Home-court advantage in the numbers
Home teams win about three in five games, but the number hides a lot. This article converts the edge between wins and points, shows how noisy a single team's home record is, and works out what it is worth in a playoff series.
Home-court advantage is the most reliable regularity in basketball statistics. It appears in every league and at every level, it has survived rule changes and new arenas, and it is large enough to be built into every rating system, every forecast and every betting line. Yet it is easy to misuse, because the number that describes it, a home winning percentage, hides several things: how it relates to points, how much it varies from team to team, how it has changed over time and what it is worth when the games are arranged into a playoff series.
This article is the numbers companion to Home-court advantage: what it is, why it exists and how big it is, which covers the causes. Here the question is what the figures say, how to translate between the ways of stating them, how much weight a single team's home record can bear, and how the effect works through a season and a series. The arithmetic is simple enough to check with a calculator, and it explains why careful analysts treat the advantage as a constant to be applied everywhere, not as a trait to be admired in particular teams.
Two ways to say the same thing
The advantage can be expressed as a winning percentage or as a margin in points, and the two are linked. Over a long stretch of NBA history, home teams have won roughly 60 percent of regular-season games, and the typical edge in scoring margin has been on the order of two to three points per game. These are the same fact seen from two sides.
The link comes from the way single-game margins are spread out. In the NBA, the margin of a game varies around its expected value by something like a dozen points. If that is so, a home team that is expected to outscore an equal visitor by 3 points has a chance of winning that you can look up on a normal curve: about 60 percent. The table shows the approximate pairings, using that spread.
| Edge in points | Home win probability against an equal team |
|---|---|
| 2 | About 57 percent |
| 2.5 | About 58 percent |
| 3 | About 60 percent |
| 4 | About 63 percent |
The spread of a dozen points is itself an approximation, so treat the table as a guide to the size of things and not an exact equivalence. The useful lesson is that a few points translates to a big difference in winning percentage. A team that is exactly average on a neutral floor is a 50 percent proposition; put it at home against another average team and it becomes a 60 percent proposition. In a league where the difference between the best and worst teams in a season is a few dozen wins, ten points of winning percentage is large.
A different era, a different number
The size of the edge has not been constant. In earlier decades of the NBA, home teams won a noticeably larger share of their games, somewhere near two-thirds in some stretches, and the figure has drifted downward since. The causes are debated. Travel has become easier and more comfortable, with charter flights and better accommodations replacing commercial trips and long bus rides. Officiating is more closely reviewed, and replay and technology make calls less sensitive to crowd pressure. Training and sleep science have improved the preparation of visiting teams. And the league has become more uniform in style, with fewer quirky buildings and fewer quirky ways of playing, so a visitor is less surprised by what it finds.
None of that has eliminated the advantage. In the 2010s, home teams still won a share of games close to 60 percent, and the effect remains robust to a very wide range of conditions. What changed is the number, not the direction. If you are reading an old text that quotes a home winning percentage, check the era before you compare it with anything modern.
The same pattern appears in other sports and, in a more muted form, in basketball at other levels. College basketball is commonly said to carry a larger edge than the professional game, on the order of three to four points, partly because student sections press close to the court, and partly because travel for amateurs is less comfortable. High school games, played in gyms with small, loud crowds, have fewer reliable records, so claims about the size of the edge deserve caution. Women's basketball also shows a home edge, but because its seasons are shorter the numbers are noisier and the estimates less settled.
What happens when the crowd disappears
The best evidence about what produces the edge came from an accident. When the NBA completed its 2019-20 season in an enclosed site in Florida, with no fans in the arenas and no travel for the teams involved, the "home" team in each game was a designation on a schedule. The results were striking: the home teams in those games won at a rate close to 50 percent, and the scoring margin in favor of the nominal home side was close to zero.
Samples that small cannot be definitive, and the conditions in the bubble were unusual in many ways. But the pattern fits what researchers had suspected, that much of the home edge depends on the presence of a crowd, together with the familiarity and the travel that go with a real home game. The causes are examined in the companion article. The numbers lesson is that the advantage is not a fixed property of basketball like the size of the hoop. It is the product of conditions that can be removed, and when they are removed the number goes with them.
How much can one team's home record tell you?
Here is the part that surprises many fans. It is natural to look at a team's home and road records and conclude that some teams are special at home. A team that went 33-8 at home and 18-23 on the road looks like a team with a huge home edge. But a single season gives very little information about that, and basic arithmetic shows why.
A team plays 41 home games. Suppose its true chance of winning each home game is 60 percent. The standard deviation of its number of home wins is the square root of 41 times 0.6 times 0.4, which is about 3.1 wins, or roughly 7.7 percentage points of winning percentage. The road record has a similar spread. The difference between a team's home winning percentage and its road winning percentage therefore varies by about 11 percentage points from pure chance, even if the team's true home edge is exactly the same as everyone else's. A gap of 20 points between home and road winning percentages, which sounds dramatic, is less than two standard deviations away from an ordinary team's expected gap.
