The Tape Room

Defenses Have Fingerprints Too

How many defenders a coordinator stacks in the box is the most repeatable thing about a defense. Sack totals are the least. And the run funnel you spotted in October will probably still be there next season.

We have written before about offensive identities, the run-gap maps and shotgun rates that follow a team from one season to the next. Defenses have fingerprints too. Some ridges persist, others smudge. Rank a defense's core traits by how much of each survives into the following season and the ladder is not the one the highlight reel suggests.

The box count is the scheme talking

The most repeatable trait on the list is the one fans look at least: how many defenders the coordinator puts in the box. Since charting began in 2016, a defense's average box count has carried over with a correlation of 0.45, positive in all nine season pairs. That makes sense once you say it out loud. Box count is a choice, not an outcome. It is the coordinator's philosophy written into the pre-snap picture, and philosophies change slower than results.

The funnel is real, and partly scoreboard

The pass rate a defense faces, whether opponents attack it on the ground or through the air, persists at 0.34, positive in 25 of 26 season pairs. Run-funnels stay run-funnels; pass-funnels stay pass-funnels. For fantasy purposes that is a genuinely useful fact, because the funnel you diagnosed in October tends to still be operating the following September.

But name the confound before you spend it: part of a funnel is just the scoreboard. Defenses on winning teams face a pass rate 1 to 3.5 percentage points above the league's, and that has been true in every season since 1999, because opponents trail against good teams and throw to catch up. Some of what looks like a defense that invites the pass is simply a defense that plays with a lead.

The run wall is steadier than folklore says

Football wisdom holds that run defense is noise and pass defense is the stable skill. The record disagrees, mildly but consistently: rush efficiency allowed carries over at 0.285, a touch above pass efficiency allowed at 0.263. Neither is a rock, but the run wall is at least as repeatable a unit as the secondary. Explosives are the sharper surprise. The explosive-play rate a defense allows persists at 0.35, and it persists slightly better than the explosive rate offenses generate, which runs about 0.29. We would have bet the offense owns the big play. The data says the defense allowing it is the steadier half of the transaction.

The traits fans quote most persist least. The sack total is the loudest number a defense produces and the flimsiest one it repeats.

Sacks are the flimsy signature

At the bottom of the ladder sits the number that leads every defensive highlight package. A defense's sack rate per dropback carries over at just 0.21, below every other core trait measured here. Even the shape of the pass rush is only weakly sticky: how concentrated a team's sacks are in its top rusher, one star versus a committee, persists at about 0.17. If your team led the league in sacks, history says pencil in regression before you pencil in a repeat.

Two smaller prints worth keeping. Defenses sack the quarterback more at home than on the road, in 25 of 27 seasons, by about 0.3 percentage points on a base of 6 to 7 percent of dropbacks; crowd noise on third-and-long is a real, small thing. And terrible defenses heal faster than terrible offenses: compare each season's five worst defenses with its five worst offenses and the defenses win the improvement race the following year more often than not.

How to use it: when you preview a defense, trust the box count, the funnel, and the explosive rate, in that order, and discount last season's sack total almost entirely. The fingerprint is in the pre-snap picture, not the box score.