Baseball extra innings rules, and the runner nobody put there

The ninth inning ends with the score level. Both teams jog off. When they come back out there is a man standing on second base who has not batted, has not run, and did nothing whatsoever to get there.

That is the centre of the modern baseball extra innings rules, and it remains, several seasons in, the single most divisive thing in the sport's rulebook. Not the pitch timer, which almost everybody now likes. Not the limits on defensive positioning. The runner.

He also does not exist in October, which is the part most people have not registered.

What changes at the start of the tenth

Nine innings is a regulation game. If the score is level once the visiting team has batted nine times and the home team has had its nine, play continues in complete innings until one club is ahead at the end of one. There is no clock, no draw, and no upper limit.

From the tenth inning onwards, in the regular season, each half-inning begins with a runner already on second base.

The identity of that runner is fixed by rule rather than chosen. He is the player in the batting order immediately preceding the leadoff hitter for that half-inning. Almost always that is the man who made the final out of the previous inning, which is why the shorthand caught on, but the shorthand is wrong often enough to matter. If the previous inning ended with a runner caught stealing, the batter who was at the plate leads off the new inning, and the automatic runner is the man in front of him, not the runner who was thrown out.

A club may put a pinch-runner in his place. That costs a bench player for the rest of the game, permanently, and in a game that might run four more innings it is a real decision rather than a free upgrade. Managers who spend a pinch-runner in the tenth and then need a catcher in the fourteenth have made a mistake they cannot undo.

Once the inning starts he is a completely ordinary runner. He can steal third. He can be picked off. He can be doubled off on a line drive. He can be forced out at third on a ground ball, which is the most demoralising way an extra inning has ever begun.

• 2Base the automatic runner starts on
• 10First inning in which he appears
• 0Automatic runners in a postseason inning
• 1Ball-strike challenges restored per extra inning

Rule-defined values. The postseason figure is zero because the automatic runner provision was adopted for the regular season only; October extra innings begin with the bases empty.

Why the regular season and the postseason play different games

This is the thing to understand before anything else, because it tells you what the rule is for.

The automatic runner arrived in the shortened 2020 season as a health and safety measure. The schedule was compressed, the rosters were unstable, and a fifteen-inning game in July was a threat to a pitching staff that had no depth to absorb it. It was a blunt instrument for a specific emergency.

It did not go away when the emergency did. The joint competition committee, on which both the clubs and the players are represented, made it a permanent part of the regular season rulebook before the 2023 season.

The stated case is workload. A handful of very long games each season does disproportionate damage: a bullpen emptied in the fourteenth inning is a bullpen that is short for the next three days, and a club that has burned six relievers on a Tuesday is fielding a worse team on Wednesday and Thursday. Across a hundred and sixty-two games, the cost of marathon games is not paid by the two clubs in the marathon. It is paid by their opponents in the following week, who face a depleted pitching staff through no doing of their own.

None of that argument survives contact with the postseason. There is no next week to protect. Rosters are reset between rounds, off days are built into the schedule, and every pitcher on the staff is available because there is no point saving anybody. The workload case simply evaporates, and with it the justification for putting a run on the board that nobody earned.

There is a second reason, less often stated and probably more decisive. The automatic runner is a device for ending games quickly, and it works, but it works by increasing the influence of chance in the innings that decide the result. That is an acceptable trade over a hundred and sixty-two games, where the standings are supposed to reflect the season rather than any one night. It is not an acceptable trade in an elimination game, and the sport has been clear that the games it most cares about should be settled the old way. Anyone reading the shape of the playoff bracket is looking at a competition that plays by a visibly different rulebook once it starts.

So the sport runs two versions of its own game, and the switch happens on the day the postseason begins.

