Nylon Calculus Week 12 in Review: Is Russell Westbrook padding his rebounding numbers at the free throw line?

Every week I try to summarize the last week in basketball by highlighting a few key news stories, performances, and players. Last time I tackled All-Star selections, so I’m a bit behind with news links, but there is a real point to this: we sometimes lose sight of big events because in the NBA cycle even two days is a long stretch of time where just two or three headlines can erase the previous stories.

I’d also like to tackle normal headlines under the guise of an analytics blog, breaking down videos and opinion pieces. If you believe in the power of numbers and objective empiricism, you can still enjoy dunks and great passes and personalities too. There’s no magic voodoo here, and the stat-inclined fans watch just as much basketball as anyone. And with that, here’s week 12 of the NBA in review.

Salt Lake Hill

One bad habit NBA fans have, and really everyone, is the inclination to completely dismiss a new fact or system because it doesn’t agree with one particular bit of information they already believe. If an NBA metric has an anomalous result, and it strongly goes against conventional wisdom, people reject and deride said metric, instead of looking more at the data and information trying to figure out why the result happened — and in the process learning more about the player and the NBA in general. That process of discovery is extremely important for becoming more knowledgeable.

Read More: Isaiah Thomas is a giant

I wonder if that’s happening with BPM and the Utah Jazz. If you asked someone who was leading the Jazz in BPM, you’d get guesses of Gordon Hayward or Rudy Gobert, split pretty evenly too. But George Hill is currently leading the team: 5.9 to Gobert’s 5.3 and Gordon’s 5.0. He’s actually leading them in PER and WS/48 as well (23.3 to 23.0 from Gordon and 21.7 from Gobert, respectively). and nearly leading the team in RPM too. This isn’t far from Hill’s career high in BPM either, which would surprise most; he hit 5.6 a couple years ago. There’s a common thread here: he had one year on the Pacers where he had a much bigger role. His usage rate was well above average, and his assist rate had climbed too. That was the season Paul George was injured, and it was pre-Monta Ellis. It’s just a shame it was a partial season and he was largely ignored. Hill has had a bigger role on the Jazz though, and it’s greatly boosted his overall numbers.

However, there’s something unsustainable about Hill’s numbers, and it’s not something that happened with his 2015 season: his 3-point percentage is abnormally high at 48.1 percent, and consequently his true shooting percentage is well above his previous career high too at 64.1. That’s why he’s leading in so many different metrics, except for ones that regress heavily like RPM, Dredge, and PT-PM. If you substitute his career 3-point percentage into his percentage this season, his BPM would drop by roughly 1.7 points — both Hayward and Gobert would leap ahead of him. But most importantly, Hill would still rate as a very good player. He has a much better role on Utah, and he should be quite valuable for them the rest of the season, even if his 3-point percentage tanks.

What are Millennials ruining now?

Last Thursday, Adam Silver mentioned a few changes the league was considering, and one of them was shortening the game. That change created a few headlines because it included a buzzword: millennials, who are apparently so transfixed by their phones and other devices we need to overhaul the NBA so they can watch. But what most articles and other comments failed to mention was that he actually said everyone’s attention span was decreasing, and only said “particularly” millennials — he wasn’t targeting just them. Plus, solutions to the slow grind at the end of close games have been proposed for years. It’s not exactly natural that it takes several minutes, on average, to play just one minute on the court. So, the NBA isn’t going to change the league just to suit young people, and it’s not like changes haven’t happened before.

Sixty years ago, the Greatest Generation had a waning interest in basketball due to its slow nature and laborious finishes. They weren’t enthralled by low score games and “stall ball” — where leading teams would simply hold the ball to kill time — was killing fan interest. Then came the invention of the shot clock and the NBA as we know it. Maybe the younger generation does want speedier games, but that’s been discussed for decades, from the shot clock to the seven-seconds-or-less Suns, and we don’t need to shame Millennials for it.

