NFL two-point conversion rules, and when to go for two

A two-point conversion is one snap from the two-yard line. That is the whole of it. No special formation, no separate rulebook, no clock: the offence gets a single scrimmage down, and if it reaches the end zone the touchdown is worth eight instead of seven.

The reason the two-point conversion rules repay proper reading is not the play. It is everything the league has bolted onto the try since, most of it in one afternoon in 2015, and the fact that the decision to attempt one is the single cleanest piece of applied probability in American sport. A coach who understands the arithmetic and a coach who does not will make visibly different choices in the same situation, and both will be able to defend themselves afterwards.

Start with what the rules say, because a surprising amount of the confusion is people arguing about a rule that changed a decade ago.

What the NFL two-point conversion rules actually say

After a touchdown the scoring team is given a try. One down, no clock, and two ways to take it.

Ask for the kick and the ball is spotted at the fifteen-yard line. Ask for the two and it is spotted at the two. The offence declares which it wants, the referee spots the ball, and from that moment the choice is made. A team that has asked for the kick and then wants the two has to spend a timeout or be given a penalty; it cannot simply line up differently and hope nobody noticed the ball is on the fifteen.

That declaration is a small rule with a large effect. It removes the possibility of disguising the intention, which means the defence knows exactly which of two completely different plays is coming, which in turn is a large part of why the two-point play is harder than its distance suggests.

• 15Yard line the kick is snapped from
• 2Yard line a two-point attempt is snapped from
• 2Points for a defensive return on a try
• 1Points if a try ends in what would be a safety

Fixed by Rule 11. None of these changes from season to season, and none of them is a measured value.

The try is one down and it is untimed. The clock does not run, the period cannot expire during it, and there is no possibility of a first down, because a try is not a series. Either the ball reaches the end zone in the offence's possession, or it does not.

The kick is a 33-yard field goal now, and the geometry is the point

Before 2015 the kick was snapped from the two-yard line. It was, in every meaningful sense, not a play. Kickers converted it so reliably that broadcasters cut to adverts during it and nobody complained.

The 2015 change moved the snap to the fifteen. Owners passed it by a vote of thirty to two, which is about as close to unanimous as that room gets, and the reason for the near-unanimity is that everyone could see the same problem: a play that succeeds essentially always is not a play, it is a delay.

Work out what the change did in yards, because the arithmetic is more interesting than the headline.

45%30%24%

• Line of scrimmage to the goal line15
• Depth of the end zone10
• Distance the holder sets up behind the line8

The kick is measured from the spot of the hold to the goalposts at the back of the end zone. Snapping from the fifteen rather than the two adds thirteen yards to what was previously a twenty-yard attempt.

Show the numbers

Where the 33 yards of an extra point come from

Item	Value
Line of scrimmage to the goal line	15
Depth of the end zone	10
Distance the holder sets up behind the line	8

Thirty-three yards. Under the old rule the same three components gave twenty. That is a genuinely different kick: not a hard one by the standards of a professional specialist, but far enough that wind, a poor snap, a slow hold or a low trajectory turn into misses rather than near-misses.

The consequence the league wanted was misses. The consequence it also got, and which mattered more, was that the expected value of a kick fell far enough for the two-point play to become a rational alternative in situations where it previously was not. Move one option's payoff down and you change every comparison that option appears in. The league did not so much promote the two-point conversion as demote its competitor.

The defence can score, and that changed the risk more than the reward

The same 2015 vote made the try a live ball in both directions. Before it, a failed conversion was simply a failed conversion: the ball was dead, the score stood at six, and everybody lined up for the kickoff. After it, the defence can take an interception, a fumble or a blocked kick the length of the field and score two points of its own.

This is the part of the rule people most often get wrong, usually by assuming a defensive return is worth one point because it followed a kick. It is not. It is two, whichever kind of try was being attempted, and the team that scored the original touchdown then kicks off as normal.

