NFL overtime rules explained, and why the toss still matters

Overtime in the NFL starts with a coin. The referee flips it, a captain calls it in the air, and the winner decides whether to begin the extra period with the ball. Every revision the league has made to its overtime rules in fifty years has been an attempt to stop that flip from deciding the game.

Sit with the problem before the mechanics, because the mechanics only make sense as answers to it. In most sports possession is a shared, shifting thing that changes hands dozens of times a minute. In American football it is a discrete asset. A team that takes the ball at its own twenty-five gets to run a private sequence of plays that the opposition cannot interrupt except by force, and if that sequence works, the team can travel forty yards and kick a field goal without the other offence ever coming off the bench.

That is what makes overtime hard to legislate. Handing one team the ball is not a small procedural courtesy. It is handing them a scoring opportunity that the other side may never get, and for thirty-six years the league did exactly that and called it fair.

Possession is the prize, and that is the whole problem

Think about what a single drive is worth. A competent offence starting in its own territory does not need a touchdown to win a sudden-death period. It needs somewhere in the region of forty yards, because a modern placekicker converts from distances that would have been comical in the 1970s. Four first downs and the game is over.

The kicking game made this worse in a way that crept up on everybody. As kickers got better, the field goal range that ends a game moved further and further back, which shortened the drive a team needed and lengthened the shadow the coin toss cast over the result. A rule written when a fifty-yard attempt was a lottery ticket was still in force when it had become a routine expectation. Rules do not have to change to stop working. The sport can move underneath them.

There is a second asymmetry underneath the first, and it is the one the current rules are built around. Football is a game of sequential, fully observable decisions. The team that goes second sees the whole of the first team's possession before it makes a single call of its own. It knows whether it needs three points or seven, whether fourth and two at the opposition thirty-eight is a punt or a conversion attempt, and whether the clock is a friend or an enemy. Information of that quality has real value, and until recently the rules made sure nobody could use it, because the game usually ended before the second team got the chance.

So the league has spent five decades trading off two things that pull in opposite directions. Give both teams the ball and you neutralise the toss, but you also lengthen games in a sport that already asks a great deal of the bodies playing it, and you hand a different kind of edge to whoever goes second. Keep it short and simple and you leave the outcome partly to a coin. Every version of the rule is a different answer to the same question, and none of them is clean.

• 10Minutes in a regular-season overtime period
• 15Minutes in a postseason overtime period
• 2Timeouts per team in regular-season overtime
• 24Club votes needed to pass a rule change

Fixed by Rule 16 and by the league's rule-change procedure rather than set annually. The postseason plays as many periods as it takes, so it has no maximum.

Sudden death, and why it lasted thirty-six years

The original rule was one sentence long. Play on, and the first score of any kind wins. No possession guarantee, no qualifications, no distinction between a field goal and a touchdown. It arrived in the championship game long before the regular season had any provision for a tie at all, and the regular season adopted it in 1974.

Sudden death had two enormous virtues, and it is worth being honest about them, because the modern rules gave up both.

It was instantly comprehensible. A child watching for the first time understood the stakes of every single snap without anybody explaining anything, and the tension of a sudden-death drive is a genuinely different quality of tension from the tension of a fourth quarter. Nothing since has matched it.

It was also short. One drive, one result, everybody goes home. A sport in which the participants collide with each other for three hours has a real interest in not adding a fourth, and the injury argument for keeping overtime brief is not a fig leaf. It is the strongest defence sudden death ever had.

What killed it was the growing gap between how the rule was designed and how the sport actually played. In the era it was written for, a team receiving the kickoff had to earn a touchdown or a shortish field goal to end it, and the defence had a real chance. As field position and kicking improved, the first possession became decisive far more often than the rule's authors had in mind, and the coin toss started to look like what it was: a fifty-fifty draw for a materially unequal prize.

The league's own response tells you it agreed. In the 1990s it moved the kickoff spot, in part to change the field position that a receiving team started an overtime drive from. That is not the behaviour of an organisation confident in its rule. It is the behaviour of one trying to patch a rule without admitting the patch is needed.

The first fix: a field goal stops being enough

The change that came in 2010 for the postseason, and in 2012 for the regular season, is the most elegant thing the league has done in this area, and it is the one people find hardest to explain from memory.

