What impact does wingspan actually have?
When evaluating NBA prospects, many things are taken as givens. Players who shoot well from three-point range in college will be able to do so at the NBA level. Point guards shorter than 6 feet will be poor defenders. Players with long arms will be better at defense than players with short arms. However, none of these assertions are universally true. Today I am tackling the third – the idea that having longer arms will make a player a better defender.
I went into my research with an open mind, ready to discover anything about the way wingspan did or did not correlate with defensive impact. Before I dove into data, I had to come up with a player pool. I decided to take all players who had debuted in the 2013-14 NBA season or later as my initial pool. The reason for this was twofold – I wanted to eliminate experience as a factor in defensive ability, focusing just on how wingspan affected defensive ability. I also wanted to use ESPN’s DRPM (Defensive Real Plus-Minus) as one of the statistics for quantifying defensive impact, and the RPM (Real Plus-Minus) database only runs back to the 2013-14 season. Taking players who had come into the league during the “RPM era” allowed me to (mostly) eliminate experience as a factor in defensive impact and focus instead on wingspan. I had to remove players from the pool for various reasons in order to create a reliable sample:
- I disqualified all players who had not played 300 minutes in the NBA (sorry to a group of guys that are too numerous to list)
- I disqualified all players who did not have easily accessible official measurements found on DraftExpress or elsewhere (sorry Georgios Papagiannis, Briante Weber, Juan Hernangomez, Yogi Ferrell, Alex Abrines, Nicolas Brussino, Travis Wear, Nemanja Bjelica, Salah Mejri, Nick Calathes, Miroslav Raduljica, Vitor Faverani, Toure’ Murry, and Gal Mekel)
- I disqualified all players who debuted at over the age of 30 (sorry Marcelo Huertas and Pero Antic)
- I disqualified a number of players due to difficulty finding DRPM data for certain seasons for them – most of these guys barely cracked the 300 minute threshold which made cutting them easier (sorry CJ Wilcox, PJ Hairston, Cleanthony Early, Markel Brown, Devyn Marble, Cory Jefferson, Tarik Black, Jakarr Sampson, Lamar Patterson, Mitch McGary, RJ Hunter, Anthony Brown, Jordan McRae, Erick Green, Furkan Aldemir, Kostas Papanikolaou, Malcolm Delaney, Isaiah Whitehead, James Ennis, Elijah Millsap, Raul Neto, Tim Frazier, Lorenzo Brown, Phil Pressey, Ray McCallum, Anthony Bennett, Archie Goodwin, Ian Clark, and Axel Toupane)
- I threw out extremely short seasons (Julius Randle 2014-15, Sam Dekker 2015-16, Sean Kilpatrick 2014-15, etc) when collecting yearly DRPM/DBPM data
After this trimming, I was left with a pool of 165 players, all of whom had a respectable sample of playing time with which to determine defensive impact.
Wingspan by itself is not sufficient for judging the sample, simply because it does not account for height. A 6’0” player with a 6’6” wingspan is completely different than a 6’7” player with a 6’6” wingspan. Thus, I came up with the simple idea of “Reach”.
I defined “Reach” as the difference between a player’s wingspan and his height (there is probably a word for this that I am not aware of). For example, Jerian Grant was measured at the 2015 NBA Draft Combine with a height of 6’4.5” and a wingspan of 6’7.5”. Thus, Grant’s “Reach” would be 3”. For reference, the average Reach of the 165 player sample was 3.37”.
Next came quantifying “Defensive Impact”. I chose to plug in values of DRPM for each season for each player and then average that out to “Average DRPM”. This is imperfect because it doesn’t account for minutes played each season. However, most players in this sample have a similar distribution of minutes as they are young and generally increasing in minutes played each year – this helps mitigate the issues caused by failing to account for minutes played. I did the same with Basketball Reference’s DBPM (Defensive Box Plus-Minus) to come up with an “Average DBPM” value for each player. I then added the Average DRPM and Average DBPM values to come up with a Defensive Impact value for each player. The averaged value for Defensive Impact among the sample was about -0.65. I’m no expert on defensive statistics, and not accounting for minutes is just one way that this is imperfect, but I considered it to be a fairly successful measure of defensive ability. For reference, here were the top 15 and bottom 15 defenders of the sample according to Defensive Impact.
