The Dallas Cowboys have no understanding of probability. You might think that has very little to do with being a quality football team, but a basic comprehension of the numbers has everything to do with playing efficiently.
Consider the old adage that it benefits offenses to “set up manageable third downs.” Even in today’s NFL, many teams still run the ball way too often, particularly on first down, to set up third downs that are easier to convert.
But stats have shown again and again that the best offenses are those that face third down the least—those that don’t try to set up manageable third downs, but rather call plays more efficiently on first and second down to avoid third down altogether. It might be easier to convert a 3rd-and-3 than a 3rd-and-7, but it’s easier to convert a single 3rd-and-7 than four straight 3rd-and-3s. Math. Nice.
One of the ways that you can see that the Cowboys don’t comprehend the numbers is that they don’t throw the football downfield nearly as much as they should. It’s the same idea as “setting up manageable third downs”—why throw downfield and risk an incompletion or interception when you can continually throw underneath and complete 70 percent of your passes?
Well, it’s kind of difficult to do something effectively over and over in the NFL because, you know, the other guys are professionals too. Even if each offensive play has, say, an 80 percent chance of being successful, the probability of running five straight successful plays is just 50 percent. So in many situations, it makes sense to aim for lower-percentage plays with greater rewards.
The Cowboys are a risk-averse team that seeks to maximize the happiness they get from a bunch of moderately successful plays instead of, you know, maximizing points. For them, a low-variance offensive strategy—a Stoic offense, of sorts—is superior to the ups-and-downs that come with a truly efficient unit. Meanwhile, the Saints are one of the teams that has embraced the use of analytics, and it shows in their offense. Let’s take a look.
An Attacking Offense
On a 3rd-and-7 at their own 34-yard line, the Saints lined up in a Shotgun Spread formation. The Cowboys showed a double-A gap blitz prior to the snap.
Dallas backed out of the blitz look, rushing only four defenders. The Saints protected well. That’s a prerequisite for getting the ball downfield, but not necessarily one for running downfield routes since they also open up underneath routes, i.e. a checkdown for Brees.
This is a look we don’t see often from Dallas—one on which every receiver threatens his defender vertically. On this play, wide receiver Kenny Stills ran a crossing route. Because of the way the Saints attacked every part of the field, he was singled up on cornerback B.W. Webb with no one else in the area.
The Saints understand that the value of throwing deep isn’t confined to their stats on those plays alone. There are indirect benefits to stretching the field, which include increased efficiency on intermediate routes like this one.
Later, New Orleans faced a 1st-and-10 at the Cowboys’ 28-yard line. They called for an empty set with three receivers lined up tight to the field.
Again, three of the receivers threatened the Cowboys’ defense vertically. Wide receiver Lance Moore was the target of Drew Brees on this play, running a seam route from the slot.
Brees threw accurately into a very tight window, but the Cowboys played this particular play well (likely because there was just 19 seconds remaining on the clock).
The result here is less important than the route combination, which the Saints used all night to attack Dallas. Instead of getting one receiver downfield and looking among four options underneath, the Saints use multiple players to attack downfield. If nothing is there, Brees has one or two underneath players to whom he can check down. The fact that this particular play was unsuccessful has nothing to do with the effectiveness of the overarching offensive philosophy.
A Stagnant Offense
Let’s compare that approach to what the Cowboys do. Coming out of halftime, the Cowboys were down by three scores and faced a 3rd-and-8 at their own 23-yard line. They used a Shotgun Trips Left formation with wide receiver Dez Bryant isolated to the boundary, as usual.
Look at the difference in how Dallas attacks defenses. Instead of receivers running vertically, everyone broke off their route even before the first down marker.
Tony Romo decided to hit Jason Witten in the flat, despite good protection, and he ended up gaining two yards.
This is a two-fold problem, in my view. First, the offense isn’t set up to really keep defenses honest downfield. If I were a defensive coordinator, I’d just double-team Dez Bryant and play everyone else underneath. Terrance Williams can potentially beat defenses deep, but Dallas doesn’t utilize him downfield as much as they should. Both he and Bryant ranked 32nd in the NFL in deep targets with 10 coming into Week 10, according to Pro Football Focus.
Here’s a visualization of the difference in aggressiveness for Dallas and New Orleans in 2013.
Romo’s average pass has traveled 8.0 yards past the line, compared to 9.1 for Brees. That’s not as small of a difference as you’d think; Brees ranks 12th in the NFL in average pass length and Romo is all the way down at 25th.
The short average pass isn’t a bad thing by itself, though. Actually, as long as the offense is using short passes in place of runs, it’s fine. But you still need to attack defenses vertically. Everyone wants to talk about play-calling balance, and they’re right that the Cowboys don’t have it right now.
That has nothing to do with the number of runs, however, and everything to do with the fact that there’s little variation in the passing game—few crossing routes, little deviation in alignments, and not enough downfield routes and throws.
So if the Cowboys want to throw the ball short again and again, it needs to be counteracted by some downfield looks. But Romo ranks only 28th in deep passing percentage with 9.5 percent of his passes traveling at least 20 yards past the line. Despite that, he ranks fifth in the NFL in deep passing touchdowns.
If you examine the other effective passers who have a relatively short average pass length—Aaron Rodgers and Peyton Manning, for example—they also offset that by throwing the ball deep much more than Romo. Both of those quarterbacks are at least 33 percent higher than Romo in deep passing percentage. They use screens and other short looks in place of most running plays, which decreases the average pass length, but they also get the ball downfield to optimize overall offensive efficiency.
For Cowboys fans, it’s difficult to retain hope. There’s no question that they’re talented, but that talent isn’t utilized in an effective manner. Worse, the Cowboys aren’t particularly good at adapting—fixing potential problems before they become major issues.
If the NFL is a microcosm of Darwinian evolution, the Cowboys and their stubborn refusal to embrace analytics are a species on the verge of extinction.