NBA Playoff Format: Seeds, Bracket and Best-of-Seven Rules
The NBA playoff format explained for fans in the United States: 16 teams, a fixed bracket in each conference, four best-of-seven rounds and 16 wins.
By CricketTaken EditorialPublished Updated NBA Season, Playoffs & Draft22 min read
- Playoff teams
- 16
- Rounds, all best-of-seven
- 4
- Wins needed for a title
- 16
- Home-court pattern
- 2-2-1-1-1
The 2017 Golden State Warriors lost one playoff game. It was Game 4 of the Finals, in Cleveland, and it is the only thing that separates that team from a perfect run through the NBA's bracket: they finished 16-1. Sixteen is the number that matters. A champion has to win four series, each of them a best-of-seven, and four times four is the smallest total that ends with a trophy.
The NBA playoff format in full: 16 teams qualify, eight from the Eastern Conference and eight from the Western. The top six in each conference get in on regular-season record, and seeds seven and eight are settled by a short play-in tournament among the teams that finish seventh through tenth. Each conference then runs a fixed bracket of 1 against 8, 4 against 5, 2 against 7 and 3 against 6, with no reseeding between rounds. Every series is best-of-seven on a 2-2-1-1-1 home pattern, and the two conference champions meet in the NBA Finals.
Sixteen teams, four rounds, 16 wins: the format in one pass
The league calls its rounds the First Round, the Conference Semifinals, the Conference Finals and the NBA Finals. The first three are played entirely inside a conference. An Eastern team cannot meet a Western team until the Finals, which is why the Finals so often pairs two clubs that have played each other only twice all year.
Each round halves the field. Eight series in the first round, four in the second, two conference finals, one Finals. That makes 15 series in all, and since each one needs between four and seven games, a full postseason contains somewhere between 60 and 105 games before anyone counts the play-in.
- Regular seasonThirty teams play 82 games each. The top six in each conference clinch a playoff place outright.
- Play-in tournamentThe teams that finish seventh through tenth in each conference play for the last two seeds.
- First Round1 plays 8, 4 plays 5, 2 plays 7 and 3 plays 6 in each conference. Best-of-seven.
- Conference SemifinalsThe four first-round winners on each side meet along fixed bracket lines.
- Conference FinalsTwo teams remain in each conference. The winner takes the conference title.
- NBA FinalsEast champion against West champion, best-of-seven, for the Larry O'Brien Trophy.
The structure in use since the play-in tournament joined the format in 2021. Round names are the league's own.
Two features set the NBA apart from the other American leagues, and both get their own section below. The bracket is fixed, where the NFL's is not. And home court in the last round is awarded on record, with seed number ignored entirely. Most confusion about the postseason traces back to one of those two points.
What the format does not include is also worth stating. There is no bye for anyone. There are no wild cards in the NFL or MLB sense, because every one of the 16 places is a numbered seed inside a conference. And since 2003 there has been no short series anywhere in the bracket, which is the main reason upsets are rarer here than in any other major American postseason.
How the field is set: six automatic seeds and the play-in
Qualifying used to be simple. The eight best records in each conference went in, and the ninth-place team went home. Since 2021 the line has been drawn in two places. Sixth place is the last guaranteed seed, and tenth place is the last position with any route into the bracket.
| Finish in the conference | What it earns |
|---|---|
| 1st through 6th | A playoff seed matching the finish, and a week of rest |
| 7th | A home play-in game against 8th, with a second chance if it loses |
| 8th | A road play-in game at 7th, then a home game if it loses |
| 9th | A home play-in game against 10th, and it must win twice |
| 10th | A road play-in game at 9th, and it must win twice |
| 11th through 15th | Nothing. The season is over |
The mechanics are short. The seventh-place team hosts the eighth, and the winner becomes the No. 7 seed. The ninth-place team hosts the tenth, and the loser is eliminated. Then the loser of the first game hosts the winner of the second, and that game produces the No. 8 seed. Seventh and eighth place need one win in two tries. Ninth and tenth need two wins in two. Our explainer on the play-in tournament format walks through each game and the arguments about it.
A version of the idea first appeared in the 2020 restart at Walt Disney World, where a ninth-place team within four games of eighth could force a play-in. The current four-team design arrived for 2020-21 and was made a permanent part of the format in 2022.
What the play-in is not
Play-in games are not playoff games. They do not count toward the 16 wins, they are not part of any series, and their statistics are kept apart from both the regular season and the playoffs. A team that wins its way in as the eighth seed starts its first-round series at 0-0 like everyone else, a day or two later and on the road.
