Why NFL Teams Go For It on Fourth Down So Much Now

Published on September 11th, 2026
Written By: Dave Manuel

For twelve consecutive seasons, from 2006 to 2017, NFL teams went for it on fourth down about ten per cent of the time. Every single year. Not drifting, not trending, just the same number over and over like somebody had set a thermostat and walked away. Then in 2018 the line went vertical, and it has never come back down. The league now goes for it just under twenty per cent of the time, and one coach has personally posted the three most aggressive seasons ever recorded. What makes this worth writing about is not that the numbers changed. It is that the maths never did. The case for going for it on fourth down was sitting in plain view for decades while an entire sport politely ignored it, and then changed its mind almost all at once. This is how that happened, who got blamed for being right too early, and why your instincts about fourth down are almost certainly wrong in a very specific direction.

Sports-King Feature

The Fourth Down Revolution: How the NFL Stopped Punting

A flat line for twelve years, then a cliff. The Belichick call everybody laughed at, the coach who took it further than anyone, and the reason nobody can guess a conversion rate correctly.

By Sports-King

Go Rate, 2006 to 201710%

Peak Go Rate19.5%

Fourth and One Converts65.7%

Tush Push Ban Failed By2 votes

Every entry below carries a line, because that is genuinely what this story is: a shape on a chart that stayed flat for a decade and then did something. Green lines are things going up that should go up. Orange marks the moment something broke. Grey is a trend that has stalled. Read them in order and you get the shape first, then the man who was blamed for it, then the maths that was right all along.

The LineTwelve flat years and then a cliff

Share of fourth downs where teams went for itRoughly 10 to 11 per cent for over a decade, then a break in the series that you can date to a single season.

2006 TO 2017 ROUGHLY 10 TO 11 PER CENT, EVERY YEAR
FROM 2018 .. CONSISTENTLY ABOVE 15 PER CENT
PEAK ...... 19.5 PER CENT IN 2021
ATTEMPTS 2017 485
ATTEMPTS 2018 539
ATTEMPTS 2025 886

Here is the whole story in one shape. For twelve straight seasons, from 2006 to 2017, NFL teams went for it on fourth down about ten or eleven per cent of the time. Not roughly. Not trending. The same number, year after year, like a thermostat. Then 2018 arrived and the line snapped upwards and has not come back down. The league now goes for it just under twenty per cent of the time, and it peaked at 19.5 per cent in 2021. In raw attempts: 485 in 2017, 539 the following season, and 886 last year. Something changed, and it changed fast enough that you can point at the season.

Belichick Was EarlyThe most criticised call of his career, and he was right

Go-for-it rate on fourth-and-2Twenty per cent when he did it. Thirty-nine per cent by 2021.

WHEN ..... NOVEMBER 15, 2009
SITUATION . 4TH AND 2, HIS OWN TERRITORY
GAME ..... PRIMETIME, LATE, AGAINST THE COLTS
OUTCOME .. FAILED. MANNING SCORED THE WINNER
REACTION . WIDELY DERIDED
THE RATE THEN 20 PER CENT
THE RATE BY 2021 39 PER CENT

On a Sunday night in November 2009, Bill Belichick went for it on fourth-and-two in his own territory in the closing minutes against Indianapolis. They did not get it. Peyton Manning took the short field and scored the winning touchdown, and Belichick spent the next week being taken apart by absolutely everybody. Here is the context nobody had at the time: in the first twelve weeks of that season, teams were going for it in that situation twenty per cent of the time. By 2021 the same situation produced a go-for-it decision thirty-nine per cent of the time. He was not reckless. He was about a decade ahead of the field, in public, on national television.

What Broke in 2018One Super Bowl, and everybody changed their mind

League-wide fourth down attempts per season485 attempts in 2017. 886 last season, even allowing for the extra game added in 2021.

