Can Centers Dunk Their Way to Efficiency?

By Daniel Lowe
Reading Time: 5 minutes

We are witnessing a change in the way the NBA is played. In the 2016-17 season, the overall rate of three-point attempts and average offensive rating were both the highest that they have been in NBA history. NBA offenses are becoming more effective, and improvements are driven by a radical shift in style of play. Teams like the Warriors and the Rockets are spreading the floor to an extent never seen before, and it’s working. It’s becoming clear that spacing, ball movement, and efficient shot selection is the way forward in the NBA.

This trend, however, poses a question: how do centers fit in the new style of offense? Traditionally, NBA centers have been utilized largely in the low post, using strength and size to get under the basket for dunks and layups. Recently, though, there’s a new breed of stretch big-men that are changing the way the position is played. Al Horford, Brook Lopez and others are all showing that big men aren’t limited to scoring from within three feet from the basket. This style of “stretch five” seems to fit well with the trend of spreading the floor for offensive efficiency. I decided to take a statistical look at the effect spacing-orientated centers have on a team’s offense, to try and determine if they are going to become the standard for the position, or if there is still room for the traditional big man in the new NBA.

The first step was to determine which centers are effective when scoring from distance and which dominate in the paint. To do this, I split the stats of NBA centers in the 2016-17 season into shots taken from 0-3 feet from the basket, and shots taken outside that range. From this data, I calculated each center’s pointers per 36 minutes (PP36) and effective field goal percentage (EFG%) under the basket, and from distance. This data is shown in the two graphs below.

A couple things stand out. First, there’s a pretty clear trend that players shoot more in the areas they are efficient in, as shown by the trend line, which was to be expected. In terms of specific players, these metrics match up pretty well with common intuition regarding which players are efficient from range and under the basket, as can be seen in the graph, and the table below. In order to make this analysis possible, I chose to combine both stats (PP36 and EFG%) into a unified metric by multiplying them. Multiplying the stats places value on players that score both consistently and efficiently from range or from under the basket, as it accounts for both scoring volume and field goal percentage. I normalized the values on a scale of 100, with 100 being Lopez’s stats for spacing, and McGee’s stats for close range, as they ranked the highest at the perspective metrics. The table below shows the top ten players in both metrics.

With unified metrics, we can compare these values to other stats in order to determine the effect a player with good spacing or close-range metrics has on an offense. I chose to look at the on/off splits of the centers for these comparisons. On/Off Splits show how the overall performance of a team’s offense and defense change when a specific player is either on or off the court. Specifically, I chose effective field goal percentage, assists percentage, and offensive rating. These three stats show how efficiently a team is scoring, how effectively they are moving the ball, and how many points they score per 100 possessions. I graphed the data against the two metrics, then assigned linear fits and used r-squared analysis to try and understand the relationships. This is a statistical measurement that describes how closes data fits a linear relationship. An r squared value of 1 indicates one variable is directly proportional to another, a lower value indicates they are loosely correlated, and a negative value indicates an inverse relationship.

Screen Shot 2017-09-20 at 9.15.56 PM.png

Screen Shot 2017-09-20 at 9.16.17 PM.png

As we can see, there is a positive trend associated with both metrics in relation to team effective field goal percentage. r-squared analysis shows that the trend is very similar, the spacing metric has an r-squared value of 0.0547 and the close-range metric has an r squared value of 0.0527. This indicates that centers that score effectively from any distance are equally likely to make their team’s offense more efficient.

Screen Shot 2017-09-20 at 9.16.28 PM.pngScreen Shot 2017-09-20 at 9.16.40 PM.png

As seen in these graphs, there is also a positive trend associated with both metrics and offensive rating. r-squared analysis shows there is actually a stronger trend associated with the close-range metric, 0.0975 as opposed to 0.0348 for the spacing metric. This indicates that centers who are effective at scoring in the paint are more likely to improve their team’s scoring than centers who effectively stretch the floor.

Screen Shot 2017-09-20 at 9.16.53 PM.png

Screen Shot 2017-09-20 at 9.17.01 PM.png

The final metric I looked at was assist percentage, which measures what percent of shots the team took were assisted. This showed the most dramatic difference. There was a very strong positive correlation between the spacing metric and assist percentage (r-squared value of 0.209), and a slight negative correlation between the close-range metric and assist percentage (r-squared value of -0.00133). This is a strong indication that centers who can space the floor improve ball movement, while paint focused centers actually make the ball movement worse, which is exactly what you’d expect.

So, to summarize the findings so far, both metrics are indicative of centers who have a positive effect on their offense in general. Centers with strong close-range metrics are more likely to improve their team’s scoring, while centers with high spacing metrics are more likely to improve their team’s ball movement. This doesn’t really help us decide if the traditional center is irrelevant compared to new breed. However, during the analysis, I did notice a further trend worth investigating. The two teams with the best spacing in the league, the Rockets and the Warriors, also had some of top centers in terms of close range metrics. The best explanation for this is that these teams are taking advantage of the space provided by their other players to create easy lobs and dunks for their centers, like Javale McGee and Clint Capela. I decided to remove these two teams from the data, and the trends became much clearer.

Without the centers from the Rockets and Warriors in the data, there was a stronger positive correlation across the board between the spacing metric and offensive stats than the close-range metric. Spacing and offensive efficiency has an r-squared value of 0.083 while the close-range metric only has a r-squared value of 0.039. The correlation for the spacing metric and offensive rating increased to an r-squared value of 0.065, while the correlation for the close-range metric decreased to 0.056. Finally, the already drastic difference in correlation with assists increased, with the correlation with the spacing metric increasing to 0.245, while the correlation with the close-range metric decreased to -0.002.

From this filtered data, we can draw some actually interesting conclusions regarding the value of different types of centers. For almost every team in the league, these trends indicate that a center that spaces the floor improves their offense more than an equally effective center that primarily scores under the basket. The two clearest exceptions to this rule are two teams that already have excellent spacing. Particularly, this data shows if coaches want to run an offense with effective ball movement, a spacing-oriented center is a must have, and a paint focused center actually appears to be a liability.

I don’t believe this is conclusive proof that the traditional big man is completely outdated. There were centers with high close-range metrics that had positive offensive stats across the board, such as DeAndre Jordan. And the traditional breed of center appears to be effective in offenses that are already heavily spacing oriented, by using the openings created by their teammates. But I believe this does show that the way centers are evaluated and used in the NBA should be shifting with the change in play style. Look for smart, analytics-focused NBA teams to acquire and utilize the new breed of stretch five in their offenses, and focus less on acquiring and playing the traditional NBA center in the future.

Posted in NBA

Leave a Reply