NBA Statistical Test: Can Team Basketball Be Measured?

There are two prevailing schools of thought among NBA commentators in today’s game. One holds that “you can’t win without a superstar”, pointing at guys like LeBron James, Michael Jordan, and Kobe Bryant as guys who put (admittedly very good) supporting casts on their shoulders and lifted them to championships. The other holds that “team basketball” is king and points to teams like the Golden State Warriors and San Antonio Spurs with a side order of teams like the 2004 Pistons and 2008 Celtics, less star-driven by one guy and more driven by a top-to-bottom team concept that wins with depth.

(Note: This piece originally ran April 21, 2017, and covers the 2016-17 season prior to the playoffs.)

But which one is better? After all, if you do star-driven wrong, you get “great stats, bad team” guys like Anthony Davis in New Orleans and Chris Bosh when he was with the Raptors. And if you do team-concept wrong, at best you get the 2017 Heat, which just took too long to come together as a team to salvage a dreadful season, and at worst you get teams like the Timberwolves, who should be Warriors Lite but instead have spent the last decade being less than the sum of their parts.

Now, you could do this by guesswork and get a reputation as a genius when you do it right that ends up totally undeserved when you don’t have MJ or Shaq and Kobe (Phil Jackson, looking squarely at you.)

But there really does need to be a way to quantify team basketball. The most obvious starting candidate is Usage Rate. If one player is using a disproportionate number of the possessions, we can probably surmise that the team is built around star-driven hero ball (Russell Westbrook just broke a record held by Kobe Bryant for highest usage rate in a season, and in each case it was in the man’s cause of dragging a D-League team to the playoffs.)

On the other hand, a team like the Spurs, while they do have a distinct star on the floor in MVP candidate Kawhi Leonard, doesn’t have one guy taking the bulk of the shots. Likewise, the Warriors share the ball better than any team in the league, possibly any team in history.

So let’s consider a testable hypothesis and do some science.

Is there a correlation between distribution of Usage Rate and winning? And if so, is that distribution consistent with “star-driven” or “team concept” numbers?

To get at the problem, we’ll look at the ten players who led the league in Usage Rate this season and see how their teams made out over the course of the year. Then we’ll look at the teams with the lowest Usage Rate among their leading player, suggesting more equal distribution. For Boogie Cousins, who was second in the league, we’ll limit this to his time in Sacramento, and along the same lines, we’ll consider Anthony Davis’s New Orleans team before the trade.

Then we’ll see if it’s better to be the United States or Denmark in terms of possession inequality. Stats via Basketball Reference.

League Leaders in USG%:

Russell Westbrook (41.7, OKC 47-35)
DeMarcus Cousins (37.5, SAC 24-33)
DeMar DeRozan (34.3, TOR 51-31)
James Harden (34.2, HOU 55-27)
Isaiah Thomas (34.0, BOS 53-29)
Anthony Davis (32.6, NOP 23-34)
Damian Lillard (31.5, POR 41-41)
Kawhi Leonard (31.1, SAS 61-21)
Kyrie Irving (30.8, CLE 51-31)
John Wall (30.6, WAS 49-33)

And who’s just below the top ten? Stephen Curry and LeBron James.

So it would seem that the NBA is a star-driven league and the numbers back it up. With the exception of Boogie and the Brow, who were on utterly atrocious teams and really need to find themselves a situation where they’re not surrounded by dysfunctional horrors (and no, playing together in a pace-and-space league where Twin Towers went out with Robinson and Duncan isn’t going to help, as the Pellies’ 10-15 finish makes obvious), this is a who’s who of great players on great teams. Hell, Cleveland’s the only team with two guys in the top 12, and they just won the title last year.

The Teams Left Behind

Now let’s look at this from the opposite perspective. In the top 20 players for USG%, 19 teams are represented (if you leave Lou Williams out of this, since his usage declined sharply in Houston after putting up a 30.6 with the Lakers that would’ve snuck him into the top ten, and really, all that tells us is that the Lakers are such a horror show that Lou Williams was the guy they trusted most with the ball.)

That means there are 11 teams who didn’t have a star player who could take command of their offense and do the heavy lifting. How’d they do in the standings? Well, here’s the list.

Phoenix (24-58)
Philadelphia (28-54)
Orlando (29-53)
Dallas (33-49)
Detroit (37-45)
Denver (40-42)
Miami (41-41)
Milwaukee (42-40)
Atlanta (43-39)
LA Clippers (51-31)
Utah (51-31)

So four playoff teams, two teams that won 50 games, out of 11. As compared to nine playoff teams out of the 11 represented at the other end of the scale, teams that if you sorted the entire league by record, would be nine of the 11 best teams in the entire league.

So What Have We Learned?

Well, we’ve learned that the conventional wisdom got one right. The NBA is a star-driven league, and teams that set up their offenses to move the ball and consistently get their best players in positions where they can make plays are teams that don’t just make the playoffs, but get home-court advantage in Game 7.

This is true even if you’re the Golden State Warriors. Steph Curry was 11th in the league in Usage Rate on a team that had Kevin Durant and Klay Thompson. Even a 67-win team has a best shooter who should have the ball in his hands.

Sure, having a player topping the league in Usage Rate isn’t itself a magic bullet; if you’re as utterly verklempt as the Kings and Pelicans, you’re still gonna suck.

But this one’s pretty well settled. “Team ball” is an illusion. Long live the superstars, for they are the team.