1. 1The ninth inning ends levelBoth halves complete, the score tied. Nothing about a regulation game has failed; the game simply has no result yet and the rules now change.
2. 2The batting order is read backwards by oneThe automatic runner is the player immediately before the man due to lead off. This is a position in the order, not a description of who made an out, and the two come apart whenever an inning ends on the bases.
3. 3The manager decides whether to substituteA pinch-runner may be sent out in his place. He is now in the game for good, occupying that spot in the order, and the bench is one player shorter for however long the game runs.
4. 4He takes his place at second baseNo error is charged, no fielder is credited with anything, and nothing appears in the play-by-play except his presence. The pitcher inherits a runner in scoring position having thrown nothing.
5. 5The scorer marks him as having reached on an errorA bookkeeping fiction with no error attached. It exists purely so that the run he might score does not land on the pitcher's earned run average, and it is the only place in the rulebook where an error is imagined rather than observed.
6. 6The infield is positionedWith nobody out and a runner on second, the defence has to decide immediately whether it is defending the single run or playing for outs. That decision is made before the first pitch of the inning and it usually settles how the inning goes.
7. 7The first pitch is thrownFrom this moment everything is ordinary baseball. The runner may be picked off, may steal, may be forced out, and the pitcher's disengagement allowance and the balk rules apply exactly as they would to a runner who had earned the base.
8. 8The half-inning ends and it all happens againBoth clubs get the same start in the same inning. The rule is symmetrical, which is the strongest thing anyone can say in its favour.

The sequence as the rules set it out. Every step is mechanical except the pinch-runner decision, which is the only genuine choice in the list and is the one that most often goes wrong later in the game.

The scoring fiction, and what it does to a pitcher

The bookkeeping deserves a moment because it is genuinely strange.

The rules require the scorer to treat the automatic runner as though he reached second base on a fielding error. No error is charged. No fielder's record is touched. The error is imagined solely so that the earned-run machinery has something to work with, because earned-run accounting was built on the principle that a pitcher should not be charged for runs that reach base through his fielders' failures, and this runner did not reach base through anything at all.

The consequence is that a pitcher who allows the automatic runner to score has an unearned run against him and no visible mark on his line. A pitcher who gives up a single to the next hitter, and watches the run come home, has surrendered a run that decided the game and will not appear in his earned run average.

Runs scored by anyone else are treated normally. A batter who reaches on his own and comes round to score is an earned run in the ordinary way. Only the placed runner carries the exemption, and it travels with him rather than with the inning.

Two things follow. The first is that earned run average, which was already an imperfect statistic, now systematically understates what happens in extra innings, and any comparison of a modern reliever's figure with one from before 2020 is comparing two slightly different quantities. Anyone building a single value number for a pitcher has to decide what to do about this, and the different public versions do not all decide the same way.

The second is more human. Relievers know the run does not count against them. Whether that changes how anybody pitches is unmeasurable and probably small, but it is a genuinely odd feature of a competitive sport that the most consequential run of the night is officially nobody's fault.

The choices a manager faces, and why most of them are bad

A runner on second with nobody out is one of the most valuable situations in baseball, and the rule hands it to both sides. What it does not hand over is a good way to use it.

1. 1Bunt the runner to thirdThe classic play. It gives up an out to move the runner ninety feet, after which a sacrifice fly, a ground ball to the right side, a wild pitch or a passed ball all score him. It also hands the defence an out it did not have to earn, and with only one run likely to result, a one-run lead in the bottom half is often not enough.
2. 2Swing away and hope for the singleKeeps all three outs and plays for a bigger inning, which is what the arithmetic usually prefers, because the visiting team needs more than one run. The risk is the strikeout followed by the double play, which leaves the runner exactly where he started and the inning over.
3. 3Steal thirdRarely attempted and often the best available idea, particularly under the current pitch timer, which limits how many times a pitcher may step off and therefore how hard he can hold a runner. A successful steal achieves the bunt's objective without spending an out.
4. 4Send up a pinch-hitterUses a bench player to improve one plate appearance in an inning that may be the last, or may be the first of five. Managers who empty the bench in the tenth are the ones taking positions players out of the bullpen in the fourteenth.
5. 5Do nothing unusual at allDefensible, and increasingly common. The single most reliable route to two runs is two hits, and every clever manoeuvre in this list reduces the number of chances to get them.

A constructed walk-through of the standard decision rather than a description of any particular game. The trade-offs are the ones that recur; the conclusions are the article's own.

The home team's problem is different and considerably easier. It bats second, so it knows exactly what it needs. If the visitors scored one, a bunt to third is far more attractive, because a single run ties the game and keeps it alive, and the defence is now obliged to defend both the tying run at third and the winning run behind it. If the visitors scored three, the bunt is unthinkable.

