The following is another amazing piece of analysis from Kevin Draper (@TheDissNBA) from the great blog The Diss, which has been serving angsty NBA analysis since 2011. I can’t help but contain the spoiler – it doesn’t appear that players changing their shot attempts has a major impact on their shooting efficiency!
In the next week, the Oklahoma City Thunder have a difficult decision to make: should they resign James Harden and if so, does he deserve a maximum contract? If Harden is not resigned by October 31st he becomes a restricted free agent at the end of the season, increasing the chances that he will leave. Everybody that fancies themselves a salary capologist has weighed in on whether or not the Thunder can afford him, and everybody that fancies themselves a talent scout has weighed in on how good Harden really is.
In a roundabout sort of way, I too have weighed in on the debate, when I examined Russell Westbrook’s value. The basic takeaway is that, while Harden takes many fewer shots than Russell Westbrook and Kevin Durant, he is a much more efficient shooter than both of them. I ended the post with an open question, one that I think is especially pertinent to Harden’s contract negotiations: if Harden increased his shot attempts, would he be able to maintain his high shooting efficiency? If the Thunder or any other team sign Harden to a max contract, it is likely they will want him to play more than the 31.4 minutes and shoot more than the 10.1 field goals per game than he did last year.
The above graph, which shows us the relationship between field goal attempts and true shooting percentage among all 260 players that shot at least five times per game and played over 350 minutes last season, illustrates how elite Harden is as a shooter. Nobody that shot as many times as Harden did last year did so at as high of a true shooting percentage as Harden did. On a macro level, the graph seems to confirm what we assume to be true of all shooters. Players with low true shooting percentages generally don’t shoot a lot because they are bad at it (Editor’s note. However once players even get close to average, all bets are off in terms of taking shots. See Jamal Crawford). As one’s true shooting percentage increases, so do their field goal attempts. At a certain point, however, to increase true shooting percentage you need to stop shooting. Maintaining a 70% true shooting percentage while shooting 20 times per game is incredibly difficult. I have looked at the data for previous years, and the shape of the graph is very similar.
If I were to describe the graph in two sentences, I would say: for bad shooters, as field goal attempts per game increase, true shooting percentage increases. For good shooters, as field goal attempts increase, true shooting percentage decreases. (Editor’s note, don’t forget about the Bell Curve)
The above graph displays the change in true shooting percentage and field goal attempts per game from the 2010-11 to the 2011-12 season. The sample is restricted to the 204 players that played a minimum amount of minutes each season (400 in 2010-11, 350 in 2011-12) and shot five times per game or more in each season. The first thing to notice is that there is no linear relationship: there is no clear pattern of the two variables increasing or decreasing at a constant rate. Now, this isn’t necessarily a surprise, as our first graph showed that the relationship between true shooting percentage and field goal attempts per game differed depending on the skill of the shooter.
This graph is the same as the one above, but only includes the 152 players who had a true shooting percentage of 50% or better in each of the years under examination (ie, the good shooters). Given the above hypotheses, we would expect to see a liner relationship here, with field goal attempts decreasing as true shooting percentage increases. The equation of the line of best fit (the line that best fits the series of data points) is indeed negative, but only barely so. The R2 value, which measures the strength of the relationship between the two variables, indicates that a relationship is non-existent.
It is always possible that either the 2010-11 season, 2011-12 season, or both seasons were an anomaly, and not evidence of a true trend (or, in this case, the lack of a trend). I gathered the data for changes in true shooting percentage and field goal attempts per game for each of the last ten seasons and the data is basically the same, with the R2 value once again indicating no relationship between the two variables.
The original motivation for this post was to understand how James Harden would perform with an increase in field goal attempts per game, so it makes sense to try to replicate the situation that he is in. The above graph shows the change in true shooting percentage among all 43 players in the last ten seasons who increased their field goal attempts by five or more from one season to another. Once again, there is no relationship between the two variables, only interesting observations. Especially impressive is David West’s 2005-06 season, when he increased his FGA/G from 5.7 to 14, and his TS% from 47.9 to 55.4.
Perhaps David West’s 2005-06 season is more than just an interesting statistical outlier. In one year he became a “good” shooter, and has been able to maintain that designation throughout his career. Perhaps James Harden can even increase his true shooting percentage with an increase in field goal attempts.
This brings up one final question: what is the year-to-year predictive power of true shooting percentage? If a player shoots well one season, what is the chance that they will be able to do it again? The answer is a mixed bag: the previous season’s true shooting percentage explains about 34% of the variation in the current season’s true shooting percentage. There is certainly a relationship between the two variables—a good shooter that makes good decisions usually won’t morph into Allen Iverson the next season—but there are a lot of other variables that impact true shooting percentage.
I wish I had the confidence to make a strong statement about James Harden’s value, but the lack of evidence doesn’t let me. In somewhat limited minutes and shot attempts, James Harden is one of the best scorers in the game. That in of itself has a lot of value, and he certainly deserves (and will receive) a large contract, that I am sure of. But how James Harden will fair with an increased role on a team is the $1,000,000 question, that is of yet unanswered.