Time to dump Defended FG%

There are few objective ways to measure individual defense, so when the NBA unveiled “defended FG%” — every shooter has a nearest defender, thanks to SportVU tracking, so you can total up every “defended” shot for each player in the league — I imagine many people, mostly NBA writers and analysts, rejoiced. There’s even an expected field-goal percentage from the shooters the specified defender has been nearest to, and a resultant difference between the expected and real rates. Thus, people then connect the dots and proclaim the players with the highest differences the best defenders.

The problem is that it’s mostly noise. as Krishna Narsu found recently that there’s virtually no correlation between Defended FG% for a player one year to the next. Thankfully, that’s not true when you only looking at shots five feet and in, but from six feet and out, it’s random. This is especially true of 3-pointers. I’ve discussed it before. It’s a junk stat, and, outside of rim protection, it’s not useful in picking out the best defenders.

You see it cited now and then among various blogs and even at bigger sites. We’ll probably hear about it more when the Defensive Player of the Year arguments heat up. But think about how it’s formulated: should a defender get credit for a jump shot taken far from his fingertips? What if the defender is behind the shooter and turned away? I know we’re all yearning for useful defensive stats, but this is not one we should be using.

Your weekly Giannis fix

Giannis Antetokounmpo looks like a lock to start in the All-Star Game, which marks another notch in his steep rise to stardom. His tremendous length has enabled him to be a defensive menace, and he’s nigh unstoppable near the rim[ I’d implore you to donate to Kevin Ferrigan’s patreon and support his work, if you haven’t already.]. Look at that play embedded below — he needs two dribbles to clear most of the court and dunks nearly from the free throw line. In hindsight, it seems obvious that his unique combination of ridiculous length and ballhandling abilities would lead to an MVP-caliber status, but history is filled with failed draft picks from guys with potential like the horizon and unreal height and wingspan dimensions coupled with guard-like skills. Giannis is the dream every GM has when watching an intentional prospect of mystery, but rarely do those players even reach half that potential, much less becoming more than we ever thought possible. Perhaps he’ll led to a few gambles in this upcoming draft — just remember there are several Nikoloz Tskitishvili’s.

Rampant NBA inflation

This has been a season like few others with a high number of 50-point games, Russell Westbrook averaging a triple double, Giannis being Giannis, and James Harden summoning his inner Magic Johnson. You can view some theories here. The increased pace has helped, as players now have more opportunities to deliver crazy stat-lines than a decade ago. But I think small ball has been helping too. Not only is the league’s obsession with downsizing and 3-pointers creating the most efficient environment since the 1980’s — and most records we care about are offensive — there are certain actions that are easier now, particularly for perimeter players. Rebounding is less difficult for guards with fewer behemoths on the court and, when said guards dive inside, the court is spread and there aren’t as many players near the rim. Maybe some people will lament for an era where 50-point games were tougher to collect, but I for one am glad we’re not living in the dead-ball era of the early 2000’s anymore. This is more exciting.

Bouncing ball

The incomparable Michael Beuoy released a long article slyly titled “Deflategate Analysis.” It’s a thorough piece on a concept known as coefficient of restitution (CoR), which is a scientific/engineering term for measuring bounciness. A coefficient of one means a ball, when released from one foot, would bounce up exactly one foot because it lost no velocity on the bounce; whereas a coefficient of 0.5 means the ball would only bounce partway up because it lost half its velocity. In the NBA, the coefficient of restitution is usually around 0.80 to 0.84. (This means that in terms of height, ignoring friction and assuming a vertical drop with no initial velocity, the ball dropped from, say, one foot would bounce back upwards two-thirds of a foot.)

