Earlier this season you may have seen a version of this graph at Regressing, the statistical arm of Deadspin. It’s a stunning illustration of just how far LeBron James and Kevin Durant have been pushing the limits of offensive efficiency this season.
The data points on the chart represent every individual player season, going back to 1979-1980, where a player played at least 1,000 minutes, with a Usage Percentage of at least 24.0 percent and a True Shooting Percentage (TS%) of at least 0.558. The black line on the chart is the product of calculations by Evan Zamir and represents, roughly, the hypothetical boundary between usage and efficiency. Over the last 34 NBA seasons, no one has ever finished farther beyond that boundary then LeBron and Durant did this season. Among high-usage scorers, the TS% marks that Durant and LeBron posted this season were literally unprecedented.
LeBron’s career-high TS% this season is the culmination of a steady march toward historically elite levels of efficiency. He came into the league directly from high school and his immediate and outsized impact was largely the product of innate physical talents. As he’s complemented those physical gifts with refined skill and awareness the results have become increasingly absurd. The graph below shows the career progression of his TS% by season (going back to the 2000-2001 season, league average is 0.532).
Except for his rookie season LeBron’s TS% has been above the league average every year, absurdly so over the last half-decade. His development into a hyper-efficient volume scorer has been widely celebrated and explanatory credit for this growth is usually subjectively parceled between his development as a shooter and his adherence to a more efficient shot distribution pattern.
His development as a shooter is eye-popping and undeniable. As a rookie LeBron made just 32 percent of the shots he took from 10 or more feet. In the last two seasons he’s made 41 percent of his shots from 10 or more feet. The animation below (built with heat maps from Basketball-Reference.com) shows how LeBron’s scoring area has changed over his time in the league.
This shows how the quantity of points he’s scored in the mid-range area has gradually shrunk and become focused around a few key spots. This mid-range area is generally regarded as the least efficient scoring area on the floor and LeBron has made a point of using it less and less. We know that LeBron’s overall efficiency has improved tremendously–we can see it reflected in his TS%. And while certain breakdowns of his shot distributions, like these heat maps, and his field goal percentages from different distances tell us he has become both a better and more selective shooter it’s hard to see the line between the two. Those factors–shot accuracy and shot distribution–are incredibly intertwined and TS% gives us no obvious ways to separate them.
So which factor is really driving the development of LeBron’s historic scoring efficiency?
Any evaluation of shooting abilities using TS%, or any other variation of field goal percentage, is muddy because the measure includes both a player’s accuracy and their shot selection. This combination is not always intuitively understood, but it’s there nonetheless. For instance, imagine if LeBron made a point of not taking any shot from outside the paint. It might collapse the entire Heat offense and severely limit the number of shots available to him, but it would undoubtedly cause his TS% to skyrocket if he was taking nothing but layups and dunks.
When we look at shooting percentages we are seeing a reflection of the shots a player made and the shots they decided to take, shot-selection and accuracy wrapped up into one number. However, there are some techniques we can use to separate the two factors and identify just how much development LeBron’s has seen in each area.
The idea is to hold constant each piece of the puzzle — accuracy and shot selection — to see how the other side changes. We can do this by substituting an average value for one piece, using LeBron’s own values for the other side, and measure the change in TS%.
To isolate and measure the value of changes in LeBron’s shot selection we need to substitute a league average value for accuracy. We can derive this by using the league average expected values for the six offensive scoring zones as defined by NBA.com/stats — Restricted Area, In the Paint (Non-RA), Mid-Range, Corner Three, Above the Break Three and Free Throws. These different locations have inherently different expected values based on the relative ease with which shots can be made from each area and the scoring value of each shot (three-pointers vs. two-pointers).
We can calculate what LeBron’s TS% would be in each season of his career if we used his individual array of shots from those locations but substitute those different values for his own shooting percentages from each zone. By using these league average expected values we can see exactly how much LeBron’s TS% is being elevated by just the inherent efficiency in his shot distribution.
The graph below shows both his actual TS% and his Accuracy Controlled TS%.
When we control for accuracy and just look at the relative value of his shot distribution we can see a general pattern of modest growth since his rookie season. You can also see the decline that occurred when he joined the Miami Heat in 2011 and he and Dwyane Wade muddled through their initial “my turn, your turn” isolation offense. Over the last two years we can see a steep increase as his actual TS% has pushed towards historic levels. However, the average annual growth rate for his Accuracy Controlled TS% across his entire career is just 0.30 percent. It’s clear by comparing the trajectory of his actual TS% and his Accuracy Controlled TS% that something else is going on.
For comparison, we can reverse this technique, controlling shot selection and isolating the development in his shooting accuracy. To do this we need to identify a league average shot distribution pattern to substitute for his own.
To arrive at this league average shot distribution pattern I tried to approximate what would be expected of LeBron offensively. This involved weeding out all players who didn’t have at least 10% of their shot attempts from behind the three-point line. Using the average shot distribution pattern of those that remained and LeBron’s own shooting percentages from each area we see the following pattern in his TS%.
Here we see a much steeper rate of growth and one that more closely matches the rate of growth in his overall TS%. The compound annual growth rate for his Distribution Controlled TS% across his entire career is 1.17 percent, nearly four times the growth rate in his Accuracy Controlled TS%.
From these measures it’s clear that LeBron’s growth as an accurate shooter has had a far bigger impact on his overall scoring efficiency than the improvement in his shot distribution. In fact if we create a simple ratio with the two average annual growth rates we find that about 74 percent of the growth in his TS% can be explained by the growth in his shooting accuracy.
This technique is not perfect. Even as much as I’ve separated them here accuracy and shot selection are still indelibly linked. But this does illustrate the huge disparity in relative value between taking better shots and doing a better job of making those shots. LeBron James is a more efficient scorer, primarily because he has become a better shooter. A more effective shot distribution is just puffing things up at the margins. But when you’re are in pursuit of historic thresholds, even improvements at the margins are important.