A Data-driven Approach to Analyze if Defense Really Wins Championships in the NBA
INTRO
Argumentum Ad Populum: A statement appealing to a crowd’s belief to influence an outcome.
Illusory Truth Effect: The phenomenon where false statements gain perceived validity through repetition.
When coupled together, these inaccurate phrases become widely known and difficult to disprove. They make for excellent slogans and political taglines, often losing their attribution due to how “true” they seem.
“Success breeds success.”
“The apple doesn’t fall far from the tree.”
“Offense wins games, but defense wins championships.”
They’re pervasive in everyday speech and their validity often goes unchecked. But with the power of data, we can more critically analyze the truth of these claims.
BACKGROUND
Basketball is a huge part of my life.
I was introduced to sports at a young age as the middle child of a family deeply invested in sports. Through hard work, I developed my talent in basketball to the point where I was able to play collegiately at New York University. After rising to captain my senior year, I continued my path by coaching at the Division 3 level for 5 years.
From there, I made the leap to pursuing a Master’s degree in Data Analytics.
As I’ve started this journey, I’ve begun questioning how I saw the game of basketball and the beliefs I’ve held. I’ve always thought my “feel for the game” and “great eye” have led to my success, but through what I’ve learned I’m curious what a more objective, data-driven approach might mean.
I’d like to start by analyzing a phrase that was instrumental in my subjective approach: “Offense wins games, but defense wins championships.” The original quote attributed to legendary University of Alabama football coach, Bear Bryant, is “Offense sells tickets, but defense wins championships.” However, the misquoted phrase is widely used by coaches, announcers and players without the historical data to back it up.
MISSION
Assess the validity of the famous quote “Offense wins games, but defense wins championships” through the lens of the NBA. I’ll look at 38 seasons’ worth of data from 1983–1984 through 2020–2021 to determine:
· Is defense more effective in winning champions?
· How has the dynamic changed over time?
At the conclusion, I aim to better understand what’s important in winning or building an NBA champion.
DEFINITIONS
· Points Per Possession (PPP): A calculation of how many points an NBA team can score per one possession
· Offensive Rating (ORtg): Team’s points per possession (PPP) multiplied by 100 possessions
· Defensive Rating (DRtg): Opponent’s points per possession (PPP) against the team multiplied by 100 possessions
· Net Rating (NRtg): ORtg minus DRtg
In my analysis, I ranked ORtg, DRtg and NRtg across teams within a season as well as within the data set as a whole (38 seasons). I classify these as DRank, ORank and NRank. After ranking, I manipulated the data in two new statistics to analyze the famous quote. The new statistics are defined below:
· Team Rating & Play Style (by season):
o Team Rating: A team’s DRank minus their ORank for a given season
o Team Play Style: I chose to define an offensive-oriented team at a Team Rating of >5 and a defensive-oriented team at a Team Rating of <-5. I consider teams with a Team Rating between -5 and 5 to have a moderate Play Style which is insignificant for my analysis.
o Example: 2021 Milwaukee Bucks’ DRank = 8 and ORank = 7. DRank (8) minus ORank (7) equals a Team Rating of 1. Since -5<1<5, their 2021 Team Play Style is moderate.
· League Rating & Play Style:
o League Rating: A sum of Team Ratings based on ranking ORtg and DRtg across all teams and all seasons
o League Play Style: I chose to define an offensive-oriented season at a League Play Style of >5,000 and a defensive-oriented season at a League Play Style of
<-5000.
METHOD
· Data: I sourced by-season data from basketball-reference.com. I manually added information tied to playoff finish.
· Coding: I used R to manipulate, browse and create visualizations and utilized the readr, ggplot2, ggeasy, dplyr, ggpubr and ggtext libraries. Some of the coding skills I learned in my data visualization course, and others I obtained by scouring Stack Overflow and Google, and from reading R documentation.
· Presentation: I took snapshots of my visualizations and combined them in Google Slides, which you can see at the bottom of this article. The summary will cover each graph individually.
SUMMARY
1. TEAM PLAY STYLE OF CHAMPIONS
I created a line graph to determine if each season’s champion had an offensive- or defensive-oriented style of play. According to my Team Rating and Play Style statistics, offensive-oriented teams and defensive-oriented teams won 11 times each. As the peaks and valleys fluctuated consistently throughout the 38 seasons, there does not seem to be a true offensive or defensive period for champions. However, the remaining 16 teams possessed a moderate Team Play Style which indicates a well-balanced offense and defense is more effective than a particularly skewed style.
