NBA Tiebreaker Rules: How Tied Teams Are Seeded
Head-to-head record breaks a two-team NBA tie; among three or more, division leaders go first. The order down to point differential, and how draft ties go.
By CricketTaken EditorialPublished Updated NBA Season, Playoffs & Draft22 min read
- First tiebreaker, two teams
- Head-to-head record
- First tiebreaker, three or more teams
- Division leader
- Last statistical tiebreaker
- Point differential
- Draft-order ties
- Random drawing
On April 14, 2024, the last day of the NBA regular season, Orlando, Indiana and Philadelphia all finished 47-35. One of them became the Eastern Conference's fifth seed. One became the sixth. One dropped to seventh and into the play-in tournament. No extra game was played to sort them out, and none ever is. A list in the league's procedures had done the job before the arenas emptied.
The list is short. When two NBA teams finish with the same record, the tie is broken by head-to-head record first. If the season series was split, a team leading its division beats one that is not. After that come division record, for teams in the same division, then conference record, then record against postseason teams in the same conference, then record against postseason teams in the other conference, and last of all point differential. A tie among three or more teams uses a slightly different list that puts division leaders at the top. If every step fails, the league holds a random drawing.
The NBA tiebreaker rules in order, from head-to-head to point differential
Every NBA team plays 82 games, so two teams tied in winning percentage have identical records. That is the only situation the procedure addresses. "Games behind" is a convenience for newspapers and has no standing in the rules.
| Step | Two teams tied | What it measures |
|---|---|---|
| 1 | Better winning percentage in games against each other | The season series, two to four games |
| 2 | Division leader wins the tie over a team not leading a division | A division title |
| 3 | Division won-lost percentage, only if both teams are in the same division | 16 games |
| 4 | Conference won-lost percentage | 52 games |
| 5 | Winning percentage against postseason teams in own conference | Games against the conference's top ten |
| 6 | Winning percentage against postseason teams in the other conference | Games against the other side's top ten |
| 7 | Point differential against all opponents | All 82 games |
There is a logic to the order. It starts with the games the two teams played against each other, widens to the opponents they have most in common, widens again to the conference, and only at the very end looks at margin of victory. The league prefers wins to points at every stage. A club that won its games by one point is treated the same as a club that won them by twenty until nothing else is left to compare.
Three things about the procedure are worth knowing before the steps themselves.
It is applied once, to the final standings. The standings pages that reorder teams every night in March are running the same list on incomplete data, which is why two teams can swap places without either one playing.
There is no tiebreaker game. Baseball settled ties with an extra game for decades and abandoned the practice in 2022. The NBA staged a few tiebreaker games of its own in its first decade, as late as 1957, and has relied on a list in the decades since.
And the same rules govern every position in a conference. They decide who is the top seed, who gets home court in a first-round series between the fourth and fifth seeds, who finishes sixth and avoids the play-in, and who finishes tenth instead of eleventh. How those seeds feed the bracket is covered in our guide to the NBA playoff format.
The two-team ladder, worked through for a sixth seed
Phoenix and New Orleans both went into the final Sunday of the 2023-24 season with a chance at sixth place in the West. The Suns won in Minnesota. The Pelicans lost at home to the Lakers. Both finished 49-33.
Step one settled it. Phoenix had won the season series, so the Suns took the sixth seed and a guaranteed place in the first round. New Orleans became the seventh seed, lost its first play-in game, won its second, and entered the bracket as the eighth seed with a banged-up roster. Same record, very different week.
Most two-team ties end there, because most season series produce a winner. The ladder below shows what happens when they do not.
- Head-to-headCompare the results of the games the two teams played against each other. A 2-2 split moves the tie down.
- Division leaderIf one team is leading its division and the other is not, the leader takes the higher seed.
- Division recordUsed only when both teams are in the same division. Sixteen games each.
- Conference recordWinning percentage in the 52 games against the team's own conference.
- Postseason opponentsRecord against the top ten in the team's own conference, then against the top ten in the other.
- Point differentialTotal points scored minus total points allowed across the season.
- Random drawingThe last resort if every statistical step is level.
The order in the NBA's published procedure. Each step is used only if every step above it leaves the teams level.