The change matters because of what it does to the distribution rather than the average. A two-point attempt used to have two outcomes worth zero and two. It now has three, worth two, zero and minus two, and the third is rare enough to ignore in a first-quarter decision and completely unignorable in a fourth-quarter one. A four-point swing at the wrong moment is a game.

There is a related rule that governs what the offence may do after losing the ball on a try, and it is unusually restrictive. A fumble on a try is treated the way a fumble in the final two minutes is treated: only the player who fumbled may recover and advance it for his own side. If a teammate recovers, the ball is dead and the try has failed. The defence is under no such limitation and may recover and advance freely. That asymmetry exists to stop the offence from turning a botched conversion into a scramble drill with eleven possible recoverers, and it means a loose ball on a two-point play is close to a defensive event by default.

The one-point safety, and why it has never happened here

Rule 11 contains one sentence that sounds like a typing error. If a try ends in what would ordinarily be a safety against either team, one point is awarded to the opponent.

A single point. In a sport whose smallest scoring unit is otherwise one point only as an appendage to six, this is the only route to a score of one on its own.

The realistic sequence runs like this. The offence attempts a two-point play. A defender intercepts it in the end zone he is defending, or recovers a fumble there, and instead of kneeling or getting out he is tackled in that end zone. The ball is dead where it lies, in the defence's own end zone, in the defence's possession, and the ordinary safety rule would award two points to the other side. On a try, the award is one.

There is a second, even less likely route in which possession changes twice and the ball ends up dead in the original offence's own end zone, which would hand a single point to the defence. It is worth knowing only because the rule is written to cover both directions rather than the obvious one.

The play has been theoretically available every week since the try became live, and the NFL has not yet recorded one. College football has, which is the usual pattern for exotic rules: the college game plays enough snaps in enough conditions that its long tail arrives first.

Do not file this as trivia and stop there. The one-point safety exists because the try was made a live ball with a change of possession available, and once a defence can carry the ball on a try, every ordinary rule about carrying the ball has to be given a value on a try as well. The rulebook is closing a hole rather than inventing a curiosity.

Penalties on a try are a small rulebook of their own

This is the section every explainer skips, and it is where officials earn their money.

A try is a scrimmage down, so all the ordinary fouls exist on it. What is unusual is enforcement, because the down cannot simply be replayed from a different distance without also changing which conversion is being attempted.

A defensive foul on a successful try. The score stands. The offence may take the points and have the penalty enforced on the ensuing kickoff, or decline the points and try again from a shorter distance. That second option is genuinely used: a defensive foul can convert a two-point attempt into one from the one-yard line, and a coach who wanted the two and got the kick instead now has a much better version of the play he originally asked for.

A defensive foul on a failed try. The offence retries with the penalty assessed. This is the most consequential category, because it turns a failure into a second attempt from closer in, and a defence that jumps offside on a goal-line conversion has effectively awarded points.

An offensive foul. The try fails, or is retried from further out at the offence's choice where the rules permit, and a fifteen-yard foul is carried forward to the kickoff unless it has already been used to order a retry or to wipe out a score.

Defensive pass interference. This is the one worth knowing. On a try, pass interference is enforced as a spot foul, which on a play starting at the two-yard line generally puts the ball at the one. It cannot be carried forward to the kickoff, and from the new spot the offence may attempt either a one-point or a two-point conversion. A defence that interferes on a conversion has given up half the field it had.

Consecutive tries. If penalties force repeated attempts and the offence switches between a kick and a scrimmage play, the enforcement is restacked and applied in order from the new spot. This is rare, it is ugly, and it is the reason the referee spends an unusually long time explaining what is about to happen.

The pattern across all of these is consistent once you see it. The rules are written so that a foul cannot be used to convert one type of try into the other for free, and so that yardage that has nowhere to go on a two-yard play ends up on the kickoff instead of vanishing.