I highlighted a few champions to provide context:
· 1989 & 1990 Detroit Pistons: Led by star Isaiah Thomas and enforcers Bill Laimbeer and Dennis Rodman, this squad matched the personality of Detroit: tough and gritty. Adding to the team’s rough-and-tumble reputation was head coach Chuck Daly’s drive to intimidate the top player in the league, Michael Jordan. Known as the “Jordan Rules,” Daly enforced a rule that if Jordan was in your vicinity, you had to “nail” him and when he went to the basket, the goal was to knock him down. While the team was well-liked by Detroit fans, others disliked their attitude and approach, ultimately leading to rule changes in the league.
· 2001 Los Angeles Lakers: The 3-peat Lakers became an offensive force in the early 2000s with the introduction of head coach Phil Jackson’s triangle offense. This system was particularly effective in maximizing the ability of dynamic duo Kobe Bryant and Shaquille O’Neal, the latter’s size and athletic ability unrivaled in the league.
· 2004 Detroit Pistons: The Pistons of the late ’80s were reborn, aided by a physicality unmatched for their era. Led by one of the most underrated players of all time, Ben Wallace’s defensive prowess had an unbelievable impact on the game.
2. LEAGUE PLAY STYLE OVER TIME
I created a second line graph to determine League Play Style per season across the distribution of 38 seasons. In particular, I aimed to identify if a champion possessed a play style opposite of the predominant play style of the league. I found five instances where this was the case: three offensive teams won in a defensive season and two defensive teams won in an offensive season.
I indicated key rule changes that may have impacted League Play Style:
· 1991: Implementation of the Flagrant Foul. Intended to protect players from injury, the league established a foul against excessive or violent contact. This appears to be a direct result of the Detroit Pistons’ style of play and the “Jordan Rules,” where the league’s top players were targeted as a means to stifle their success on the court. Though the rule didn’t create a major shift in play style, it did prevent the league from becoming increasingly violent in a defensive-dominant space.
· 2002: Legalization of Zone Defense. Due to the unstoppable force of Shaquille O’Neal’s offense, the league was compelled to create a better way for teams to defend against quick-on-the-perimeter and powerful-in-the-post players. Enter zone defense, which enabled teams to assign defenders to zones as opposed to people, eliminating the one-for-one speed and skill required to defend against league superstars. Zone defense gave an advantage to the defensive style of play, but as this was a new allowance it took time for teams to realize the benefit.
· 2005: Hand-checking is outlawed. Previously common practice, hand-checking allowed defenders to place a hand on the hip of the ballhandler, thwarting the ballhandler’s ability to quickly get past them. To counter the defensive favor of zone defense, the hand-checking rule promoted a more offensive play style as players were able to score more efficiently.
· 2021: Flopping made illegal. As players find new and inventive ways to “sell fouls” to referees, defenders have an increasingly difficult time guarding them. The ability to flop at a high level has only increased the offensive-oriented teams’ advantage. In response to this, the league officially banned flopping. As the act and ruling is highly subjective, it may take time for referees to determine how to effectively call it and restore balance toward a moderate play style.
3. MAXIMUM OFFENSIVE RATING OVER TIME
I built a line graph to show the effectiveness of top offensive play styles in efficiency and its positive impact on Offensive Rating over time.
I highlighted the following:
· A dip in ORtg in the late 1990s-early 2000s, related to a trend in “slowing the game down.” This shift was intended to be a strategic move for offenses to tire out their opponents by driving the shot clock down. However, teams often wasted the time rather than using it to execute scoring actions, degrading their shot quality as they scrambled to beat shot clock violations. In turn, this failure to capitalize on offensive efficiency meant offensive rating took a hit.
· The introduction of head coach Mike D’Antoni’s “7 seconds or less” offense with the 2004 Phoenix Suns. Brought over from Europe where D’Antoni played, the speedy style of play enabled dynamic point guards like Steve Nash to dominate. Though revolutionary to the league at the time, other teams soon followed suit in adopting similar quick-paced approaches. Because of this shift, we see an increase in offensive rating.
The NBA lockout of 2011, and its impact on offensive performance. The league went dark for 161 days, which meant that when playing resumed teams were competing within a compressed schedule and no preseason preparation. Without the ability to refine their strategies and hone their cohesiveness, teams struggled to regain the growing offensive dominance of the prior years.
4. MINIMUM DEFENSIVE RATING OVER TIME
In this line graph, I show the impact of the 3-point revolution on defensive effectiveness over time. As shooters found success farther from the net, defenses struggled to effectively guard with increased spacing.