EAGLES 2017 20 OF 29 ON FOURTH DOWN
RATE ..... 69 PER CENT
POINTS GAINED 74
POINTS LOST 3
IN THE PLAYOFFS CONVERTED ALL THREE ATTEMPTS
THE NEXT SEASON ALL BUT EIGHT TEAMS GOT MORE AGGRESSIVE
THAN THEIR OWN THREE-YEAR AVERAGE

So what actually happened in 2018? The season before, Doug Pederson's Eagles went 20 of 29 on fourth down - a 69 per cent conversion rate, more attempts than anybody - and won the Super Bowl. One analysis put the value of that aggression at 74 points gained against three points lost, which is not a marginal edge, it is a rounding error away from free. They converted all three fourth-down attempts in the playoffs, two of them in the Super Bowl itself, including a fourth-and-goal from the one where a running back took the snap, pitched to a tight end, and threw to the quarterback in the end zone. Nick Foles became the first quarterback in Super Bowl history to catch a touchdown pass, and he had suggested the play himself. None of the underlying maths was new - the Eagles were running numbers fed to them by an outside firm through Pederson's headset as green, yellow and red lights. What the ring provided was cover. The following season all but eight teams in the league were more aggressive than their own three-year average.

Dan CampbellThe man who took it further than anyone

Highest single-season go-for-it rate on recordWhen Detroit hired him in 2021, the record was 25.4 per cent. He now owns the top three marks in history.

HIRED .... DETROIT, 2021
RECORD AT THE TIME 25.4%, THE 2020 EAGLES
2021 LIONS 29.5 PER CENT, 41 ATTEMPTS
2022 LIONS 29.6 PER CENT
2023 LIONS 30.5 PER CENT
RANKING .. THE TOP THREE SEASONS EVER RECORDED
FOUR-YEAR TOTAL 151 ATTEMPTS, A LEAGUE HIGH
NOBODY ELSE HAS REACHED THAT IN FOUR YEARS SINCE 1991

Every revolution needs somebody who takes it too far, and in this case that is Dan Campbell. When Detroit hired him in 2021, the most aggressive single season anybody had ever produced was the 2020 Eagles at 25.4 per cent. Campbell's Lions then posted 29.5, 29.6 and 30.5 per cent in three consecutive years. He does not just hold the record - he holds the top three, and he set them consecutively. In his very first season he went for it 41 times, which was itself a record at the time, and across his first four years he kept the offence on the field on fourth down 151 times - a total no coach has reached in a four-year span since at least 1991. For scale, the most aggressive team of the entire 1980s went for it thirty times in a season. Campbell's Lions treated thirty as a slow start.

Short Yardage Went FirstAnd it is not close

Go-for-it rate on fourth-and-1Fourth-and-one is now a go-for-it situation more often than not, which would have been unthinkable fifteen years ago.

2019 4TH AND 1 54 PER CENT
2020 4TH AND 1 67 PER CENT, THE BIGGEST SINGLE JUMP
4TH AND 1 .. 69 PER CENT GO RATE IN 2021
4TH AND 2 .. 39 PER CENT
4TH AND 3 .. 19 PER CENT
TWO-POINT TRIES 11 PER CENT, ALSO A RECORD
4TH AND 3 OR LESS ATTEMPTS DOUBLED FROM 2017 TO 2021

The revolution did not arrive everywhere at once. It started in short yardage and worked outwards, and you can date each stage. The single biggest jump came in 2020, when the fourth-and-one go rate leapt from 54 per cent to 67 in one season. Longer distances lagged badly behind - by most accounts coaches did not really get over the hump on fourth-and-two and fourth-and-three until 2024. In 2021 the league set modern records across the board: a 69 per cent go-for-it rate on fourth-and-one, 39 per cent on fourth-and-two, 19 per cent on fourth-and-three, and an 11 per cent two-point conversion rate. Attempts on fourth-and-three-or-less simply doubled between 2017 and 2021. Fourth-and-one has now flipped from an automatic punt or kick into a situation where going is the majority decision, which is about as complete a reversal of received wisdom as a sport is ever likely to produce.

Your Gut Is WrongAnd so is everybody else's, in the same direction

Fourth-and-4: what people think, and what happensResearchers asked people to estimate the conversion rate. The average guess was 36 per cent. The actual figure is about 47.