The rule said the game ends on the first possession only if that possession produces a touchdown. A field goal on the opening drive did not end it. The other team got the ball, and if it matched the field goal the period continued as sudden death, where anything wins.

The logic is precise. Two objections had been made to sudden death: that the first drive too often ended the game, and that it did so on the strength of a kick from distance rather than a drive into the end zone. The 2010 rule answered both at once without adding a guaranteed second possession, because it targeted the specific outcome that had become too easy rather than the structure around it. A team that wanted to end the game on its first possession still could, but the price went up from forty yards to the whole field.

It also created something the old rule never had, which was a real decision. A team facing fourth and three at the opposition twenty-five in the old format kicked without thinking. Under the modified rule, kicking meant handing the ball back to an opponent who now knew precisely what was required, so the fourth-down call became a genuine problem with two defensible answers. Football only got interesting there because the rule stopped making the choice for the coach.

What it did not solve was the case everybody eventually complained about. If the first drive ended in a touchdown, the second team still never touched the ball. And touchdowns on opening overtime drives are not exotic. The fix had narrowed the unfairness without removing it, which is a normal outcome for a compromise and an unstable one, because the remaining cases are always the loudest.

The second fix: both teams get the ball

The postseason moved first, in 2022, to a rule guaranteeing each team an opportunity to possess the ball regardless of what happened on the first possession. A touchdown on the opening drive no longer ends a playoff game. The scoring team kicks off and the other team gets its turn.

The regular season stayed where it was for three more seasons, which produced the odd situation of a league running two different overtime rules depending on the month. That ended in 2025, when the clubs approved a proposal put forward by Philadelphia to align the two, and the possession guarantee became universal. The regular season kept its ten-minute period; the postseason kept its fifteen-minute periods and its refusal to end in a tie.

Follow the sequence and you can see the league working through the problem in stages, each one answering the failure the last one left behind.

Pure sudden death answered nothing except the need to produce a winner, and it left the toss too powerful. The 2010 rule answered the cheapness of the winning field goal and left the touchdown case open. The 2022 rule answered the touchdown case in the games that mattered most and left the regular season inconsistent with the postseason. The 2025 alignment answered the inconsistency. Each fix was aimed at the most visible remaining complaint, which is how sports rules are almost always written, and it is why they end up layered rather than designed.

Note what has not changed at any point. The coin toss survived all four versions. Nobody has removed it, and the reason is that no proposed replacement has been able to command three-quarters of the room, which is the bar a rule change has to clear. Twenty-four of the thirty-two clubs have to agree, and a coin has the singular advantage of being a mechanism nobody can accuse of favouritism.

How NFL overtime rules work now, from the toss onwards

Regulation ends level. There is a short intermission, and the referee conducts a coin toss with the visiting captain calling it.

The winner of the toss chooses one of three things: take the ball, kick off, or pick the goal to defend. In the postseason, where more than one period may be played, the choice can be deferred. Those options are the same ones available at the start of a game, which is deliberate: the league treats the extra period as a fresh miniature contest rather than a continuation of the fourth quarter.

Then the period runs, and it has three phases.

The first phase is the receiving team's guaranteed possession. It ends when that team scores, punts, turns the ball over, or fails on fourth down. If it scores a touchdown, the extra point is attempted, because the margin still matters. The game does not end there.

The second phase is the other team's guaranteed possession, played with full knowledge of what the first phase produced. This is where the informational asymmetry lives, and it deserves its own section below.

The third phase is sudden death in its original form. Once both teams have had their opportunity, the next score of any kind wins, including a field goal, including a defensive touchdown. The rule reverts to what it was in 1974 and stays there until the period expires or somebody scores.

Two exceptions cut across all of this, and they are the parts that catch people out. A safety ends the game whenever it is scored. If the team that kicked off tackles a ball carrier in his own end zone on that first possession, it wins by two points with its own offence still standing on the sideline, and the rule says so explicitly. A defensive touchdown on the first possession ends it too, because the change of possession has already satisfied the guarantee for both sides, and the score arrives after it.

