By: Tyler Chia & Aidan Horng
Introduction
As the 2023–2024 NBA regular season comes to an end, we will take a look at one of the most talked about awards for this year — The NBA Defensive Player of the Year Award. According to most NBA analysts and sports betting books, Rudy Gobert is the clear frontrunner for the award this year. However, many fans and certain members of the media believe that Victor Wembanyama, AKA “Wemby”, is more deserving of the DPOY award this year. We will take a look into who truly deserves the award based on past DPOY winners.
As of recent years, many argue that the end-of-season awards such as Most Valuable Player, Defensive Player of the Year, and 6th Man of the Year should go to the best player in that category on the team with the or one of the best records. For example, many fans are arguing that Rudy Gobert is deserving of the DPOY award since he is the best defensive player on the Minnesota Timberwolves, one of the best teams in the league this year. Others argue that Victor Wembanyama is clearly the best individual defensive player in the NBA this season, even though his team’s defense does not reflect it.
Note: For the purpose of the study, we will only consider data starting from the 1983 season since that is when the DPOY award started.
Data Source
The data used in the analysis is sourced from this Kaggle data which scrapes NBA data from Basketball Reference.
Variables
Below is a list of the variables that will be included in the study. It is critical to understand the meaning of each variable as they will be used in the analysis later on.
season: season year
player: player name
pos: position
age: age
tm: team
g: games played during regular season
w: games won during regular season
mp: total minutes played during the season
stl_percent: percentage of defensive possessions that player got a steal
blk_percent: percentage of defensive possessions that player got a block
drb_percent: percentage of defensive possessions that player got a rebound
dws: defensive win shares — estimates the number of wins a player contributes to their team through their defensive performance. It is part of the win shares metric, which is widely used in basketball analytics to quantify a player’s overall contribution to their team’s success. DWS takes into account various defensive statistics such as blocks, steals, defensive rebounds, and individual defensive rating (points allowed per 100 possessions while the player is on the court). It also considers the player’s playing time and the team’s overall defensive performance. The formula for Defensive Win Shares is complex and proprietary, but it generally aims to attribute a player’s defensive contributions to their team’s success in terms of wins. A higher DWS indicates that a player has made a significant defensive impact on their team’s performance, while a lower DWS suggests a lesser impact.
dbpm: defensive box plus minus — estimates a player’s overall defensive impact on the game. It’s a component of the box plus/minus (BPM) metric, which is part of the advanced statistics used in basketball analytics. DBPM is calculated by attempting to measure a player’s contribution to their team’s defensive performance while they are on the court, relative to league average. It considers various defensive statistics such as blocks, steals, defensive rebounds, and personal fouls, along with the player’s playing time and the team’s performance while the player is on the court. A positive DBPM indicates that a player contributes positively to their team’s defense, while a negative DBPM suggests that the player’s defensive impact is below average. The scale is relative to average performance, so a DBPM of 0.0 represents league-average defense.
d_rtg: defensive rating — measures the number of points a team allows per 100 possessions. It’s a way to evaluate how effective a player or team is at preventing their opponents from scoring.
opp_e_fg_percent: opposing team effective field goal percentage — measures the efficiency of a team’s defense by considering the value of each field goal attempt made by their opponents. It accounts for the fact that three-pointers are worth more than two-pointers by giving them extra weight in the calculation. A lower eFG% for the opposing team signifies better defensive performance, indicating that the defense is effectively contesting shots and limiting the opponent’s scoring efficiency. Teams with lower eFG% numbers typically force their opponents into more challenging shots, resulting in a decreased overall shooting percentage for the opposing team and reflecting strong defensive play.
pf_per_game: personal fouls per game
stl_per_game: steals per game
blk_per_game: blocks per game
pf: season total personal fouls
stl: season total steals
blk: season total blocks
Individual Statistics Exploration
We will begin our analysis by exploring the individual statistics of the past DPOY winners and compare them to Victor Wembanyama’s and Rudy Gobert’s statistics this season.
Steal and Block Percent for DPOY Winners
We see that Victor Wembanyama has the highest block percentage out of all past NBA DPOY winners as well as Rudy Gobert this season. His steal percentage seems to be average amongst the past DPOY winners, but is still significantly higher than that of Rudy Gobert. Since Wembanyama and Gobert are both centers, we will filter the data so that we see where they land amongst only centers. We will do this because most guards who have won DPOY will likely have a higher steal percentage than that of big men.
Steal and Block Percent for DPOY-Winning Centers
When we filter the data to only Defensive Player of the Years who were centers, we see that Victor Wembanyama has one of the higher steal percentages and still has the highest block percentage. Rudy Gobert is on the lower end of both steal and block percentages amongst centers.