Carrying the example down the list
Suppose the facts had been different. This is an illustration, not a real season. Two Western teams finish 46-36, one from the Pacific Division and one from the Southwest, and they split their four meetings 2-2.
Step two asks whether either team is leading its division. Neither is. Step three is skipped, because it applies only to teams in the same division. Step four compares conference records. Say one club went 30-22 against the West and the other 28-24. The first team is seeded higher.
Notice what that implies. Both teams won 46 games, so the club with the weaker conference record must have been better against the East, 18-12 against 16-14. The procedure does not care. Games inside the conference count for more in a tie than games outside it, on the reasonable ground that conference opponents are the ones both teams faced and the ones they are competing with for seeds.
A real step-three case came on that same April day in 2024. Sacramento and Golden State both finished 46-36 and had split their season series 2-2, with three of the four games decided by a single point. Neither led the Pacific Division. The Kings had the better record inside the division, so they were ninth and the Warriors tenth, and the play-in game between them was played in Sacramento. The Kings won it.
Why three or more tied teams use a different list
Head-to-head works for two teams. With three it can fall apart. Team A wins its series with Team B, B wins its series with C, and C wins its series with A. Nobody has an edge that the other two would recognize, and comparing pairs gets the league nowhere.
So the multi-team procedure pools the games. It asks one question of each tied team: what was its winning percentage in all the games played among the tied group?
| Step | Three or more teams tied |
|---|---|
| 1 | Division leader wins the tie from a team not leading a division |
| 2 | Better winning percentage in all games among the tied teams |
| 3 | Division won-lost percentage, only if all the teams are in the same division |
| 4 | Conference won-lost percentage |
| 5 | Winning percentage against postseason teams in own conference |
| 6 | Point differential against all opponents |
Two differences from the two-team list stand out. The division-leader test has moved to the top, ahead of any results on the floor. And the head-to-head step has become a combined record.
That combined record is a percentage for a reason. Tied teams have not necessarily played each other the same number of times. A club might have met one of the others four times and the third team three times, for seven games in the pool. Another might have played eight. A 5-3 mark is a .625 percentage and beats 4-3, which is .571, even though the second team has fewer losses.
One caution on the later steps. The league's standalone procedures document gives the multi-team list six steps, running from postseason opponents in the team's own conference straight to point differential. Some published versions add the other-conference step that appears on the two-team list. The difference is nearly academic, since a tie among three teams that survives division leaders, pooled head-to-head and conference record is close to a statistical freak.
The restart rule when a multi-team tie is only partly broken
Here is the part that trips people up. The multi-team list is not run from top to bottom until every team has a place. It is run until the first step that creates any separation at all, and then it stops.
The procedure calls a tie either completely or partially broken, and a partial break is enough to halt the list. The team or teams no longer tied are assigned their seeds. The remaining teams go back to the beginning. If two are left, they start at the top of the two-team list, and that means their own head-to-head record, not the pooled one.
The three 47-35 teams from 2024 show it cleanly.
- Three teams levelOrlando, Indiana and Philadelphia finish with identical records, competing for seeds five, six and seven.
- Division leader testOrlando won the Southeast Division. Neither of the others led a division, so the Magic take the fifth seed.
- RestartIndiana and Philadelphia are still tied. The procedure begins again with only those two teams.
- Two-team head-to-headThe Pacers held the edge in the games between the two clubs and take the sixth seed.
- ResultPhiladelphia is seventh and goes to the play-in. Indiana goes straight into the first round.
Orlando, Indiana and Philadelphia each finished 47-35 in 2023-24. The sequence follows the league's multi-team procedure and its restart rule.
An illustration with letters makes the point. Teams A, B and C are tied and none won a division. A went 6-2 in games against the other two, and B and C each went 3-5. A gets the highest seed. B and C are still level on pooled record, so the two-team list is applied to B and C alone.
The restart has a consequence that surprises people every April. A team's standing against one rival can depend on whether a third team is tied with them. Take away Orlando from the 2024 example and nothing changes, because Indiana's head-to-head edge over Philadelphia would have decided a two-way tie as well. But arrange the results differently and it matters. A club can hold the two-team tiebreaker over a rival, lose a three-way tie to a division leader, and find the order below it rearranged by pooled results. Published scenarios in the last week of the season often run to dozens of lines for exactly this reason.