I included relevant context to the data:
· 1994–1997: The NBA moves the 3-point line 21 inches closer to the net. Prior to this, the league heavily favored 2-point shots for the perceived efficiency of their distance to the rim. This move enabled high 3-point scorers like Dell Curry to explode, encouraging teams to invest in their 3-point shots.
· 2003: Data analysis of the Celtics uncovered they were the first NBA team to take at least 30% of shots from 3-point range.
· 2010: The Orlando Magic became the first team to take at least 40% of shots from 3-point range.
· 2018: The Houston Rockets became the first team to take at least 50% of shots from 3-point range.
Though in theory, moving the 3-point shot line would increase the amount of 3-pointers taken and would result in offensive success, the benefit to shooters and schemes was stifled by the league’s philosophy of wasting the clock and the incidental decrease in spacing. In line with D’Antoni’s shift toward a fast offense, 3-point shots quickly gained momentum from 2003 to present day.
As of the 2021 season, all but 2 NBA teams took at least one third of their shots from 3-point range.
5. COUNT OF CHAMPION BY TYPE
After my analysis of the four points above, I did not find evidence to support either offense or defense as the key factor in winning championships. Rather, the data stated they’re equally important. So, I chose to look elsewhere to gain insights.
In this bar graph I looked at the top-ranked teams in ORtg, DRtg, NRtg, and Win/Loss Percentage.
Here’s what I found:
· Top-ranked ORtg and DRtg teams had an insignificant difference in total champions
· Nearly half of the champions (16 of 38) had the top NRtg, which proved to be the key indicator in predicting a champion
· Overall dominant teams are most likely to win it all at playoff time
6. MAXIMUM NET RATING OVER TIME
In this line graph, I show how consistent top NRtg teams stayed over time. While some teams proved overtly dominant with a Net Rating >10 (ie. the championship-winning 1992 Chicago Bulls and 2017 Golden State Warriors), the data otherwise points to evenness in NRtg.
To note: while there is a slight increase in Net Rating over time, this is likely due to the expansion of the NBA from 23 teams to 30.
7. OVERALL TOP 100 NET RATINGS AND THEIR PLAYOFF FINISH OVER TIME
In this line graph, I show how consistent top NRtg teams stayed over time. While some teams proved overtly dominant with a Net Rating >10 (ie. the championship-winning 1992 Chicago Bulls and 2017 Golden State Warriors), the data otherwise points to evenness in NRtg.
To note: while there is a slight increase in Net Rating over time, this is likely due to the expansion of the NBA from 23 teams to 30.
FINDINGS
1. I could not find evidence to support that “defense wins championships.” A better indicator to predict championship winningness is a team’s Net Rating rank.
2. The NBA is on an upward offensive swing, due in part to the rising popularity of the 3-point shot. However, this may change if current NBA commissioner Adam Silver establishes a 3-point line further away from the net.
3. Through analysis, I recommend coaches aim to maximize a team’s ORtg while minimizing their DRtg. Finding balance between offense and defense rather than focusing only on one is likely to garner a higher NRtg.
4. Following this project, I’d like to delve deeper into the context surrounding the data points I analyzed. Through the brief investigation into various rule changes and philosophies, I found that context adds color and insight into the numbers.
5. My final thoughts on the famous Bear Bryant quote I analyzed: Overall, it’s misleading. While it’s a nice reminder to young athletes who are often exposed to the flashy side of offense, both offense and defense are hallmarks of a championship team. Pride in both is necessary to be great.
DESIGN PRINCIPLES
Gestalt design principles played an integral role in my visualizations.
· Proximity: Used in annotations. Proximity allows the viewer to identify which part of the visualization I’m commenting on.
· Similarity: Present in all visualizations through the use of a consistent color scheme. All terms, including offense, defense, championships and Net Rating, are colored-coded for ease of understanding each graph individually and in the context of the whole.
· Enclosure: Used in the shading of offensive and defensive areas on the graphs, which is effective in visually separating data for the viewer.
· Continuity: Apparent in the elimination of the graphs’ axis lines, which declutters the visualization and enable viewers to focus on relevant data instead.
· Closure: Leveraged by the lack of borders or a background to enclose the individual visualizations. Viewers do not need bounding boxes to register each graph as a cohesive entity.
· Connection: Utilized via the arrows connecting information to years on the x-axis. Although the data is labeled by year, the arrows promote a quick visual association for the viewer.
CONCLUSION
I greatly enjoyed this project, and it vastly aided me in learning how to create visualizations and master R. The best tools at my disposal were Google, YouTube, Stack Overflow, and R documentation — they were reliable in helping me find solutions to my problems.
My code is below so you can view my methods to create and customize these graphs.
