Illinois Wesleyan University
Digital Commons @ IWU Digital Commons @ IWU
Honors Projects Economics Department
2014
Estimating the Effect of Home Court Advantage on Wins in the Estimating the Effect of Home Court Advantage on Wins in the
NBA NBA
Jason Kotecki
Illinois Wesleyan University
Follow this and additional works at: https://digitalcommons.iwu.edu/econ_honproj
Part of the Economics Commons
Recommended Citation
Kotecki, Jason, "Estimating the E8ect of Home Court Advantage on Wins in the NBA"
(2014).
Honors Projects
. 124.
https://digitalcommons.iwu.edu/econ_honproj/124
This Article is protected by copyright and/or related rights. It has been brought to you by Digital
Commons @ IWU with permission from the rights-holder(s). You are free to use this material in any
way that is permitted by the copyright and related rights legislation that applies to your use. For
other uses you need to obtain permission from the rights-holder(s) directly, unless additional rights
are indicated by a Creative Commons license in the record and/ or on the work itself. This material
has been accepted for inclusion by faculty at Illinois Wesleyan University. For more information,
please contact [email protected].
©Copyright is owned by the author of this document.
Estimating the Effect of Home Court Advantage on Wins in the NBA
Jason Kotecki
Faculty Advisor: Robert Leekley
April 25, 2014
Illinois Wesleyan University
Home Court Advantage 2
Abstract
What is the effect of home court advantage in the National Basketball Association
(NBA)? Based on the Economic Theory of Professional Sports and the concept of
shirking, teams should perform better at home than they do on the road. Descriptive
statistics support this expectation. It is hypothesized that a home court advantage is due
to fan attendance, field goal and free throw percentages, and fouls called by the referee.
Following every NBA team and every game played over a three-year span (2008-2011),
this paper estimates the probability of producing a win at home based on the
aforementioned variables. Using a logit regression analysis, it is found that a one
standard deviation increase in attendance increases the home team’s chances of winning
the game by 2.7%. A referee bias is also found, increasing the home court advantage for
NBA teams.
Keywords: Home Court Advantage, Basketball, Attendance, Referee-Bias
Home Court Advantage 3
I. Introduction
During the 2012-13 National Basketball Association (NBA) season, the Houston
Rockets compiled a record of 45-37 and were the eighth seed out of eight in the Western
Conference playoffs. The Rockets had a road record of just 16-25, but went 29-12 at
home. The Utah Jazz won 30 games at home last year, but only 13 on the road. Cooper
(2013), a freelance writer for the Atlanta Hawks, found that in the 2013 playoffs,
“Through the first 20 games of the First Round (heading into games of Friday, April 26),
the home team had won 17 times.” How can this major difference between records be
explained? Home court advantage is the answer. Carron et al., (2005) define home court
advantage as, “the consistent finding that home teams in sport competitions win over
50% of the games played under a balanced home and away schedule.” Each NBA team
plays 41 games at home and 41 games on the road each year. Each team also plays each
team at least once at home and once on the road. Based on Carron’s definition of home
court advantage, each team is expected to win at least 21 games at home each year.
But, where does this home court advantage come from? This paper seeks to
answer that question by using logit regressions to analyze the determinants of winning at
home. The dependent variable is a dummy variable for wins. I hypothesize that home
court advantage exists, and that it can be explained mostly by fan attendance,
performance statistics, and referee bias.
Section II examines previous literature on the subject of home advantage. Section
III lays out the theoretical framework, while Section IV defines the empirical models.
Section V gives descriptive statistics and Section VI reports the results. Finally, Section
VII offers concluding thoughts.
Home Court Advantage 4
II. Literature Review
The home advantage is a well-established concept in the literature. There is no
debate over whether or not it exists; rather much of the literature examines the causes and
effects of home court advantage. Carron et al. (2005) offer a conceptual framework for
analyzing home court advantage. Most of the variables in the models presented in this
paper come from the Carron et al. article. They present variables representing game
location factors, critical psychological and behavioral states, and performance outcomes.
This study builds on their work.
A. Game Location Factors
Game location factors, as defined by Carron et al. (2005), are crowd factors,
learning/familiarity factors, travel factors, and rule factors. Crowd factors acknowledge
that generally competitors at home have more support from spectators than do visitors,
and thus have a greater home court advantage. Schwartz and Barsky (1977) compare
home advantages among baseball, football, hockey, and college basketball. They find
that the home advantage is greatest in indoor sports and primarily has to do with support
of the home crowd. The literature dealing with crowd factors and attendance is extensive
(Forrest et al., 2005; Greer 1983; Nevill 1999; Nevill et al., 1996; Smith 2005). All of
these studies report that fan attendance has a positive effect on wins. Nevill et al. (1996)
specifically find that absolute crowd size is positively related to home advantage.