1. 1The coin is tossedThe visiting captain calls it, and the winner takes the ball, kicks off, or picks an end to defend.
2. 2The first team's guaranteed possessionIt ends on a score, a punt, a turnover or a failed fourth down. A touchdown here no longer ends the game, so the extra point is still attempted.
3. 3A safety or a defensive score ends it hereThose are the two exceptions. A safety always wins the game, and a turnover returned for a touchdown finishes it because both teams have then had the ball.
4. 4The second team's guaranteed possessionPlayed knowing exactly what is needed. Trailing by seven means touchdown or nothing, and every fourth down is priced against that.
5. 5Level after both possessions, and it becomes sudden deathThe next score of any kind wins. A field goal is enough again, and so is a defensive touchdown.
6. 6Ahead after both possessions, and the game is overThe moment the trailing team's possession ends without equalising, the result stands. No extra point is attempted on a touchdown that wins it.
7. 7The ten minutes expireWhoever is ahead wins. If the score is level the game is a tie, and if the first drive consumed the whole period the second team simply never got its turn.

The sequence is the same in the postseason except for the period length, the timeout allowance and the fact that a level score at the end of a period leads to another one instead of a tie.

The clock deserves emphasis, because it is the part of the rule most people assume works differently from how it does. In the regular season the possession guarantee is an opportunity, not a right. Ten minutes is the hard boundary, and a first drive that eats all ten minutes ends the game without the opposition taking a snap on offence. That is not a loophole. It is the rule doing what the rule says, and the drive that achieves it has earned the win on any reasonable view.

Inside the period, timing follows the same logic as the end of a half. From the five-minute mark the timing rules that govern the closing five minutes of the fourth quarter apply, and there is a two-minute warning. In the postseason the two-minute warning falls near the end of the second overtime period rather than the first, because the two periods are treated as a half. That single detail explains a lot of apparently arbitrary clock behaviour in long playoff games, and the exact wording of it lives in the published rulebook rather than in anybody's memory.

Each team has two timeouts in regular-season overtime, three per half in the postseason. Coaches' challenges are gone. Every review in overtime is initiated from the booth by the replay official, on the same principle that governs the final minutes of a half, and the reason is straightforward: a challenge system depends on a coach having timeouts to risk, and the risk is not comparable when the game can end on the next snap.

Why NFL overtime rules still allow a tie in the regular season

A tie is a strange thing for an American sport to permit, and the instinct is to read it as a leftover. It is not. It is a deliberate choice, and the argument for it is stronger than it looks.

Start with what a tie costs. In the standings it counts as half a win and half a loss, so it is not a void result. It moves a club's winning percentage, it feeds into the tiebreakers that decide playoff seeding, and it feeds into the strength-of-schedule calculation that helps set the order clubs pick in during the draft. A tie is a real outcome with real consequences that the season absorbs without difficulty.

Now consider what abolishing it would cost. A seventeen-game regular season is played by rosters that are, at any given moment, carrying more injuries than the injury report suggests, and every additional period is more snaps at full speed for players who will be asked to do it again in six days. The league has spent two decades legislating in the direction of fewer hits, and adding open-ended overtime to a September fixture would run directly against all of it. Ten minutes and a tie is a rationing decision about player workload dressed up as a competitive rule.

There is a fairness argument too. Regular-season games are individually less decisive than playoff games, because the season has seventeen of them and a single tie almost never settles anything on its own. Spending an extra half-hour of football to convert an outcome that the standings can already represent is a poor trade. In the playoffs the calculus inverts completely: one team must advance, there is no next week for the loser, and the sport accepts whatever length is required.

That is why the two sets of rules differ, and the difference is now narrower than at any point in decades. The mechanism is identical. Only the clock and the consequence of the clock running out have been left apart, because those are the two things the regular season and the postseason genuinely need to treat differently.

The team that goes second knows things the first team could not

This is the part almost nobody explains properly, and it is the reason the strategic layer of overtime changed completely when the possession guarantee arrived.

The first team plays blind. It is trying to score without knowing what score will be sufficient, which forces it to play a generic optimum: take the points that are available, protect the ball, do not do anything exotic. It has no target, only a preference for more.

The second team plays with the answer sheet. Every decision it faces has a defined objective, and defined objectives change the arithmetic of the whole drive.

If the first team kicked a field goal, the second team knows a field goal ties and a touchdown wins outright. That makes the drive one of the clearest strategic problems in the sport. Fourth and one at the opposition thirty-five stops being a matter of taste, because the team can weigh a conversion attempt that keeps a touchdown alive against a kick that ties and hands the ball back into sudden death. Both are respectable. The point is that the team can actually calculate them, and the first team could not.