That asymmetry existed before the rule. The automatic runner sharpens it, because both clubs now start every half-inning from a position where a single run is genuinely reachable, and the information advantage of batting second becomes the difference between playing for one and playing for several.

The defence has a matching dilemma. Bringing the infield in stops the run on a ground ball and turns routine outs into base hits. Playing back concedes the run to get the out. With nobody out and the runner ninety feet away, most clubs play back on the first two batters and come in with two outs, which is a compromise nobody likes and everybody uses.

There is one further pressure that did not exist before the timer. A runner on second faces a pitcher who may only disengage from the rubber twice per plate appearance, and a third unsuccessful attempt is a balk. The rules on what a pitcher may and may not do with a runner on base suddenly matter every single inning rather than occasionally, and the limits the timer places on holding runners are being tested in the highest-leverage situations of the season.

The intentional walk came back

One tactic the rule revived almost overnight is the deliberate free pass, and the mechanism is worth spelling out because it is pure geometry.

A runner on second base leaves first base empty. An empty base is an invitation. Put the batter on first and three things change at once: there is now a force play at third, which means a ground ball anywhere on the infield can retire the lead runner without a tag; there is a double play available, which is the only way to erase two of the three outs the defence needs; and the defence has swapped a hitter it fears for one it does not.

The cost is that the potential winning run is now standing on first base, and the argument about whether that matters is one of the oldest in the sport. For the visiting team's half of the inning it usually does not matter much, because a runner on first with two out scores on very little. For the home team's half it matters enormously, since a run that reaches the plate ends the game on the spot.

The tactic got cheaper in 2017, when the deliberate walk stopped requiring four thrown pitches and became a signal from the dugout. Removing the pitches removed both the time and the small but real chance of a wild throw, which had been the only thing making a manager think twice.

Take the situation one further and it becomes the standard extra-innings picture. A runner on second, a bunt or a ground ball moves him to third, a walk fills first, and now there is a force at every base with the infield drawn in. That is a defensive alignment built entirely on the certainty that a runner will be standing on second when the inning begins, and before 2020 it appeared a handful of times a season. It now appears most weeks.

What the rule does to the shape of an inning

A runner on second with nobody out is one of the most productive situations in baseball. It is not, however, productive in the way the visiting team needs.

The distinction is between scoring at all and scoring enough. Starting a half-inning in scoring position makes it likely that a run will come home. It does not make it much more likely that three will, because the extra base does nothing for the second and third runs, which still require the offence to string together the same hits it always did. The rule therefore inflates the floor of an inning far more than it raises the ceiling.

For the home team, a raised floor is exactly what it wants, since one run ends the game. For the visiting team, a raised floor is close to worthless. It bats first, it does not know what it needs, and a single run puts it in the position of defending a one-run lead against a club that will begin its half in scoring position too. The visitors have to score more than the home side, from an identical start, with less information.

That is the hidden asymmetry in a rule which is symmetrical on its face. Both clubs get the same runner in the same place. Only one of them gets to bat knowing the target, and the automatic runner makes that knowledge more valuable than it has ever been, because a target of one run is now genuinely reachable and a manager can plan a whole inning around reaching it.

The other visible effect is on how games end. A walk-off used to require either a home run or a sequence of events beginning with somebody reaching base. With the runner already at second, a single ends it. Extra-inning games now finish on ordinary hits to the outfield far more often than on anything dramatic, which is another way of saying the rule has traded spectacle for brevity. It was designed to do exactly that and it should not be criticised for succeeding.

Who is left to pitch

Extra innings are a resource problem before they are a tactical one, and the rules governing who may take the mound are tighter than most viewers realise.

• 3Batters a relief pitcher must face
• 8Runs behind at which a position player may pitch
• 10Runs ahead in the ninth at which he may
• 0Runs required to use one in extra innings

Rule-defined thresholds. The position player provisions were tightened after clubs began using them to protect large leads rather than only to survive blowouts; extra innings remain an exception at any margin.

The three-batter minimum is the constraint that bites hardest. A reliever who enters must face at least three batters or finish the half-inning, which removed the one-batter specialist from the game entirely. In extra innings that means a manager cannot bring in a left-hander purely for the dangerous left-handed hitter and take him straight back out; he is committing that pitcher to the next two hitters as well, whoever they turn out to be.