Salminen (1993) is the only study that finds fans’ cheering for the home team is not
related to greater home team success.
Another aspect of Carron’s game location factors is travel. Ashman et al. (2010)
find that when playing games at home on consecutive nights the home team plays poorly
Home Court Advantage 5
in the second game when the visitor has one or two days rest. But, Nutting (2010), finds
that game frequency itself has a negative impact on wins, so the home factor does not
matter as much as the frequency. Therefore, a days of rest variable, measuring the
amount of rest each home team has before each competition, is included in the models.
B. Critical Psychological and Behavioral States
The next set of variables that Carron et al. (2005) present in their conceptual
framework are critical psychological and behavioral states. These two are related and
deal with how coaches, competitors and officials affect the outcome of the game. An
interesting example is the possibility of a referee bias. Do referees call fewer fouls on the
home team, and are they influenced to make calls based on the home crowd reaction?
Page and Page (2010) study the roles of referees in determining home field advantage in
European soccer. They find that there is a significant impact of the referee on home field
advantage. This means that some referees cave under the pressure of a large and
boisterous crowd, giving the home team more of an advantage with certain calls.
Moskowitz and Wertheim (2011) also find this referee bias. They study all five major
professional sports (basketball, baseball, football, hockey, soccer) and agree that the
home field advantage in virtually all sports is largely due to the bias of officials toward
the home team.
C. Performance Outcomes
The final variables Carron et al. (2005) present in their conceptual model are
variables measuring performance. These performance outcomes are statistically based
variables that in this study are field goal and free throw percentages. By studying college
basketball teams, Harville and Smith (1994) find that the advantage of playing at home
Home Court Advantage 6
(in relation to playing on a neutral court) is estimated to be 4.68 ± 0.28 points.
Continuing with performance based home court advantage, Cao et al. (2011) find that
being the home team has a positive effect on free throw performance. The authors state
that this is because the home fans may be able to distract shooters from the away team.
D. Control Variable
One variable that Carron et al. (2005) do not include in their model, but other
literature states must be included, is a control variable measuring the quality of the
visiting team. This is included in the model because Nevill (1999) states, “In order to
correctly calculate the home advantage of individual teams, the ability of the opposition
must also be included.” A variable controlling for the quality of the home team,
however, is not necessary in determining home court advantage. This is because the
quality of the team is already being controlled for due to the balanced nature of the
competition. A balanced competition is one where teams play at least one home game
and one away game versus each opponent. Nevill (1999) states that for sports like
basketball, the quality of the team at home “is effectively eliminated by counterbalancing
the game location.”
III. Theory
Stefan Kesenne’s “Economic Theory of Professional Sports” states that
professional sports teams can either be profit maximizing or win maximizing (Kesenne
2007). If teams are win maximizing, then they will do everything they can to produce
more wins and create an advantage over their opponents. One way teams can get this
advantage is through creating a larger home court advantage. A home court advantage
Home Court Advantage 7
produces wins, and thus a higher home court advantage produces more wins. Ultimately,
a production function is being proposed.
But, why do teams play better at home? One explanation could be rationalized
through the fans. Katie Stankiewicz (2009) explores shirking in Major League Baseball
(MLB). Shirking is when a player purposefully does not perform to the best of his
ability. Stankiewicz does not find any evidence that players in the MLB shirk.
Stankiewicz explains lack of player shirking through fan monitoring. Players are less
likely to shirk in front of their home fans because they do not want to lose fan approval.
Fans express their approval or disapproval by attending games, cheering or booing at
games, or buying a player’s jersey. Attendance and merchandise sales are a large part of
a player’s salary. So, a player is going to make sure he performs especially well at home
to keep the fans happy and his salary high. Thus, being at home should have a greater
chance of producing a win than being on the road.
Referee biases can be explained through psychological theory that people want to
be liked and to be confirmed in their judgments. Referees do not like to be booed, and
therefore will base some of their decisions on crowd reaction. If the home crowd is loud
and boisterous, the referee is more likely not to call a foul or infraction on the home team.
But, in the same situation, the referee is more likely to call a foul on the away team and
receive cheers from the home crowd.
IV. Empirical Model
Data for the models are retrieved from the game logs of regular season NBA
games from the 2008-09 season through the 2010-11 season. Every game during those
seasons is accounted for, except the first games of each season for which DAYS REST is
Home Court Advantage 8
missing. Thus the sample size is limited to 3,642 game entries. Performance based data
statistics such as field goal percentage, free throw percentage, fouls, and win percentage
are retrieved from basketball-reference.com (Kubatko, J., 2014). The attendance, win
percentage, and days of rest data are obtained from nba.com (NBA Stats, 2014). Using
data from the selected time period allows for a recent analysis while avoiding lockout
years in the NBA.
