The Numbers Game: The Fantasy Value of Usage
Usage rate is a popular metric with a simple purpose – determining what percentage of a team’s plays are ‘used’ by a given player. It is based upon a players’ total field goal attempts (FGA), free throw attempts (FTA) and turnovers (TO), and it’s tethered to the time the player was on the court. The goal of this column is to determine how strongly usage rates correlated to overall fantasy values last season, if at all.
Usage is only counted for the minutes a player is on the court, which means that reserve players, especially PGs and SGs, sometimes have very high usage rates — Tony Wroten, Gerald Green, Jamal Crawford, Isaiah Thomas, Dennis Schroder and Lou Williams all had usage rates above 26.6% percent last season.
For perspective, the mean usage rate among the top-200 8-cat values last season was 21.4%. Russell Westbrook easily led the way at a whopping 37.2%, followed by D-Wade (34.5%), Kobe Bryant (34.3%), DeMarcus Cousins (33.9%) and LeBron James (32.4%). The lowest usage rates among top-200 players belonged mostly to big men like Tyson Chandler (12.8%), though Andre Iguodala, P.J. Tucker, Kyle Korver, Nicolas Batum and Harrison Barnes were all below 15.0%.
Given those two lists of the highest-usage and lowest-usage, we can already infer that this offensive-minded metric might positively correlate with fantasy values — and it does.
[I gratefully acquired all raw statistics from NBA.com and Basketball-Reference.com, which I then manipulated in Systat and Excel].
The figure 0.55024 in this table is the multiple correlation coefficient, or Pearson’s r, which measures the linear correlation between two variables: usage rate and 8-cat values.
In plain English, it shows that usage rates and 8-cat values have a strong positive correlation — when one increases, the other tends to also increase (at least among the top-200 players). The implication is that, generally speaking, you want players with high usage rates on your teams.
Here is a chart mapping the top-200 players’ usage rates and overall values, with a trendline which supports the findings above.
You can follow me on Twitter @Knaus_RW.
That peak far to the right is, of course, Russell Westbrook. The lowest point among the elite players, far to the right, is Chris Paul — his usage rate was a modest 23.4% last year. If you’re on board with the argument I’m making here, that usage rate is strongly correlated with usage rate (if not necessarily a predictor), then it makes sense to identify players with high usage this preseason.
Here are the top-20 players, by usage rate, during the exhibition season (to qualify, they must have played a total of 40+ minutes):
DeMarcus Cousins: 37.0%
Damian Lillard: 37.0%
Anthony Davis: 34.8%
LeBron James: 34.5%
James Harden: 34.1%
Jahlil Okafor: 33.1%
Paul George: 32.8%
Gerald Green: 32.6%
Blake Griffin: 33.6%
Kawhi Leonard: 31.5%
Kyle Lowry: 31.1%
Kobe Bryant: 31.0%
Emmanuel Mudiay: 30.8%
Tyreke Evans: 30.6%
Nikola Vucevic: 30.0%
Russell Westbrook: 29.9%
Stephen Curry: 29.4%
Derrick Williams: 29.4%
Lou Williams: 29.1%
John Jenkins: 29.0%
Note that Westbrook’s usage is more than seven points lower than last season — Kevin Durant has a solid 28.0% usage rate himself, so the drop-off was inevitable. The Spurs are making good on their promise to unleash Kawhi Leonard this preseason, and he’s currently 8.5% above his previous career-high for usage. Even if that tails off he’s primed for a career-best year offensively.
Career-high usage rates abound in that list, including DeMarcus, Ant Davis, Lillard, Blake, Kawhi, Lowry, Vuc, Tyreke, D-Will and Jenkins, the last of whom has been unleashed as one of the Mavs’ only viable scorers given their injury woes.
I was also curious about the potential correlation of usage rate to offensive efficiency, as gauged by True Shooting % (factoring in 3-pointers and free throws) and Effective Field Goal % (only factoring in 3-pointers). High-volume shooting seems like a probable partner for lower percentages.
The result was a Pearson’s r of -0.1441 for True Shooting %, and of -0.3448 for Effective FG %. In other words, a negligible correlation between TS% and usage, and a moderately negative correlation between eFG% and usage. I found this interesting because it suggests that although increased usage may often be accompanied by decreased 2-point and 3-point percentages, it might not have much impact on FT shooting.
To be sure, I ran the numbers and they showed that usage actually had a weak positive correlation with FT% last season, and (of course) a very strong positive correlation with FT attempts.
Turnovers are another obvious positive correlation, given that they’re literally part of the usage-rate equation – last season it was a strong positive correlation (0.6563). For this reason expected usage and overall values to have a weaker positive correlation in 9-cat leagues, where increased usage becomes a damaging factor in the turnover category.
Usage is, in fact, weaker when applied to 9-cat values, shifting from 0.55024 in 8-cat leagues to 0.41003 in 9-cat. It’s something to keep in mind if you consider usage when drafting or working the waiver wire this season.
I may discuss usage again at some point this season, but you can get a weekly fix by seeking out Mike Gallagher’s ‘Use It or Lose It’ series of columns, wherein he explores usage and related concepts from every angle.
If you have any questions or insights, you can find me on Twitter @Knaus_RW.
For those who find this stuff interesting, note that in future weeks I’ll return to the idea of correlation with other advanced metrics: how do pace, TS%, Player Impact Estimates, age and playing time correlate to fantasy value? We’ll find out.
This entry passed through the Full-Text RSS service – if this is your content and you’re reading it on someone else’s site, please read the FAQ at fivefilters.org/content-only/faq.php#publishers.