Position players are allowed to pitch in extra innings regardless of the score. That is deliberate. A club that has genuinely run out of pitchers in the sixteenth inning has to put somebody out there, and the alternative is forfeiting. Outside extra innings the thresholds are strict, because clubs had started using shortstops to protect comfortable leads and the league decided that was a step too far toward farce.

Roster limits are the background constraint on all of this. Clubs carry a capped number of pitchers, and the rules on how a roster may be constructed and changed mean a bullpen emptied on Tuesday cannot simply be replaced on Wednesday. This is the real argument for the automatic runner, restated: the rule exists because the roster rules make long games expensive in a way that lasts.

The games that used to happen

The record for a major league game is twenty-six innings, played in 1920 between Brooklyn and Boston at Braves Field. It finished level at one apiece and was called for darkness, which is to say it did not finish at all. Both starting pitchers, Leon Cadore for Brooklyn and Joe Oeschger for Boston, threw the entire twenty-six innings. The whole thing took three hours and fifty minutes, which is a comment on how the sport was played then and how it was played a hundred years later.

The postseason record is eighteen innings, and it has been reached five times. One of those, the third game of the 2018 World Series, ran seven hours and twenty minutes and was ended by a home run from Max Muncy in the eighteenth. Los Angeles won it 3-2. The 2025 World Series produced another of the five.

A regulation game9innings

First inning with an automatic runner10innings

Longest postseason game on record18innings

Longest major league game on record26innings

Innings played. The regulation figure and the first automatic-runner inning are rule-defined. The two records are the longest major league game on record, in 1920, and the longest postseason games, a mark reached five times.

Show the numbers

How long a baseball game can actually run

Item	Innings
A regulation game	9innings
First inning with an automatic runner	10innings
Longest postseason game on record	18innings
Longest major league game on record	26innings

Look at the gap between the second bar and the fourth. That distance is what the rule is trying to remove, and there is an honest case that removing it costs the sport something irreplaceable. Nobody remembers a tenth inning. Everybody who watched the 2018 game remembers where they were, and the reason it is remembered is precisely that it went on far longer than it should have.

The rule does not make an eighteen-inning game impossible in the regular season. It makes it very unlikely, because a runner starting on second converts a large share of extra innings into scoring innings. What it makes impossible is the specific thing those games had, which is the sense that the two clubs had run out of ideas and were now simply enduring each other.

Games that cannot be finished

A tie is not a result in the major leagues, and this is where the rules get unexpectedly tidy.

If a game cannot be completed, usually because of weather, it is not replayed from the beginning and it is not recorded as a draw. It becomes a suspended game and is resumed later from the exact situation it stopped in: the same inning, the same count, the same runners on the same bases, the same batting order. A game halted with two out in the eleventh and a runner on first restarts with two out in the eleventh and a runner on first, possibly weeks later, possibly with several of the original players no longer on either roster.

That principle is why the 1920 game is a curiosity rather than a template. It was called for darkness and stood as a tie, which the modern rules would not permit. Any game that reaches the point of being unplayable is put on ice rather than thrown away, and the automatic runner comes back with it: a suspended game resumed in the twelfth inning begins that inning with a runner on second, because the rules of the inning travel with the inning.

The appeal budget refreshes too

One quiet interaction is worth knowing about. Under the challenge system for balls and strikes, each club starts the game with two appeals and keeps every one it wins. In extra innings, a club that has none left is handed one back at the start of each extra inning.

The reasoning is the same as the reasoning for the runner, pointed in the opposite direction. A long game would otherwise reach the innings that decide it with both clubs unable to question a single call, which is exactly backwards, since those are the innings where a ball-strike call carries the most weight. So the automated challenge system tops up rather than running dry, and it does so one inning at a time rather than accumulating, which keeps the resource scarce without letting it disappear.

That produces a small tactical curiosity. A club that goes into the tenth with a challenge in hand gets nothing extra, so there is a mild argument for spending a challenge late in the ninth of a tied game rather than carrying it into extras where it would have been replaced anyway. Nobody has admitted to doing this on purpose. Somebody is doing it.

