In Defense of Russell Westbrook

“As for me being a gunslinger, I’ve just got this one granddaddy Paterson Colt and a borrowed belt to stick it in. But I also got an appetite for greater things. I hoped by joining up with you, it’d put me that much closer to getting them.”
-Robert Ford, The Assassination of Jesse James

It’s official: the basketball world has turned against Russell Westbrook.

Honestly, it’s been a long time coming. Westbrook wears his flaws on his sleeve, and his tendency to freeze out the offense while forcing bad shots is understandably alienating. It’s hard to watch Westbrook when he shifts into “hero mode,” as he did in the Thunder’s Game 4 loss — a performance that catalyzed the criticism of his game and his ability to function alongside Kevin Durant over the long term. Among that critical chorus was our own Zach Harper, who laid out his stat-supported case for Westbrook’s negative impact on Durant’s play:

[Westbrook] can win you basketball games. But it comes with a price and that price is the production of Kevin Durant.

Kevin Durant is the best player on the Thunder. Nobody should quibble with this fact. It’s unquibbable (made it up). And yet, you have people wondering if Westbrook might be the Thunder’s best player. The reason people are thinking this is because they fail to see how Westbrook negatively impacts what Kevin Durant does on the court….The reason Durant is less efficient [this season] seems to be that Russell Westbrook might be the most erratic star point guard since the fabled Stephon Marbury-Steve Francis era. You never know what he’s going to do on the court. Is he going to run the offense or is he going to awkwardly pull up on his jumper and show you what it would look like if Andre Miller actually elevated while shooting?

This is where Durant suffers. Yes, KD has issues with getting separation from his defenders, but the bigger problem is that the way he gets the ball is so inconsistent…The reason Durant gets the ball so inconsistently is because Westbrook is still trying to toe the line between point guard and “holy shnikes, I think I can get by everyone and get my own shot.”

That line that Harper describes is very real. We’ve seen Westbrook attempt to run some semblance of a cooperative offense, and we’ve also witnessed his mental shift into a realm where he is a lone gunman. Those two states are blatantly evident in Westbrook’s play, and I wouldn’t dare argue against the fact that he can act as a detriment to his team when at his shot-hunting worst (Game 4 serves as the most recent example, but is only one of many relevant ones).

Still, before we condemn Westbrook too much for his indiscretions against his teammate and the game of basketball, we should all take a deep breath. Westbrook and Durant are both stars, and played as such this season. Both ranked in the top 10 in PER, so being highly critical of their synergy is, to some degree, making a mountain out of the pile of dust that remains after OKC demolishes their opponents. This is the highly successful core of a highly successful team, and while there’s nothing wrong with pondering the influence of one star on another, I think the case against Westbrook isn’t quite as well-supported as one may think.

In Harper’s piece, for example, he cites Durant’s regular season performance when Westbrook was on the court in comparison to when Westbrook was on the bench (courtesy of the invaluable NBA.com StatsCube):

We can see that Durant has significant gains in points, field goal attempts, three-point attempts, free throw attempts, rebounds, assists, and +/- on a per-minute basis with Westbrook on the bench, while maintaining essentially the same shooting percentages.

For comparison’s sake, let’s look at a few other examples. First, on how Kobe Bryant plays with Pau Gasol on and off the court:

Or Paul Pierce, with Kevin Garnett either in the game or on the bench:

Or perhaps the most famous pair of stars around, LeBron James and Dwyane Wade:

We see the same general production trends across the board even on those teams like the Lakers, Celtics, and Heat. Shooting percentages either go up or maintain their previous levels once one star goes to the bench, while shot attempts, points, and other production increase. These are the natural dynamics of star players sharing the court: there are only so many shots, assists, and boards to go around, and playing with another highly productive player naturally curbs some box score stat accumulation. Removing one star from the equation opens the door for more opportunity to produce.