If the first team scored a touchdown, the second team's decision tree collapses to a single branch. A field goal is worth nothing. Fourth and eight from the opposition twenty-nine is not a kick, it is a play. Every fourth down inside the opposition half is an attempt, punting is close to pointless, and the whole drive is played at a level of aggression that the first team could not justify. That is simpler, and simpler is not the same as easier. The second team is playing a harder task with better information.

If the first team punted or turned the ball over without scoring, the second team is in the strongest position available anywhere in the structure. Any score wins the game outright, because the moment its possession ends with a lead, the game is over. It has the coin toss loser's consolation prize and it is not a small one.

1. 1The first team puntedAny score wins. A field goal ends the game the moment the drive finishes, so the second team plays for position and takes the points.
2. 2The first team kicked a field goalThree ties it and seven wins it. Fourth and short in kicking range becomes a genuine choice between extending the game and ending it.
3. 3The first team scored a touchdownOnly a touchdown matters. Field goal range is irrelevant, punting is close to pointless, and every fourth down inside the opposition half is an attempt.
4. 4The touchdown is scored, and now the tryKick the extra point to tie and go to sudden death, or attempt two points to win the game on one snap and lose it on a failure.
5. 5The second team fails to answerThe game ends the instant the possession does. There is no clock to run out and no further phase to reach.
6. 6The second team equalisesSudden death begins with the first team due to receive next, which is the strongest argument for having taken the ball in the first place.

A constructed illustration of the decision the possession guarantee creates. No real drive, game or probability is being described.

Once the second team's advantage is stated plainly, the coin toss looks different. The winner of the toss is not choosing between an advantage and nothing. It is choosing between two different advantages, which is exactly what a rule designed to neutralise a coin flip should produce.

Receive or kick, and why the right answer moved

Under pure sudden death the decision was not a decision. Take the ball. A team that kicked off was volunteering to lose without touching the ball, and no amount of confidence in a defence made that arithmetic work.

Under the 2010 rule it was still nearly automatic. A touchdown ended the game, so the receiving team held both the first chance to end it outright and the option of playing for the touchdown rather than the field goal.

The guarantee changed the shape of the question. A team that kicks off now has a defence on the field that can end the game with a takeaway returned for a score or a safety, and if it makes a stop, its own offence takes over knowing precisely what is required. That is a coherent case, and it rests on three things that are all real.

Information, as above. Field position, which is genuinely underrated: a stop followed by a punt often produces better starting position than a kickoff return, and in bad weather the difference grows. And the ability to win without your offence doing anything at all, which is a small probability attached to an outcome of infinite value.

Against that sits the strongest argument for taking the ball, and it is not the obvious one. It is the third possession. If both teams score equally in their guaranteed possessions, the period reverts to sudden death, and the team that received first is the team due to have the ball. Taking the ball first is therefore taking the ball twice, in every sequence where the game does not resolve early. Anyone weighing the toss decision and ignoring the sudden-death phase is solving half the problem.

The clock modifies this in the regular season, and modifies it in a direction people get backwards. A ten-minute period makes the third possession less likely to happen at all, because two full drives can consume most of it. That erodes some of the value of receiving first, which is why the regular-season and postseason answers to the same question are not necessarily identical even now that the possession rule is identical.

What has not changed is the behaviour. Coaches take the ball. The theoretical case for kicking has existed since 2022 and has been acted on rarely, for reasons that have very little to do with analytics and everything to do with accountability. A coach who takes the ball and loses has followed the convention. A coach who kicks and loses has invented a way to lose, and will be asked about it for a fortnight. That asymmetry is not irrational on the coach's part. It is simply a different objective function from the one the win probability model is solving.

The two-point conversion that only exists because of the rule

Here is a decision the old rules could not produce, and it is the sharpest thing in the modern format.

The first team scores a touchdown and kicks the extra point, leading by seven. The second team drives the length of the field and scores a touchdown of its own. It is now behind by one point with the try to come, and it has two choices.

Kick the extra point, tie the game, and go to sudden death with the opposition due to receive. Or attempt a two-point conversion from close range and win the game immediately on a single snap, with a failure ending the game just as immediately.

