MLB playoff format explained: byes, seeds and short series

Baseball plays 162 games because it has to. Over a single game the better team beats the worse one barely more often than not, and the sport spends six months accumulating enough evidence to separate a genuinely good club from a lucky one. Then it takes the twelve survivors and settles the championship in a set of short series where luck is very nearly back in charge.

That contradiction is the whole subject. The MLB playoff format is explained easily enough in a paragraph of seeds and byes, and almost every account stops there. The interesting question is what the format is actually measuring, and the answer is that the regular season and the postseason are measuring two different things and always have been.

Twelve teams, six seeds, one fixed bracket

The field is twelve clubs, six from each league.

Seeds one, two and three are the three division winners in that league, ranked by regular-season record. Seeds four, five and six are the three best remaining clubs in the league, ranked the same way. Every division winner qualifies, whatever its record, and every division winner is seeded above every wild card, whatever their records.

That second rule has a consequence people notice every year. A division winner with a worse record than a wild card team is still the higher seed, still hosts, and still gets the bracket position. It is deliberate. Winning a division is meant to be worth something, and the something is a guaranteed seed above the teams that did not win one.

The top two seeds in each league receive a bye directly to the Division Series.

Seeds three through six play a best-of-three Wild Card Series, with three against six and four against five. All three games are at the higher seed's park. There is no travel and no split of the home dates.

There is no reseeding. The bracket is drawn before a pitch is thrown: the one seed waits for the winner of four against five, the two seed waits for the winner of three against six. If the six seed wins its series, the two seed inherits it. If the five seed wins, the one seed gets a club that may well be stronger than the one the second seed drew. That asymmetry is a known cost of a fixed bracket, accepted because a fixed bracket can be printed, sold and scheduled in advance.

Ties are broken mathematically. The one-game playoff is gone. Where two clubs finish level, head-to-head record is the first tiebreaker, then intradivision record, then records against progressively wider groups of opponents. Nothing is decided on the field. The trade was an extra game of drama for a postseason that starts on a fixed date, which television requires and the calendar increasingly demands.

• 12Clubs in the postseason field
• 2First-round byes per league
• 11Wins needed to take the title from a bye
• 13Wins needed to take it from the Wild Card round

Counts defined by the published postseason format. The win totals are the sum of the wins required in each round on each path, not an estimate of anything.

Eleven wins against thirteen is the clearest statement of what a bye is worth. Two extra wins is not a rounding error at this level. It is an entire extra series in which a good team can be beaten by a slightly worse one.

Round by round, and who hosts what

The four rounds are not the same shape, and the differences matter more than the names.

1. 1The regular season sets six seeds per leagueDivision winners take seeds one to three by record, the three best remaining clubs take four to six. Ties are resolved by head-to-head record and then by record against progressively wider groups, with no play-off game.
2. 2Wild Card Series, best of threeSeed three hosts six and seed four hosts five, with all three games at the higher seed's park. Two wins ends it, so a single bad afternoon by a starting pitcher can end a ninety-win season.
3. 3Division Series, best of fiveThe bye teams enter here. The higher seed hosts games one, two and five, the lower seed hosts three and four. Three wins advances, and the higher seed's edge is one extra home date plus the decider.
4. 4League Championship Series, best of sevenTwo, three, two. The higher seed hosts the first two and the last two, the lower seed hosts the middle three. Four wins takes the pennant.
5. 5World Series, best of sevenSame two, three, two pattern, but seeding no longer applies across leagues. Home field goes to the pennant winner with the better regular-season record, a rule adopted in 2017 to replace the All-Star Game result.
6. 6The bracket never reseedsThe one seed always meets the four against five survivor and the two seed always meets the three against six survivor, whatever happens in the opening round.

The published bracket and home-field patterns. Series lengths and hosting rules are fixed by the format rather than negotiated round by round.

Notice what the home-field reward actually amounts to once you count it.