4TH AND 4 .. CONVERTS ABOUT 47 PER CENT
4TH AND 5 .. CONVERTS ABOUT 44 PER CENT
AVERAGE GUESS 36 PER CENT
WHO GUESSED LOW FANS, ACADEMICS AND AT LEAST ONE
NFL FRONT-OFFICE EMPLOYEE

This is the part I find genuinely great. Researchers at Wharton pulled apart a dataset of more than 600,000 plays and found that fourth-and-four converts about 47 per cent of the time and fourth-and-five about 44. Then they asked people to guess. The average guess was 36 per cent. They were not asking idiots, either - the low guesses came from academics, a fantasy league full of obsessives and at least one person who works in an NFL front office. Everybody's intuition is miscalibrated, and crucially everybody's is miscalibrated the same way: we all think fourth down is harder than it is. Which is precisely how a whole sport punts its way through a decade without anybody noticing.

The SneakThe most efficient play in football, and it is not close

Expected points added on fourth-and-1, by play typeA quarterback sneak produces roughly double the expected points of any other call on fourth-and-one.

4TH AND 1 CONVERSION 65.7 PER CENT OVERALL
SAMPLE ... MORE THAN 3,000 ATTEMPTS
SNEAK EPA . ROUGHLY DOUBLE OTHER PLAY TYPES
THE EDGE .. ABOUT 0.77 POINTS PER PLAY
WHY ...... NO HANDOFF, NO READ, NO TIME TO REACT

Underneath the strategy sits a very boring play that happens to be enormously effective. Across more than three thousand fourth-and-one attempts, teams converted 65.7 per cent of the time. But break it down by play call and the quarterback sneak produces roughly double the expected points of anything else - an edge of about three quarters of a point every time it is run. There is no handoff to be timed, no read to be made and no time for a defence to react. It is one man falling forwards behind five other men. And it is, statistically, the single most efficient thing an offence can do with a yard to gain.

The BotA newspaper built a robot to shame coaches, and it worked

Public tolerance for punting, roughly speakingIllustrative. The New York Times built its fourth down bot in 2014 to analyse decisions live and call out coaches who punted against the numbers.

BUILT .... 2014, BY THE NEW YORK TIMES
WHAT IT DID ANALYSED DECISIONS IN REAL TIME
AND THEN .. PUBLICLY CALLED OUT PUNTS
STATUS ... NOW DEFUNCT
WHY IT MATTERED IT MOVED THE ARGUMENT INTO PUBLIC VIEW

One underrated actor in all of this is the media, and specifically a small robot. In 2014 the New York Times built a fourth down bot that analysed every decision in real time and cheerfully announced when a coach had punted against the numbers. It is defunct now, which is the highest compliment you can pay it - its entire job was taken over by the sport itself. What the bot did was move the argument out of front offices and onto television. Once a fan watching at home could see a graphic saying the model said go, the professional cost of punting flipped. Coaches no longer got quietly criticised for gambling. They got loudly criticised for not.

The Tush PushThe logical endpoint, and the fight over it

Eagles conversion rate on the play, by season92.6 per cent in 2022, 83.3 in 2023, 82.1 in 2024. At one point they had converted 96.6 per cent across 116 attempts.

PROPOSED BAN BY GREEN BAY, MAY 2025
VOTES NEEDED 24 OF 32
VOTES RECEIVED 22
MARGIN ... FAILED BY TWO
AGAINST .. EAGLES, RAVENS, BROWNS, LIONS, JAGUARS,
DOLPHINS, PATRIOTS, SAINTS, JETS, TITANS
IN 2026 .. NOBODY EVEN PROPOSED IT AGAIN

Follow the logic far enough and you arrive at the tush push, which is a quarterback sneak with people pushing from behind, and which Philadelphia converted at 92.6 per cent in 2022, 83.3 in 2023 and 82.1 in 2024. Green Bay proposed banning it in May 2025. The vote needed 24 of 32 and got 22. It failed by two. Jeffrey Lurie brought Jason Kelce into the meeting room about ten minutes before the session started, which is the single best detail in this entire article. The Eagles then posted "Push On" and uploaded twenty-six consecutive minutes of tush push footage to their own YouTube channel. In 2026 nobody bothered proposing a ban at all.

The Scale of ItEleven teams doing what one team used to do

Teams going for it 30 or more times in a seasonThe most aggressive team of the entire 1980s went for it thirty times in a season. Last year eleven teams did.