Two models are presented in this study. Both have a dummy dependent variable
representing a win as 1 and a loss as 0. As such, a logit model is used. The logit model
enables a more accurate analysis than ordinary least squares (OLS) regression because
the dependent variable is either a 0 or a 1. Problems occur when using OLS with a
dependent dummy variable because the estimated probability of a win can turn out to be
less than 0 or greater than 1 (which is not possible for probabilities). This could result
with an inaccurate best-fit line. A logit model avoids this problem by limiting estimated
probabilities to be between 0 and 1. Logit also fixes the heteroscedasticity problem of
OLS regression with a dependent dummy.
Model A contains six independent variables: LN ATTENDANCE, HOME FG%,
HOME FT%, FOUL RATIO, AWAY WIN% OF VISITORS, and DAYS REST. Model
B uses those same variables except for HOME FT% and FOUL RATIO. Attendance
could have an effect through these two variables; therefore the two models are used to
show the effect attendance has before and after controlling for them. Table 1 lists all of
these variables with descriptions and their expected sign.
Home Court Advantage 9
Table 1: Variables List
Variable Description Expected Effect
WIN
Dependent dummy indicating a home
win
LN ATTENDANCE
Natural log of attendance for home
games
Positive
HOME FG% Field goal percentage at home Positive
HOME FT% Free throw percentage at home Positive
FOUL RATIO
Ratio of fouls called on visiting team
over fouls called on home team
Positive
AWAY WIN% OF
VISITORS
Control variable of the visiting team’s
win percentage on the road
Negative
DAYS REST
Number of days of rest the home team
has before each competition
Positive
LN ATTENDANCE is expected to be positive, meaning the more fans in the
arena, the greater the probability of the home team winning. HOME FG% and
HOME FT% are also expected to be positive, meaning higher percentages give greater
chances of the home team winning.
FOUL RATIO is a made up metric that measures the fouls called on the away
team over the fouls called on the home team. This variable determines if there is a
referee bias that leads to a home court advantage. If there is a referee bias, this
coefficient will be positive and greater than one. This means that the referee calls more
fouls on the visitors than he does on the home team, and thus creates a home court
advantage. Therefore, this coefficient is expected to be greater than one and have a
positive effect on producing a win.
AWAY WIN% OF VISITORS is a control variable that measures how good the
visiting team is. It measures the visiting team’s winning percentage on the road. This
control coefficient is expected to be negative, meaning the higher the away team’s
winning percentage, the less of a chance the home team has of winning the game.
Home Court Advantage 10
DAYS REST measures how many days of rest the home team has before each individual
game. Again, this variable is missing for the first games of each year. The coefficient is
expected to be positive, with more rest giving teams a greater chance of winning.
Finally, an error term is included in both models.
Model A
ln(
) = β
1
+ β
2
(LN ATTENDANCE) + β
3
(HOME FG%) +
β
4
(HOME FT%) + β
5
(FOUL RATIO) + β
6
(AWAY WIN% OF VISITORS) +
β
7
(DAYS REST) + e
Model B
ln(
) = β
1
+ β
2
(LN ATTENDANCE) + β
3
(HOME FG%) +
β
4
(AWAY WIN% OF VISITORS) + β
5
(DAYS REST) + e
V. Descriptive Statistics
Table 2 lists descriptive statistics for each variable in the models. Looking at
AWAY WIN% OF VISITORS and WIN, it is easy to see that some home court
advantage is happening. On average, between the years 2008-2011, home teams won
60.5% of the games during the regular season. There is a wide range in the data for
AWAY WIN% OF VISITORS, showing that there are some very good teams and some
very bad teams. The minimum value of AWAY WIN% OF VISITORS is .073, while the
maximum ranged up to .707. There is also a lot of variation in ATTENDENCE with data
ranging from 8,866 people to 23,129 people.
On average home teams shoot 46.7% from the field and 76.5% from the free
throw line. By way of comparison, visiting teams shoot 45.5% from the field and 76.3%
Home Court Advantage 11
from the free throw line. FOUL RATIO has a mean of over 1, suggesting a referee bias.
This means that on average, the referees call more fouls on the visiting team than they do
on the home team. Finally, DAYS REST shows that the average NBA team has 1.25
days of rest between games. The most days of rest are 11 over the all-star break, and the
fewest is 0, with teams often playing games on back-to-back nights.
