Skip to content
CricketTaken

Explainer

NFL challenge and replay rules: what New York can fix

NFL challenge and replay rules from the rulebook: the two challenges, how a third is earned, what the booth reviews alone, and what replay assist may fix.

By CricketTaken EditorialPublished Explainer22 min read

How this is written and checkedReport an error

A coach throws a red flag. The referee picks it up, walks to a monitor at the edge of the field, and puts on a headset. Whatever happens next, he is not the one deciding. The ruling will be made by the officiating department at league headquarters, on a screen the referee may not even be shown, and he will announce a result he did not reach.

That is the part of the NFL challenge and replay rules almost nobody has updated in their head. The pictures still look like 1999. The authority moved years ago.

What follows is the whole system as it currently stands: who may start a review, what may be looked at, what may not, what the officiating department can now fix without anybody asking, and where the money is on a challenge decision. Every provision here comes from Rule 15.

What a challenge buys, and what it costs

Each team is permitted a minimum of two challenges. The head coach, or a designee named before the game, initiates one by throwing a red flag onto the field of play before the next legal snap or kick. That timing requirement is absolute: once the next play begins, the previous one is closed forever.

Each challenge requires an available team timeout. A team out of timeouts may not challenge at all. A team that has used all its available challenges may not challenge. And a team that throws the flag anyway when it has exhausted its timeouts is penalised fifteen yards, enforced as a foul between downs.

Fifteen yards is a startling number for a procedural foul, and the reason for it is straightforward. A team with no timeouts remaining has nothing to lose by challenging every marginal ruling, so the rule has to invent something for it to lose.

If a challenge is unsuccessful, the team is charged a timeout. If it is successful, nothing is charged. And a challenge counts as successful if any reviewable aspect of the play is changed, not merely the one the coach was complaining about.

Then the provision that changed the arithmetic of the whole system. A team is permitted a third challenge if it has been successful on at least one of its challenges. A fourth is never permitted.

That sounds like a small adjustment. It replaced a rule under which a third challenge required both of the first two to have succeeded, which was so demanding that it functioned mostly as a piece of trivia. Requiring one success rather than two turned the third challenge from a rarity into something a coach can reasonably plan around, and it changes the calculus on the first flag of the afternoon. Winning a challenge in the first quarter no longer just fixes one play. It buys an extra option for the fourth quarter.

The fixed numbers in the challenge system
  • 2Challenges each team starts with
  • 1Successful challenges needed to earn a third
  • 15Yards for challenging with no timeouts left
  • 10Seconds run off a changed ruling inside two minutes

All four are set by Rule 15. Everything else about a challenge, including whether it is worth throwing the flag, is judgement.

One further restriction catches teams out. A team that commits a foul preventing the next snap can no longer challenge the previous play. The non-fouling team still can, and both teams get the benefit of whatever the review finds. So a false start or a delay of game after a doubtful ruling does not merely cost five yards. It closes the window.

Worked example: challenges and timeouts through one half
  • Timeouts remaining
  • Challenges remaining
Start of the half32
After an unsuccessful challenge21
After a successful challenge21
After a third challenge that fails10

A constructed sequence applying the current rule. A successful challenge costs nothing and unlocks a third; an unsuccessful one costs a timeout. There is no fourth challenge under any circumstances.

Show the numbers
Worked example: challenges and timeouts through one half
ItemTimeouts remainingChallenges remaining
Start of the half32
After an unsuccessful challenge21
After a successful challenge21
After a third challenge that fails10

Read the second and third rows together and the incentive structure is visible. The failed challenge cost a timeout and a challenge. The successful one cost nothing and replaced itself. A coach who is right is playing with house money; a coach who is guessing is spending two resources at once, in a phase of the game where both are about to matter.

The plays a coach is not allowed to challenge

A large proportion of the game is off limits to the red flag, and not because those plays are unreviewable. It is because somebody else already has them.

Only the replay official, or the Senior Vice President of Officiating or a designee, may initiate a review of a play that begins after the two-minute warning of either half, throughout any overtime period, on which points are scored by either team, or that is a Try attempt of any kind, successful or not.

The same booth-only rule applies when the on-field officials rule an interception, a fumble or backward pass recovered by an opponent or going out of bounds through the opponent's end zone, a failure to reach the line to gain on fourth down, possession by the kicking team at the end of a kick, or a player disqualification.

The pattern is obvious once you see it. Every high-stakes category has been taken away from the coaches and given to the booth, because the plays where a challenge matters most are exactly the plays where a coach is most likely to be out of timeouts, or on the wrong sideline to see it, or simply too late.