THE 1980s BENCHMARK THE 1984 ATLANTA FALCONS
THEIR TOTAL 30 ATTEMPTS, MOST OF THE DECADE
LAST SEASON 11 OF 32 TEAMS HIT 30 OR MORE
PER GAME .. ABOUT 1.4 ATTEMPTS PER TEAM
SUCCESS .. ABOUT 56.8 PER CENT

If you want one comparison that carries the whole thing, it is this. Across the entire decade of the 1980s, the single most aggressive team in the league was the 1984 Atlanta Falcons, who went for it thirty times in a season and were presumably considered lunatics for it. Last season, eleven of the thirty-two teams did that or more. The average team now attempts about 1.4 fourth-down conversions a game and succeeds around 56.8 per cent of the time - better than a coin flip, on a down that used to be treated as the end of a possession by definition.

Has It Peaked?The honest answer is probably

League-wide go-for-it rate, recent seasonsThe peak was 19.5 per cent in 2021. It has hovered rather than climbed since, which suggests the easy gains have been taken.

PEAK ..... 19.5 PER CENT, 2021
SINCE THEN HOVERING, NOT CLIMBING
WHY ...... THE CLEARLY CORRECT GOES ARE ALREADY TAKEN
WHAT IS LEFT GENUINELY MARGINAL DECISIONS
THE RISK .. AGGRESSION AS IDENTITY RATHER THAN MATHS

So does the line keep going up? Probably not much. The rate peaked at 19.5 per cent in 2021 and has hovered rather than climbed since, and there is a decent reason for that: the obviously correct decisions have all been taken. What remains are the genuinely marginal ones, where the models are close enough that the confidence intervals overlap and going for it is not clearly better. There is also a newer failure mode worth naming. A few coaches now appear to go for it because being aggressive is part of their brand rather than because the situation calls for it, which is the same mistake as reflexive punting, just wearing better clothes. The revolution won. The next argument is about the leftovers.

The Numbers That Moved

Everything that changed, and roughly when.

Measure	Before	After	Note
League go-for-it rate	About 10%	Just under 20%	Flat 2006 to 2017, then rose
Attempts per season	485 in 2017	886 in 2025	An extra game was added in 2021
Fourth and one go rate	Low	69% in 2021	Now the majority decision
Fourth and two go rate	20% in 2009	39% in 2021	The Belichick situation
Fourth and three go rate	Low	19% in 2021	Modern record
Two-point attempt rate	Low	11% in 2021	Also a modern record
Single-season record rate	25.4%, 2020 Eagles	30.5%, 2023 Lions	Campbell holds the top three
Teams with 30+ attempts	One, in the 1980s	Eleven last season	The 1984 Falcons were the outlier

The Whole Thing in One Picture

Fourth down decisions, before and after the break

There are not many charts in professional sport with a break this clean. Twelve seasons of a sport doing the same thing, then a single year where the whole league changes its mind at once.

Bits and Pieces

The Maths Was Never the ProblemNothing in the underlying analysis changed in 2018. Academics had been publishing versions of this argument since the 1970s, and the play-by-play data everybody now uses goes back to 1999. The numbers were sitting there the whole time. What changed was the cost of acting on them - once one team had won a Super Bowl being aggressive, a coach who went for it and failed was following a trend rather than losing his mind. This is a story about professional risk, not about arithmetic.

Everyone Guesses LowThe Wharton finding is the one to remember, because it explains decades of institutional behaviour in one line. Asked to estimate the conversion rate on fourth-and-four, people average about 36 per cent. The real answer is around 47. Nobody guesses high. And when everybody in a building shares the same directional error - players, coaches, executives, fans, broadcasters - the collective decision is wrong in the same direction for thirty years and nobody in the room notices, because the room agrees with itself.

The Team That RefusedThe best illustration of how uneven the shift was comes from San Francisco in 2018. According to one analysis using a win probability model, the 49ers did not go for it on a single fourth down in a one-score game outside the final four minutes that season - despite the model recommending it eleven separate times. That is not a coach making eleven close judgement calls. That is a coach who had decided the question was already answered. Within a few years that position had become untenable across the league.

Defence Has Not Caught UpOne genuinely open question. The whole revolution has happened on the offensive side of the ledger, and there is no equivalent tactical response available to a defence. You cannot decline a fourth down. The best a defensive coordinator can do is keep more bodies near the line in short yardage, which creates its own problems if the offence throws. Offences got a new option. Defences got a new problem.

