It is often speculated by NBA fans and talking heads alike that the impact of home-court advantage has become less significant in the modern era than it was in the days of old. Various arguments have and continue to be made in this regard. For one, many cite the increasing marketability of the league, via ESPN, national TV broadcasts, and modern technology in general, as one reason fan attendance has become less regional and more “bandwagon”-oriented. Fans, nowadays, are less inclined to merely cheer for teams in their own cities; unlike, say, thirty years ago, when games broadcast in a certain city would include the home team almost exclusively, ESPN, TNT, ABC, and NBA League Pass have allowed for fans all across the world to watch whichever team interests them just about whenever they like. This, in theory, would create a phenomenon in which larger and larger proportions of fans in cities are likelier to attend games to root for (and, in turn, prop up) the opposing team, and diminish the inherent home-court advantage that the home team holds.

Furthermore, many cite the rise of hubs such as “SeatGeek” and “Stubhub”, sites that have made ticket distribution far more frequent and easy, as a reason fan attendance has become a lot less homogenous. It is a whole lot easier to use one of these databases not only to find a favorable deal, but to book an adjacent flight, plan ahead of time, and even distribute tickets to fans in other cities. This kind of secondary distribution of tickets can take the authority out of the hands of team officials and employees and into the hands of independent buyers and sellers, for whom money trumps all. This would further trumpet the idea that crowds, on any given night (and particularly for games defined as especially ‘important’, for whatever reason), have become far less homogenous and regional.

Another theory states that home-court advantage has become less significant because the NBA, as a whole, has become a whole lot more three-point heavy. This idea stems from the fact that, for the first time ever, recent NBA seasons have seen teams take more three-pointers than free-throws on various monthly occasions. Overall, the rate of three-pointers being taken in the NBA is greater nowadays than it has ever been, by an uncanny margin. And on top of this, the rate of free throw attempts is down (largely because threes, quite obviously, are less likely to attract fouls than drives to the rim) by a significant margin. This is important because one of the main “advantages” of having home-court has often been speculated (and supported by numerous professional studies) to be via referees, who may be unintentionally affected by home crowds. Thus, with long shots meaning less contact and fewer fouls, there is less chance for the basic humanity of referees to influence the outcomes of games.

Lastly, one theory also attributes home court advantage to the advance in technology, but for a wholly different reason. Because modern technology allows a far more advanced and analytical approach to tracking player health, teams nowadays can be smarter about resting players and putting them on better, more analytically sound regimens crafted towards preserving their season-long health. Also to this point, because the grind of a season has become quite a bit less grinding (via the development of commercial planes (vs charter flights), biometrics, an increase in mechanisms that work towards the immediate rest and recovery of players, etc), the road is a whole lot less taxing on players than it once was. As a result, players are not as often at such a disadvantage, in terms of mental and physical health, when playing on the road, both because it has become far less daunting and because they are better prepared for it.

Ultimately, these theories, in amalgam, have contributed to widespread speculation that home-court advantage has dwindled to a statistically significant degree. And while these theories are certainly heavily researched and thought out, I decided to set out and research whether they are indeed affecting teams’ home performance as significantly as speculated. To do so, I researched NBA teams’ home and road winning percentages from two different five-year periods. The first of these two five-year periods included all NBA seasons from 1985-86 to 1989-90. I chose this period because it was firmly before networks such as ESPN started broadcasting NBA games on such a mass level, before NBA League Pass was a thing, before three-pointers began dominating NBA offenses, and before hubs such as “SeatGeek” and “Stubhub” took off. The other five-year period I collected data from was simply the most recent one: 2010-11 through 2014-2015.

I determined that my question of interest would be the following: “Given various technological advancements and societal changes that have affected the day-to-day proceedings of NBA players and personnel alike, has home-court advantage in the modern NBA declined to a statistically significant and notable degree?” To investigate this question, I would try to find out whether the performance of NBA teams at home in the modern area, as opposed to thirty or so years ago, is deviant enough to suggest that teams’ abilities are different. As such, my null hypothesis for the study states that the ability of NBA teams at home, on a league-wide scale, is no greater or lesser in the modern era than it was from 1985-1990. On the other hand, my alternate hypothesis states that NBA teams from 1985-1990 had a greater ability to win games at home than did teams in the modern era (2010-2015).