When the try is not attempted at all

A try is skipped when a successful try could not affect the outcome of the game. That is why a touchdown that wins an overtime period is never followed by anything, and why a game can end on a score with no conversion attempted.

The precise wording carries more weight than it appears to. The exception is written around a successful try, and the reason is the 2015 change: an unsuccessful try can now produce two points for the defence, so a try that cannot help the scoring team may still be capable of changing the score. The rule is drafted narrowly to keep that possibility from being legislated away by accident.

In practice this catches out fans more often than coaches. A team that scores a walk-off touchdown in the fourth quarter and does not line up for the kick has not forgotten. The officials have applied the rule, and the final score is one point lower than the scoreboard arithmetic people are used to.

The break-even maths of going for two

Here is the whole calculation, and it fits on one line.

Let q be the probability that the kick is good and p the probability that the two-point play works. The expected points from kicking are q. The expected points from going are 2p. Going for two is the better bet whenever 2p is greater than q, which is to say whenever p is greater than q divided by two.

Because q is high but not one, the break-even sits a shade under half.

Arithmetic rather than measurement. Expected points are q for the kick and 2p for the two, so the break-even two-point rate is q divided by two. The kick reliabilities along the axis are illustrative round numbers, not observed values.

Show the numbers

The two-point rate you need, given how reliable the kick is

Item	Value
Kick good 90	45%
Kick good 92	46%
Kick good 94	47%
Kick good 96	48%
Kick good 98	49%

Read the line and the first useful conclusion arrives immediately. The break-even is almost flat. Across any plausible range of kicking reliability, the two-point play needs to succeed a little under half the time to be worth attempting, and it does not much matter which end of the range the kicker sits at. Arguments about whether a particular kicker is reliable are therefore mostly beside the point. The decision hinges on the two-point play, not on the kick.

The second conclusion is more uncomfortable. If a team believes its own short-yardage offence converts around half the time from the two, then in expected-point terms the two is roughly a coin flip against the kick for the entire first three quarters of every game. Not obviously better. Not obviously worse. Which means the expected-point calculation, on its own, almost never settles anything.

That is not a failure of the method. It is the method telling you that the answer lives somewhere else.

Why the marginal point is worth almost nothing, until it is worth everything

Football scores are lumpy. They cluster on multiples of three and seven, because those are the two ordinary ways to score, and a lead is not a continuous quantity so much as a position relative to a small set of thresholds. Three points behind is qualitatively different from four points behind. Eight behind is qualitatively different from seven.

So the value of the marginal point is not its expected contribution. It is whether it moves the team across one of those thresholds. This is the same logic that drives fourth-down decisions and it is measured the same way, in win probability rather than in points, using the sort of situational modelling that also produces expected points added.

Work the cases through by hand and the picture is clearer than any chart.

After the touchdown you trail by	A kick leaves you	Two points leaves you	What changes
1	level	one ahead	A kick guarantees at least extra time. The two buys a lead that a field goal erases anyway.
2	one behind	level	The cleanest case there is. The try itself can tie the game rather than merely shortening it.
4	three behind	two behind	Three behind means a field goal ties. Two behind means a field goal wins.
5	four behind	three behind	Four behind still needs a touchdown. Three behind needs only a kick.
9	eight behind	seven behind	Eight behind means a touchdown and another two. Seven behind means a touchdown and a kick.
12	eleven behind	ten behind	Both require two scores of exactly the same kind. Nothing has changed.

The same exercise from in front is shorter and less well known.

Lead by five after the touchdown and the kick puts you six ahead, where a touchdown and a kick beats you by one. The two puts you seven ahead, where the same drive only ties. Lead by twelve and the kick puts you thirteen ahead, where two converted touchdowns win; the two puts you fourteen ahead, where two converted touchdowns tie and the opposition has to find a conversion of its own somewhere. Both are real, and both are routinely passed up by coaches who treat a comfortable lead as a reason to stop thinking.