Table 2: Descriptive Statistics
Variable Mean Std. Deviation Min Max
ATTENDENCE 17,305.2 2,840 8,866 23,129
HOME FG % .467 .057 .279 .675
AWAY FG % .455 .055 .289 .658
HOME FT % .765 .096 .364 1
AWAY FT % .763 .099 .357 1
FOUL RATIO 1.08 .280 .379 3
AWAY WIN% OF VISITORS .397 .167 .073 .707
DAYS REST 1.25 .980 0 11
WIN .605 .489 0 1
Table A1, in Appendix A, lists more descriptive statistics pertaining to each
individual NBA team between 2008-2011. It shows each team’s record at home and on
the road, as well as each team’s field goal percentage at home and on the road, and the
plus/minus statistic of average points scored at home and on the road. This table shows
an overwhelming support of the home court advantage thesis. More insight on this table
is given in Appendix A.
VI. Results
Table 3 shows the initial results of both models. A distinct home court advantage
is found in the NBA. All coefficients except for DAYS REST are significant at the 1%
level and have the correct signs. However, while these coefficients can indicate
Home Court Advantage 12
Table 3: Model Results
Variable Model A Model B
Coefficient Std. Error Coefficient Std. Error
LN ATTENDANCE 0.816*** 0.224 1.02*** 0.212
HOME FG% 22.1*** 0.903 19.5*** 0.830
HOME FT% 2.88*** 0.427
FOUL RATIO 2.66*** 0.170
AWAY WIN% OF VISITORS -1.92*** 0.248 -2.04*** 0.234
DAYS REST -0.022 0.041 -0.0047 0.039
Sample Size 3642
Pseudo R^2 0.2468
***Significant at the 1% level
**Significant at the 5% level
*Significant at the 10% level
significance, their values are log odds and difficult to interpret. Tables B1 and B2 in
Appendix B show the calculations of predicted probabilities at various values for the
explanatory variables. Table 4 summarizes the most important results from those
calculations.
Beginning with Model A and Table 4 below, with all variables at their mean
values, the predicted probability of the home team winning the game is 66.6%. The
probability column then shows the predicted probability of the home team winning the
game when increasing each variable by one standard deviation, in turn, and the change in
probability column shows the difference in the probability between the results of the
Table 4: Probability Results
Variable Model A Probability: .666 Model B Probability: .643
Probability
Change in
Probability
Probability
Change in
Probability
LN ATTENDANCE .693 .027 .678 .035
HOME FG% .875 .209 .846 .203
HOME FT% .724 .058
FOUL RATIO .808 .142
AWAY WIN% OF VISITORS .591 -.075 .562 -.081
DAYS REST .656 -.010 .641 -.002
Home Court Advantage 13
original variables and the results of the standard deviation increased variables. Therefore,
Table 4 shows that increasing LN ATTENDANCE by one standard deviation increases
the probability of the home team winning by 2.7%.
It is important to recognize that this effect of attendance does not include the
effect fans have through the FOUL RATIO and HOME FT% because those variables are
held constant in the model. As previously stated, attendance could have an effect through
a referee bias and through the free throw percentage because fans can influence the
referee’s decisions as well as affect free throws. To see how much these variables add to
the effect, Model B is created to remove the variables through which attendance can have
an effect. The results of Model B show that a one standard deviation increase in
LN ATTENDANCE increases the probability the home team wins by 3.5%. This means
that part of the effect that attendance has comes through FOUL RATIO and HOME FT%.
The rest of the coefficients in Model B are similar to those in Model A so the rest of the
analysis will be concerned just with the results of Model A.
Increasing the HOME FG% by one standard deviation increases the probability of
the home team winning by 20.9 %, the largest effect. HOME FT% also has a positive
effect on producing a win. By increasing HOME FT% by one standard deviation, the
probability of the home team winning the game increases by 5.8%. FOUL RATIO shows
that a referee bias is evident in the NBA, and it creates a greater home court advantage.
By increasing FOUL RATIO by one standard deviation, the probability of the home team
winning the game increases by 14.2%.
AWAY WIN% OF VISITORS is the final significant variable. This variable
shows that the better the visiting team is on the road, the less of a chance the home team
Home Court Advantage 14
has at winning. Specifically, for an increase in this variable by one standard deviation,
the home team’s chances of winning the game decreases by 7.5%. The final variable,
DAYS REST, is not significant at all, nor does it yield the expected effect, and therefore
will not be discussed further.
VII. Conclusion
Initial descriptive statistics show a home advantage. Home records are
consistently better than away records with home teams winning 60.5% of the games
between 2008-2011. Field goal percentages at home are also consistently better than
away field goal percentages. The plus/minus statistic also show a home court advantage
with teams scoring more and allowing fewer points at home than they do on the road.
This all agrees with the theory that teams will not shirk in front of their home fans.
Regression results also show a home court advantage through the selected
variables. All variables except for DAYS REST are significant at the 1% level.