• Games if the series goes the distance
• Hosted by the higher seed

Wild Card Series33

Division Series53

League Championship Series74

World Series74

Rule-defined counts for a series that goes the full distance. In the World Series the host is the pennant winner with the better regular-season record rather than a seed.

Show the numbers

Games in each round, and how many the higher seed hosts

Item	Games if the series goes the distance	Hosted by the higher seed
Wild Card Series	3	3
Division Series	5	3
League Championship Series	7	4
World Series	7	4

The Wild Card Series is the outlier and it is the only round where hosting is a real weapon, because the lower seed never gets a home game at all. Everywhere else the advantage is one extra date out of five or seven, which is a thinner edge than the phrase home-field advantage suggests.

The two patterns are not arbitrary either, and the difference between them is worth a moment. Two-two-one means one round trip: the series opens at the higher seed's park, moves once, and comes back if it needs a fifth game. Two-three-two also means one round trip, which is why it is used for the longer rounds, where a strict alternation would put four flights into a seven-game series. Both formats are shaped by travel first and competitive design second.

They do distribute the advantage differently, though. Two-two-one hands the higher seed the decisive fifth game at home, which is the cleanest form of the reward. Two-three-two hands it games six and seven but concedes three straight in the other park in the middle, so a higher seed that loses the opening two is in genuine trouble and a lower seed that wins the middle three never has to come back. The longer series is the fairer one overall, and it is also the one that gives the underdog a real window.

What the bye is worth, and what it costs

The bye is the single largest prize the regular season awards, and its value is easy to state precisely: it removes a round in which you could be eliminated. Nothing else the format hands out comes close.

It also allows a manager to arrange his pitching. Coming out of a bye, a club sets up its best starter for game one of the Division Series on full rest, its second starter for game two, and can plan a third-time-through plan for the series rather than improvising after a coin-flip round. A team arriving from the Wild Card Series has usually already spent its best pitcher, and possibly its second, and arrives with a bullpen that has worked on consecutive days.

Against that sits a genuine cost, and the debate about it is not settled.

A bye means roughly five days without competitive baseball, and hitting is a timing skill that decays fast. Simulated at-bats against a coach are not the same as facing a postseason starter, and the club coming out of the Wild Card round arrives having just played high-leverage baseball with everything sharp. Whether that layoff is worth more or less than the round of elimination risk it avoids is the kind of question that would need many more postseasons than have been played under this format to answer. What can be said cleanly is that the two effects run in opposite directions, that they are of very different sizes, and that eliminating a whole round of coin flips is a large, certain benefit while rust is a small, uncertain cost.

There is a second-order effect that gets discussed less. A bye changes what a club does in September. If the difference between seeds two and three is the difference between resting your rotation and playing a best-of-three, then a club that has clinched a spot but not a bye has every reason to keep pushing, and a club that has locked up a bye has every reason to stop. The format quietly gave the last fortnight of the regular season a stake it did not have under a ten-team field, which is one of the better arguments for it.

Series length is the only thing that protects the better team

This is the piece of arithmetic that everything else in the argument rests on, and it is worth doing rather than asserting.

Take an invented example. Suppose the better of two clubs would win 55 per cent of the games between them, played at neutral sites. That is a large edge in baseball terms, larger than most postseason matchups. Now play a series.

Over one game the favourite wins 55 per cent of the time. Over a best-of-three, 57.5 per cent. Over a best-of-five, 59.3 per cent. Over a best-of-seven, 60.8 per cent. Those are computed from the assumed 55 per cent, and the assumption is invented; the arithmetic is not.

Look at what that says. Extending a series from one game to seven, which is the longest format the sport uses, moves the better team from winning 55 times in a hundred to winning 61. Six percentage points, for six extra games. The better team is still losing two series in five.

Now run the same arithmetic with a per-game edge of 65 per cent, which is roughly what a serious mismatch looks like in basketball and essentially never happens between two qualified baseball teams.

