0 yards = 0 yards: Incomplete passes and the Jets running game

Much has been said about New York Jets Quarterback Zach Wilson’s completion percentage (51.4%) from last Sunday’s matchup against the Detroit Lions. Indeed, the general rule of thumb typically places the threshold for good at 60%, which Zach fell several completions shy of.

However, in watching the game I came away thinking that the offensive struggles could hardly be attributed to the passing game, which would be supported by a comparison of the yards per pass attempt (9.1) and yards per rush attempt (2.3).

Despite this clear divergence in the effectiveness of the two main methods of moving the football while on offense (passing and rushing), the fault for Sunday’s loss has been placed on Zach Wilson by many. In support, we can check out this very informal poll conducted on Twitter by Mets legend turned SNY analyst Steve Gelbs, which at the time of my writing has Zach Wilson leading the poll with 36% of ~5200 votes cast.

Given that Zach’s yards per attempt were not the issue (his YPA was the second highest among all QBs this week), I settled on the idea that the completion % must be the issue from a statistical perspective. What I then thought would be interesting would be to compare the “completion %” (if you will) of the running game as a comparison point, which I did using the following logical steps:

1. I landed on the assumption that incomplete passes are undesirable in a vacuum is because they lead to 0 yards.

2. Given this, the reason we care about completion % as a raw statistic is because it tells us about the rate at which a QB threw a pass that did not generate 0 yards.

3. We can just as easily create a similar statistic for the running game by dividing the number of runs that achieved 1 yard (or less since a -1 or -2 yard run doesn’t really do much more harm than a 0 yard run. It also felt disingenuous to compare that kind of small yardage loss to a QB sack that typically loses 5-10 yards)

So that’s exactly what I did using ESPN.com’s play-by-play data (https://www.espn.com/nfl/playbyplay/_/gameId/401437911).

More specifically, I grabbed every run and coded it 1 if a yard or more was achieved (a positive play) and 0 if a yard was not achieved. I also only counted runs by the Jets Running Backs given that they accounted for 17 of the Jets 19 designed “traditional” runs. The other two “non-traditional designed runs” were a “Tight End sneak” by Jets TE Tyler Conklin and an end around by Jets Wide Receiver Elijah Moore, which I did not feel adequately represented the Jets running game any more than a RB pass would represent the Jets passing game.

In total, this left 17 run plays for analysis. Of these runs plays, 10 went for positive yardage, reflecting a 58.8% “positive running play” rate. Notably, this % is a bit higher than Zach’s 51.4% completion percentage, but not nearly higher by enough to warrant the tradeoff in yards per play (9.1 for a passing attempt and 2.3 for a rushing attempt)

As a follow up, I built off my logic that a -1 or -2 yard loss did not move the needle much and decided to apply that same logic to a +1 or +2 yard gain. To support this decision, 2^nd and 8 is typically still considered 2^nd and long and thought to be “behind the sticks” so to speak. In this analysis, any 1- or 2-yard running gains were now coded as 0s (essentially net neutral outcomes) rather than 1s (essentially positive outcomes). Using this logic, two previously “positive” runs were now coded as ‘akin to an incompletion,’ and the “positive running play” rate dropped to 47.1%; this is slightly lower but approximately in the ballpark of Zach Wilson’s completion percentage.

Overall, based on this rather rudimentary analysis, I’m left with the conclusion that the passing game was far from the issue offensively. Could they have been better? Certainly and I don’t think anyone would argue otherwise. But the squeaky wheel designation almost certainly falls to the run game given the lower yards per attempt and approximately equal positive play rate.

So what do you think? Is the passing game being unfairly scapegoated (relative to the run game) for Sunday’s loss to the Detroit Lions? If not, why not?

And would you like to see more analysis in this mold? If so, what would you like to see changed or added? (As a note, I would be open to writing a statistical script to automate this kind of analysis and then providing the output weekly in an article as I think it would be interesting, but I don’t want to take the time to do so if there isn’t really much interest beyond myself. If we went that route, I’d plan to also perform similar analyses on the passing game on a play-by-play basis to allow for a more apples to apples comparison between the running and passing game. If there is interest please let me know in the comments)