Spare a Thought for PuntersThere is a hard number attached to this and it is brutal. One analysis found that the average contract value of the ten highest-paid punters had not moved at all since 2021, across a period in which the salary cap rose by more than twenty per cent. Over roughly the same span, the average pay of the ten highest-paid head coaches rose about thirty per cent. The people making the decisions got more valuable. The people whose job the decision eliminates did not. Punting has not disappeared and never will, but it has been repriced as the thing that happens when the model says no.

Sports-King's Note

Now for the fine printGo-for-it rates vary between sources depending on how fourth downs in garbage time, kneel-down situations and end-of-half desperation are filtered, and the gaps are not trivial. The 2023 Lions, for example, are cited at 30.5 per cent on one widely used basis and at 34 per cent by ESPN Stats and Information, which described that as the highest rate of any team this century. Both figures describe the same season. The figures here follow the most commonly cited analyses of the nflreadr play-by-play dataset, which covers every play since the 1999 season. Specifically: attempt rates of roughly 10 to 11 per cent from 1999-2000 through 2017-18, consistently above 15 per cent from 2018-19, and a peak of 19.5 per cent in 2021-22; league-wide attempts of 485 in 2017, 539 in 2018 and 886 in the 2025 regular season, the last of which includes the effect of the seventeenth game added in 2021; 2021 modern records of 69 per cent on fourth-and-one, 39 per cent on fourth-and-two, 19 per cent on fourth-and-three and an 11 per cent two-point attempt rate; single-season team records of 25.4 per cent by the 2020 Eagles and 29.5, 29.6 and 30.5 per cent by the Lions in 2021, 2022 and 2023; an average of about 1.4 attempts per team per game in 2024 at a 56.8 per cent success rate; a fourth-and-one conversion rate of 65.7 per cent across more than three thousand attempts between 1998 and 2016; and Wharton estimates of roughly 47 per cent on fourth-and-four and 44 per cent on fourth-and-five against an average public guess of 36 per cent. Dan Campbell went for it 41 times in 2021, a record at the time, and 151 times across his first four seasons, a four-year total no coach has reached since at least 1991 according to Sportradar. Philadelphia went 20 of 29 on fourth down in 2017 for a 69 per cent conversion rate, with one published analysis valuing the aggression at 74 points gained against three points lost. The Belichick game was played on 15 November 2009. The tush push ban proposal failed at the May 2025 spring meeting, receiving 22 of the 24 votes required. Charts in the entries are shaped to illustrate the direction and scale of each trend rather than plotted point-by-point from a full season series; the underlying figures are stated in the text and the fine print. Nothing here is betting advice.

One Last Thing

What I keep coming back to is that nobody discovered anything. There was no breakthrough, no new statistic, no piece of tracking data that suddenly revealed the truth about fourth down. The maths in 2006 was the same as the maths in 2018. Teams were leaving points on the field every single week, in plain view, on national television, and the reason they carried on doing it was not stupidity - it was that the first coach to stop would be the one who got destroyed for it. Belichick found that out in 2009. Pederson won a Super Bowl and made it safe. Campbell took it somewhere nobody had been. And now a play where a quarterback falls forwards while his teammates shove him is so effective that the league spent an entire offseason arguing about whether to make it illegal, and could not find the two votes.

The short version, for the road: NFL teams went for it on fourth down about 10 to 11 per cent of the time every season from 1999-2000 through 2017-18, then rose consistently above 15 per cent from 2018-19 and peaked at 19.5 per cent in 2021-22. League-wide attempts went from 485 in 2017 to 539 in 2018 to 886 in the 2025 season. When Bill Belichick famously went for it on fourth-and-two in his own territory in November 2009, teams were going for it in that situation 20 per cent of the time; by 2021 it was 39 per cent. The 2021 season set modern records of 69 per cent on fourth-and-one, 39 on fourth-and-two, 19 on fourth-and-three and an 11 per cent two-point rate. Dan Campbell holds the three highest single-season rates on record at 29.5, 29.6 and 30.5 per cent, having taken over a record of 25.4 per cent set by the 2020 Eagles. Fourth-and-one converts at 65.7 per cent overall, with a quarterback sneak worth roughly double the expected points of any other call. Wharton research puts fourth-and-four at about 47 per cent against an average public guess of 36. And the proposal to ban the tush push failed in May 2025 by two votes, receiving 22 of the 24 required.