By Brendan Kent
Home courts arenāt the fortresses they once were in the NBA. In a piece from 2015, ESPN NBA Insider Tom Haberstroh highlighted a general decline in home winning percentage since 1975, and a sharp decline since the 2011/12 season, with home win percentage falling from 61.2% in 2011/12 to 53.7% in 2014/15.
Haberstroh presents three theories for the decline:
Theory 1: The NBA is more 3-point heavy, less dependent on referees
Theory 2: Technology has allowed teams to handle the road better
Theory 3: Home crowds are no longer the sixth man
Theory 1 is grounded in the idea that home-court advantage stems at least partially from referee bias towards the home team. As Haberstroh notes, āNumerous studies have shown that referees may be involuntarily influenced by the home crowd.ā Most fouls are called in the paint, so referees have the most influence when the game is played in the paint. When a team takes a higher percentage of its shots from beyond the arc, referees have fewer opportunities to call fouls and favor the home team.
In determining whether Theory 1 has any clout, we can look at the relationship between a teamās home-court advantage and its 3-point attempt rate. For the purposes of this study, we will define both as such:
Focusing on the 26 seasons since 1989/90, we can visualize the decline home-court advantage:
Further, we can visualize the rise in 3-point attempt rate in the same time frame (the bump in 3-point attempt rate from 1994/95-1996/97 is due to the leagueās temporary shortening of the distance to the arc during that period):
The trends appear to coincideāas 3-point attempt rates rose from 1989/90 to 1996/97, home-court advantages decreased, and when 3-point attempt rates leveled off from 1997/98 to 2002/03, so did home-court advantages. 3-point attempt rates have been rising again since the 2003/04 season and so have home-court advantages. Ultimately, however, we need to determine if individual teams with higher 3-point attempt rates have a lower home-court advantages. We test this by regressing home-court advantage on 3-point attempt rate for all teams from 1989/90 to 2015/16. The results are plotted below:
The plot shows what is a statistically significant, negative relationshipāthat is, teams with a higher 3-point attempt rate tend to have a lower home-court advantage. Via the regression model, as 3-point rate increases by 1 percentage point, home-court advantage decreases by about 0.2 percentage points (i.e. the difference between Home Win % and Overall Win % decreases by about 0.2 percentage points). The p-value on this relationship 2.49e-12.*
The rise of the 3-pointer, it seems, is at least partially responsible for the decline in home-court advantage. Take the game out of the paint, take the influence out of the ref, take some advantage out of the home team.
*[Edited] Note: The relationship remains statistically significant when we control for season
Your conclusion (“The rise of the 3-pointer, it seems, is at least partially responsible for the decline in home-court advantage.”) is not justified from the data you present.
This is like the classic “pirates cause global warming” graphs: https://www.google.com/search?q=pirates+vs+global+warming&safe=off&espv=2&source=lnms&tbm=isch&sa=X&ved=0ahUKEwjJrLTUr_nSAhUKSSYKHTsmAc8Q_AUIBigB&biw=1745&bih=900
Note that I’m not saying you’re *wrong*, necessarily. But these graphs don’t show you’re correct.
More convincing would be a model that controls for season, showing that even after you account for the decreasing HCA baseline over time, teams with higher 3-point rates had lower HCA.
Hi David,
Great point – running the regression while controlling for season still yields a relationship of about the same magnitude with a p-value of 2.85e-05.
Awesome; much more convincing!