• A 55 per cent per-game favourite
• A 65 per cent per-game favourite

Constructed example. Both curves are calculated from an invented and fixed per-game win probability, ignoring home field, pitching order and fatigue. The point is the shape, not the values: the flatter curve is baseball.

Show the numbers

What a longer series buys the better team

Item	A 55 per cent per-game favourite	A 65 per cent per-game favourite
1 game	55%	65%
Best of 3	57.5%	71.8%
Best of 5	59.3%	76.5%
Best of 7	60.8%	80%

The gap between the two curves is the whole reason baseball's postseason behaves differently from basketball's, and it has nothing to do with the format. It is about the per-game number that goes into it. A sport where the better team wins two games in three converts a seven-game series into an 80 per cent proposition. A sport where the better team wins eleven games in twenty converts the same series into a coin flip with a lean.

Which means the format's choices, best-of-three here and best-of-five there, are adjusting a number that was never going to get far from 50 anyway. Lengthening every round would help the better teams a little. It would not make October predictable, and the games would still be in November.

Why home field matters less here than almost anywhere

The home team in baseball gets three things, and none of them is large.

It bats last, which is worth something real in a close game because the home side always knows exactly what it needs. It gets the last change, so the manager can wait to see a pinch hitter announced before choosing the reliever. And it plays in a park it knows, which in baseball is a genuinely distinctive object rather than a standard rectangle, since the dimensions and conditions of each ballpark change how the game plays in ways no other major sport permits.

What the home team does not get is the thing that drives home advantage elsewhere. There is no crowd effect on a defence trying to hear a call. The strike zone has historically been the one place where crowd pressure could plausibly move a game, and the sport is in the process of removing even that, with the challenge system that lets players contest a ball or strike call taking the judgement out of the decisive pitches.

Add it up and baseball's home advantage sits at the low end of team sport, well below basketball's. Which is why the reward structure is arranged the way it is: the format does not try to make hosting decisive, because it cannot. It gives the higher seed a whole round instead.

October is a different sport for pitchers

Nothing separates the regular season from the postseason as much as what happens to pitching staffs.

A regular season is a six-month endurance problem. Games are played nearly every day, a rotation turns over every five, and the constraint on every decision is the next three weeks rather than tonight. A club that burns its best relievers in April pays for it in August, so managers husband arms, accept losses, and let a struggling starter work through an inning because somebody has to.

The postseason has off days. Travel days between the second and third games of a round, and between rounds, mean the calendar stops being daily. That single scheduling fact reshapes everything.

A four-man rotation becomes a three-man rotation, and in a five-game series the third starter may not be needed at all. Starters who would never appear on a throw day in June come out of the bullpen in a decider. Relievers who worked one inning at a time all summer work two, on consecutive days, because there is no next month to protect. The leash on a starter shortens dramatically: a manager who would have let him face the order a third time in July pulls him in the fifth in October, because the alternative is a reliever who is not available to him again for four months.

The net effect is that the quality of pitching a hitter faces in the postseason is markedly higher than what he faced all year, and that the identity of the pitchers matters more. Depth, which is what wins 95 games, becomes close to worthless. Concentration of talent, which is a liability over 162, becomes the thing that decides a series.

That inversion is the reason for a familiar October complaint: the team that wins is not always the team that was best. Both statements can be true. It was best at the thing the regular season measures, which is being good every day for six months, and the postseason measures something else, which is being good on four specific evenings with your top two starters available.

The roster in October is not the roster that won 95 games

Postseason rosters are set separately for each round. A club submits 26 players before a series begins and can change the mix before the next one, subject to eligibility rules that generally require a player to have been in the organisation by 31 August, with defined provisions for replacing the injured.

Two things follow, and they are how postseason squads come to look so different from what a supporter watched all summer.

The first is that a manager can build a roster for a specific opponent. A club facing a lineup with a particular handedness split carries an extra left-hander it did not need in the regular season. A club facing a team that runs carries a defensive catcher. Nothing about that is available over 162 games, where the roster has to survive all opponents at once.

