What is an NBA Players True Offensive Value?
SDSU Sports MBA program co-founder builds a scheme to re-evaluate the true measure of an NBA player’s total contributions to the team’s total offense.
Do NBA players get all the credit they deserve for non-scoring offensive play? They don’t, according to Jim Lackritz, co-founder of the Sports Business MBA program at the Fowler College of Business at San Diego State University.
Lackritz and Ira Horowitz, emeritus professor of statistics from the University of Florida, created a point-sharing scheme that determines player’s direct contribution to a team’s overall point total by evaluating the team’s offensive statistics weighted against the number of possessions and the probability that those possessions fed into that point total.
The researchers' point scheme offers players credit on offensive plays made that lead to all points scored, including free throws.
Research Evolves Into a Point Sharing Scheme
What sets this research apart from previous research is that Lackritz and Horowitz’s methodology assigns value to the overall points made by a team, including points made at the free throw line. “We acknowledge that while we aren’t the first to set forth the argument for a point-sharing scheme,” said Lackritz. “Unlike previous studies, this scheme places a value on a player’s direct contribution to the total points in the model, not just field goals made.”
For example, if Paul George of the Los Angeles Clippers made a perfect above-the-rim lob to teammate Kawhi Leonard who, in turn, stuffed the ball through the hoop, George would be credited with the assist. On the other hand, if George made the same pass and Leonard was fouled on his way to the basket, George would not get credit for the assist, even if Leonard made both of his resulting free throws. The scheme proposed by Lackritz and Horowitz would assign George points for the implied assist since he initiated the play that eventually led to the team’s additional overall point total.
Creating the Model
To create the model, the researchers looked at three seasons of NBA statistics, starting with the 2016 – 17 season through the 2018 – 19 season and they assessed how many field goals were made. However, unlike previous research, they also added free throws into their calculations since their objective was to get a clearer picture of plays leading up to all points instead of just field goals.
Next, researchers developed their scheme with the idea that they would assign value to each player based on the statistical probability that some free throws are a result of fouls, steals or offensive rebounds. Their formula is based on the percentage of field goals made as a result of an assist. For example, if 62% of field goals were made from assists, the researchers assumed that 62% of points from free throws were made from plays where an assist was involved. Since offensive rebounds and steals are counted in a player’s statistics, the researchers used a similar formula to calculate the likelihood that these plays led directly to points. They next assigned a point structure to the assists (2.383), offensive rebounds (0.588) and steals (0.530) which was based on the calculation that each of these plays generated that number of points. The coefficients for offensive rebounds and steals are lower, in that not all rebounds and steals translate into points due to missed shots and turnovers. In addition, there will be plays where there will be an offensive rebound and/or a steal that is part of an assisted basket.
In the final step of the scheme, the researchers distribute their point system (1.0) among the players on a team’s roster. Since the scorer should get most of the credit for each possession, each point scored during the season is credited between .60 - .80. For example, if each point scored is worth .75 and Player 1 scored 1,400 points during the season, the researchers give him credit for 1,050 total adjusted points. If the same player made 142 assists during the same season, the researchers would credit him for 25% of the points generated by each assist, giving him 142 x 2.383 .25 = 84.6 points. The same methodology for the assists is used to calculate steals and offensive rebounds as well.
Jim Lackritz is the co-founder of the SDSU Sports MBA program
To put it in a simplified formula, if a player had a single game total of 20 points, 3 assists, 5 offensive rebounds and 2 steals, the researchers’ adjusted point total would look like this:
20 (.75) + 2.383 (3) (.25) + 5 (0.588) (.25) + 2 (.530) (.25) =
15 + 1.8 + .7 + .3 = 17.8 adjusted points
Applying the Model To Real World Player Stats
The researchers then applied the model to the statistics of 39 of the NBA’s top players to determine their overall offensive effectiveness. The researchers determined Giannis Antetokounmpo, who was voted the league’s MVP during the 2018 – 19 season, had the highest adjusted point total. However, Kevin Durant, who finished eighth in MVP voting, had the second highest adjusted point total. It is also interesting to note that Nikola Jokic, who averaged slightly over 20 points per game that season, finished among the leaders in the researchers’ adjusted point system due to his high number of assists and offensive rebounds.
Were Durant and Jokic undervalued during the MVP voting? It would seem so based on the research. “We anticipate that, at some point, these statistics and their values are going to be translated into a salary model or introduced by agents and/or basketball executives into salary discussions or negotiations. It could have the same impact for the NBA that the book, Moneyball, had on salary research in baseball,” said Lackritz. “This model demonstrates how the right metrics can be used to offer a fair assessment of players’ contributions to their teams.”
NOTE: The author(s) received no financial support for the research, authorship, and/or publication of this article.
