Feb 15, 2014; New Orleans, LA, USA; A view of the NBA logo at the NBA All Star Jam Session at the Ernest N. Morial Convention Center. Mandatory Credit: Bob Donnan-USA TODAY Sports

Work Smart, Not Hard: NBA Analytics Edition


The two best players in the National Basketball Association are, unequivocally, Lebron James and Kevin Durant.  Qualitatively, our eyes tell us this.  And quantitatively, a myriad of different advanced metrics—including John Hollinger’s ubiquitous Player Efficiency Rating (“PER”)—back up what we observe anecdotally.  But who is the third best player?  A savvy floor general like Chris Paul?  Or might it be a guy like Stephen Curry, whose 2013-2014 numbers compare favorably with Derrick Rose’s 2010-2011 MVP season.  Has a vastly-improved Blake Griffin made the leap into contention for the number three spot?  What about Paul George—already a top five NBA wing defender and the offensive leader for arguably the most impressive team in the NBA in 2013?

In studying PG24 throughout the last two seasons especially, what has always stood out to me is the frenetic pace to his game.  Although now slightly less magnified due to the emergence of teammate and fellow freak athlete Lance Stephenson, Paul George always looks like he is running faster and for longer distances and durations than his other teammates and his opponents.  Thanks to Stats LLC NBA SportVU player tracking technology being installed in every NBA arena this season, we can actually confirm this.  George runs an average of 2.53 miles per game and has traveled 161.9 miles during NBA games this season, as of  March 12, 2014.  Comparative, another young, incredibly athletic player who also makes me tired just by watching him—Andre Drummond—averages just 2.22 miles per game and has logged only 142.1 miles in 2013 regular season games.  There are no games with numbers being played here: Drummond and George each have appeared in 64 games this season, each play big minutes, and each compete with a significant amount of observable effort, but Paul George has still somehow managed to run nearly 20 more miles than Drummond has during games.

The popular expression “work smart, not hard” is on its face a bit misguided, but the underlying logic and spirit of the saying is sound.  Working smart and working hard are not mutually exclusive; it is obviously important to work hard and to also work intelligently.  When I hear the phrase “work smart, not hard,” I think about truths such as there being better ways to get six pack abs than doing thousands of crunches every day, or the notion that spending hundreds of hours cramming for an important test is likely not the best use of your time.  Ultimately, the “work smart, not hard” mantra is all about acknowledging the point of diminishing returns—the notion that when variable X is increased while other factors remain the same, the benefit derived from variable X will eventually stop increasing proportionally and may actually begin to decrease.

Might there be a point of diminishing returns with regard to Paul George’s effort and energy?  I assumed that running long distances on the basketball court is not directly correlative to success as a basketball player.  I assumed that after a certain distance, running longer distances is directly correlative with increased injury potential, as well as other negative variables.  To be sure, anyone who has played pickup basketball knows that the longer you go without an injury, the greater the likelihood you will come down on somebody’s ankle on a jump shot or rebound attempt.

With these assumptions in mind, I used all NBA rosters for the 2013-2014 season as my base sample size, and then I isolated only players appearing in at least 45 games.  From this sample, I created a sub-sample of the top 20 players in PER.  The player efficiency rating (PER) is a rating of a player’s per-minute productivity.  In the words of John Hollinger, “The PER sums up all a player’s positive accomplishments, subtracts the negative accomplishments, and returns a per-minute rating of a player’s performance.”  It is a particularly useful way to compare which players are “better” than others, since it takes into account per minute performance—thus allowing comparisons between players with disproportionate playing times.  After amassing this list, I catalogued each player’s miles traveled on the basketball court during the regular season, as well as each player’s average distance per game.  Finally, for each player I divided total distance traveled by that player’s PER—the quotient being dubbed “mileage efficiency.”

