Skip to main content
Menu
After This Weird NBA Season, We Have A Better Idea Of How Much Fans Matter

When NBA teams began reopening their stadiums to fans this winter and spring, players and coaches identified the crowd’s energy as the missing ingredient from empty arenas. From December to May, every team except for the Oklahoma City Thunder gradually permitted fans to return at limited capacity, but each team’s timing differed. As a result, some teams continued to play at home to an empty arena, while on the road they faced rivals with fans in the stands.

Players felt it. According to All-NBA guard Damian Lillard, an empty home arena was the reason for the Portland Trail Blazers’ mediocre performance at home:

“We come in these other buildings and just having that fan energy, that real energy in the building, it feels like a competitive NBA game when the crowd is there. It’s fun. It feels normal. It’s how the game is supposed to be played. Then we come home and it’s an empty, dead building. It’s a noticeable difference. You can feel it.”

Lillard’s Portland Trail Blazers were among the last teams to permit fans to return when they welcomed 1,939 fans back into the Moda Center on May 7 for their game against the defending champion Los Angeles Lakers. The two teams were tied for sixth in the Western Conference with six games to play. The game’s winner would take the season series and own the tiebreaker if the teams finished the season with the same record. Up until that game, the Trail Blazers were just 16-16 at home compared with 21-13 on the road.

The Trail Blazers beat the Lakers 106-101 and went on to win four of their five remaining games, including all three games in their home arena. And it turned out that they needed every single win, because the Lakers finished the regular season with a five-game winning streak. As a result of that May 7 home win, the Trail Blazers finished with the sixth seed, forcing the Lakers into the play-in tournament. “When they told us it was going to be 10 percent (capacity) I was like, I don’t know how much of a difference it’s going to make [in] such a huge building,” Lillard reflected after playing his first game in front of a home crowd. “But … as soon as we came out to warm up and the fans, you could feel how excited they were to be there. … It was a major, major difference.”

The Trail Blazers were representative of a leaguewide trend in which fans in the stands created a stronger home-court advantage. In 440 games played in empty arenas during the regular season, the home team won by 0.39 points on average — an advantage that is not statistically distinguishable from no home-court advantage at all. In 507 games with fans present, the average home-court advantage increased to 2.13 points, which is both significantly different from zero and very much in line with the 2.55-point home-court advantage observed in the prior six seasons.

For social scientists interested in explaining why home teams tend to outperform away teams, the 2020-21 regular season was an ideal setting to explore how the presence of home fans affects professional athletes. Social psychologists refer to this phenomenon as emotional contagion: the tendency to catch the feelings of others. When our peers become joyful, angry or anxious, our emotional state changes in response. Stronger emotions in our environment tend to elicit stronger emotional responses. Professional athletes are in this respect just like the rest of us. When surrounded by a sea of excited supporters, NBA players respond with greater energy on the court, leading to better results for the home team.

While this might seem like an obvious explanation for why teams tend to perform better at home, the mechanism is tricky to study empirically. Because better teams tend to attract more fans, it is nearly impossible to measure the impact of fan attendance on performance except in settings where attendance is artificially restricted. This is a statistical problem called endogeneity. In a model that uses fan attendance to predict home-court advantage, the analyst must always be concerned that teams that are more likely to win also draw more fans.

But the 2020-21 NBA regular season was a rare case when this statistical problem could be circumvented. And because arenas reopened at different times and at different capacities, we can make apples-to-apples statistical comparisons between teams playing in arenas with fans versus teams playing in empty arenas. 

In order to test the hypothesis that fans produce home-court advantage, my co-author Kieran Allsop and I collected data from Basketball-Reference.com for all of the games played during the 2020-21 regular season. Our measure of home-court advantage is the difference between the number of points scored by the home team and number of points scored by the away team.1 We compiled data on game attendance from a number of sources in order to account for potential discrepancies in reporting. Specifically, we cross-checked figures from Basketball-Reference.com with attendance reported in the NBA official game summaries and local news reports. This led us to exclude all home games for the Indiana Pacers, Miami Heat and Sacramento Kings, which continued to report zero attendance even after their arenas reopened. We also found 25 games for which the game summaries reported zero fans despite photographic, video or other documentary evidence that fans were present. These games were taken into account in the season’s overall average margin, but not in either fan-dependent calculation. 