The row worth staring at is the one where you trail by nine, because that is the situation a team is in after scoring its first touchdown when fifteen points behind. Going for two there and failing leaves you nine behind rather than eight, which feels worse and is barely different. Going for two and succeeding turns the rest of the game into a normal one-score chase. The reason to go early is that a failure still leaves time to get the points back a different way, whereas the same decision deferred to the last touchdown has no recovery at all. Take the variance when you can still absorb it.

Why the chart exists, and what it is really doing

Every coaching staff has a two-point chart. The idea is usually credited to Dick Vermeil, who built one as a college coach long before anyone was modelling win probability, and the modern versions are computed rather than intuited.

What the chart is actually for is not the decision. It is the timing of the decision.

A coach standing on the sideline after a touchdown has perhaps twenty seconds, a crowd, a headset full of voices and a defensive coordinator asking about the next series. That is not an environment in which anyone reasons carefully about thresholds. The chart moves the reasoning to Wednesday, when it can be done slowly, and leaves the sideline with a lookup rather than a calculation.

Which is also why the criticism of charts misses. Nobody who uses one believes it captures every input. It does not know that the opposing defence has lost two interior linemen, or that the wind at that end has been ruining timing routes all afternoon, or that the quarterback has been visibly labouring since the second quarter. The chart is a default, and its purpose is to make deviating from it a deliberate act rather than a panicked one.

Why the two-point conversion rules changed the coaching more than the scoring

Moving the kick to the fifteen did not turn the NFL into a two-point league. What it did was remove the excuse for not thinking.

Under the old rules, the kick was so close to automatic that going for two required a coach to argue that his short-yardage offence would beat something approaching certainty. That argument was almost impossible to win in front of an owner, and the incentive structure of a head coach's job means the choice that is easiest to defend afterwards usually wins. Kicking was the defensible default because the kick almost never failed.

Once the kick could fail, the defence of the default weakened. A coach who kicks and misses now looks like a coach who made a choice, not a coach who was unlucky. That symmetry is what pushed the decision into the analytics department, and the analytics department was already there because of what had happened to fourth-down thinking over the previous decade.

The same rule change quietly altered late-game clock management too. A team down by fourteen with time for two possessions now has a live decision on the first touchdown rather than the second, and the two-minute drill is planned backwards from which conversion result it is playing for. Deciding to go for two after the first score, rather than the second, gives the drill a defined target for the whole of the remaining game.

How the play itself is designed, and why it is hard

Two yards sounds like nothing. It is the hardest two yards in the sport, for a reason that has little to do with the defence being close to the line.

A conversion is played in a box roughly twelve yards deep and fifty-three wide, and the back of the end zone acts as an extra defender. Every route has a ceiling. The vertical threat that stretches a secondary on any other snap does not exist, so a defence can play every receiver with inside leverage and safety help without conceding anything behind it. There is no deep third to respect and therefore no deep third to occupy.

What is left is a small number of families of answer.

Pick and rub concepts, where two receivers cross and one legally obstructs the defender covering the other. These are the reason conversion plays look like basketball, and the reason offensive pass interference is a live argument on roughly every one of them.

The fade, throwing high to the back corner where only a tall receiver can reach it. Cheap to install, respectable against man coverage, and the lowest-percentage of the common answers because a well-thrown fade is still a contested catch.

Run-pass options and quarterback runs, which are effective for the same reason they are effective anywhere: a defence that has to account for the quarterback as a runner is a defence one man short in coverage.

Motion and formation stress, used to force the defence to declare man or zone before the snap so the quarterback knows which of two answers he is running.

The good staffs treat the conversion as a separate installation, practised against a look they expect once a fortnight. The tell is the play call itself: an offence that runs its ordinary goal-line concept on a two-point try has not prepared one, and it usually shows.

How the defence plays it

Defending a conversion is a mirror of attacking it. Play everything with inside leverage, take away the crossing concepts, and dare the offence to throw the fade.