Mileage efficiency = Distance traveled (miles) / Player Efficiency Rating

While it is a tremendously great thing that most sports franchises are adopting an analytical approach to human resources and in-game strategy, the funny thing about advanced metrics in sports is that most of the quantitative analyses utilized are the antithesis of advanced.  Oftentimes, it is nothing more than an application of middle school pre-algebra.  Mileage efficiency is no different.  Although PER is somewhat complicated at first glance, there is nothing difficult about dividing that value by a variable number of miles.

In effect, a lower mileage efficiency is desirable.  Every player has to move to be productive on a basketball court.  But given injury risks and the desire to stay as fresh as possible for the postseason, it is probably optimal for each basketball player to maximize his impact on the game while also minimizing the per game physical wear and tear on his body.  So, if a hypothetical basketball player has a PER of 22, but that player also runs an improbable 3.5 miles every game to obtain that PER, that player is going to have a poor mileage efficiency score.  Theoretically, that player has ample room to improve in so far as being able to impact a game without running 3.5 miles every game in the form of short, sometimes violent sprints and leaps.

Consequently, mileage efficiency, and NBA speed and distance tracking analytics in general, are each in their respective infancies.  As a result, there is a strong possibility that a low mileage efficiency, despite the presence of a high PER, might correlate with a lack of optimal levels of effort and, in turn, signify an already good player having room to positively and efficiently grow due to a reasonable, rational effort increase.

I have included the name of DeAndre Liggins, a seldom-used Miami Heat guard, highlighted in yellow in the results table below.  Liggins demonstrates one of the many limitations to mileage efficiency.  Liggins checked in during garbage time for the Heat once this season, logged one minute of playing time, and made 1 out of 1 field goals and grabbed a rebound.  As a result, he has a PER of 128.24.  Kevin Durant, meanwhile, who leads the NBA in PER not including junk data like that of DeAndre Liggins, has a PER of 30.03.  And because Liggins ran only .1 miles, his mileage efficiency comes out to .000779788—for all intents and purposes an irrelevant, garbage value needing to discarded for the sake of accurate rankings, much like Liggins’s PER.

The Chart below is ordered from best to worst in terms of Mileage Efficiency.  In the interests of absolute clarity, Chris Paul has a lower, and better, mileage efficiency than Paul George.

Screenshot 2014-03-12 15.05.17

Rather than write generally about these findings, and then allude to other patterns stemming from the mileage efficiencies of players not in the top 20 in PER, I will instead point out three telling one on one comparisons from the aforementioned top 20 sub-sample.

Screenshot 2014-03-12 19.23.37

At this point, we have all practically run out of things to praise Tim Duncan for.  He is literally the greatest power forward to ever play professional basketball.  His excellence spans two decades.  Since being drafted by the San Antonio Spurs, Duncan has never missed the playoffs.  Even today at 37-years-old, Duncan is still a dominant power forward and anchors a legitimate contender for the Larry O’Brien trophy.  But take a look at his mileage efficiency!

Old man Tim Duncan has only ran 110 total miles during games this season and does not even run 1.9 miles per game.  He checks into games, paces himself, dominates his younger, faster, more athletic competition, and somehow leaves the court with a player efficiency rating of 21.1.  That PER tally is actually higher than that of Paul George—the guy many think has a legitimate claim to being the third best player in the NBA.  George runs .64 miles per game more than Duncan, which assuredly puts him in more positions to positively benefit his team, yet his 20.7 PER score is slightly less than Duncan’s.  With a mileage efficiency of 7.821256039, compared to Duncan’s 5.213270142 number, George clearly has ample room to grow as a player and learn to more efficiently dominate games.  Whether it has something to due with fatigue due to his distance numbers or simply putting himself in less than ideal positions due to his amazing exuberance, expect Paul George to tweak his game in the years to come—thus increasing his PER and decreasing his mileage efficiency.

Now, let’s take a look at a comparison of two younger big men: University of Kentucky Alumni DeMarcus Cousins and Anthony Davis.