The corollary matters just as much. Inside two minutes, throwing a red flag achieves nothing except a fifteen-yard problem if the timeouts are gone. The booth is already looking. The coach's job in that window is to shout at somebody upstairs, not at an official.

One further detail sits in the rule and is regularly misunderstood: these plays may be reviewed regardless of whether a foul was committed on the play that, if accepted, would negate the on-field ruling. A flag on the play does not shut the review down.

Once a review starts, the whole play is in scope

This is the provision that turns a challenge from a question into a gamble.

Once a review is initiated, all reviewable aspects of the play may be examined and are subject to change, even if they were not the reason for the challenge. A coach challenging whether a receiver got a foot down has also, without meaning to, opened up whether the ball came loose afterwards, whether the runner was down by contact, where the dead ball spot was, and whether the down was correct.

Alongside it sits a rule that cuts the other way. Any aspect of a ruling that is not changed is treated as a correct ruling for the purposes of reviewing the play. Reviews do not cascade into a general re-officiating of the down. Each element is examined, and whatever is not overturned is now, formally, right.

The standard for changing anything is one sentence long: an on-field ruling will be changed only when the officiating department determines that clear and obvious video evidence warrants a change.

That standard, and not the quality of the officiating, explains the outcome that irritates viewers most. A review that finds genuinely ambiguous evidence produces no change. The referee announces that the ruling stands, which is the convention for "we could not tell", as distinct from confirmed, which is the convention for "we could tell, and it was right". Neither announcement changes anything on the field. The distinction exists purely so that everybody knows which kind of not-changing just happened.

The route a disputed ruling takes

From the whistle to the restart, when a ruling is disputed
  1. The play ends and a ruling exists on the fieldEverything that follows is a question of changing that ruling, never of making a fresh one. The on-field call is the default and it wins every tie.
  2. Is this a booth-only category?Inside two minutes, in overtime, on any score, on a Try, or on a ruled interception, opponent recovery, fourth-down stop, kicking-team possession or disqualification, only the replay official or the league office may start a review. A coach's flag is worthless here and may be expensive.
  3. If not, does the coach throw the flag in time?The red flag must land before the next legal snap or kick, thrown by the head coach or the assistant he named before kick-off. The team must have a timeout available and a challenge left, and must not have just committed a foul that prevented the next snap.
  4. Meanwhile, is the officiating department already helping?Separately from any review, the replay official or the league office may advise the on-field crew on specific objective aspects of a play where clear and obvious video evidence exists. No challenge is spent, no timeout is at risk, and often nobody watching realises a correction has been made.
  5. The review opens, and the whole play comes with itEvery reviewable aspect of the down is now in scope, not only the one that was challenged. Anything not changed is treated as correct.
  6. New York decides, against a clear and obvious standardThe ruling is made by the Senior Vice President of Officiating or a designee, in consultation with the replay official and the referee. An expedited review can happen without consulting the referee at all, and in that case the video is never shown to him on the field.
  7. If the ruling changes, the clock is rebuiltThe game clock is reset to the moment the ball should have been declared dead. If the changed ruling produces a running clock inside two minutes, ten seconds come off unless a team spends a timeout to stop it.
  8. The challenge is settled, and so is the coach's afternoonA successful challenge costs nothing and, if it is his first success, earns him a third. An unsuccessful one costs a timeout he will want back in about eleven minutes.

The three routes into a review, and what happens once one opens. Only one of them costs a team anything, which is why the other two have quietly grown.

What is reviewable, in ten categories

Rule 15 groups the reviewable universe into ten headings, and reading them as a list is more useful than any summary.

Plays involving possession, which covers whether a pass was complete, whether a loose ball was legally recovered, whether a passer's hand was going forward or the ball was fumbled, and whether a runner lost the ball before he was down. Plays involving touching, either of the ball or of the ground, which covers down by contact, whether a pass was touched, whether a kick was touched and where. Plays governed by the goal line. Plays governed by the boundary lines. Plays governed by the line of scrimmage. Plays governed by the line to gain. The number of players on the field. Game administration, meaning penalty enforcement, the proper down, the spot of a foul and the status of the game clock. Disqualification of a player. And a final catch-all of other reviewable plays, including whether a pass was forward or backward, whether a field goal or Try crossed inside the uprights, and whether a loose ball touched a scoreboard or a guide wire.

Two of these repay attention because they behave differently from how people assume.

The line to gain has a rule of its own. The dead ball spot is reviewable to determine whether it was short of, at, or beyond the line to gain. Following the review the ball is placed at the correct spot, but the challenge is successful only if the ruling about whether a new series was awarded actually changes. Move the ball six inches closer to the sticks without moving the chains and the coach has still lost his timeout.