Therefore, through a logit regression analysis a home court advantage is found through
attendance, field goal percentage, free throw percentage, and a referee bias. In agreement
with theory and most of the previous literature, attendance is a factor in creating a home
court advantage. The more fans a team has in the stadium, the greater chance that team
has of winning the game. Specifically, with a one standard deviation increase in
LN ATTENDANCE, the home team’s chances of winning the game increase by 2.7%
while controlling for other factors through which attendance can have an affect. This
gives team executives a greater incentive to fill their stadiums. Not only will they
increase their revenue, but they will also increase their team’s chance of winning the
game, because more people equals a greater chance of winning. This also could imply
Home Court Advantage 15
executives should lower ticket prices to fill more seats. However, sufficient data on
ticket pricing to do a full analysis on this theory is not included in the dataset. It is also
worth mentioning again that LN ATTENDANCE is significant despite controlling for
performance-based variables, through which attendance can have a factor. The results of
Model B show that attendance has an effect through referee bias and through the free
throw percentage because the fans influence referees and players.
In agreement with theory, FOUL RATIO also suggests a home court advantage
for NBA teams. For a one standard deviation increase in FOUL RATIO, the home
team’s chances of winning the game increases by 14.2%. This means that there is a
referee bias towards home teams, which can be explained by fan influence. The referee
does not want to make the home fans boo him, so his calls are more favorable toward the
home team. Referee biases could lead teams to change the way they play. If they know
there is a referee bias, home teams could be more likely to attack the basket on offense
and try to draw fouls. On the defensive side of the ball, teams could be more aggressive
because they know that referees would be less willing to call a foul. This result agrees
with both Page and Page (2010) and Moskowitz and Wertheim (2011). This referee bias
could also be troubling news for the NBA. Ideally, the NBA wants unbiased referees and
this study shows that referees show favor to the home team when making decisions. At
the same time, this study speaks out to fans. They really can affect the outcome of
winning the game.
Future research could include analyzing individual teams instead of a combined
NBA model. This could yield interesting results as it could show if there is a greater
home court advantage for some teams than others. More travel factor variables could be
Home Court Advantage 16
added into the model. Visiting teams undergo the inconvenience of some travel;
therefore a variable such as a measurement of distance traveled or even game start time
could influence the effect of home court advantage. For example, compare the start times
for games on the east coast and the west coast. A 9 p.m. Eastern game start time in
Boston for a Celtics/Clippers game has a relatively low impact for both teams. However,
a 9 p.m. Pacific game start time for those same teams in Los Angeles would greatly affect
the Celtics because that same start time in Boston would be midnight or later. This factor
might be a better variable than days of rest. Finally, applying this model to sports such as
baseball, soccer, or football could yield interesting comparisons to the effect of home
court advantage between different sports.
Home Court Advantage 17
Appendix A
Again, Table A1 below lists descriptive statistics for every team for the three-year
span followed in this study. Every team has a better record at home than on the road
except for the 2009-10 Boston Celtics, the 2008-09 Minnesota Timberwolves, the 2009-
10 Philadelphia 76ers, and the 2010-11 Sacramento Kings. And, the differences in home
and away records for these four teams are very small. In support of home court
advantage, there are some drastic differences between home and away games in other
team’s records. The 2008-09 Atlanta Hawks won 31 games at home, but had a sub-.500
record of 16-25 on the road. The Denver Nuggets had a record of 34-7 at home in 2009-
10, but they had a sub-.500 record of 19-22 on the road. The 2010-11 Washington
Wizards won 20 games at home, just under a .500 record, but they won measly three
games on the road.
Finally, the home and away plus/minus statistic is a great indicator of home court
advantage. A positive number indicates more points scored than points given up, while a
negative number signifies more points given up than scored. For example, the 2008-09
Chicago Bulls outscored their opponents at home by an average of 4.1 points per game.
But, on the road, they were outscored on average by 4.7 points. Even more drastic
differences than this can be found. The 2010-11 Denver Nuggets outscored opponents by
10.6 points on average at home. But, on the road, they were outscored by an average of
1.0 points per game. Only five times in Table A1 do teams have a better plus/minus
statistic on the road than at home. And these differences, like the home and away
records, are very small.