Screenshot 2014-03-12 19.36.02

Despite both being one-and-done big men from the University of Kentucky, Cousins and Davis are very different players.  Cousins is a heavy, punishing bruiser in the paint; Davis is 220 lbs soaking wet and would prefer to finish with finesse jams on the pick and roll.  Cousins is at best incredibly emotive and at worst resembling a crazy person at times on the basketball court; Davis is demure both by comparison and in reality.  Cousins has normal facial hair; Davis has a unibrow.  Perhaps most distinctively, Cousins seemingly takes pride in strategically not crossing the half court line nonsensically when his teammates have a fast break; Davis, meanwhile, is almost always trailing looking for a dunk in transition.

Many critics of DeMarcus Cousins’s game like to label him as “lazy.”  That might be true, but if that is the case, he doesn’t let any instances of sloth affect his 25.7 player efficiency rating.  Anthony Davis is the antithesis of lazy; he is often almost reckless in his hustle, and his body sometimes takes a toll for it.  But despite his unwavering effort, Davis’s PER is at 26.3, just .6 higher than Cousins’s PER.  DeMarcus runs 2.007272727 miles per game; Davis 2.4309091.  And Cousins, through 55 games, has ran 110.4 miles on the year compared to 133.7 miles for Davis in the same amount of games.

What do these numbers mean?  With a mileage efficiency of 4.295719844, Cousins is drastically more efficient than Davis’s at 5.08365019.  With more effort leading to more average distance ran per game, might Cousins become an even better player?  Or will Anthony Davis grow as a player as he learns to rely less on athleticism and hustle and more on intelligence and skill?  A less than one year sample size of mileage efficiency does not provide us with that answer, but it does lay the groundwork for possible future adaptive growth for these two players.

Finally, what basketball comparison article would be complete without a 2013 MVP breakdown between Lebron James and Kevin Durant.

Screenshot 2014-03-12 20.04.20

King James versus KD35 aka The Servant aka The Slim Reaper aka Lil B’s worst enemy aka the guy who needs a better nickname.  While Lebron should probably win the MVP every year he is even remotely close to his prime, 2013-2014 has been Durant’s best season to date, and his league leading 30.3 PER vouches for that assertion.  In Lebron’s defense, his second place 29.5 PER is hardly something that warrants a loss of sleep.  But Durant and James are different players at this stage in their careers.  Kevin Durant is playing like a younger Lebron James—attacking the rim with ferocity and playing end line to end line with reckless abandon.  Lebron James’s game, however, is all about efficiency these days, and his mileage efficiency aids in illustrating just that.

Lebron is averaging  about 2.4 miles traveled per game while Durant has eclipsed him at about 2.52.  On the season, Durant has accumulated 158.6 miles ran to Lebron’s 141.7.  Even controlling for him playing four more games than James, Kevin Durant is just a more active player in 2013 than his rival Lebron.  Lebron gets the edge in both raw mileage efficiency at about 4.8 AND at mileage efficiency extrapolated to assume that James played the same number of games as Durant at about 5.12—Durant’s mileage efficiency is at around 5.23.

What does this tell us about Lebron and KD35?  Not a whole lot.  Both players are phenomenal.  Perhaps Durant is expending too much energy during the regular season, as evidenced by his speed and distance numbers.  Or maybe Durant’s slender frame can handle ridiculous amounts of wear and tear.  With such a small sample size, it is really too early to know at this point.

Mileage efficiency is not revolutionary. And in all candor, it isn’t that useful yet with less than an entire year of data encompassing its potential sample size.  As the seasons go by and more and more SportsVU data is at our disposal, NBA front offices, journalists, and fans alike will have a better understanding of the interrelation between basketball effectiveness and on-the-court mileage efficiency.  Expect distance stats like mileage efficiency, optimal production per quarter mile, ideal speed to become important indicators of productivity and injury mitigation.  But at this point, one thing is for certain: if you’re running a lot on the basketball court, you’d better be doing so hyper-efficiently, and if you’re picking and choosing your spots on the basketball court, you’d also better be doing that in efficient fashion.

Tags: NBA