Steal and Block Averages Per Game for DPOY Winners
From the plot above, Victor Wembanyama clears Rudy Gobert in both steals and blocks per game as well as being in the higher tier for blocks amongst past DPOY’s.
Total Season Steals and Blocks for DPOY Winners
Once again, we see that Wembanyama clears Rudy Gobert in both total season steals and blocks. Wembanyama also has one of the highest block totals amongst past DPOY winners in recent years.
Advanced Defensive Statistics for DPOY Winners
We see that Gobert has higher Defensive Win Shares than Wemby. This shows that Gobert contributes more wins to his team through his defensive performance than Wemby does. While this may be the case, we cannot ignore the fact that the Timberwolves have 34 more wins than the Spurs do. A more accurate representation would be to show the percentage of wins that player contributed to based on their defense instead of their total wins.
We see above that Victor contributed to a higher percentage of wins than Rudy did this season. We did this by dividing the defensive win shares by their team wins.
An argument that we must acknowledge is that some fans have been saying Gobert is so dominant that players don’t even want to go near him since they are afraid of getting their shot blocked. They argue that because of this, Gobert’s stats do not reflect his actual impact. While this may be true, I believe that the same can be said for Wemby, if not more. I personally believe that Wembanyama has even more dominance since he can guard any position on the floor with his speed, length, and IQ. To debunk this argument, we can simply look at their respective defensive box plus-minus statistics. We see that for the Defensive Box Plus Minus, Victor is higher than Gobert, showing that he contributes more to the Spur’s defense than Gobert contributes to the Timberwolves’.
Total Personal Fouls For DPOY Winners
We see that Gobert has had significantly more total fouls throughout the course of the season compared to Wembanyama.
Personal Fouls Per Game For DPOY Winners
Once again, we see that Gobert averages an entire extra foul per game compared to Wembanyama.
Team Statistics Exploration
We will continue our analysis by exploring the teams of past DPOY winners and compare them to the San Antonio Spurs (Wembanyama’s team) and Minnesota Timberwolves (Gobert’s team) this season.
Total Season Wins for DPOY Teams
We see that no player in the history of the NBA has won the DPOY award with nowhere near 20 wins like the current San Antonio Spurs and Victor Wembanyama have. All of the winners were on some of the best teams in the NBA for that particular year. This is the main argument as to why Gobert deserves the DPOY over Wemby.
Defensive Rating for DPOY Teams
From the graph above, our statements made earlier are reiterated, showing how terrible the Spurs’ defense is. We see that the Spurs are allowing an absurd amount of points per game.
Opposing Team Effective Field Goal Percentage for DPOY Teams
Once again, we see how terrible the Spurs’ defense is. They are allowing opposing teams to score at a crazy efficient rate on a nightly basis.
Machine Learning Model
We will finalize our analysis by creating a machine-learning model that takes in the following predictors: steal percent, block percent, defensive rebound percent, and defensive win shares. The data also contains a binary variable, “DPOY_winner”, which denotes whether or not that player has won the DPOY award during that given season. The model is trained on data which include these factors for all NBA players from the 1983 season. The model is used to predict which players in the 2023–2024 NBA season have the most likely chance of winning the award based on the statistics of previous winners.
We chose to use the steal, block, and defensive rebound percentages over the totals or per-game averages since we wanted to see who is truly the most dominant per 100 possessions. Using the totals and/or per-game averages would skew the data since total games and minutes played vary by player and would also cause issues with multicollinearity.
We decided not to include defensive box plus-minus since the calculation is very similar to that of defensive win shares and we did not want to overfit the model. In addition, a player who does not get many minutes may have a high defensive box plus-minus solely based on “garbage-time” minutes, which does not accurately represent their defensive impact in the case of this study.
We decided not to include defensive box plus-minus since the calculation is very similar to that of defensive win shares and we did not want to overfit the model.
From the output of the model, we see that Victor Wembanyama should be the clear frontrunner for the Defensive Player of the Year Award for the 2024 season.
Conclusion
Although the model and virtually all of the plots predict that Victor Wembanyama should be the Defensive Player of the Year for the 2023–2024 season, the award is still up to the discretion of the panel of media members selected by the NBA. Even though almost every graph and the model show that Wemby deserves to be the DPOY just based on the individual statistics, there may be other intangible factors that play into the voting for end-of-season NBA awards.
As stated previously, we see that in the modern NBA, awards are almost always given to players on teams that have winning records. Unfortunately, for Wembanyama, the Spurs only have 22 wins, making them one of the worst teams in the league. On the other hand, Gobert and the Timberwolves ended the season with 56 wins, making them one of the best teams in the league. Even with team wins being accounted for in the model, Wembanyama’s individual statistics are so dominant that the model still predicted him to be the DPOY winner, despite his abysmal team.