Home Court Advantage 18
Table A1: Individual Team Descriptive Statistics
Team Year
Home
Record
Away
Record
Home
FG%
Away
FG%
Home
Points
+/-
Away
Points
+/-
Atlanta
Hawks
2008-09
31-10 16-25 47.0% 44.7% 5.7 -2.5
2009-10
34-7 19-22 47.2% 46.4% 8.5 0.8
2010-11
24-17 20-21 46.5% 45.9% -0.9 -0.8
Boston
Celtics
2008-09
35-6 27-14 49.7% 47.4% 10.1 4.9
2009-10
24-17 26-15 48.9% 47.6% 3.4 3.9
2010-11
33-8 23-18 50.3% 46.9% 8.2 2.6
Charlotte
Bobcats
2008-09
23-18 12-29 45.4% 45.6% 2.2 -4.7
2009-10
31-10 13-28 46.1% 44.5% 6.6 -3.6
2010-11
21-20 13-28 46.0% 44.3% -0.6 -7.4
Chicago Bulls
2008-09
28-13 13-28 46.0% 45.3% 4.1 -4.7
2009-10
24-17 17-24 45.2% 45.1% 0.7 -3.9
2010-11
36-5 26-15 46.4% 45.9% 10.2 4.4
Cleveland
Cavaliers
2008-09
39-2 27-14 47.8% 45.9% 14.4 3.5
2009-10
35-6 26-15 49.7% 47.4% 8.9 4.2
2010-11
12-29 7-34 43.4% 43.5% -5.7 -12.3
Dallas
Mavericks
2008-09
32-9 18-23 47.9% 44.5% 6.5 -2.5
2009-10
28-13 27-14 46.1% 46.8% 2.2 3.2
2010-11
29-12 28-13 48.2% 46.7% 6.2 2.2
Denver
Nuggets
2008-09
33-8 21-20 47.5% 46.5% 7.1 -0.3
2009-10
34-7 19-22 47.7% 45.9% 9.0 -0.8
2010-11
33-8 17-24 48.8% 46.3% 10.6 -1.0
Detroit
Pistons
2008-09
21-20 18-23 46.2% 44.7% 0.2 -1.2
2009-10
17-24 10-31 43.8% 45.2% -3.3 -6.9
2010-11
21-20 9-32 47.6% 44.3% -0.5 -6.7
Golden State
Warriors
2008-09
21-20 8-33 47.5% 44.1% 1.4 -8.9
2009-10
18-23 8-33 47.3% 46.5% 0.8 -8.0
2010-11
26-15 10-31 47.3% 45.0% 2.7 -7.4
Houston
Rockets
2008-09
33-8 20-21 45.6% 45.0% 8.8 -0.8
2009-10
23-18 19-22 45.6% 43.8% 1.5 -2.3
2010-11
25-16 18-23 45.1% 45.8% 4.7 -0.3
Indiana
Pacers
2008-09
25-16 11-30 45.5% 45.5% 2.5 -4.8
2009-10
23-18 9-32 44.9% 43.6% 1.9 -7.9
2010-11
24-17 13-28 45.2% 43.3% 3.5 -5.6
Los Angeles
Clippers
2008-09
11-30 8-33 44.2% 44.1% -7.6 -10.0
2009-10
21-20 8-33 46.2% 44.8% -2.8 -10.0
2010-11
23-18 9-32 46.4% 45.0% 1.2 -7.4
Los Angeles
Lakers
2008-09
36-5 29-12 47.5% 47.3% 10.1 5.2
2009-10
34-7 23-18 45.9% 45.6% 8.5 0.9
2010-11
30-11 27-14 46.6% 45.9% 8.7 3.6
Memphis
Grizzlies
2008-09
16-25 8-33 45.9% 44.9% -3.3 -7.7
2009-10
23-18 17-24 46.2% 47.5% 0.0 -3.1
Home Court Advantage 19
2010-11
30-11 16-25 49.2% 44.9% 7.2 -2.5
Miami Heat
2008-09
28-13 15-26 46.6% 44.7% 4.2 -3.7
2009-10
24-17 23-18 45.6% 46.0% 3.4 1.1
2010-11
30-11 28-13 47.9% 48.3% 9.1 5.8
Milwaukee
Bucks
2008-09
22-19 12-29 45.5% 43.6% 2.9 -5.1
2009-10
28-13 18-23 43.9% 43.3% 3.3 0.1
2010-11
22-19 13-28 43.4% 42.6% 2.0 -3.7
Minnesota
Timberwolves
2008-09
11-30 13-28 44.1% 44.2% -5.8 -4.0
2009-10
10-31 5-36 45.6% 44.2% -5.6 -13.6
2010-11
12-29 5-36 43.9% 44.2% -3.3 -10.0
New Jersey
Nets
2008-09
19-22 15-26 44.4% 45.1% -0.9 -4.0
2009-10
8-33 4-37 42.3% 43.5% -7.2 -11.0
2010-11
19-22 5-36 44.3% 43.6% -3.2 -9.2
New Orleans
Hornets
2008-09
28-13 21-20 46.3% 45.2% 4.6 -1.5
2009-10
24-17 13-28 46.5% 46.4% 0.9 -5.8
2010-11
28-13 18-23 47.0% 44.8% 3.6 -1.8
New York
Knicks
2008-09
20-21 12-29 44.9% 44.1% -1.0 -4.3
2009-10
18-23 11-30 46.0% 44.9% -0.7 -7.0
2010-11
23-18 19-22 47.2% 44.3% 1.9 -0.3
Oklahoma
City Thunder
2008-09
15-26 8-33 45.3% 44.1% -3.7 -8.5
2009-10
27-14 23-18 47.6% 44.8% 5.6 1.3
2010-11
30-11 25-16 47.3% 45.5% 6.3 1.3
Orlando
Magic
2008-09
32-9 27-14 45.8% 45.5% 9.9 3.5
2009-10
34-7 25-16 48.3% 45.7% 11.6 3.3
