Nylon Calculus: Is Russell Westbrook producing more than Michael Jordan?

Michael Jordan inducted Russell Westbrook into the Oklahoma Hall of Fame in November. During the ceremony, Jordan reflected on how he saw a lot of himself in Westbrook:

"“I am truly a fan of his. If you can ever say – being as we’re so many years apart – that when I watch him play, I see a lot of resemblance of his passion for the game of basketball, the way I played the game of basketball.”"

I, too, see those similarities and have written about them multiple times. I once described Jordan as “a 6-6 Russell Westbrook”, and later wrote that Russell’s relentless approach to basketball made him more similar to Jordan and Allen Iverson than to Oscar Robertson or Magic Johnson, despite his and the latter duo’s shared penchants for triple-doubles.

But as the calendar turns to 2017, Westbrook is in the midst of an offensive season the likes of which has never been seen. Jordan, despite having a strong case to be the greatest basketball player that’s ever lived, never had an offensive season as strong as the one that Westbrook is currently turning in.

Read More: Charles Barkley was one of the original stretch-4s

I know that statement is borderline basketball blasphemy to many, so let’s dig further into it, starting with the box score numbers. While Jordan may have maximized his greatness in the 90s as a champion, his most impressive individual offensive seasons came in the late 80s when he was trying to carry the Bulls by himself. Here is how current Westbrook shapes up against 1988 MVP Jordan, at his statistical peak:

1988 Jordan vs 2017 Westbrook, per 100 possessions

Jordan: 43.6 pts (60.3% TS), 7.4 asts, 3.8 TOs, 2.1 off. Reb.
Westbrook: 44.3 pts (54.7% TS), 15.3 asts, 7.6 TOs, 2.8 off. Reb

Jordan was a marvel of individual volume and efficiency. As I showed previously, none of the high-scoring wings of the past 30 years could match his scoring volume, and to do it with such a high true shooting percentage without even really utilizing the 3-point shot is amazing. Then, add 7.4 assists per 100 possessions with a just under 2-to-1 assist/turnover ratio, and we have a combination of individual scoring and team offense creation that has never before been seen from a shooting guard.

However, Westbrook’s individual volume dwarfs Jordan’s. He doesn’t come close to Jordan’s scoring efficiency, but he produces similar scoring volume and surpasses Jordan with a just over 2-to-1 assist/turnover ratio but at twice the volume. So, while Jordan was more efficient at creating points for himself than Westbrook is, Westbrook is slightly more efficient at creating offense for his teammates than Jordan was, and Westbrook directly creates a lot more offense for his teammates than Jordan did for his.

Impact of offense creators vs. efficient scorers

In June of 2015, Jeremias Englemann (creator of the RPM stat on ESPN) released a 15-year RAPM study from 2000 to 2015 that characterized players according to their offensive, defensive, and overall RAPM. Here are the top 20 Offensive RAPM scores from that list, indicating the 20 players whose presence correlated with the biggest improvements in their teams’ offenses.

Note that 11 of those top-20 offensive impact players are point guards.

Of the non-point guards, six (LeBron James, Kobe Bryant, James Harden, Manu GInobili, Dwyane Wade and Tracy McGrady) all played a lot of on-ball, lead guard for their respective offenses. The remaining three, Ray Allen, Peja Stojakovic and Dirk Nowitzki are among the greatest long-range shooters of all-time, very arguably the best long-range shooters at their respective positions in history.

Thus, the top-20 consists of three all-time shooters/floor spacers and 17 players that create a lot of team offense with respect to their positions. Many of those 17 are not hyper-efficient scorers, yet they made a large positive offensive impact through their ability to create for the team.

This supports the notion that creating team offense is one of the biggest ways to have a large offensive impact. Andrew Johnson argued that shot-creation was more valuable to a given individual than shooting. I demonstrated that for a list of the highest scoring wing-forwards of the past 30 years, assist percentage had a greater correlation to higher impact as measured by WOWYr than did scoring efficiency.

Carrying the offensive load while creating team offense: Usage% and Assist%

The archetype position responsibilities would have the point guard as the team offense creator/floor general and the shooting guard/wing as the scorer. Oscar Robertson and Jerry West in the 60s were two of the first and best to combine both responsibilities at heavy volume. Jordan pushed that envelope further in the 80s and 90s, scoring at unprecedented volumes while still playing a large part in creating offense for his teammates. The generation after Jordan wanted to “Be like Mike”, so there were a plethora of high-scoring/distributing wings that have come since Jordan. But in general, point guards still handle the lion share of the distributing while the wings carry more of the scoring load, even if there is overlap. Let’s look further at this, with a test dataset.