The game clock is reviewable only in one direction. Time can be restored; time cannot be taken off. A clock operator who starts the clock when it should have stayed stopped can be corrected, provided it happens before the next snap. And a ruling that time expired at the end of a half is reviewable only when the video shows the clock should have stopped with two or more seconds remaining, and in the second half only when the next play would be a snap by a team trailing by eight points or less, or by either team with the score level. That last clause is a piece of housekeeping most fans have never heard of and it quietly decides whether a game gets one more play.

What is not reviewable, and the one that still stings

The non-reviewable list is short and every item on it is deliberate. Whether an erroneous whistle sounded. Whether a ball was illegally batted or kicked. Whether a passer intentionally grounded a pass. Whether an ineligible receiver was downfield before a pass. Whether a receiver was illegally contacted. Whether a block was illegal. And, as a sweep, any aspect of a play not listed as reviewable.

Pass interference is not on the reviewable list, which makes it non-reviewable by that last clause.

It was reviewable once, for exactly one season. After a heavily criticised missed defensive interference call in a conference championship game following the 2018 season, the league made both offensive and defensive interference subject to challenge and to booth review for 2019. The experiment was not renewed for 2020, and the reasons given were the ones anybody who has watched a rugby or football review could have predicted: interference is a judgement about whether an opportunity was significantly hindered, and slow motion does not resolve judgement, it only makes contact look worse.

That episode is the single most instructive thing in the history of this system, because it establishes the boundary. Replay in this sport is for facts. It has been tried on judgement, once, formally, and it was withdrawn. Everything the league has built since has stayed on the factual side of that line, which is a discipline the video review systems in other sports have found much harder to hold. The contrast with a system that reviews only geometry, such as cricket's ball-tracking and edge-detection apparatus, is instructive in the other direction: the narrower the question, the better replay performs.

None of which stops the criticism, because the interference foul remains the most expensive penalty in the sport and it is the one thing nobody may look at twice.

Replay assist, and the quiet expansion nobody voted against

The most significant change to this system in years was not a challenge rule. It was the growth of a mechanism that costs teams nothing and that most viewers have never had explained to them.

Replay assist began as a compromise. Proposals for an eighth official in the booth with authority to intervene had been circulating for years and kept failing. What passed instead, from 2021, was a much narrower thing: the replay official or the league office could tell the referee in real time about a clear and obvious mistake on defined objective matters, before anybody had to spend a challenge.

Rule 15 now carries that authority as a full section. The replay official or designated members of the officiating department may provide information on the correct application of the rules, may advise or assist the on-field officials on specific objective aspects of a play where clear and obvious video evidence is present, and may address game administration issues. The listed examples run to a paragraph: penalty enforcement, the proper down, the timing or spot of a foul, the game clock, possession, whether a pass was completed or intercepted or thrown forward, the touching of a loose ball or a line, the location of the ball or a player relative to a line or the pocket, and whether a player was down by contact when he was not ruled down on the field.

Then, from 2025, the part that changed the texture of watching games. Replay assist may cause a flag to be picked up for seven specific fouls, each with its own precise trigger:

Roughing the passer called only for a hit to the head or neck, where the video clearly shows no contact with the head or neck. Intentional grounding, where the video clearly shows the passer was outside the pocket. Unnecessary roughness for hitting a runner out of bounds, where he clearly was not out of bounds. Unnecessary roughness for hitting a defenceless player's head or neck, where there clearly was no such contact. A facemask, where the hand clearly never touched the facemask or helmet opening. A horse-collar tackle, where the hand clearly never contacted the inside collar or the shoulder pads or the jersey at the name plate or above. Tripping, where the leg or foot clearly never touched the opponent. And roughing or running into the kicker, where the defender who contacted the kicker clearly touched the kick in flight.

Read the list carefully and the design becomes clear. Every one of these is a factual precondition rather than a judgement. Not "was that roughing" but "did he touch the head at all". Not "was that a horse collar" but "did the hand reach the collar". The league has not given replay the power to referee. It has given replay the power to check whether the thing the official thought he saw was physically there.

The other half of the design is the direction of travel. Replay assist could pick a flag up. It could not put one down. An official who missed a foul entirely was, and for ordinary fouls still is, beyond correction.

The 2026 change: New York can now put a flag on the field

That one-way rule has a hole in it as of the 2026 season, and it is a genuinely new kind of power.

The replay command centre may now consult with the on-field referee on disqualifications for flagrant football acts and non-football acts, and may cause a flag to be thrown where none was. The competition committee's own description of it acknowledged that this was the first time the league office had ventured into putting a flag on the field rather than removing one.