The following table displays all of the data relevant to my study:

League Year League Home Record League Home-Win-%
1985-86 617-326 65.43%
1986-87 627-316 66.49%
1987-88 640-303 67.87%
1988-89 694-331 67.71%
1989-90 713-394 64.41%
2010-11 743-487 60.41%
2011-12 580-410 58.59%
2012-13 752-477 61.19%
2013-14 714-526 58.05%
2014-15 707-523 57.48%

Additionally, the following three visuals (dotplot, boxplot, chart of summary statistics) display the information, with “Past” denoting the five-year period from 1985-86 through 1989-90 and “Present” denoting the five-year period from 2010-11 through 2014-15:

As the chart of summary statistics shows, the mean for the ’85-90 period was 7.238% higher than that of the ’10-15 period (66.382% to 59.144%), the median was 7.90 percentage points higher (66.49% to 58.59%), each of the minimum, maximum, Q1, and Q3 values were 5-7% in favor of the “past” period, and the IQR, calculated by subtracting the Q1 value from the Q3 value, was 2.28 for the ’85-90 period, and 1.82 for the ’10-15 period, suggesting more variability for the earlier period.

Based on the dotplot, the distribution of home win-percentages for the past looks slightly skewed to the left (which is supported by the median being ever so slightly greater than the mean) and the distribution of home win-percentages for the present period looks fairly uniform (even with the mean a tiny bit greater than the median). Because the center (and the mean) of the past distribution is higher than the center of the present distribution (and even more simply, the minimum value of the past distribution is higher than the maximum value of the present distribution), the home win-percentages for teams in the ’85-90 period were clearly and typically higher than that those of teams in the ’10-15 period.

To actually parse this data and test my question of interest, however, I ran a simulation via an online statistical applet, using the following as my test statistic: difference in means (present-past)= -7.238.

I conducted one hundred trials of the simulation, assuming the ability of NBA teams to win games at home was the same in both eras, and the difference in means for each trial was recorded on the dotplot provided on the next page:

The dotplot shows the possible differences in mean performance that could occur simply by random chance, assuming that the ability of teams to win home games was the same in both eras. Because there was no simulated differences equal to or less than -7.238 percentage points, the p-value comes in at 0%. Assuming teams had the same ability to win games in both eras, this tells us that there is just about no possibility of obtaining a difference in mean performances less than or equal to -7.238 by mere random chance.

Since the p-value is 0, I reject the null hypothesis that NBA teams in both eras had the same ability to win home games. Instead, I have convincing evidence that teams held a better ability to win games at home in the 1985-90 time period.

Admittedly, perhaps the main issue with this investigation is the fact that my scope of inference is fairly limited. I have extremely strong evidence that teams were better at home in that specific time period (1985-90) than they are in the present, but I cannot quite say that the league held a superior home-court advantage in the past altogether (to do that, I would have to randomly select the years that qualify as the “past” and compare it to randomly selected years that qualify within the “modern NBA”–in other words, I’d have to run an experiment). This kind of conclusion is nonetheless valuable, however, as I have conclusive evidence in support of the fact that the home-court advantage acting all throughout the NBA has dwindled to a statistically significant degree in the past twenty-five years.

Lastly, though I have convincing evidence that NBA teams were better at home in the ’85-90 time period, I still don’t know for a fact the cause of the diminished advantageousness of home-court, as it presently stands. It could be any of the theories outlined in the introductory paragraphs, or it could be some kind of combination of them all. Even still, I can conclusively infer that these forces, in one form or another, are at play, and, because of them, the modern NBA is a noticeably different place just twenty-five years later.


Leave a Reply

Your email address will not be published. Required fields are marked *