2010-11
29-12 23-18 46.6% 45.6% 9.0 1.9
Philadelphia
76ers
2008-09
24-17 17-24 46.0% 45.8% 4.0 -3.9
2009-10
12-29 15-26 46.2% 45.8% -4.8 -3.0
2010-11
26-15 15-26 46.7% 45.5% 5.1 -2.1
Phoenix Suns
2008-09
28-13 18-23 52.0% 48.8% 6.3 -2.5
2009-10
32-9 22-19 49.8% 48.6% 9.4 0.4
2010-11
23-18 17-24 46.9% 47.1% 0.8 -2.6
Portland
Trailblazers
2008-09
34-7 20-21 47.5% 45.4% 10.3 0.3
2009-10
26-16 24-17 46.0% 46.2% 4.8 1.9
2010-11
30-11 18-23 45.3% 44.1% 5.6 -2.6
Sacramento
Kings
2008-09
11-30 6-35 44.6% 44.9% -6.0 -11.5
2009-10
18-23 7-34 46.3% 44.9% -1.2 -7.5
2010-11
11-30 13-28 44.4% 45.5% -3.4 -7.2
San Antonio
Spurs
2008-09
28-13 26-15 48.1% 45.0% 4.9 2.6
2009-10
29-12 21-20 49.2% 45.4% 8.3 1.8
2010-11
36-5 25-16 48.8% 46.2% 9.9 1.5
Toronto
Raptors
2008-09
18-23 15-26 47.1% 44.5% 0.5 -6.1
2009-10
25-16 15-26 48.1% 48.4% 1.3 -4.9
2010-11
16-25 6-35 48.2% 45.0% -2.6 -10.0
Utah Jazz
2008-09
33-8 15-26 48.6% 46.3% 9.5 -4.2
2009-10
32-9 21-20 51.1% 47.1% 9.5 1.2
Home Court Advantage 20
2010-11
21-20 18-23 47.1% 45.9% 0.8 -4.5
Washington
Wizards
2008-09
13-28 6-35 45.1% 44.8% -4.7 -10.2
2009-10
15-26 11-30 45.7% 44.1% -2.8 -6.8
2010-11
20-21 3-38 44.6% 43.9% -2.0 -12.8
Home Court Advantage 21
Appendix B
Tables B1 and B2, below, show the calculation behind the probabilities in Table
4. Looking at Table B1, the first column in each table multiplies each coefficient by the
mean of each variable. The means for all variables are found in Table 2. Summing all
the results of that action together yields log odds (0.690) of the home team winning the
game. Taking the antilog of the log odds gives us the odds ratio (1.994) of a win. This
odds ratio is changed into a probability by dividing the odds ratio by one plus the odds
ratio. This probability (0.666, or 66.6%) is the probability the home team wins the game.
The remaining columns in the tables show how the probability of a win changes by
increasing a variable mean by one standard deviation. A one standard deviation increase
in LN ATTENDANCE, for example, increases the probability of a win to 69.3%, an
increase of 2.7%. Or a one standard deviation increase in HOME FG%, for example,
increases the probability of a win to 87.5%, an increase of 20.9%. Table B2 is interpreted
in the same manner as Table B1.
Table B1: Model A Probability Table
Variable
Average
(Mean x
Coeff)
Change
LN ATT;
Otherwise
Average
Change
HOME
FG%
Otherwise
Average
Change
HOME
FT%
Otherwise
Average
Change
FOUL
RATIO
Otherwise
Average
Change
AWAY
WIN%
Otherwise
Average
Change
DAYS
REST
Otherwise
Average
LN ATTEND 7.963 8.087* 7.963 7.963 7.963 7.963 7.963
HOME FG% 10.316 10.316 11.575* 10.316 10.316 10.316 10.316
HOME FT% 2.203 2.203 2.203 2.480* 2.203 2.203 2.203
FOUL RATIO 2.873 2.873 2.873 2.873 3.618* 2.873 2.873
AWAY WIN% -0.762 -0.762 -0.762 -0.762 -0.762 -1.083* -0.762
DAYS REST -0.003 -0.003 -0.003 -0.003 -0.003 -0.003 -0.048*
CONSTANT -21.9 -21.9 -21.9 -21.9 -21.9 -21.9 -21.9
Log Odds 0.690 0.814 1.949 0.997 1.435 0.340 0.645
Odds Ratio 1.994 2.257 7.024 2.629 4.120 1.447 1.906
Probability 0.666 0.693 0.875 0.724 0.808 0.591 0.656
Change in Prob 0.027 0.209 0.058 0.142 -0.075 -0.010
* Changed Variable by Std. Dev.
