> ## Content Index
> Fetch the complete content index at: https://www.statsdigest.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Run Plays In the 4-Minute Offense Were Football's Worst Play-Calls in 2025
- URL: https://www.statsdigest.com/4-minute-offense-run-epa-collapse-2025/
- Published: 2026-08-25T12:23:32.000Z
- Updated: 2026-08-25T12:23:32.000Z
- Description: In 2025's clock-killing 4-minute offense, the run play NFL teams called most was by far their worst option.
- Author: Bill Seifert
- Tags: 4-minute offense, EPA, NFL analytics, play calling, run game, 2025 NFL

The NFL's 4-minute offense — the possession you get when you're up a score with the clock inside five minutes — is supposed to be the game's most controlled situation. You have the ball, you have the lead, and you have a play sheet built around one job: don't screw it up. In 2025 the league screwed it up on a scale we haven't seen in the play-by-play era.

Filter every 2025 play down to fourth quarter, game clock at or under 5:00, offense leading by 1 to 8 points, and a real snap (not a spike or kneel). You get 568 plays. On those plays, offenses called a designed run 398 times and a pass 170 times — a 70.1% run rate. And the runs averaged **−0.246 EPA per play, 3.15 yards per carry, and a 15.3% first-down/TD conversion rate**. The passes on the same down-and-distance situations averaged **+0.068 EPA, 6.82 yards, and a 35.9% conversion rate**.

That gap — a third of a point of EPA per play between the two options, with the worse option getting called more than two-thirds of the time — is the biggest reason 4-minute lead offense was, by EPA, the single most inefficient offensive segment in football in 2025.

## Table of Contents

1. How bad was the 2025 run?
2. It wasn't always this bad
3. The stuff rate is doing the damage
4. Every down looks broken
5. The one honest counterargument
6. What I'd watch for in 2026

## How bad was the 2025 run?

To put −0.246 EPA/play into context, here's every 2025 rush attempt bucketed by whether it came in 4-minute lead offense or literally anywhere else:

| Bucket             | Plays  | EPA/play | YPC  | Conv % |
| ------------------ | ------ | -------- | ---- | ------ |
| 4-min lead offense | 398    | −0.246   | 3.15 | 15.3   |
| Every other rush   | 14,376 | +0.013   | 4.57 | 27.0   |

The same league that ran for 4.57 yards a pop everywhere else in 2025 ran for 3.15 in the situation where clock-and-lead management is supposedly the whole point of the play sheet. The conversion rate — how often the run actually moved the sticks or scored — was cut nearly in half.

Compare that to the passing game in the same 4-minute spot: 6.82 yards per play and a 35.9% conversion rate, both meaningfully better than the season-long passing baseline (5.99 EPA-neutral pass plays across all situations produced 0.014 EPA/play and 32.0% conversion). Whatever defenses were doing to defeat 4-minute rushes, they were doing the opposite to 4-minute pass calls.

## It wasn't always this bad

The angle here isn't that runs underperform passes in 4-minute offense. That's been true for years — defenses stack the box, everyone in the stadium knows what's coming. The angle is that 2025 was a step-change worse than what came before it.

Full 2019–2025 window, same 4-minute lead definition:

| Season   | Run EPA/play | Run YPC  | Run conv % | Pass EPA/play | Run share |
| -------- | ------------ | -------- | ---------- | ------------- | --------- |
| 2019     | −0.198       | 3.70     | 17.7       | +0.195        | 70.5%     |
| 2020     | −0.071       | 4.87     | 21.3       | −0.035        | 67.6%     |
| 2021     | −0.112       | 4.07     | 21.7       | +0.007        | 67.7%     |
| 2022     | −0.137       | 4.25     | 19.8       | −0.113        | 73.7%     |
| 2023     | −0.192       | 3.76     | 18.3       | +0.015        | 72.7%     |
| 2024     | −0.130       | 4.37     | 18.1       | +0.066        | 67.4%     |
| **2025** | **−0.246**   | **3.15** | **15.3**   | **+0.068**    | **70.1%** |

2025's run EPA/play is the worst of the seven-year window by nearly five one-hundredths of a point — larger than the gap between second-worst (2019) and fifth-worst (2021). Yards per carry (3.15) is the lowest in the sample by more than half a yard. Conversion rate (15.3%) is also the lowest.

Passing in the situation didn't get worse. If anything it drifted a little more efficient. The collapse is entirely on the run side.