Two things make the defensive job easier than it looks. The first is the declaration: the ball on the two tells the defence a conversion play is coming, so there is no need to defend a kick. The second is the shortened field, which removes the coverage's worst nightmare, and lets a defensive coordinator use bodies in the flat and on the goal line rather than holding them deep.

The one genuine risk is exactly the one the 2015 change introduced. A defence that intercepts or recovers now has ninety-eight yards of open field and a two-point score at the end of it, and the players in position to make that return are, by construction, the ones who have been sprinting towards the line of scrimmage. The return is chaotic, badly blocked and rarely rehearsed, which is why the sensible instruction on many staffs is to secure the ball and go down.

How the rest of football does it

The two-point conversion is not an NFL invention, and the NFL was late to it.

College football has had it since 1958, from the three-yard line rather than the two. The American Football League used it throughout its existence in the 1960s. When the leagues merged, the rule did not survive the merger, and it took until 1994 for the NFL to adopt it. Tom Tupa, a punter operating as the holder, converted the first one in league history in the opening week of that season for Cleveland against Cincinnati, which is a properly football way for a rule to arrive.

College also uses the conversion structurally rather than tactically. In overtime, the alternating-possession format requires teams to go for two from a defined period onwards, and in the later periods the possessions themselves reduce to nothing but two-point plays. That is a deliberate device to end games, and it is the sharpest available contrast with the way the NFL settles a tie through its own overtime procedure, which keeps ordinary scoring and limits the time instead.

The difference in yard line is not cosmetic either. A yard, at that distance, is a large proportion of the task, and any comparison between college and professional conversion rates has to account for it before it means anything.

What happens on a try, start to finish

The sequence below is the whole rule in order, and it is worth having in mind because most of the argument about conversions happens in the middle three steps.

1. 1The touchdown is scoredThe scoring team is awarded a try. It is skipped entirely if a successful try could not affect the outcome of the game.
2. 2The offence declaresAsk for the kick and the ball is spotted at the fifteen. Ask for the two and it is spotted at the two. Changing after that costs a timeout.
3. 3One untimed down is playedThe clock does not run and the period cannot expire. There is no first down available, because a try is not a series.
4. 4The offence reaches the end zoneTwo points for a scrimmage play, one for a successful kick. The scoring team kicks off.
5. 5The defence gains possessionThe ball stays live. A return to the other end zone is worth two points to the defence, after which the team that scored the touchdown kicks off.
6. 6The ball is downed in the wrong end zoneIf the try produces what would ordinarily be a safety, the other team is awarded a single point. It is the only way to score one point on its own.
7. 7A foul changes the pictureA defensive foul can force a retry from closer in. Fifteen-yard fouls that have nowhere to go are carried forward to the kickoff.

One untimed down, four ways it can end, and a live ball throughout. The defence has been able to score on this play since 2015.

What to look for

Four habits will tell you more about how a coaching staff thinks than any post-match press conference.

Watch when they decide, not what they decide. A staff that goes for two after the first touchdown of a fifteen-point deficit has planned. A staff that goes after the second has been forced.

Watch the leads. Going for two from five ahead, or twelve ahead, is the tell of a staff working from thresholds rather than instinct, and it is far less common than the trailing cases.

Watch the play call. A designed conversion concept, with motion that forces a coverage declaration, is a staff that practised this. A goal-line run out of a heavy set is a staff that did not.

Watch what the defence does after an interception on a try. Securing the ball is the coached answer and returning it is the instinctive one, and which of the two you see tells you whether that detail was ever addressed in a meeting.

The rule itself is four sentences long. Everything that makes it interesting sits in what teams do with a play whose value cannot be settled by expected points, only by where the scoreboard happens to be standing. More on the rules that shape the way the sport is actually coached is collected in the American football archive, including what the league did to the kickoff for very similar reasons.