The limit is important. This is not general foul detection. It is confined to conduct serious enough to eject a player, which is the category where a missed call does the most damage to the sport's credibility and where nobody has ever seriously argued that the on-field crew should have the last word.

Approved alongside it was a one-year provision, drafted with its own expiry built in, allowing the officiating department to correct clear and obvious misses by replacement officials in the event of a work stoppage. It says something about how the league now thinks about this system that its contingency plan for losing its officials is to route more authority through the command centre.

Fouls, the clock, and the things a review can create

A review does not only change what happened. It changes what the penalties for it were.

Penalty administration is expressly reviewable, including the number of the down, the yardage, the number of the fouling player and the spot of a foul. A foul is nullified when a necessary aspect of it is changed in replay. And a foul can be created following a review, in two circumstances: where the reviewable aspect itself creates the foul, or where the referee announced before the review that there was no foul because of a specific ruling that then changes.

That second circumstance is worth understanding, because it is the reason referees sometimes make an oddly specific pre-review announcement about a foul that was not called. They are preserving the option. Without that announcement, a ruling that changes cannot bring a foul into existence with it.

There is also a tidy-up rule for the mess a changed ruling leaves behind. When a ruling changes, any foul that occurred after the ball should have been dead is disregarded, except for personal fouls and unsportsmanlike conduct. Pre-snap fouls that happened before the review, other than those two categories, are ignored if the previous ruling changes. And a head coach may change a decision to accept or decline a penalty that he made before the review, which is only fair, since the facts he based it on have just moved.

Then the clock. When a ruling is changed, the game clock is reset to the time when the ball should have been declared dead. If the ruling is not changed, the clock is not reset at all.

If a changed ruling produces a running clock after the two-minute warning, or a clock that would otherwise have restarted before the next snap, ten seconds run off. The play clock resets to thirty, and both clocks start on the ready-for-play signal. Neither team may decline the runoff. Either team may avoid it by taking a charged timeout, in which case the play clock goes to twenty-five and the game clock starts on the snap instead.

And if fewer than ten seconds remain in the half or the game, the half or the game is over.

Worked example: what a ten-second runoff costs
  • Clock after the reset
  • Clock after the runoff
1:40 remaining100sec90sec
0:35 remaining35sec25sec
0:12 remaining12sec2sec
0:08 remaining8sec0sec

A constructed illustration applying the runoff provision to four invented clock readings. The clock is reset to the moment the play should have ended, then ten seconds come off. Below ten seconds the period simply ends, unless a team spends a timeout to prevent it.

Show the numbers
Worked example: what a ten-second runoff costs
ItemClock after the resetClock after the runoff
1:40 remaining100sec90sec
0:35 remaining35sec25sec
0:12 remaining12sec2sec
0:08 remaining8sec0sec

The bottom row is the one that ends games. A team correctly awarded a ruling in its favour, inside the last ten seconds, can find that winning the review ends the half. This is why the timeout provision exists and why coaches with a timeout in hand and eight seconds on the clock are being told, urgently, by somebody upstairs, exactly what to do with it.

Awarding possession, and the whistle that should not have blown

One of Rule 15's least-known provisions is also one of its most consequential.

When the on-field ruling made the ball dead, a score, a player down by contact, an incomplete pass, and a review then determines that possession was actually lost before the ball should have been dead, possession may be awarded to a player who clearly recovers the loose ball in the immediate continuing action. A loose ball that touches out of bounds counts as a clear recovery by whoever last had it. And the rule states expressly that it supersedes the ordinary treatment of an inadvertent whistle.

That last clause is doing an enormous amount of work. Under the general rules, a whistle blown by mistake kills the play and the ball goes back. Rule 15 says that in this specific situation, it does not, and the recovery counts. Even a preliminary on-field indication of which team recovered can constitute a clear recovery for these purposes.

There is a matching limit on the other side, and it is the wrinkle that most often produces a team winning a review and gaining nothing. Where a ruling of incomplete is changed to a catch, the ball is placed where it was when the receiver's second foot or other body part came down with control, and no advance is given for anything he did afterwards. The same principle governs several other reversals. Replay corrects the ruling; it does not simulate the football that did not get played. Anyone reading the completed pass rule in detail will find that this single sentence explains most of what looks like unfairness in reversed incompletions.

When the phone line dies

There is a rule for what happens if the technology fails, and it is short enough to enjoy.

If communications between the officials and the league office are lost, the referee stays at the field-level monitor for one minute while they are restored. If they are not restored within that minute, the referee conducts the review himself in consultation with the replay official. If the connection comes back during the review, the league office takes over as normal.