The order of finish itself comes from winning percentage, and ties are common enough that the league maintains a full procedure for them. Head-to-head record is the first test. The rest is set out in our guide to the NBA's tiebreaker rules, which matters more now than it once did, because the gap between sixth and seventh is the gap between a week off and an elimination game.
The fixed NBA playoff bracket and why the league does not reseed
In the first round the pairings are 1 against 8, 2 against 7, 3 against 6 and 4 against 5. What happens next is decided before a single game is played. The 1-8 winner meets the 4-5 winner. The 2-7 winner meets the 3-6 winner. Those lines never move.
The NFL does it differently. In the divisional round the top remaining seed is always matched with the lowest remaining seed, so the bracket rearranges itself after every upset. The NBA has never done that under the 16-team format.
The spring of 2023 shows what the difference means. Miami, the eighth seed in the East, eliminated top-seeded Milwaukee in five games. The four teams left on that side were seeds 2, 3, 5 and 8. A reseeding league would have sent the Heat to Boston, the best team remaining, and paired Philadelphia with New York. The NBA's bracket instead gave the Heat to the fifth-seeded Knicks, who had beaten Cleveland in the 4-5 series, and left the Celtics to deal with the 76ers over seven games. Miami beat New York, then beat Boston in seven, and reached the Finals.
So an upset in a fixed bracket benefits someone other than the team that pulled it off. Whoever sits in the same quarter inherits an easier path. Fans of the second seed tend to notice this about once a decade.
Why the league keeps it this way
Two reasons are usually given, and one is better than the other.
The practical one is scheduling. A fixed bracket lets arenas, broadcasters and traveling parties plan the second round before the first is over, because each team knows the two possible opponents it could face. With seven-game series that end on different nights, that is worth something.
The second is that the bracket is easy to follow. A fan can fill out all 15 series in April. That argument is more about marketing than fairness, and the league has not pretended otherwise. A reasonable judgment is that reseeding would be slightly fairer to the top seed and that the NBA has decided the difference is too small to give up a bracket people can read at a glance.
Home-court advantage and the 2-2-1-1-1 pattern
Every series gives one team four of the seven possible home dates. That team hosts the first two games, travels for the next two, and then the series alternates one game at a time.
| Game | Host | Played |
|---|---|---|
| 1 | Team with home-court advantage | Always |
| 2 | Team with home-court advantage | Always |
| 3 | Opponent | Always |
| 4 | Opponent | Always |
| 5 | Team with home-court advantage | If necessary |
| 6 | Opponent | If necessary |
| 7 | Team with home-court advantage | If necessary |
Within a conference, home court belongs to the team with the better regular-season record. Because seeds have been assigned strictly by record since 2015-16, with tiebreakers applied to any teams that finish level, that is now the same thing as saying the higher seed hosts. It was not always the same thing, as the next section explains.
Game 7 is the prize. The team that earned the advantage plays the deciding game in its own building, and home teams have won a clear majority of the Game 7s in league history. That is the practical value of finishing one line higher in April: not the first two games, which the visitor can split, but the last one.
The Finals are decided on record, not seed
The Finals is where seed numbers stop mattering. Two conference champions meet, and the one with the better regular-season record gets Games 1, 2, 5 and 7. A second seed can host a first seed. It happened in 2019, when Toronto, seeded second in the East at 58-24, opened at home against a Golden State team that had been the West's top seed at 57-25. In 2022 the Warriors, a third seed at 53-29, held home court over second-seeded Boston at 51-31.
When the records match, the league breaks the tie, beginning with the two regular-season games the teams played against each other. The 1998 Finals is the textbook case. Chicago and Utah both finished 62-20, the Jazz had won both meetings, and so the series opened in Salt Lake City with the Bulls as the road team.
What a division title is worth after 2015-16
Not much. A division championship earns a banner and a line in the tiebreaker procedure, and that is the extent of it.
It used to be worth a great deal more, and the history explains a lot of old brackets that look wrong today. Through 2003-04 each conference had two divisions and their winners took the top two seeds. When the league went to three divisions per conference in 2004-05, the three winners were given seeds one through three regardless of record. That produced the 2006 Western bracket, in which Dallas won 60 games, finished behind San Antonio in its own division, and was dropped to the fourth seed. The two best teams in the conference met in the second round.
