Who Won the Trenches?

Since 2021, college players can get paid for their name, image and likeness, and they can transfer without sitting out a year. So here's a popular theory: the richest programs scoop up all the best pass rushers, and defensive lines have never been scarier. More sacks, more tackles for loss, quarterbacks running for their lives. It's a great theory. We checked it. The pass rushers didn't get any scarier, and the first few years of the new era actually belonged to the blockers.

Count the right way first

Here's a trap: you can't just count sacks per game. College games have fewer plays than they used to: about 64 a game in 2016–18 (not counting garbage time), and about 60 now. Fewer plays means fewer sacks even if defenses are exactly as good. So everything below is a rate: how often something happens per chance to do it.

The pass rush: flat as a pancake

Our sack and tackle-for-loss counts start in 2016, which gives us three clean seasons before the new rules (2016–18) and five after (2021–25). Here they are, side by side:

That's about as close to “nothing changed” as real data gets. If NIL money built super pass rushes, they're hiding it well.

The run game: blockers won round one

Now the plot twist. A “stuff” is when a run gets stopped at or behind the line of scrimmage, the clearest sign a defensive line won the play. Our run data goes back to 2005, so we can see the whole story:

Stuffs were already sliding for years: 23.8 per 100 runs in 2005, about 20 by the late 2010s. Offenses have been getting better at blocking, or at least better at scheming runs where the blocking works. But 2021 fell off a cliff: 17.0, when the old trend said 18.7. 2022 and 2023 stayed down at 17.3 and 17.2. Those are the three lowest seasons in our data. Blockers also won more ground per carry. “Line yards” estimate how much of each run the blockers earned (the first few yards count fully, longer gains less), and they hit 3.04, 3.05 and 3.06 in those years, the three highest we've got.

Then the defense punched back. In 2024 stuffs were right back on the old trend (17.9, trend 18.0), and in 2025 they jumped to 19.1, well above where the trend said they'd be.

Update: 2026 so far

We first wrote this before 2026 had much football in it, so here's the first look, after five weeks. To keep it fair, we only compared the same five weeks of every season, and only games between two FBS teams (early-season blowouts of smaller schools would skew it).

The comeback stalled. Through five weeks, 2026 looks almost exactly like 2024: better for defenses than the 2021–23 low, but a full point behind 2025's pace, and 9th of the 12 seasons we compared.

Five weeks is early, but early weeks are a surprisingly good preview. From 2014 through 2025, the first five weeks lined up with the full season almost perfectly (a correlation of 0.98). If 2026 follows that pattern, it'll finish somewhere between 17.5 and 18.4 stuffs per 100 runs, short of 2025's 19.1. So “the defense took it back” looks more like one big year than a new normal. We'll check again when the season's over.

So why 2021?

Here's the problem: three big things hit college football at the same time, and this data can't pull them apart.

• Name, image and likeness pay, starting in July 2021.
• Free transfers: from 2021, a player could switch schools once without sitting out a season.
• The COVID bonus year: every player in 2020 got an extra year of eligibility. So 2021 through 2023 were full of fifth- and sixth-year players.

Our favorite suspect is the third one. Offensive line is the position where experience matters most: five guys who have to move together like one. Suddenly teams had lots of veteran linemen who'd been blocking together for years. When those super-seniors finally ran out of eligibility, the defense got its edge back. That fits the timing well, but fitting isn't proving, and 2026's early slip back toward the 2024 level is a hint it isn't the whole story.

Are the rich teams hoarding pass rushers?

That's the sharpest version of the NIL theory, so we checked it directly: compare power-conference defenses (SEC, Big Ten, Big 12, ACC, the Pac-12 through 2023, and Notre Dame) with everyone else. If big money is stacking up the best rushers, that gap should grow.

The gap shrank. In 2021 and 2022, the power-conference fronts actually made slightly fewer havoc plays than everyone else (by 0.21 and 0.20 per 100 plays). But watch the last two years: that havoc gap was 0.60 in 2024 and 0.76 in 2025, the widest of any season we can measure. Two seasons isn't a trend yet. Ask us again in a year or two.

One more suspect: the offenses

Sacks don't just depend on the pass rush. Offenses now throw a bit more (48.8% of plays in 2021–25, up from 47.2% in 2014–19), and lots of those throws are quick, get-it-out-now passes built to beat the rush. A defense could be getting more pressure without getting more sacks, and our data doesn't track pressure. So the fairest way to say it: defensive fronts aren't finishing more plays behind the line than they used to.

Data: season stats for every FBS team from CollegeFootballData.com: sacks, tackles for loss and opponents' passes and runs (2016–2025), and play-by-play based advanced stats with garbage time left out: stuff rate and line yards (2005–2025) and front-seven havoc (2016–2025). 2019's sack and tackle-for-loss counts look broken (Navy is credited with 45 sacks, after 16 and 10 the two seasons before), so that season is left out of those numbers; 2020, the COVID season, is left out of everything. The 2026 update uses per-game advanced stats for weeks 1–5 of every season since 2014 (games between two FBS teams, garbage time left out). The rule changes: NIL policy, one-time transfers and the 2020 eligibility waiver. The full analysis is in dl_nil_era.ipynb in the project.