Sixty seconds of waiting, and then the system reverts to what it used to be. It is the only place in Rule 15 where the referee gets to be the decision-maker, and it exists purely as a backstop.

A system that has been rebuilt four times

The current arrangement did not evolve smoothly. It was demolished and rebuilt.

A limited replay system arrived in 1986 and remained in place through 1991. It was unpopular, slow, and run entirely from the booth with no involvement from the teams, and it was abandoned.

Replay returned in 1999 in the form that is still recognisable: a coach's challenge, a red flag, a timeout at risk, and a referee under a hood at the side of the field. From 2004, a third challenge became available if both of the first two had succeeded.

From 2014, reviews were overseen centrally at league headquarters, where the command centre the league calls Art McNally GameDay Central watches every game from the preseason to the Super Bowl. The league's published description of it lists sixteen stations with 58-inch touch-screen monitors running synchronised camera angles, and three decision-makers able to follow as many as twelve games at once on 90-inch displays. The hood at the side of the field survived as an image long after it stopped being where the decision was made.

From 2022 the expedited review appeared, allowing the replay official or the league office to correct an obvious error through the referee's headset. From 2021 replay assist began, from 2024 the third challenge became available on one success rather than two, from 2025 the flag pick-up list arrived, and from 2026 the command centre can put a flag down for a disqualification.

Read as a sequence, the direction is unmistakable. Authority has moved steadily away from the sideline and away from the field, towards a room in New York, and the coach's challenge has been quietly demoted from the main mechanism to the residual one. It now covers the cases the booth is not already watching, which is a smaller share of the game every few years.

Whether that is progress depends on what you think replay is for. It has made the system faster and more consistent, and it has taken a decision that used to be visibly made by a man everyone could see and moved it somewhere nobody can. The referee still announces it. He is now, on most reviews, the messenger.

When to throw the flag

For anyone actually making the decision, or arguing with somebody who is, the rule reduces to four questions.

Is this play in a booth-only category? Inside two minutes, in overtime, on a score, on a Try, on a ruled turnover, on a fourth-down stop. If yes, put the flag away. Somebody is already looking, and throwing it can only cost you.

Is the evidence clear and obvious, or merely favourable? A ruling that looks wrong at full speed and ambiguous in slow motion will not be changed, and the timeout is gone. The question is never whether you are right. It is whether the video proves it.

Does changing the ruling change the outcome that matters? Moving the ball without moving the chains does not count as a successful challenge on a line-to-gain review, and a spot that improves without changing the down has cost a timeout for nothing.

And have you already won one? If a challenge has succeeded, a third is available and the downside of a marginal second one is smaller than it looks. If not, the flag you are holding is one of two, and both of them are timeouts.

Get those four right and you will out-manage most sidelines, which is a strange thing to be able to say about a rule that exists to remove human error. The rest of the professional game's rules and how they interact rewards the same kind of reading: the interesting decisions are almost never where the broadcast is pointing.

Common questions

How many challenges does an NFL coach get?

Each team starts with two, and each one requires an available timeout. A third is granted if the team has been successful on at least one of its first two, and no fourth challenge exists under any circumstances. Before 2024 the third challenge required both of the first two to have succeeded, which made it close to theoretical.

What happens if a coach challenges with no timeouts left?

The challenge is not permitted and the team is penalised fifteen yards, enforced as a foul between downs. This is one of the heaviest procedural penalties in the rulebook, and it exists because a team with no timeouts left would otherwise have nothing at all to lose by throwing the flag on every doubtful ruling.

Can pass interference be challenged in the NFL?

No. Pass interference was made reviewable for the 2019 season only, following a heavily criticised missed call in a conference championship game, and the provision was not renewed. The current rulebook lists whether a receiver was illegally contacted among the non-reviewable items, and interference does not appear anywhere in the reviewable list.

What is replay assist, and how is it different from a review?

Replay assist is the officiating department advising the on-field crew on objective aspects of a play where clear and obvious video evidence exists, without any review being formally initiated and without any team spending a challenge. It can pick up a flag that has already been thrown for a defined list of fouls, and since 2026 it can also cause a flag to be thrown for a disqualification. It cannot manufacture ordinary fouls that were never called.

Why do so many replay reviews end with the ruling standing?

Because the standard is not whether the ruling was right but whether clear and obvious video evidence warrants changing it. A review that finds the evidence genuinely ambiguous produces no change, which is announced as the ruling standing rather than being confirmed. The distinction between confirmed and stands is a matter of announcement convention rather than a difference in outcome.

Filed under American Football·nfl · rules · officiating · replay · penalties