The league adjusted for 2006-07. Division winners were still promised a place in the top four, but the top four were then ordered by record, so a strong second-place team could climb as high as the second seed. Home court in any series went to the better record whatever the seed numbers said.
That compromise lasted until 2015. Portland won the Northwest Division that year at 51-31 and was seeded fourth, ahead of two 55-win teams, Memphis and San Antonio. The Trail Blazers did not even get home court out of it, since Memphis had the better record and hosted the series. A seed that carried no advantage and still distorted the bracket was hard to defend, and the Board of Governors scrapped the guarantee before the 2015-16 season. Since then all eight seeds in a conference have been ordered by record alone.
What remains is a clause. If two teams finish with the same record and head-to-head does not separate them, a division leader wins the tie over a team that did not lead its division. In a tie among three or more clubs, the division-leader test is applied first. It is a small thing, and it has decided real seeds. The five-team groups themselves are laid out in our guide to the league's conferences and divisions.
The postseason calendar: mid-April to June and a 28-game ceiling
The regular season ends on a Sunday in the middle of April. The play-in occupies the following Tuesday through Friday, and the first round begins that weekend. The Finals is played in June, on dates fixed long in advance with the league's broadcast partners. From the first play-in game to a seventh game of the Finals is roughly two months.
Inside that window the schedule is looser than fans expect. There is no rule that a series must be played every other day. First-round series are staggered across the week so that television can show every game, which means one series may get two days off between games and another only one. The modern playoffs do not use back-to-backs, and a change of city normally brings at least a day for travel.
Later rounds do not wait for a fixed date if they do not have to. When both series feeding a matchup end early, the next round can start sooner, though the Finals itself holds its announced dates. A team that sweeps can sit for a week or more, and whether that is rest or rust is an argument revived every May.
- 4Rounds to win
- 16Wins required
- 28Most games a champion can play
- 26Most games a champion has played
Counts playoff games only; play-in games are separate. The 26-game figure belongs to the 2008 Boston Celtics and stands through the 2026 playoffs.
Four rounds of seven games gives a maximum of 28 for a champion. Nobody has needed all of them. The 2008 Celtics came nearest, going seven games with Atlanta, seven with Cleveland, six with Detroit and six with the Lakers for a 16-10 record. At the other end, the shortest title run since best-of-seven became universal is the Warriors' 17 games in 2017. For a team already carrying an 82-game regular season, a long postseason adds the equivalent of another third of a year.
The NBA Finals and the 2-3-2 experiment of 1985 to 2013
For 29 years the last round was played to a different pattern from the rest of the bracket. From 1985 through 2013 the Finals went 2-3-2: two games in the city with home court, three straight in the other, then two back where the series began.
The reason was the airplane. In the early 1980s the Finals kept sending an East Coast team to Los Angeles, Philadelphia in 1980, 1982 and 1983 and then Boston in 1984. Teams flew commercial, and a 2-2-1-1-1 series that went the distance meant crossing the country four times in two weeks. Red Auerbach of the Celtics complained about it, the league's new commissioner David Stern listened, and the format changed for 1985. It cut the maximum number of cross-country trips in half.
It also changed the series, and not in a way the better team liked. Under 2-3-2 the club with home-court advantage could split the first two games and then spend eight days or so on the road, playing three in a row in a hostile building. Game 5, the pivotal game of a tied series, was played on the lower seed's floor. On the other side of the ledger, the team without home court had to win all three middle games to avoid going back on the road needing a win there, and sweeping the middle three turned out to be hard. Only three home teams did it in the 2-3-2 era: Detroit in 2004, Miami in 2006 and Miami again in 2012.
Nothing else about the Finals differed then or differs now from the rounds before it. The games are the same length, the rosters are the ones fixed in April, and a series level after six goes to a seventh game on the floor of the team with the better record. The pace does change. With one series left there is no other matchup to schedule around, and the league spaces the games out, commonly with two days between them and an extra day when the series changes cities. A Finals that goes seven can take more than two weeks.
By 2013 charter flights had made the travel argument thin. In October of that year the owners voted unanimously to return the Finals to 2-2-1-1-1, and the 2014 series between San Antonio and Miami was the first played that way since 1984. Every round has used the same pattern since. Franchise totals from all of those Finals are collected in our list of NBA championships by team.
Who is eligible to play: March 1, two-way contracts and the roster deadline
A playoff roster is narrower than the group of players a fan saw in uniform during the winter. Three rules do most of the narrowing.