That is one play for the whole result, and it is a real decision rather than a stunt. The case for going for two rests on the alternative being worse than it looks: tying the game does not get you back to a coin toss, it gets you to sudden death with the other team holding the ball, and the value of that position to them is not far off the value of the whole game. The case against is that a two-point attempt from short range is a single, well-defended snap against a defence that knows what is coming, while sudden death still contains a stop, a punt, and everything that can happen afterwards.

There is a mirror image of the same decision at the other end, and it is more subtle. The first team, scoring a touchdown on its guaranteed possession, could in principle attempt two points itself to lead by eight and force the opposition to convert twice. Analysts have generally been unimpressed by that idea, because the first team would be risking a point it already has in order to complicate a task the other team has not yet started, and it is doing so with less information than anybody else in the stadium. The team with the answer sheet should take the risk. The team without it should bank the point.

The rule also removes one attempt entirely. If a touchdown wins the game outright, the try is not attempted, because there is nothing left for it to affect. Players who have spent a career kicking that extra point simply walk off the field.

Playoff overtime, and the game that will not end

Postseason overtime is the same machine with the safety catch removed.

The periods are fifteen minutes rather than ten. The possession guarantee is honoured even if the first period expires before the second team's drive is complete, because play carries into the next period rather than stopping. There is a two-minute intermission between periods, teams change ends, and there is no half-time break of the ordinary kind. Each team has three timeouts per half, and the first two overtime periods are treated as a half for that purpose and for the two-minute warning.

If the score is still level after two periods, the game keeps going. The rulebook sets out how the choice is allocated at the start of the third period, with the captain who lost the original toss getting the choice rather than the referee flipping again, unless the winner of the original toss deferred. This is one of the details worth reading in the published rulebook rather than trusting to memory, because it has moved more than once and because it almost never comes up.

What is worth understanding is the design principle. The league has decided that a playoff game will be settled by football rather than by a tiebreaker of any kind, and it has accepted the consequences, including the possibility of a game that runs deep into a fourth hour. There is no shootout, no penalty competition, no reduction in players on the field, no reversion to a scoring contest of the kind other sports adopt when they get tired. Ice hockey takes exactly the opposite view in the regular season and then agrees with football in the playoffs, which is why the way a playoff series is actually decided on the ice is a useful comparison rather than a distant one.

The cost of that principle falls on the players, and it is not hypothetical. Snaps accumulate, and the eventual winner then plays again the following week against a team that may have finished its own game in regulation. The league has never pretended this is fair. It has decided that it is more acceptable than the alternative, which is deciding a playoff game by something that is not football.

The alternatives the league keeps taking seriously

Four alternatives come up every time this is discussed, and each has been proposed formally at some point. None has reached twenty-four votes. It is worth taking each on its merits rather than dismissing them, because the reasons they fail are more interesting than the proposals.

Alternating possessions, as in college. Each team gets the ball at a fixed spot, usually the opposition twenty-five, with no clock and no kickoffs, and the teams alternate until one of them outscores the other in a matched pair of possessions. Later periods require two-point conversions after a touchdown, and later still the whole thing reduces to alternating two-point plays. It has a great deal to recommend it. It is perfectly symmetrical, it is short, and it produces an obvious winner. Anybody who follows the expanded college postseason and how its bracket resolves has seen it work under real pressure.

The objection is that it is not the same sport. Field position is one of professional football's central variables, and an overtime that begins every possession in field goal range deletes it. There is no punting, no kickoff coverage, no defence pinning an offence at its own eight, no drive of eighty yards. Special teams, a third of the game, stop existing. The professional game has consistently declined to abolish a third of itself in order to settle a tie, and that is a defensible position rather than a stubborn one.

Spot and choose. One team names a yard line, and the other team chooses whether to take the ball at that spot or to play defence and give the ball away there. Baltimore put a version of this to the clubs. It is the divide-and-choose principle from cake-cutting applied to a football field, and its theoretical properties are lovely: the team naming the spot is forced to name one it would genuinely be content with on either side of the choice, so the advantage of choosing first is self-limiting.

It failed for two reasons that both matter. It is difficult to explain to a stadium, and a rule that a casual viewer cannot follow in real time carries a cost the elegance does not cover. And it introduces a strategic decision of a type football does not otherwise contain, made by a coach with a headset in about fifteen seconds, in which a small misjudgement of the equilibrium spot becomes a visible and enduring error. Coaches were not enthusiastic about being handed a game theory problem in front of eighty thousand people.