Two housekeeping rules sit alongside the restart. Ties that determine a division winner are broken before any other tie. And the result of that particular tie-break is used only to name the division winner. It is not carried over as an answer to any seeding question, which is then worked out from scratch.
The division-winner clause, the last job divisions do in seeding
A division title in the NBA no longer guarantees anything. Until 2015 it guaranteed a top-four seed. Before that, a top-three seed. What survives is a place on the tiebreaker lists: second for two teams, first for three or more.
The different positions produce different outcomes. Between two tied teams, results on the floor come first. If one club won the season series, it does not matter that the other won a division. The clause is reached only after a split.
Among three or more, the order flips. A division leader is lifted out of the tie before anyone looks at head-to-head results. Orlando in 2024 did not need to have beaten Indiana or Philadelphia. Winning the Southeast, the weakest division in the conference that year, was enough.
This is messy, and there is no elegant defense of it. A team can lose a season series to a division champion, be tied with it alone, and win the tie on head-to-head. Add a third club to the same tie and the division champion jumps to the front of the line. The inconsistency is real. It persists because ties among three teams are uncommon, because division titles need to mean something if divisions are to exist at all, and because nobody has made changing it a priority.
Two details of wording matter. The clause refers to a division leader, which at the end of the season means the division winner. And on the multi-team list the league specifies that it applies whether or not the tied teams are in the same division. A Southeast champion gets the benefit against clubs from the Central and the Atlantic.
The clause does nothing when both tied teams won their divisions, or when neither did. Then the list simply moves on. The six divisions and their membership are set out in our guide to how many teams are in the NBA.
Record against playoff-eligible teams and why the late steps are rarely reached
Steps five and six on the two-team list ask how each team did against good opponents. The group is defined by the final standings. As the league now words it on its standings pages, it means teams eligible for the postseason, including any team that finished tied for one of the ten best records in its conference. Play-in teams count.
The parenthetical is doing careful work. Without it there would be a circularity: the tiebreaker depends on who qualified, and who qualified might depend on the tiebreaker. By counting every team tied for a place, the league makes the group knowable before any tie is broken.
Step five looks at those opponents in the team's own conference. Step six looks at the equivalent group on the other side of the league. A Western club's games against the East's top ten are few, usually about 20, so that step turns on a small sample.
- 16Games against division rivals
- 52Games against own conference
- 30Games against the other conference
- 2Meetings with any one team from the other conference
Totals for one team in a standard 82-game NBA regular season.
These steps are rarely reached for a simple reason. To get to step five, two teams must have the same overall record, a split season series, no division leader between them and an identical conference record. That last condition is the hard one. Same overall record and same conference record means the same record outside the conference too, a coincidence across three separate columns.
Point differential, step seven, is rarer still. It exists so that the list has a statistical end before the drawing. No one inside the league plans around it, and that is a good thing. A tiebreaker based on margin gives teams a reason to keep starters on the floor in games that are already decided, which is the behavior the league discourages everywhere else.
Uneven season series: three games against some opponents, four against others
Step one assumes a season series has a winner. Whether it can fail to have one depends on how many times the teams met, and the NBA schedule does not treat every opponent alike.
A four-game series can finish 2-2. A three-game series cannot. So for any pair of conference opponents that met three times, head-to-head always produces an answer and the tie never goes further down the list.
That has an awkward side. In a three-game series one team hosts twice. The club with the extra home date has a real, if modest, advantage in the single most important tiebreaker, and it got that advantage from the rotation in the scheduling formula, not from anything it did. Coaches know which opponents they see three times. The formula behind those numbers is explained in our guide to the 82-game schedule.
The four-game split is where the lower steps earn their place. Division rivals always meet four times, so a tie between two teams from the same division is the likeliest to reach step three, division record. Sacramento and Golden State in 2024 are the proof.
Since 2023 there has been one more wrinkle. NBA Cup group games and the knockout quarterfinals and semifinals count in the regular-season standings, so they count in season series too. Which opponents a team meets a fourth time, and where, can depend in part on the Cup draw. The Cup final is the exception. It is not a regular-season game and has no effect on any tiebreaker.