The March 1 waiver cutoff. A player who is waived after March 1 cannot appear in that season's playoffs for a different team. He can still sign somewhere and play out the regular season, but he is ineligible once the postseason begins. The date that matters is the day he is released, not the day he signs, which is why the last week of February fills with contract buyouts. Our guide to the buyout market and the March 1 deadline covers how those deals are negotiated. A free agent who was not on any NBA roster after that date, such as a veteran returning from overseas, stays eligible provided he signs before the regular season ends.
Two-way players. A player on a two-way contract, the hybrid deal that splits time between an NBA club and its G League affiliate, is not eligible for the playoffs at all. He may practice and travel. He may not play. To use him in the postseason, the team has to convert the deal to a standard NBA contract, or sign him to one, before its final regular-season game, and it needs an open spot among its standard roster places to do so. That is why April brings a run of conversions from contenders. The contract type itself is described in our piece on how two-way deals work.
The roster deadline. Playoff rosters are set the day after the regular season ends. On the league's published calendar that is a mid-afternoon deadline, Eastern time, on the Monday. After it, a team cannot add a player for the postseason. The trade deadline passed back in February, so the only moves available by then were signings and conversions anyway.
One consequence is easy to miss. A team that loses a player to injury in the first round has no way to replace him from outside. The limits on standard and two-way places that shape these choices are set out in our guide to NBA roster sizes.
What history says about seeds, from the eighth seeds to the 1995 Rockets
Top seeds almost always survive the first round. From 1984, when the bracket grew to 16 teams, through 2026 there were 86 series between a first seed and an eighth. The eighth seed won six.
| Year | Eighth seed | First seed beaten | Result | Series length |
|---|---|---|---|---|
| 1994 | Denver Nuggets | Seattle SuperSonics | 3-2 | Best-of-five |
| 1999 | New York Knicks | Miami Heat | 3-2 | Best-of-five |
| 2007 | Golden State Warriors | Dallas Mavericks | 4-2 | Best-of-seven |
| 2011 | Memphis Grizzlies | San Antonio Spurs | 4-2 | Best-of-seven |
| 2012 | Philadelphia 76ers | Chicago Bulls | 4-2 | Best-of-seven |
| 2023 | Miami Heat | Milwaukee Bucks | 4-1 | Best-of-seven |
A curiosity sits in that table. The first round was lengthened in 2003 partly on the theory that a longer series protects the better team, yet four of the six upsets have come since then, in fewer than 50 series, against two in 38 under best-of-five. The sample is too small to prove anything. It does suggest that an eighth seed good enough to win three games from a top seed is usually good enough to win a fourth.
Two eighth seeds kept going. The 1999 Knicks, who finished 27-23 in a season cut to 50 games by a lockout, reached the Finals and lost to San Antonio in five. The 2023 Heat, who had lost their first play-in game and needed the second just to enter the bracket, got there too and lost to Denver in five. No eighth seed has won the title.
The one bar on the right is Houston. The 1994-95 Rockets were the defending champions, finished 47-35, and entered as the sixth seed in the West. They beat Utah after trailing two games to one in a best-of-five, beat Phoenix after trailing three games to one, beat top-seeded San Antonio, and swept Orlando in the Finals. They did not have home-court advantage in any of the four series. No team seeded lower than third has won a championship in the 16-team era apart from them.
Series leads in numbers: 2-0, 3-1 and the 3-0 wall
Seven games sounds like room to recover. The record says there is less room than it sounds.
- 0Teams to win a series after trailing 3-0
- 4Teams to force a Game 7 after trailing 3-0
- 15Teams to win a series after trailing 3-1
- 1Comebacks from 3-1 down in the Finals
League history through the 2026 playoffs. The totals rise whenever another comeback happens.
Start with the wall. More than 150 teams have taken a 3-0 lead in an NBA series and every one of them has advanced. Baseball and hockey have each seen a 3-0 lead lost. Basketball has not. Four teams have at least dragged the series to a seventh game: the 1951 Knicks against Rochester in the Finals, the 1994 Nuggets against Utah, the 2003 Trail Blazers against Dallas and the 2023 Celtics against Miami. All four lost Game 7.
A 3-1 lead is nearly as safe. Through 2026 it had been overturned 15 times in the league's history. The list begins with the 1968 Celtics against Philadelphia and includes the only one in the championship round, Cleveland over Golden State in 2016. Denver did it twice in the same postseason in 2020, inside the Disney World bubble where home court meant nothing. The two most recent came in the first round in 2026, when Philadelphia beat Boston and top-seeded Detroit escaped Orlando after being a game from joining the table in the previous section.