## The stuff rate is doing the damage

The mechanism that killed the 2025 4-minute run wasn't a missing homer — those plays weren't built for homers. It was the disappearance of the *acceptable* run. Look at how often 4-minute rushes ended at or below zero yards, and how often they ended at or below two yards (i.e. gave up second-and-long or third-and-long):

| Season   | Stuffs (≤0 yd) | ≤2 yards  |
| -------- | -------------- | --------- |
| 2019     | 24.8%          | 51.0%     |
| 2020     | 21.0%          | 45.7%     |
| 2021     | 20.9%          | 49.9%     |
| 2022     | 21.7%          | 46.9%     |
| 2023     | 20.5%          | 51.4%     |
| 2024     | 23.3%          | 45.8%     |
| **2025** | **25.9%**      | **54.0%** |

Both the stuff rate (25.9%) and the ≤2-yard rate (54.0%) were the highest of the seven-year window. More than one in four 4-minute lead runs in 2025 lost yardage or gained none; more than half went for a first-down manslaughter of two yards or fewer.

## Every down looks broken

There's no down where the 2025 4-minute run functioned. Broken out:

| Down | Plays | Avg togo | EPA/play | YPC   | Conv % |
| ---- | ----- | -------- | -------- | ----- | ------ |
| 1st  | 188   | 9.97     | −0.306   | 3.28  | 9.0    |
| 2nd  | 133   | 7.55     | −0.136   | 3.84  | 15.0   |
| 3rd  | 69    | 6.04     | −0.169   | 2.23  | 30.4   |
| 4th  | 8     | 2.25     | −1.341   | −3.25 | 37.5   |

First-down runs — the least defensible spot to run into a stacked box because you still have three downs to make it — went for 3.28 YPC and moved the sticks 9% of the time. Third-down runs went for 2.23 YPC. The four 4th-down runs, at an average of 2.25 to go, averaged −3.25 yards. That's not "physical football wearing them down." That's a phone-booth fight the offense keeps losing.

Even restricting to 1st and 2nd down only — cutting out the third-and-short and fourth-down calls where the run rate is naturally highest — the pattern holds: 321 first/second-down runs at −0.236 EPA/play and 11.5% conversion, versus 77 first/second-down passes at +0.139 EPA/play and 31.2% conversion.

And this is a league-wide pattern, not a team pattern. The 10 most run-heavy 4-minute lead teams by usage all cleared 70% run rate; seven of them had negative rush EPA in the situation, several of them badly (HOU −0.50, DAL −0.51, WAS −0.53, ATL −0.49 on their 4-minute lead runs).

## The one honest counterargument

The one number in this dataset that does *not* condemn the coaches is win probability added. Averaged across the full 2025 4-minute lead sample:

- **Run:** +0.0040 WPA per play
- **Pass:** −0.0008 WPA per play

Because WPA properly prices in the runoff on the game clock, a stuffed run with the lead is not the disaster its EPA implies — clock burn is a real asset when you're ahead late. This isn't a 2025 quirk either; runs have graded better than passes by WPA in this situation in every season of the 2019–2025 window I pulled.

So the tension is real: the play call the analytics community grades as bad by EPA is actually grading fine — even slightly positive — by the win-probability model built to score exactly this game state. If you believe the WPA framing, the "bad" 4-minute run is just doing its job as a clock-eater.

The honest read is that both things can be true. Running the ball keeps the clock moving, and that alone justifies some of the calls. But 2025 pushed past the point where the tradeoff pencils out cleanly. YPC dropped nearly a full yard from 2024\. Conversion rate fell to 15.3%, meaning more than four in five of these possessions handed the ball right back to a trailing offense — often with the clock at 4:00 or 3:20 rather than :00\. And the pass game in the same situation was more efficient by EPA than the *league-average* pass call across all situations (+0.068 vs +0.015 for all 2025 passes).

If the run is only defensible by clock burn, then the run has to actually be sustainable enough to burn clock. In 2025 it wasn't.

## What I'd watch for in 2026

I'd expect defenses to double down on this — front-loading run fits when the tendency data screams run, betting that the offense will accept the negative expected-points call because it protects the win probability. Which, at 70% run rate, most offenses did.

I'd also watch to see whether any of the more aggressive play-callers rebalance. On the observed 2025 splits, a coach who moved his 4-minute lead pass rate from \~30% to \~40% would have been trading calls worth −0.246 EPA for calls worth +0.068\. On roughly 400 league-wide plays a season, that's not a small number. Whether they'll accept the variance is a different question — but the numbers say the tradeoff is on the table.

The 4th-down conversation ate the 2010s, and it was worth having. The 4-minute offense conversation is the one that's quietly been sitting there in 2025\. It's not the coaches being too conservative when to go for it. It's what they're calling to protect leads once they have them.