The second is that the marginal roster spots invert. The long reliever who ate innings in blowouts, the fifth starter, and the utility player who covered days off are all far less useful in a five-game series than a specialist who might be needed for four high-leverage batters. Those choices are governed by the same active-roster machinery that runs all year, and the roster and eligibility rules that decide who is even available do a lot of quiet work in October that nobody sees.

There is a competitive-balance point buried in this too. Building a roster for a short series rewards front offices that have depth to select from, and depth costs money, which is where the postseason connects to the tax system that governs how much clubs can spend before penalties bite. Spending does not buy October, because nothing buys October. It buys more attempts at it.

Why baseball's playoffs are less predictable than basketball's

The usual explanation is that baseball is a game of luck and basketball is a game of skill, which is both insulting and wrong. The real reason is structural, and it comes down to how much of a game one excellent player can take.

A basketball team's best player can be on the floor for four fifths of the game and involved in a third of the possessions while he is there. If he is the best player in the series, he applies that advantage over and over, hundreds of times, and the accumulation is what turns a talent gap into a large per-game win probability.

A baseball team's best hitter bats in one slot of nine. Over a game he comes up four or five times out of the roughly thirty-eight plate appearances his team takes. He cannot bat when it matters most by choice; the order decides. He cannot demand the ball. If the game turns in the eighth with two on and he is due fifth, he watches.

The one player who can dominate a baseball game is the starting pitcher, and the format limits him too. In a best-of-five he starts twice at the very most. In a best-of-three, once. A truly great starter is worth an enormous amount in the games he pitches and precisely nothing in the others.

So the mechanism by which talent converts into wins is throttled at both ends. The best hitter gets an eleventh of the opportunities. The best pitcher gets a third of the games. Against that, defence is spread across nine players, and the outcome of any individual event depends on a ball travelling a few feet either side of a fielder.

Put the throttling together with the arithmetic above and the result follows without any appeal to luck as a mystical force. Talent gaps between playoff-qualified baseball teams are compressed by the structure of the sport, small per-game gaps stay small even in a seven-game series, and the format then adds a best-of-three round at the front. The postseason is unpredictable because the sport is unpredictable at the scale a series operates on. The comparison worth making is not with basketball but with the way American football seeds and rewards its own playoff field, a sport that solved the same problem by playing single-elimination games and simply accepting the variance rather than pretending to remove it.

The honest case that the format devalues the regular season

The complaint deserves stating at its strongest, because it is usually made badly.

The strong version is not that good teams lose. It is that the format has decoupled the reward from the achievement. A club that wins its division by a distance and a club that scrapes the final wild card spot both arrive in the same bracket, and the only structural difference between them is a bye and some home dates. The gulf between them, built over six months and hundreds of games, is compressed into an advantage that the arithmetic above says is worth a handful of percentage points.

There is a sharper version still. Because all three division winners qualify and are seeded above every wild card, and because divisions are unequal, the seeding does not even reliably rank the field by quality. A three seed can be worse than a five seed and still host. The bracket rewards the shape of your division as much as the quality of your season.

And there is a behavioural consequence that concerns front offices more than supporters. If a 92-win team and a 100-win team have similar championship odds once the bracket starts, the marginal value of the eight extra wins collapses, and so does the case for the spending or the trades that would produce them. The rational club builds to qualify rather than to dominate. That is an argument about incentives rather than about fairness, and it is the one that should worry the sport most.

The honest case that it does not

The counter-argument is stronger than its critics allow.

The regular season decides who gets in at all. Eighteen clubs are eliminated before October and no amount of postseason variance rescues them. That is by far the largest filter in the system and it is entirely a function of 162 games.

Within the field, the regular season decides seeding, and seeding decides byes, home dates and bracket position. Eleven wins against thirteen, as the numbers above show, is a substantial structural gap even if it does not feel like one.