The result is that most of the variation in teams' home-versus-road splits in any one season is noise. A team that has a big split this year will, on average, have a smaller one next year. This is the regression to the mean described in Regression to the mean: why extreme seasons rarely repeat, and it is the reason sensible models give every team the same home bonus, or shrink any team-specific estimate heavily toward the league average. Some places probably do have a larger edge; altitude in Denver and the noise and size of certain college arenas are the classic candidates. But the evidence for any individual team in any individual season is weak, and you would need many seasons of data to separate a genuine effect from luck. The warning in Small samples in basketball: when a hot streak is noise applies in full.
Home court inside a rating
Every rating system has to deal with the advantage, and the mechanics are instructive. In an Elo system, the home team's rating is increased by a fixed number of points before the expected result is calculated, a bonus of the order of 100 points in public basketball versions, which turns a game between equals into a roughly 64 percent proposition for the home side. In a margin-based system, such as the type discussed in Net rating: the best single number for a team, each home game's margin is reduced by the home edge, and each road game's margin is increased, before team strength is estimated. The adjustment matters when schedules are lopsided. A team that played an unusual number of its early games at home looks better than it is unless the home edge is removed. The idea is the same one used to build a rating from scratch in How to build a simple team rating at home.
The adjustment is also built into betting lines, since the market prices the home edge into every spread. This article is not about wagering, and nothing here is a suggestion to bet. But the fact is a useful check on the size of the effect: when an evenly matched game is priced with the home team as a modest favorite, it reflects the same few points that the data show.
What the edge does to a schedule
Because the edge is the same for everyone, it matters most when schedules are uneven. Take an evenly matched team, one that would win half its games on a neutral floor. Give it a stretch of ten games in which seven are on the road. Using 60 percent at home and 40 percent on the road, its expected wins are 3 times 0.6 plus 7 times 0.4, or 4.6, rather than the 5.0 that a balanced stretch would produce. Four-tenths of a win is not much, but over a season the league's schedule makers deal out different mixes of back-to-backs, road trips and long homestands, and each team's run of dates has its own flavor. The construction of the calendar is described in How the NBA schedule is made: 82 games, arenas and television.
The edge also has a hidden effect on how to read a hot or cold month. A team that goes 9-1 with eight home games has done less than a team that goes 9-1 with eight road games, because an average team would be expected to win about 5.6 of the first ten and only about 4.4 of the second. Adjusting for it turns raw results into something closer to a measure of strength, which is the point of every rating system. And it explains a recurring pattern in the standings: early in a season, a team that has played many home games tends to have a better record than its quality warrants, and the record drifts back as the road games arrive.
What it is worth in a series
Now apply the number to the NBA playoff format. In a best-of-seven series, the higher seed has home court, which in the 2-2-1-1-1 format means it plays four of the seven games at home and the lower seed plays three. The format is described in How the NBA playoffs work: seeding, series format and home court.
Suppose two teams are exactly equal in strength and home teams win 60 percent of games between equals. A calculation over all the ways the series can unfold shows that the team with home court wins the series about 53 percent of the time. If the home edge were smaller, at 55 percent per game, the series edge would be about 51.5 percent; if it were larger, at 65 percent per game, it would be about 55 percent. So even a large home edge in each game produces a modest edge in the series, because the advantage largely cancels out: each team plays at home in some games and on the road in others, and the only net benefit comes from the one extra home game.
That is a useful correction to a common instinct. Home court is not decisive in a series between equals. It is worth roughly three percentage points of series probability. It matters more when the teams are close, because it can tip a coin-flip series, and it matters less when one team is clearly stronger. And the game-by-game edge still matters inside the series, because a crucial seventh game is played in one team's building. Home teams have won the large majority of Game 7s in NBA history, which is why the stories in Game 7 classics: the greatest decisive games in NBA history so often involve a road team's upset, which stands out because it is rare.
Reading home-court numbers well
A handful of rules will keep you from misreading the figures.
- Compare like with like. A home winning percentage from the 1980s is not a home winning percentage from the 2010s, and a college figure is not an NBA figure.
- Do not credit a team with a special home edge on one season's record. Use several seasons, or assume the league average.
- Remember that good teams win more everywhere. A team's home record is high partly because the team is good. The home edge is the difference between home and road performance after accounting for quality.
- Treat the edge as a constant when adjusting for schedule. A team's road-heavy month is harder than a home-heavy one, and a rating that ignores that is misleading.
- Keep the series number in proportion. Home court in a seven-game series is worth about three percentage points between equals, not a guarantee.
What to take away
Home teams win about three games in five in the long run, which corresponds to an edge of two to three points a game. The number has drifted down over the decades, shrank to almost nothing when the crowd was removed and varies far less from team to team than raw home-and-road splits suggest. In a rating it is a constant; in a seven-game series between equals it is worth about three percentage points. For a fan the practical meaning is humility: the building matters, but a lot less than the quality of the players in it. You can see how the season's home and road results are adding up on the standings page and in the scores.