Everyone else has a different answer

The major leagues are not the only competition that had to solve this, and the range of solutions is wider than most American fans realise.

The idea itself is not American in origin. Placing runners to force a result came into baseball from international softball, where tournaments played to a fixed schedule cannot allow a game to run indefinitely, and it spread from there into international baseball. The international version is more aggressive than the major league one, typically placing two runners rather than one, which produces a result faster and looks even less like ordinary baseball. It is used because a tournament with flights booked and a stadium handed back at midnight has no other option.

The professional leagues in Japan and South Korea took the opposite route. Both have long been willing to cap a regular-season game at a set number of innings and record a tie, which the American game has never seriously contemplated. That solution has an obvious elegance: it protects pitching staffs completely, it requires no invented baserunners, and it costs nothing except the idea that every game must have a winner. American baseball's refusal to consider it is cultural rather than practical, and it is the reason the automatic runner exists at all.

Below the majors, the affiliated minor leagues use the automatic runner, which is unsurprising given that they exist to develop players for a competition that uses it. College and high school baseball have their own arrangements, and amateur competitions that must finish a tournament in a weekend have used tiebreaker rules of one kind or another for decades without anybody writing angry columns about them.

The pattern across all of it is consistent. The more a competition is constrained by a schedule, the more artificial its method of ending a tied game. The major leagues sit in the middle of that spectrum, with a long schedule but no hard deadline, and the rule they landed on is a compromise that satisfies almost nobody completely.

The case against, taken seriously

The objection is not nostalgia, or not only nostalgia.

The strongest version is that the rule changes what a win means. A baseball season is a hundred and sixty-two attempts to establish which clubs are better than which, and the standings are the sport's answer. Extra-inning games decided by a rule that begins each half-inning in scoring position are decided by a smaller sample of baseball than the innings that preceded them. Two or three wins a season shifting from the better club to the luckier one is not nothing when playoff places turn on one game, and the sport's own decision to switch the rule off in October is a tacit admission that it agrees.

The second objection is that the rule solves the wrong problem. Games got long because there were more pitching changes, more time between pitches and more strikeouts, and the sport has since addressed all three directly and effectively. A tenth inning played under a pitch timer is a far shorter thing than a tenth inning played without one. The emergency measure has outlived the emergency and now sits on top of a set of rules that were doing the same job better.

The defence to both is symmetry and honesty about purpose. Both clubs get the identical advantage in the identical inning, so the rule does not favour anyone. And the sport is not pretending the runner is earned. It has written into the scoring rules that he arrived by a fiction, which is an unusually candid thing for a governing body to do.

There is a third objection which is not really an argument at all, and it is the one with the longest reach. The rule acquired a mocking nickname within weeks of arriving, built out of the commissioner's surname, and it stuck so thoroughly that broadcasters use it without thinking. A rule that the sport's own audience has given a joke name is a rule the sport has not persuaded anybody to love, and no amount of workload data will change that. Names are the part of a rulebook that supporters actually own.

Where you land probably depends on whether you think a regular season game is a contest or a data point. If it is a contest, the runner is an insult. If it is one of a hundred and sixty-two measurements, and the object is to take those measurements without destroying the pitching staffs that produce them, the runner is a reasonable piece of engineering.

What to watch for in the tenth

Four things tell you most of what is about to happen.

Whether the visiting club bunts. A bunt in the top of the tenth says the manager has decided one run is what he is playing for, which also tells you what he thinks of his bullpen for the bottom half.

Whether either club pinch-runs. Spending a bench player on the automatic runner is a statement that this inning is the last one, and if the game continues, that decision will constrain everything that follows.

Where the infield sets up on the first pitch. In means the run matters more than the out. Back means the opposite. It is the clearest single signal of a manager's intentions available anywhere in the sport, and it is visible before a pitch is thrown.

Who is warming up. Count the pitchers left. In the twelfth inning, that number is the game, and the club that runs out first will be watching a shortstop throw an eighty-mile-an-hour fastball to decide a game it had led in the eighth.

None of this applies in October, when the tenth inning starts with nobody on and both clubs go looking for a hit like everyone did for a century and a half. That contrast, running side by side in the same sport in the same season, is the most interesting thing about the rule, and it is the part almost nobody mentions.