Defensive Impact is by no means a perfect stat – I’ll be the first one to say it’s flawed – but I like it because it attempts to combine DRPM and DBPM. DRPM focuses more on on/off stats, while DBPM focuses more on individual stats. Defensive Impact is an attempt to synthesize the idea of on/off impact as well as individual output. Based on the results and ranking of players, Defensive Impact did what it was supposed to.
Once I had collected all my data on Reach and Defensive Impact, it was time to answer the titular question; how does wingspan affect defensive impact? Once I had “zeroed” the data – to make axes accurate – by subtracting the average Reach value from every player’s Reach value and adding the average DI value to every player’s DI value, I had a graph with which to work.
I was very surprised when my data spit out this graph. This graph suggests that there is basically no correlation between Reach and Defensive Impact. The worst defender according to DI (Shabazz Muhammad) had a Reach of 4.75”, while the 2nd best defender (Nerlens Noel) had a shorter Reach, at 4”. The player with the greatest Reach (Montrezl Harrell, at 8.75”) was only slightly above average in terms of DI (-0.56), while the player with the least Reach (Kelly Olynyk, at -2.25”) was well above average with a DI of 1.36. Cody Zeller ranked 7th in DI among the sample despite a Reach of -1.5”.
What this all suggests is that Reach should not be a factor when discussing a prospect’s ability to play defense in the NBA. Many draft analysts, myself included, were quick to write off the defensive ability of 2017 prospects like Luke Kennard, John Collins, Dillon Brooks, and others due to (among other reasons) their poor Reach. This isn’t to say that those players will be good defenders in the NBA – all could still be bad defenders (their stats and performance would suggest that) – but citing “short arms” as the sole reason a player will be a bad defender is not legitimate.
This is a double-edged sword; using “long arms” as justification for a player’s defensive chops doesn’t work either.
It is very likely that OG Anunoby and Jarrett Allen were overrated in this year’s class due to misconceptions that their elite Reach values would lead to them both being high-level defenders. While both posted solid defensive stats in college, analysts using Anunoby and Allen’s Reach values to project them as high-level defenders were using the wrong justification. Instead, they could have pointed to Allen’s good-but-not-great shot blocking stats (5.0 Block Percentage) or his solid DBPM (4.6). In the case of Anunoby, they could have highlighted his 3.2 Steal Percentage, his 5.5 Block Percentage, or his 5.1 DBPM. Beyond stats, they could discuss Allen’s quick leaping ability or Anunoby’s jacked frame, multipositional projectability, and freakish athleticism.
In short, one should look factors besides Reach when evaluating a prospect’s defensive ability. Analyze Steal Percentage – a stat that many draft models show to be a good indicator of defensive ability – or any other readily available advanced stats, like Block Percentage, DBPM, or Defensive Rating. Beyond that, one could simply watch the player play to gain an understanding of defensive ability. Defensive metrics are still young and improving, and in many ways watching guys is still the best way to gauge their defensive ability.
This isn’t to say that Reach doesn’t help players. Defensive stars like Rudy Gobert, Draymond Green, and Kawhi Leonard all benefit from having ridiculously long arms. If this data included Kawhi, he’d be a dot way in the top right corner of the graph with his ridiculous Reach of 8” and his defensive play that has won him two Defensive Player of the Year awards. It would certainly seem that Reach helps to raise the defensive ceiling of a player – eight of the top ten DI values from the sample came from players with above-average Reach values. The idea that elite Reach raises defensive ceiling is encouraging for a prospect like OG Anunoby who, prior to his injury, showed great defensive signs outside of his 6.5” Reach.
However, this data suggests that Reach does nothing to raise the defensive floor of a player – just ask Dennis Schroder or Shabazz Muhammad, two awful defenders with great Reach values. Jarrett Allen struggles with mobility, effort, and understanding as a defender, and all these could lead to his failure despite his 6.25” Reach. Reach is simply a way for athletic, smart, strong, high-effort players to reach (no pun intended) another echelon defensively. But without the athleticism, quick-twitch ability, effort, strength, understanding, and nuance that goes into playing solid defense, long arms mean nothing.
Evaluating the defensive ability of players is tough – particularly how their defense will translate from the NCAA or Europe to the NBA – but don’t get caught up on whether a player has short or long arms when measuring his ability to succeed on defense.
You can view the full rankings of Reach, Defensive Impact, and more here: https://docs.google.com/spreadsheets/d/1zHf-TKuMrka–NniPUNBHzq0KSnW8GfKaplZZuxa__g/edit?usp=sharing