A 2-0 lead is the most common of these situations and the least decisive, though only by comparison. Teams that win the first two games of a best-of-seven have gone on to take the series better than nine times in ten. Most of those teams were at home for both games and were the stronger side to begin with, so the number says as much about seeding as about momentum.
The reason basketball's leads hold better than hockey's or baseball's is not mysterious. A basketball game has around a hundred possessions per team, which leaves little room for luck, and there is no hot goaltender or ace pitcher who can swing three straight games alone.
How the format grew from short series to best-of-seven throughout
The NBA did not begin with a bracket like this one. Through the 1950s and 1960s a handful of teams qualified from two divisions, opening series were best-of-three or best-of-five, and the first-place team sometimes sat out the first round altogether.
| Seasons | Playoff teams | Opening round |
|---|---|---|
| 1975 and 1976 | 10 | Best-of-three, between the fourth and fifth teams in each conference |
| 1977 to 1983 | 12 | Best-of-three, with first-round byes for the top two seeds in each conference |
| 1984 to 2002 | 16 | Best-of-five, no byes |
| 2003 to 2020 | 16 | Best-of-seven |
| 2021 onward | 16 | Best-of-seven, with seeds seven and eight decided by the play-in |
The modern shape dates from 1984. That year the field went from 12 teams to 16, the byes disappeared, and the first round became best-of-five. A 23-team league was now sending more than two-thirds of its membership to the postseason, a proportion that drew criticism at the time and has fallen as expansion brought the league to 30.
The three-game "miniseries" of the late 1970s and early 1980s had been a genuine hazard. A good team could lose one at home and be out two nights later. Best-of-five reduced that risk and best-of-seven, adopted for the first round in 2003, reduced it further. The league's stated reasoning was competitive: a longer series is a better test. It was also true that every first-round series now offered broadcasters up to two more games, and nobody involved pretended that was a coincidence.
One side effect is permanent. Championship records from before 2003 are not directly comparable with those after it. The 2001 Lakers went 15-1, one loss like the 2017 Warriors, but needed only 15 wins because their first round was best-of-five. The 1983 76ers, whose center Moses Malone predicted three sweeps and nearly got them, won the title at 12-1 in a format with a first-round bye.
The 1-to-16 argument and why the conference brackets survive
In 2014 Phoenix won 48 games and missed the Western playoffs. Atlanta got into the Eastern bracket that spring at 38-44. Every time the conferences drift that far apart, the same proposal comes back. Take the 16 best records in the league. Seed them 1 through 16. Ignore conferences.
The appeal is obvious. The Finals would more often match the two best teams, since nothing would stop them from landing on opposite halves of the bracket. Weak qualifiers would disappear. The WNBA made a version of this change in 2016, dropping conference seeding in favor of the eight best records overall, and the details are in our guide to the WNBA's playoff format.
The NBA has looked at it and stopped short. Commissioner Adam Silver said in 2018 that the idea had received serious attention in the league office and that the league was not ready to adopt it. The version he described would have kept eight qualifiers from each conference and then seeded those 16 teams on one ladder. The obstacles he named were travel and the schedule.
Travel first. A conference bracket keeps early-round series regional: Eastern teams fly to Eastern cities. Seed the league 1 to 16 and a first-round series could pair Boston with Portland, with a 2-2-1-1-1 pattern sending both teams back and forth across three time zones. Late tipoffs on one coast and early ones on the other are a television problem as well as a physical one.
The schedule is the deeper issue. NBA teams play 52 of their 82 games inside their own conference. Records built on different sets of opponents are not strictly comparable, and ranking all 30 teams on one ladder would be fair only if everyone played a balanced schedule. Balancing it means more cross-country trips in the regular season, which runs against a decade of effort to cut travel and back-to-backs.
There is a third consideration the league mentions less. Conference finals are valuable properties, with trophies and histories of their own and two guaranteed series a year to sell. A single 16-team ladder would replace them with generic semifinals.
The fair summary is that 1-to-16 would produce a somewhat better bracket in the years the conferences are lopsided and a worse travel calendar every year. The league has judged that trade not worth making, and the imbalance between East and West has shifted often enough to keep the pressure from building. More on how the league is organized sits on our basketball hub.