For years now, analysts have created awesome new all-in-one metrics to quantify a player’s overall value. It started back with John Hollinger and PER, and has evolved since into metrics like BPM (Box Plus-Minus) and PIPM (Player Impact Plus-Minus). And while, as Seth Partnow mentioned in his Q&A for The Athletic, these metrics can be extremely useful, they don’t paint a full picture of a player’s skillset. They are the best guess we have at the “what” of a player’s impact but don’t contain much, if any, info about the “why.”
To focus on the why, I looked at what I thought were the ten most important factors for a basketball player (other than off-ball movement, which is hard to properly quantify without access to tracking data). The players are grouped by generic position (Guard, Wing, Big) from Cleaning the Glass’ denominations. I then standardized their scores for each of the ten factors to create the percentile ranks. Now, using the tool, we can see, visually and quantitatively, how a player compares to his peers in the same position.
We might not need these charts, say, to tell us that Kawhi Leonard is one of the most complete two-way players in today’s game, but they can be helpful in unearthing some aspects of play that may otherwise go unnoticed. Take Jayson Tatum, for one. While the Celtics wing endured a well-publicized sophomore slump on the offensive end last season with his decreased efficiency and proclivity for contested mid-rangers, his impressive defensive performances went relatively under the radar last year. As noted by the bars in yellow, the third-year wing didn’t rank below the 79th percentile in any of the three defensive metrics used.
But the tool becomes even more useful when one is trying to compare an individual player, not against everyone else in the same position but, instead, against a specific player. (One important caveat when comparing players from different positions — since the percentiles are based off positions, if a big has a higher 3-point shooting percentile than a guard, it doesn’t necessarily mean he is a better shooter, just that he is better compared to fellow bigs than the guard is compared to fellow guards).
For example, let’s examine the case of D’Angelo Russell and Donovan Mitchell. While Russell is entering his fifth season and Mitchell just his third, both are 23 years old and led their teams to the playoffs last season. Could the Nets have made the playoffs last season with Mitchell as their lead guard instead of Russell? Granted, Mitchell played as a shooting guard last season, but he was the Jazz’s main initiator and, at 6-foot-3, is actually shorter than the Warriors’ new addition.
In the chart below, Mitchell is the player in focus and so his charts are displayed while Russell’s rankings are indicated by the black lines. The two of them weren’t extremely efficient shooters last year, as both ranked around the 40th percentile in terms of true shooting percentage amongst guards. Much of that, though, could be attributed to the extremely high workload that both carried throughout the season. Mitchell was essentially the lone creator on that Jazz team last season and endured a miserably inefficient start to the season before turning it around after the turn of the year.
For 3-point shooting, I used Mike Bossetti’s Adjusted 3-Point Percentage which accounts for closest-defender distance. Russell put the league on notice last season with his ability to get hot in the blink of an eye with a barrage of 3-pointers, but while Mitchell is more known for his athleticism, the Jazz guard was surprisingly efficient from downtown on a high volume of attempts. Both should also enjoy some better looks from distance this season, as Russell will have Steph Curry to attract defenders while Mitchell now has three solid shooters surrounding him in the Jazz starting lineup in Bojan Bogdanovic, Mike Conley, and Joe Ingles.
Mitchell’s biggest advantage over Russell comes in his ability to go downhill and attack the basket, and that difference is shown in the charts in his better Finishing and Drawing Contact figures. One of the biggest critiques of Russell’s game has been his low free throw totals. While he showed flashes of an improved ability to get to the basket during his time in Brooklyn (especially during his incredible 44-point effort against the Kings), DLo has developed a lethal floater and mid-range game which he predominantly uses in the pick-and-roll. On the other hand, Mitchell is far more comfortable attacking the basket, and, when he got there, finished at a much more efficient rate than Russell. Despite being just 6-foot-3, his 36.5-inch standing vertical leap allows him to elevate with opposing bigs and either finish or draw a foul.
But while Mitchell often puts his head down and attacks the basket, Russell is as much a playmaker as he is a scorer. Coming out of Ohio State, his passing ability was one of his most talked-about traits, and while he caught the headlines for his shooting, Russell has also been an excellent passer during his time in the league. Last year, he ranked in the 87th percentile amongst guards in terms of assist points created, a lot of which came out of the pick-and-roll.
One of the most interesting storylines going into next season is how D’Angelo Russell will fit alongside Steph Curry, Draymond Green and, eventually, Klay Thompson, for the Warriors. Russell spent a lot of time in the pick-and-roll last season for Kenny Atkinson in Brooklyn, finishing an average of 11.4 possessions a game as the ball-handler, behind only Kemba Walker in the entire league. While he had always been an above-average pick-and-roll guy, Russell’s pull-up proficiency coupled with his playmaking ability coming off screens established him as one of the elite players in those situations in the league. The Warriors, on the other hand, ranked last in pick-and-roll possessions per game with their motion offense. Russell’s superior playmaking ability should still be evident next season, but it remains to be seen whether he will be as effective a shot-creator without the benefit of a screen.
As Seth and many members of the media have noted recently, judging defense is hard. In measuring both defensive activity (deflections, blocks, charges) and shot-contesting using 538’s DRAYMOND figures, I tried to capture both on and off-ball defense, but truly quantifying defensive ability still requires some work.
While Mitchell showed up as being more disruptive on the defensive end, a lot of his numbers are skewed by the fact that he had Rudy Gobert to protect the rim for him. With the ability to usher offensive players into the lane and into Gobert, he was able to gamble more and look for deflections.
As Ben Dowsett noted in his insightful article about Mitchell’s defense, he will be operating in this system for the foreseeable future and Mitchell’s future role for the team could be as something of a “ball hawk,” where he uses his 6-foot-10 wingspan to generate turnovers. The Jazz guard’s DRAYMOND rating was low, but again, a lot of that has to do with the fact that, in Quin Snyder’s system, he is happy to direct players into Gobert and thus might not be as close on the shot-contest.
Russell has great size for the position, at 6-foot-5, and with a 6-foot-9 wingspan, should also have the potential to be more of a disruptor on the defensive end of the floor. But his lack of elite athleticism makes it difficult for him to stay in front of quick guards, and while he performed well in 538’s rating last year, Russell will have his hands full next to Stephen Curry on the defensive end in what will be an undersized backcourt. The Nets were able to hide him for the most part on that end of the floor last season, but without Klay the Warriors will need to see improvement from Russell on the defensive side of the ball.
Determining which one of these players are better at the current moment is extremely subjective. They both have their positives and negatives, but with these bar charts, my main goal was to visually display all-round information about a player’s skillset and allow the user to make a more informed analysis.Other features
While the primary use for these charts will likely be to compare players, there are a few other potentially useful features. For one, the similar players tool on the right-hand side displays 15 similar players in the same position as the highlighted one (this can be seen on the Mitchell vs. Russell chart). The similarity is based solely off percentile values, but the tool allows us to highlight players with similar skillsets who could potentially be used to replace someone else in free agency. Finally, there is a view which lets us compare a player to the rest of his teammates — both from last season and this season. For example, we can see above how Mike Budenholzer and Bucks front office have surrounded Giannis with above-average shooters over the last two seasons to best maximize his skillset.
I have attached the full file below, so feel free to explore and compare anyone who played over 1000 minutes in the NBA. You can also access the interactive visualization here.