Matt Johnson compiled single-season PI RAPM studies from 1998 thru 2012 in one spreadsheet (studies from either Jeremias Engelmann or Across The Court). He normalized and scaled each study using study using standard deviations from a multi-year APM study to create a dataset of single-season RAPM scores that are comparable on a single scale.

Let’s choose the top-30 guard/wing offensive seasons from that dataset (only one season per player) from 1999 – 2012 by that scaled, normalized ORAPM (ORAPMsn), and compare each season using a measure to estimate how much of an offensive load a player is carrying (Usage%) and an estimate for how much team offense a player is creating (Assist%).

I added 1988 Jordan, 2017 Westbrook and fellow 2017 MVP frontrunner James Harden to the plot and highlighted them in red. The point guard seasons are highlighted in blue. What immediately jumps out is that the point guards are all clustered in the high assist/relatively low usage area of the plot, indicating that they are still mainly fulfilling the point guard archetype.

The wings vary by responsibility/style, with shooters like Reggie Miller, Jeff Hornacek, Thunder-version James Harden, Jason Richardson, Stephen Jackson, Ray Allen and Big-3 era Paul Pierce more in the lower assist/lower usage area of the plot. This makes sense as they were making a big part of their offensive impact through spacing the court for teammates.

Higher up on the usage axis you’ve got your high-scoring wings that vary from more pure scorer (Carmelo Anthony, early Kevin Durant) out to high-scoring team creators (LeBron James, Tracy McGrady, Dwyane Wade). Jordan from 1988 falls right into this continuum, and current James Harden pushes the envelope with the highest assist percentage among the high-scoring wings.

And then there’s Westbrook. He is out on an absolute island, with by-far the highest assist-percentage and by-far the highest Usage-percentage of any of the offensive impact players in this data set. Westbrook is carrying an offensive load, both as an individual scorer and a team distributor, that no one in NBA history has ever carried before.

Jordan 1988 vs. Westbrook 2017, offensive impact via OBPM estimate of ORAPM

The best estimates of impact that we have are the family of +/- stats, but we don’t have that data for 1988. Thus, the only way to compare 1988 Jordan with 2017 Westbrook on the same scale is through composite boxscore stats that estimate impact. Basketball-reference offers two such stats for offense only, Offensive Win Shares (OWS) and Offensive Boxscore Plus Minus (OBPM).

1988 Jordan: 0.22 OWS/48; 9.8 OBPM
2017 Westbrook: .154 OWS/48, 10.1 OBPM

OWS is strongly correlated with scoring efficiency, so it is not surprising that Jordan measures out better in that stat. But OBPM was derived specifically to approximate actual offensive +/- results, so it’s interesting that Westbrook currently measures better than Jordan there. Let’s check how well OBPM correlates with Offensive single-season prior informed RAPM using the same Johnson dataset described above. Here is a plot of the top-30 individual offensive seasons (only one season per player) from 1999 – 2012 by that scaled, normalized ORAPM (ORAPMsn) vs. the OBPM from those same seasons:

There is a general linear correlation between the ORAPM and OBPM scores. A line with a slope of 1 shows that where the estimates differ, in almost every case it is OBPM under-estimating the ORAPMnc. But we need to know by how much, so the following plot shows the (ORAPM – OBPM) vs true shooting percentage for all 30 seasons:

There is a general negative correlation between the difference (ORAPM – OBPM) and true shooting percentage. Generally, the higher the player’s true shooting percentage, the smaller difference between their box score estimated impact (OBPM) and their actual offensive +/- calculations (ORAPMns).

Bringing it back to Jordan and Westbrook, these graphs indicate that their OBPM scores likely correlate with but actually underestimates what their ORAPM scores would be…but that since Westbrook’s true shooting percentage is lower, his OBPM score likely underestimates his actual impact to a larger degree than Jordan’s OBPM score underestimates his actual impact.

Thus, while Westbrook (10.1) already has the slightly higher OBPM than 1988 Jordan (9.8), if anything we would estimate that Westbrook’s actual offensive impact is even higher.

Next: A first pass at Dredge statistics for 2016-17

Michael Jordan in 1987-88, and Russell Westbrook to date in this season, both put together some of the most remarkable offensive numbers in league history. Jordan is the more efficient scorer, but Westbrook is creating at a volume (combining individual and team production) never seen before. Team offensive creation tends to correlate with larger individual offensive impact than does scoring efficiency, and the best boxscore-based estimates of offensive impact also indicate that Westbrook is currently having a larger offensive impact than peak Michael Jordan. While Jordan himself feels that Westbrook is a lot like him, in some ways it is possible that Westbrook may just surpass him.