Teams from opposite conferences meet twice, once in each building. A 1-1 split is common. It matters in only one place, described below: the Finals.
Why the play-in raised the stakes of a tiebreaker
Before 2021 a tiebreaker in the middle of a conference decided a seed number and sometimes home court. The difference between sixth and seventh was a different first-round opponent. Now it is the difference between a week of rest and a game that can start a slide out of the postseason.
| Line in the standings | What the higher team gets | What the lower team gets |
|---|---|---|
| 6th over 7th | A guaranteed first-round series | The play-in, though with two chances to win one game |
| 7th over 8th | The first play-in game at home | The same game on the road |
| 8th over 9th | Two chances to win one game | An elimination game, and a second one if it wins |
| 9th over 10th | The elimination game at home | The elimination game on the road |
| 10th over 11th | A place in the play-in | The end of the season |
Every one of those lines can be drawn by a tiebreaker, and in 2024 the sixth-seventh line was drawn that way in both conferences. Indiana over Philadelphia in the East. Phoenix over New Orleans in the West, where Sacramento over Golden State at ninth-tenth also put an elimination game in one building instead of another. The structure of those games is set out in our explainer on how the play-in tournament works.
The line with the most at stake is tenth against eleventh. One team plays on. The other is finished, with the same record, because of a game it lost in November.
This has changed how the last month of the season is played. A late-March meeting between two teams sitting eighth and ninth is worth more than one win in the standings, because it may also decide the season series. Announcers call these "two-game swings," and the phrase undersells it. A team that wins the game and the series gains a game in the standings and, in effect, a half-game cushion for the tie it now cannot lose.
The play-in has also made the tiebreakers more visible. A casual fan once had little reason to learn them. Now a tie that decides who rests and who plays on can come down to the final day, as it did in both conferences in 2024.
Home court in the Finals between teams from different conferences
The Finals is the one series in which the seeding procedure does not directly apply, because the two teams were never in the same standings. Home-court advantage goes to the team with the better regular-season record. Seed number is irrelevant. A second seed with 58 wins hosts a first seed with 57.
When the records are identical, the league breaks the tie, and the first test is the familiar one: the regular-season games between the two clubs. There are only two of them, one in each arena, so the series is either a sweep or a split.
The best-known case is 1998. Chicago and Utah both finished 62-20. The Jazz had won both regular-season meetings, so the Finals opened in Salt Lake City and the Bulls, with the best player in the world and five titles in seven years, were the road team. They won in six anyway, clinching in Utah.
What comes after a 1-1 split is less well documented, and it is better to say so than to guess. The seeding lists are written for teams in the same conference. Division record and conference record mean nothing when the clubs come from opposite sides of the league, and the division-leader test would usually be a wash, since both finalists have often won their divisions. Accounts of the Finals procedure generally give record against the opposite conference as the next comparison. That is plausible. It measures each team against the other's half of the league over 30 games. Treat it as likely, not certain.
The reason this rarely comes up is that identical records across conferences are uncommon, and identical records with a split series rarer again. The reason it matters when it does is Game 7. The team that wins the tie would host the last game of the season.
The other end of the standings: draft lottery ties and the random drawing
Everything above rewards the team with the better case. At the bottom of the standings the league refuses to make a case at all. Teams that finish with identical records are ordered for the draft by a random drawing.
The drawings are held at the league office shortly after the regular season ends, under the supervision of an outside accounting firm, and the results are announced the same day. In April 2026 six separate ties were broken that way, including a three-team tie.
Under the lottery used through the 2026 draft, the drawing did two things for teams tied inside the lottery. Their chances were pooled and divided evenly, so that neither club was favored in the lottery itself. And the drawing fixed which of them would pick first if the lottery did not move either one up. For tied playoff teams, whose picks are not subject to the lottery, the drawing simply set the order. In the second round the order among tied teams is reversed, so the club that loses the drawing gets something back.
The reason for a drawing is easy to see once the alternative is spelled out. Any tiebreaker based on results would hand the better draft pick to the team with the worse head-to-head record, or the worse conference record, or the worse point differential. That would give two bad teams meeting in April a reason to lose to each other. Chance removes it.
