You are here
Running the Numbers: Callahan & Unexpected Play-Calls in 2013
Football is a game of competing minds. The Cowboys’ coaches must not only figure out what their players do best, but they need to determine how they can play to their players’ strengths while knowing that the opposing coaches are trying to limit the effectiveness of those strengths. In many situations, that can lead to counterintuitive thinking; the Patriots, for example, do such a great job of taking away an opponent’s best option that, in many situations, it’s optimal to build the game plan around “sub-optimal” players. Quite the paradox.
The study of this sort of decision-making – the type that’s governed by rational minds, often in a zero-sum game (just like football) – is known as game theory. The best way to make sense of the concepts behind game theory is with an example.
Last week, I published an article detailing why I think the running game philosophy is changing under Bill Callahan. In it, I mentioned that the Cowboys have historically had the most success running to the weak side and from spread formations, i.e. where they have the fewest blockers. It’s not inherently advantageous to run where there aren’t any blockers, of course, but since defenses typically have fewer defenders in those areas, the net effect is positive.
So why don’t the Cowboys just run to the weak side all the time? Well, that would obviously be a problem since defenses would catch on quite quickly, negating the advantage the Cowboys once possessed. So NFL offenses need to find some sort of balance, running optimal plays as much as they can before defenses defend them differently. The point that will maximize overall efficiency is known as the Nash equilibrium. In regards to run location, that point is where the efficiency of weak-side runs matches that of strong-side runs. And since NFL teams are still much more successful when running to the weak side, it follows that they should be doing it more often.
NFL coaches obviously don’t need to concern themselves with the specifics of equilibriums and complex decision theory, but they should be doing two things: 1) calling the “unexpected” in an effort to remain one step ahead of the opponent, and 2) determining when it’s best to play to your strengths and when you should be focused on exploiting opponent weaknesses (or, preferably, both).
Expect the Unexpected
Thus far through two preseason games, we’ve already seen a change in the Cowboys’ play-calls. Although they’ve expectedly stayed “vanilla,” the ’Boys have still run the ball way more often to the perimeter and thrown more screen passes than what we saw in 2012. Those play-types suggest that Callahan is at least taking opponents’ actions into account when picking plays – a rudimentary form of game theory.
So without further ado, here are the “unexpected” play-types, the plays that might be sub-optimal in a vacuum but most beneficial in practice, that the Cowboys should consider running more often in 2013. Read
There’s no single type of play that I’d like to see more in 2013 than the counter. Counters take a long time to develop and there’s a bigger opportunity for negative yardage than on a dive or other quick-hitting run, but there’s also a (much) larger probability of hitting a home run. From 2009 to 2011, Dallas averaged over 7.0 yards per carry (YPC) on counters! They ran only six of them last year, however, likely due to concerns about the offensive line.
And as mentioned, weak-side runs are usually best. The Cowboys have actually increased their rate of weak side runs, which was as low as 19.5 percent, every year since 2009. In 2012, it was nearly double that. Read
Everyone seems to believe that the Cowboys need to run the ball more often to take pressure off of Tony Romo (which is a form of game theory itself), but why not run more screens? Screens are high-percentage passes that can hold back pass-rushers, generating the same effect you’d want from a handoff (risk minimization), but with higher upside.
Dallas ran only 24 screens in all of 2012, and 16 of those were to receivers. Those aren’t the screen passes I’m talking about (although they can be useful at times). Rather, the Cowboys look like they’re going to run more running back screens under Callahan, which shouldn’t be difficult considering they called an average of one every two games last season. With exceptional pass-catching backs, there’s no reason that we shouldn’t see more screens in 2013. Read
- Spread Runs and Tight Passes
One of the great things about having versatile tight ends is that you can split them out wide and utilize spread formations with heavy personnel. That allows you to keep your best blockers on the field when you want to run, yet still spread out the defense. And that sort of strategy works. Read
The Cowboys have always been most successful when calling a play that is “unexpected” based on the formation. Again, it might be best to run from tight formations in theory, but in practice, it’s typically most beneficial to run from spread formations (and pass from tight ones) due to the defense’s preconceptions and actions. Read
- Play-action passes
If you follow me on Twitter (@BalesFootball), you know I talk quite a bit about play-action passes during the season. That’s because they’re underutilized by just about every NFL team. Of the 27 quarterbacks who took at least half of their team’s snaps in 2012, only five of them had a lower yards per attempt (YPA) on play-action passes than straight dropbacks.
And Romo was one of the league’s best. He completed 66.2 percent of his play-action looks, averaging 8.6 YPA and generating a 109.1 passer rating. He recorded a passer rating of 88.3 on all other passes.
Despite that, the Cowboys had the lowest play-action pass rate in the NFL, by far. Romo attempted a play-action pass on just 10 percent of his dropbacks. The difference between Romo and the second-lowest quarterback, Eli Manning, was larger than the difference between Manning and the next 11 quarterbacks!
And I know it’s popular to argue that you can’t run play-action passes without an effective running game, but that’s just not true. Defenses play situations, not necessarily past rushing efficiency, so defenders typically bite up on play-action passes based on the down-and-distance, not whether you’re averaging 5.0 YPC or 3.5 YPC.
We saw Romo have success on play-action without a running game, and the same was true for passers like Matt Ryan, Ryan Tannehill, Ben Roethlisberger, Peyton Manning and Andrew Luck, all of whom ranked in the top 10 in play-action passer rating with rushing offenses ranked 20th or worse in efficiency. Actually, of the 10 most successful play-action passers in 2012, only two, Cam Newton and Robert Griffin III, played on offenses that ranked in the top 10 in rushing efficiency (and that’s really just because of their own rushing prowess).
So the Cowboys’ rate of play-action passes should really increase in 2013, regardless of whether or not the rushing game improves. Read