And a competition is not only a device for identifying the best team. If it were, the answer would be to award the title to the club with the best record in September and skip October entirely, which is roughly what league football does and which nobody watching baseball has ever proposed. The postseason exists because elimination baseball is a different and more intense product, and because a sport that finished with a coronation would have nothing to sell in autumn.

The expanded field also does real work in the standings. A twelve-team bracket keeps more clubs mathematically alive into September, which keeps more markets engaged and gives more front offices a reason to buy at the deadline rather than sell. Whether that is a benefit to the sport or a subsidy to mediocrity is a matter of taste. That it changes club behaviour is not.

The problem the format did not plan for

One current issue is worth setting out precisely, because it exposes a structural assumption nobody examined.

Since 2023 the schedule has been balanced: every club plays every other club in the sport during the season, with interleague games rising to 46 of the 162. That was done for good reasons, mostly so that supporters everywhere would see every star and so that unbalanced divisional schedules stopped distorting the standings.

But the playoff structure is still organised by league, and half of a club's opponents are now in the other one. If one league is materially stronger than the other in a given year, its clubs beat up on the weaker league's clubs in interleague play, records in the strong league cluster near .500 and records in the weak league spread out. The playoff field in the weaker league is then drawn from a pool whose won-lost records look worse than they are, and the field in the stronger league from a pool whose records look better.

Push that far enough and a club with a losing overall record qualifies for the postseason, having been perfectly respectable against the opponents its own league provided. Nothing in the rules prevents it, and the balanced schedule made it more likely rather than less.

The fixes under discussion all have costs. Refusing a playoff place to a sub-.500 club shrinks the bracket unpredictably, which broadcasters have contracted against. Qualifying on intraleague record only ignores nearly a third of the season. Abolishing divisions and seeding purely on record would solve it cleanly and would end the rivalries that the divisional structure exists to create. Reducing interleague play would undo a change made deliberately and recently. There is no free option here, which is usually a sign that the underlying tension is real rather than administrative.

What actually predicts October

If the format cannot be relied on to advance the better team, it is fair to ask what does help.

A bye. The single biggest identifiable edge in the bracket, and the only one that removes elimination risk rather than nudging a probability.

Two starting pitchers who can be trusted in a decider. Not five. In a five-game series the third starter may pitch once, and a rotation's depth is close to irrelevant. What matters is the top of it, and whether the top of it is rested.

A bullpen deep enough to work on consecutive days. October usage patterns break relief staffs that were built for a daily grind. The clubs that cope carry more arms that can go multiple innings and fewer that can only go one.

Defence, more than the summer suggests. Against better pitching, runs get harder to score, games get closer, and the marginal play on the infield decides more of them. That is also why the defensive value inside an all-in-one metric like wins above replacement tends to be underweighted by supporters right up until the postseason, when every misplaced ball costs a run in a game with four of them.

What does not predict it, as far as anyone can demonstrate, is momentum, September form, or being battle-tested by a wild card series. Those are stories applied after the fact to a distribution that would look much the same without them.

What to watch when the bracket is set

Three questions, and they will tell you more than the seeding will.

Who has a bye, and by how much did they miss or earn it? The difference between the second and third seed is an entire round of survival, and it is frequently decided by one or two games in a season of 162.

Who is starting game one, and on what rest? A club emerging from the Wild Card Series into a Division Series two days later is not the club that won 95 games. It is that club without its best two pitchers available on schedule.

Which bracket half is loaded? Because there is no reseeding, one half can carry three of the four best teams in the league. The bracket is drawn before anybody knows that, and it does not adjust.

Baseball's postseason is regularly accused of being a lottery. It is not. It is a tournament in a sport that compresses the difference between good teams to a level a short series cannot reliably detect, run as a deliberate choice to sell the most intense version of the game rather than the most accurate one. If you want the sport's judgement on which club was actually best, that verdict was delivered over six months and it is sitting in the final standings. October is a separate competition that happens to award the trophy, and the sport has never seriously pretended otherwise. Anyone following the wider American game and how its competitions are structured is watching a league that decided, quite consciously, that certainty is less valuable than October.