The lottery itself is changing. In May 2026 the Board of Governors approved a new system for the 2027, 2028 and 2029 drafts. It widens the lottery to 16 teams by tying entry to play-in position: the ninth and tenth seeds in each conference and the losers of the two games between seventh and eighth are in, along with the ten clubs that miss the play-in altogether. Each team gets three, two or one lottery balls depending on where it finished. Under it the three worst records receive two balls, one fewer than the teams just above them. That makes the line between third-worst and fourth-worst a meaningful one, and a tie across it will have to be resolved. The long-standing method for that is a drawing, which keeps results on the floor out of it.
Our explainer on how the draft lottery works covers the drawing itself, and the rest of the selection process is in our guide to how the NBA draft works.
NBA Cup group tiebreakers, a separate list built on point differential
The in-season tournament the league introduced in 2023 has its own tiebreakers, and they are not the ones described above. Each team plays four group games. With so few results, ties are routine, and the long regular-season list would have nothing to work with.
| Order | NBA Cup group play | Regular-season seeding, two teams |
|---|---|---|
| 1 | Head-to-head record in the group | Head-to-head record |
| 2 | Point differential in group games | Division leader |
| 3 | Total points scored in group games | Division record, same division only |
| 4 | Regular-season record from the previous season | Conference record |
| 5 | Random drawing | Record against postseason teams, then point differential, then a drawing |
Point differential is second. In the regular season it is seventh. That single change produced the strangest basketball of the first tournament in 2023, with teams leaving starters in and pressing for baskets in the final minute of games they led by 20. Opposing benches did not enjoy it.
The league made one adjustment for 2024. Points scored in overtime no longer count toward the differential or the points-scored tiebreaker. A group game that goes to overtime is recorded as a margin of zero for both teams, which removes the incentive to run up a score in an extra period.
It is fair to ask why margin is acceptable here and kept at arm's length in the regular season. Necessity is the answer. Four games cannot produce division records or conference records worth comparing. Head-to-head is frequently circular in a five-team group. Something has to separate teams, and points are the only thing a four-game schedule generates in quantity.
The two systems do not touch each other. A group game counts as a win or a loss in the regular-season standings and as one game in a season series, like any other. Its margin counts toward the season-long point differential that sits at step seven. The Cup's own ranking is used only to decide who advances to the knockout round. The tournament's format is laid out in our guide to the NBA Cup.
How to read April's standings: clinched tiebreakers and magic numbers
By the first week of April most season series are complete, so most two-team tiebreakers are already known. A series is clinched the moment one team cannot be caught in it: three wins in a four-game set, or two in a three-game set.
- 83Starting figure for a magic number without the tiebreaker
- 82Starting figure when the tiebreaker is already held
- 3Wins that clinch a four-game season series
- 2Wins that clinch a three-game season series
Definitions, not statistics. The magic number is the combined total of one team's wins and a rival's losses that guarantees the first team finishes ahead.
The magic number is where a clinched tiebreaker shows its value. To finish ahead of a rival outright, a team needs more wins than the rival can possibly reach. The usual formula is 83 minus the leading team's wins minus the trailing team's losses. If the leader already owns the tiebreaker, finishing level is good enough, and the starting figure drops to 82.
An illustration, with invented records. A team at 45-30 leads a rival at 43-32. Without the tiebreaker its magic number is 83 minus 45 minus 32, which is 6. Any combination of its own wins and the rival's losses adding up to six ends the race. With the tiebreaker the number is 5. Holding the tiebreaker is worth exactly one game, every time, in a two-team race.
Multi-team races do not reduce to a single number. Because of the restart rule, a team's position can depend on which other clubs finish level with it, and that can depend on games it is not playing in. This is where the league's own material is more reliable than arithmetic done on a napkin. In the final weeks the NBA's communications office publishes clinching scenarios before each night's games, listing the exact results that would lock in a seed, a division title or a play-in place. Team beat writers reprint them. They are worth reading in full, since the last line is often the one that happens.
One habit helps. When two teams are close in the standings, look first at whether their season series is finished and who won it. If it is unfinished, the remaining game between them is the most important date either has left. If it is split, check whether they share a division. Those two questions answer the large majority of ties before the long list is ever needed.
More guides to American leagues and their rules are collected on our United States hub.