Home Court Advantage 22
Table B2: Model B Probability Results
Variable
Average
(Mean x
Coeff)
Change
LN ATT;
Otherwise
Average
Change
HOME
FG%;
Otherwise
Average
Change
AWAY
WIN%;
Otherwise
Average
Change
DAYS
REST;
Otherwise
Average
LN ATTEND 9.954 10.109* 9.954 9.954 9.954
HOME FG% 9.125 9.125 10.239* 9.125 9.125
AWAY WIN% -0.810 -0.810 -0.810 -1.083* -0.810
DAYS REST -0.00059 -0.00059 -0.00059 -0.00059 -0.010*
CONSTANT -17.68 -17.68 -17.68 -17.68 -17.68
Log Odds 0.589 0.744 1.702 0.248 0.579
Odds Ratio 1.802 2.104 5.487 1.281 1.784
Probability 0.643 0.678 0.846 0.562 0.641
Change in Prob 0.035 0.203 -0.081 -0.002
* Changed Variable by Std. Dev.
Home Court Advantage 23
References
Ashman, T., Bowman, R., & Lambrinos, J. (2010). The role of fatigue in NBA wagering
markets: the surprising “home disadvantage situation.” Journal of Sports
Economics, 11(6), 602-613.
Cao, Z., Price, J., & Stone, D. (2011). Performance under pressure in the NBA. Journal
of Sports Economics, 12(3), 231-252.
Carron, A., Loughhead, T., & Bray, S. (2005). The home advantage in sports
competitions: Courneya and Carron’s (1992) conceptual framework a decade
later. Journal of Sports Sciences, 23(4), 395-407.
Cooper, J. (2013). Home court advantage. Retrieved from http://www.nba.com/
hawks/features/ home-court-advantage
Forrest, D., Beaumont, J., Goddard, J, & Simmons, R. (2005). Home advantage and the
debate about competitive balance in professional sports leagues. Journal of Sports
Sciences, 23(4), 439-445.
Greer, D. L. (1983). Spectator booing and the home advantage: a study of social
influence in the basketball arena. Social Psychology Quarterly, 46(3), 252-261.
Harville, D. A., & Smith, M. H. (1994). The home-court advantage: how large is it, and
does it vary from team to team? The American Statistician, 48(1), 22-28.
Késenne, Stefan. The economic theory of professional team sports: an analytical
treatment. Cheltenham, UK: Edward Elgar, 2007. Print.
Kubatko, J. (2014). Basketball-Reference.com – Basketball statistics and history.
Retrieved from http://www.basketball-reference.com/.
Home Court Advantage 24
Moskowitz, T., & Wertheim, L. J. (2011). Scorecasting: The hidden influences behind
how sports are played and games are won. New York, NY: Crown Archetype.
NBA Stats. (2014). Retrieved February 2014 from
stats.nba.com/scores.html?ls=iref:nba:gnav
Nevill, A. M., Newell, S. M., & Gale, S. (1996). Factors associated with home
advantage in English and Scottish soccer matches. Journal of Sports Sciences, 14,
181-186.
Nevill, M. (1999). Home advantage in sport: an overview of studies on the advantage of
playing at home. Sports Medicine, 28(4), 221-236.
Nutting, A.W. (2010). Travel costs in the NBA production function. Journal of Sports
Economics, 11(5), 533-548.
Page, L., & Page, K. (2010). Alone against the crowd: individual differences in referees’
ability to cope under pressure. Journal of Economic Psychology, 31(2), 192-199.
Salminen, S. (1993). The effect of the audience on the home advantage. Perceptual and
Motor Skills, 76(3c), 1123-1128.
Schwartz, B., & Barsky, S. F. (1977). The home advantage. Social Forces, 55(3), 641-
661.
Smith, D. R. (2005). Disconnects between popular discourse and home advantage
research: what can fans and media tell us about the home advantage phenomenon?
Journal of Sports Sciences, 23(4), 351-364.
Stankiewicz, '09, Katie, "Length of Contracts and the Effect on the Performance of MLB
Players" (2009). Honors Projects. Paper 103.
http://digitalcommons.iwu.edu/econ_honproj/103