A fixed extra period played out in full. Play the whole ten or fifteen minutes with no sudden death at all, exactly like a fifth quarter, and whoever leads at the end wins. This is what association football does with extra time, and it has the great virtue of being completely symmetrical: both teams have the same clock and the same opportunities, and the coin toss decides only who starts with the ball in a period long enough for that to matter very little.

The objection is length, and it is the same objection that constrains everything else. A guaranteed full extra period is longer than the current format's typical outcome, and it adds the most snaps of any proposal on the table. There is also a quieter competitive objection: a fixed period rewards the team with the ball last in a way a regulation quarter does not, because the final drive is played with perfect knowledge of what is needed, which reproduces the asymmetry the rule was meant to remove.

Abolishing overtime in the regular season. Play seventeen games, let ties be ties, and save the extra football for the playoffs. Nobody proposes this seriously in the United States, but it is the norm in most of the world's team sports and it deserves a place on the list. It is the only proposal that reduces player workload rather than increasing it, and its defeat is a commercial fact rather than a competitive one. A tie is a poor television product, and the current rule is already a compromise in that direction.

How other sports settle a tie

Every code has faced the same problem and answered it differently, and the answers are more revealing side by side than any one of them is alone.

Association football plays thirty minutes of extra time in two halves, then decides on kicks from the penalty mark, having abandoned both the golden goal and the silver goal after finding that a sudden-death rule made teams play more cautiously rather than less. Ice hockey plays a short overtime with fewer skaters, which manufactures scoring chances rather than waiting for them, then goes to a shootout in the regular season and to full sudden-death periods in the playoffs. Basketball plays repeated five-minute periods until somebody is ahead, which works because a basketball possession is short and the sport's scoring rate makes a level score at the end of a period genuinely rare.

Tennis went the furthest of any sport, replacing an unbounded contest with an invented mechanism designed to end it, and the reasoning behind the tiebreak and the scoring system that made it necessary is the closest analogue anywhere to what the NFL has been doing since 1974. Boxing does not have overtime at all in the ordinary case, and a drawn scorecard stands, which is only tolerable because the way a bout is actually scored is built to produce a winner in almost every round.

NFL regular season10min

NFL postseason period15min

NHL regular season5min

NHL playoff period20min

Soccer extra time30min

Rule-defined period lengths only. College football overtime has no clock at all, so it cannot be shown here.

Show the numbers

How long the extra period runs, by code

Item	Minutes in the extra period
NFL regular season	10min
NFL postseason period	15min
NHL regular season	5min
NHL playoff period	20min
Soccer extra time	30min

Two patterns run through all of it. Sports that permit ties in league play do so because the season is long enough to absorb them, and sports with short seasons and heavy physical demands ration extra time hardest. And every sport eventually distinguishes between a game that must produce a winner and one that need not, which is precisely the line the NFL draws between January and September.

What to watch when the referee flips the coin

The overtime rule is short enough to hold in your head, and holding it there changes what you see. Four things are worth watching for, and all of them are visible from the sofa.

Watch the choice, not the flip. The interesting information is what the winning captain does with the toss, not what the coin did. A team electing to kick is telling you something real about how it rates its defence, the weather, and the value of information, and it is doing so knowing it will be asked to justify the decision if it loses.

Watch the first team's fourth downs. The team playing blind has no target, and the tell is whether it behaves as though it does. A first-possession team that turns down a makeable field goal is playing for the touchdown that puts the pressure on, which is a real strategy and a risky one.

Watch the second team's field goal range, then ignore it. If the first team scored a touchdown, the yellow lines the broadcast draws for kicking distance are decoration. Nothing inside forty yards matters, only the end zone, and the drive should look nothing like a normal drive.

Watch the clock against the possessions. In the regular season, ten minutes is the boundary that outranks everything else, including the possession guarantee. Once the second team's drive begins with under two minutes showing, the third phase of the period is not going to happen, and both teams already know it.

Read the period that way and the coin toss goes back to being what it should always have been, which is the first decision of the extra period rather than the last decision of the game. That is the whole point of fifty years of revision, and if it takes a rulebook section, three phases and two exceptions to get there, it is still a better answer than a coin. Understanding it properly is part of understanding how the professional game is put together, where nearly every rule that looks arbitrary turns out to be an old argument somebody finally won.