better. I can start off with something that’s not that great and make it amazing. Further, others can build on my work, and I can do the same with their work. Analytics are scientific.
But how do we improve film study? More hours in the film room? It’s a really difficult question to answer because the conclusions drawn from traditional film study aren’t falsifiable—the true mark of science. When a scout tells you that a player “has heart” or “possesses great hips,” what does that stuff even mean? How would we falsify it?
Because film study usually isn’t falsifiable, it leads to all sorts of differing opinions within the scouting community. This is analysis of the same player—Riley Cooper—by two different well-respected scouts.
Because stat analysis is efficient and adaptive, it’s scalable. It can be improved upon and built up in a way that traditional scouting cannot. At its core, traditional scouting/film study is inefficient and inaccurate. Even worse, it’s dogmatic.
If the Cowboys are going to draw conclusions about their opponents from watching film of a few hand-selected games without tracking the results, they might as well get their opponent insights from the Flying Spaghetti Monster. Film study has the potential to be extremely useful, but only through a union with stat analysis.
Using numbers from PFR and Pro Football focus, I charted some interesting numbers for Washington—specifically quarterback Kirk Cousins—heading into Week 16. You know, just to get the team started.
Throwing on First Down
With Robert Griffin III at quarterback, the Redskins have loved running the ball on first down. But that changed last week with Cousins at the helm.
The Redskins threw the ball on at least two-thirds of their first downs in every single quarter last week. It’s just one game, but it’s interesting because it shows the coaches have a lot of confidence in Cousins. The Redskins were winning the game at halftime, for what it’s worth, so the Cowboys can expect the ‘Skins to come out firing.
Throwing over the Middle
Even though his career is still young, Cousins has established a habit of looking over the middle. Here’s how often he throws between the hashes, as compared to Tony Romo.
Cousins’ pass rate over the middle is about 33 percent higher than Romo’s. The Redskins don’t have a dominant tight end, either. We’re working with a large enough sample size that there really could be something there.
Despite his penchant for passing over the middle, Cousins has been far better when throwing outside.
His passer rating when throwing to both the left and right is far superior to that over the middle. We see the same effect with completion rate...
...and with touchdown-to-interception ratio.
These numbers might suggest the Cowboys should do what they can to properly defend the perimeter of their defense. You don’t want to just let players run free over the middle, of course, but it could be smart to play zones that make it difficult to get the football out wide.
But don’t take that advice just yet, Cowboys. I still need to turn on the film.
Football Dallas: Taking Cowboys Content Further. Download it for free here!