A few days ago, ESPN Dallas published the painful proof that the Cowboys aren't anywhere near where they need to be on third down:
- The Cowboys rank 30th in the NFL in third-down conversion rate (32.8 percent, 38-of-116).
- Romo's third-down QBR (19.8) ranks 29th in the NFL.
- Romo ranks 30th in the NFL in average yards per attempt on third downs (5.74).
- Romo's third-down passer rating (57.6) ranks 32nd in the NFL.
- Romo's third-down completion percentage (47.1) ranks 34th in the NFL.
I actually have the Cowboys at 39-of-116 on third downs, but either way, the 'Boys certainly do have some awful third down numbers. So let me preface this argument by saying that 1) third downs are important and 2) Dallas needs to improve in that area of their offense. They've been poor, and that sucks.
But that doesn't mean that the Cowboys must obsess over their third down offense.
Third down stats are noisy. By that, I mean that it's really difficult to tell if the Cowboys' 39-of-116 third down success rate is because they're actually poor on offense, or just due to a small sample.
Even at this point in the season, third down stats are fragile. If Dallas had converted only a half-dozen more third down plays, they'd jump from 30th in the NFL all the way up to 16th in the NFL in third down conversion rate. When a stat is that susceptible to fluctuations, it's a good sign that it's not indicative of anything of substance (or at least not to the degree first assumed).
So we can grade the Cowboys and Tony Romo solely on one subsection of their offense which comprises fewer than one-in-five plays, or we can just use all of the plays we have. Which makes more sense?
Fooled By Importance
I acknowledged that the third down plays within each game are really important to each team. We all already know that, and we don't need stats to prove it.
But that doesn't mean that the Cowboys need to obsess over their third down offense. One reason is related to the noise, i.e. if we can't decipher a poor third down offense from randomness through Week 11, there's no reason to get consumed by the numbers.
Second, teams shouldn't be focused on improving their performance in all areas of a game that are important. Yes, you read that correctly. Instead, they should be concerned about improving in the areas in which they can actually improve.
As an example, consider fumble recoveries. Recovering fumbles is extremely important to teams; if there are four fumbles in a game, recovering three or more can send your chances of winning skyrocketing.
But fumble recoveries aren't predictable. If you're an NFL coach and you spend excessive time working on recovering fumbles, you probably didn't increase your team's chances of winning at all, even though fumble recoveries are of extreme importance.
Third downs aren't as unpredictable of fumble recoveries, of course, but the point is that you should only focus on particular football situations as much as they can be improved.
And the way to improve a third down offense is through the same means by which the entire offense improves, regardless of the situation. If Dallas improves their passing game (which currently ranks 12th in the NFL in net-YPA), for example, they'll have better numbers across the board, including on third down.
Third Downs Aren't Standardized
The most important reason that third down stats need to be taken with a grain of salt is that they aren't standardized. The Cowboys, for example, have had an average of 7.45 yards to go on their third downs this year. The Colts, on the other hand, are all the way down at 6.32 yards to go. The Giants are all the way up at 8.46 yards to go.
Shortening the distance to go on third down of course aids offenses in converting on third down, but here's the hidden downside: they face more third downs. Third down offense is important, but first and second down are just as vital since, you know, they're the downs that come before third down. If you want to improve your offense as a whole, you should try to avoid third down altogether.
And the Cowboys have actually done that pretty well this year. They've faced 116 third downs the fewest in the NFL. They haven't run that many plays, but third downs represent only 19.6 percent of all of their offensive snaps. That's a good number, and one that suggests the Cowboys' offense is way better than their third down numbers show.
Coaches who obsess over third down stats and guide their offense to set up 'manageable third downs' are really doing a disservice to their teams. The goal shouldn't be to do everything possible to convert on third down; it should be to do everything possible to avoid third down.
None of this means that the Cowboys' third down offense has been satisfactory or that they're doing everything they need to do correctly on first and second down. Actually, they should stop running the ball so much on first down early in games in order to see even fewer third downs.
If you really want to perform the silly task of grading entire offenses based solely on third downs, then you should consider not only their third down conversion rate, but also the average distance to go for a first down and the frequency with which they face third downs.
Third down conversions are important, but they're more the result of a quality offense rather than the cause of it. Chasing a high third down conversion rate is synonymous to running the ball a whole lot because rushing attempts are correlated with winning; both confuse the cause with the effect.