However, there is one substantial difference between those other star duos and the Durant-Westbrook pairing: the Thunder’s plus-minus is a disappointing +2.2 (per 48 minutes) when Westbrook and KD are on the court together, but a tremendous +9.4 (per 48 minutes) when Durant plays while Westbrook sits. On face value, this does seem to support the notion that the two aren’t so compatible after all.

But before we go any further with that thought, we should look at how Westbrook and Durant have performed in the postseason thus far. After Westbrook’s Game 4 meltdown, the plus-minus numbers should be consistent with the eyebrow-raising regular season marks, right?

It’s a small sample size, but the trend appears to have completely reversed while strengthening in magnitude. In this year’s playoffs, the Thunder are 16.8 points better per 48 minutes with both Westbrook and Durant on the court than they are with Durant playing sans Westbrook. But why are the numbers so radically different between the regular season and the postseason?

Unfortunately, as is usually the case when quantifying what held back the Thunder, the problem seems to be Jeff Green. It’s almost cruel to pick on Green at this point, but his +/- acts as a sufficient sandbag for the otherwise successful performance of Westbrook and Durant as an on-court duo.

The Thunder’s regular season on/off data is greatly affected by the fact that the core of Westbrook, Durant, and Green played a ton of minutes together, while Durant and Eric Maynor (Westbrook’s backup) only played limited minutes with Green in the lineup. If we look at every single Thunder lineup that saw more than 25 minutes* of action this season (via BasketballValue, natch), five of them featured the full trio of Westbrook, Durant and Green, and those lineups played a grand total of 1,116.7 regular season minutes. During that substantial sample, those lineups were collectively 1.6 points worse than their opponents per 100 possessions. Green isn’t a horrible player, but he was a horrible fit. OKC’s former starting lineup was a substantial part of the problem, as the Westbrook – Thabo Sefolosha – Durant – Green – Nenad Krstic combination was 0.9 points worse per 100 possessions than their opponents over a huge, 542-minute sample.

However, if we look at all of the lineups that played more than 25 minutes this season featuring Westbrook and Durant without Green, we see a completely different result. 10 different lineup combinations that fit that criteria totaled 921.6 minutes of playing time over the course of the regular season, and with Green out of the mix, those lineups were collectively 7.8 points better per 100 possessions than their opponents. In contrast to the old starters, the new starting lineup is doing quite well; the Westbrook – Sefolosha – Durant – Serge Ibaka – Kendrick Perkins lineup was 5.8 points better than their opponents per 100 possessions in 271.4 regular season minutes.

The Green Effect holds true even if we replace Westbrook with Maynor; qualified lineups featuring Maynor and Durant without Green were 12.7 points per 100 possessions better than their opponents in a small, 93.5-minute sample, while the one qualified lineup featuring Maynor, Durant and Green was 12 points worse than their opponents over 100 possessions. The only difference is that Westbrook played more minutes with Green and Durant both (1,116.7) than he did with just Durant (921.6)**, while Maynor played far more minutes with Durant alone (93.5) than he did with Durant and Green at the same time (28).

Westbrook is far from perfect, but he’s also no monster. His positive impact far outweighs his detriment, and if we filter through some of the noise in the lineup data available, we find that he and Durant actually work quite well together, so long as Jeff Green isn’t around. Westbrook doesn’t set up Durant as well as he should and sometimes takes shots even when he shouldn’t, but the Thunder are still performing at an elite level in spite of those hiccups. There are no omens in Westbrook’s errors, only lessons; he’s figuring out the best ways to be effective with every step he takes along the way — even the ones in the wrong direction. He’s learning. He’s adapting. He won’t develop Chris Paul’s court vision, but frankly he doesn’t have to; Westbrook is Westbrook, and while he may not fit neatly into the point guard mold, he and Durant can remain the core that ushers the Thunder to greatness.

———-

*25 minutes may admittedly be a somewhat arbitrary endpoint, but it’s close enough to BasketballValue’s own 33.4 minute qualifier while also allowing some of the lesser used Maynor lineups to come into play.

**These minute totals reflect only the sum of all lineups that qualified under the 25-minute criteria.