In addition to comparing the average point margin for the home team with and without fans, we estimated a linear regression model to predict the marginal impact of the number of fans on home-court advantage. The model includes controls for the home team, the away team, the home team’s point margin in their most recent home game and the month in which the game took place. The model also takes into account the potential for endogeneity by using the maximum allowable attendance in the home arena as an instrument for the actual attendance at each game.2

The results of the instrumental variables regression model support the claim that the fans in the stands have a large impact on home-court advantage. According to the model, every additional 1,000 fans above zero predicts a 1.4-point increase in the home team’s margin of victory, an effect that is statistically significant.3

The setting also makes it hard to come up with alternative explanations for the observed effect. An away team had to travel just as far with and without fans in the arenas. And it’s hard to imagine that arenas with a few thousand cheering fans are interfering with a visiting team’s on-court strategy and communication as much as an arena packed with 20,000 screaming fans might. Justifications for home-court advantage that emphasize how home fans bias the judgment of referees in favor of the home team similarly seem more credible in games with full stands and deafening noise than in games in which attendance is limited to a small percentage of total capacity. Further, when we examine the same instrumental variables regression models described above, but replace the point margin with the difference in the number of fouls called, each additional thousand fans has a minuscule and statistically insignificant effect.

Just how important are the fans to home-court advantage? A thought experiment can help put the impact into context. On average, a home team with fans beat its opponent by 1.74 more points than a home team without fans during the 2020-21 regular season. According to FiveThirtyEight’s RAPTOR plus-minus rating, New York Knicks guard RJ Barrett was more-or-less an average NBA player this season.4 Imagine that the following offer was made to the Knicks front office before their last home game of the season against Boston, a game they needed to win to secure home-court advantage in the first round of the playoffs: You can replace Barrett with All-Star guard Zach LaVine of the Chicago Bulls, but in exchange, you must play in an empty Madison Square Garden. Given that the Knicks had played 40 percent of their home games to that point without fans, they might be tempted to add the All-Star. But according to the data, the Knicks would be better off rejecting the deal.

What if, instead, the Knicks could choose to play in Boston’s TD Garden in front of Celtics fans and, in exchange, play All-NBA guard Luka Dončić of the Dallas Mavericks instead of Barrett? Same answer. Rather than giving their 2.13-point home-court advantage to the Celtics, the Knicks should stick with Barrett and their home fans.

Does the effect of fans on home-court advantage carry over into the playoffs? It is hard to say. Conclusions drawn from empty and near-empty arenas during the 2020-21 regular season may not apply to sold-out arenas in pressure-packed playoff series. However, despite home teams winning just 16 of their first 32 games in this year’s playoffs, they have gone 27-21 since. Through the first 80 games of the postseason, the average home margin has been 2.93 points, which is within 2 points of the average home margins in eight of the past 10 non-bubble playoffs.

In short, the 2020-21 regular season proved what any passionate fan already knows: The fans matter. A lot. Players on the court feed off the energy of the crowd. If Game 7 of the NBA Finals ends up being won by the home team by just a point or two, we cannot definitively conclude that the fans in the stands made the difference. But we can say with confidence that they did their part in bringing their hometown team the 2020-21 NBA championship. 

Check out our latest NBA predictions.

Footnotes

  1. For the purposes of this analysis, we treated the Toronto Raptors’ games at Amalie Arena in Tampa, Florida, as home games since, due to Canadian COVID-19 restrictions, the team elected to relocate its 2020-21 “home” games to that stadium.

  2. The logic behind instrumental variables regression is as follows: If a researcher is concerned that there is an unobservable omitted variable — for example, the expected performance of the home team — that could generate a spurious correlation between an endogenous independent variable (home game attendance) and a dependent variable (the home team’s point margin), the researcher must identify a different exogenous variable that is correlated with the endogenous independent variable and uncorrelated with the omitted variable. This variable is called an instrument. The instrumental variables regression model uses the fact that the exogenous variable is uncorrelated with the omitted variable to produce an unbiased estimate of the impact of the endogenous independent variable on the dependent variable. In our case, the fact that the maximum allowable capacity at arenas during the 2020-21 season was strongly correlated with the actual number of fans in attendance, but attendance caps were unrelated to the expected performance of the home team in any individual game, makes it a good instrument to use in our model.

  3. At the 0.05 level.

  4. His regular-season RAPTOR was -0.02, while his full-season mark was a league-average 0.0.

Scott Ganz is a research fellow in economic policy studies at the American Enterprise Institute and a visiting assistant professor at Georgetown University’s McDonough School of Business.

Comments