Why does this matter? A higher coefficient means a ball bounces harder and, thus, is worse for shooters who can benefit from shots that rattle in. You can then check for interesting patterns in the number. For example, the CoR is suspiciously higher in Utah, where people have actually stated the Jerry Sloan-era Jazz would inflate the ball as much as they could legally because it worked well with their style — they were a free throw heavy team who scored inside. Perhaps the arena’s crew got used to the high inflation principle, and it remains today. There’s another explanation too: the CoR is a relationship between two objects that collide, i.e. the floor and the ball, and arena floors may actually have different properties. Regardless, it’s an interesting article, and it’s great to see the seldom-referenced coefficient of restitution principle applied appropriately.

Discrimination and stability of NBA stats

I know academic, formal papers scare people, but they’re more welcoming than most people think and you can just skim to find some interesting tidbits you wouldn’t have otherwise known. This paper, released back in the fall and tweeted out by Mara Averick, discussed multiple sports, but its section on the NBA has some useful info.

For instance, the stability of a stat, which I have actually discussed ad naseum before, is about how well the stat stays consistent with the same players. Then there’s discrimination, which is about how well a stat differentiates players from each other — a stat that’s always one for every player is certainly stable but isn’t very useful. Stats with high discrimination and stability include rebound percentages, assist rate, and block rate — those are the “best” in a sense. However, one stat was lowly rated in stability and discrimination, and it’s by a mile: 3-point percentage. That’s something, again, I’ve talked about endlessly, but it’s for a good reason. It’s not a good statistic, at least when you’re only looking at a single season (and it’s even worse with half a season.)

Malcolm in Milwaukee

The 2016 draft class has been unusually quiet so far, as all the rookie of the year talk has followed a 2014 draft pick, Joel Embiid. But the best rookie so far from the most recent draft may be one taken in the second round: Malcolm Brogdon [He’s leading all rookies in RPM, ignoring Joel Embiid, as far as I can tell.]. He was taken 38th and not rated highly by any publication I could find. Kevin Pelton, for instance, ranked him 73rd, and he’s far down this statistical projection list from Nick Restifo — he wasn’t even discussed when Nick rated the Bucks’ draft day. He was a four-year senior with the kind of scouting report that worried about his athleticism and praised his high character, which isn’t a good sign for a prospect. He was a fantastic shooter though, hitting in the high 80’s from the free throw line his entire career and hovering around 40 percent from the 3-point line his senior season.

Brogdon’s sweet shooting has translated well to the NBA level, where he’s hitting 43 percent of those shots from behind the arc and 89.5 percent from the line. He’s already a solid NBA scorer, and he has a good assist rate too for a shooting guard; he’s played the role of secondary ball-handler next to Giannis Antetokounmpo. He’s been starting ever since Matthew Dellavedova got injured — and he’s still starting now that Matthew’s back.

Malcolm had a great game against the Spurs last week when the Bucks upset them despite only nine minutes from Giannis. You can watch all his field goal attempts here; he’s more than just a spot-up shooter. He uses his driving ability to set-up players too, like this drive-and-dish to Miles Plumlee. He does need to improve his free-throw rate though; there are too many weak missed layups where he doesn’t draw enough contact like this one. But he does have the ability to beat guys off the dribble and get to the rim with ease, which is aided by his excellent shooting stroke — you don’t want to give this guy space. The Bucks have had a quietly remarkable season, beating expectations even though they’ve been without their second best player Khris Middleton. This is yet another boon: a nice guard who’s good enough to start on a super cheap contract.

Free throw rebounds: 2017

Basic stats still dominate NBA conversation, so when a player is tackling a historic milestone, like Russell Westbrook and his triple-double average, the concurrent narratives are hard to dodge. We all knew Westbrook could score, and he can pile up the assists like few others, but his new rebounding prowess is shocking. He was always a good rebounder for his position, but this is unprecedented. Even Jason Kidd didn’t capture this many rebounds per opportunity. Check the graph below — Westbrook is far from any previous point guard season since 1974. Jason Kidd was the previous defensive rebounding king for his position, but Westbrook this season has surpassed him by a monster margin. Also take note of the underappreciated Fat Lever; his versatility was special.

But not all rebounds are the same, and it’s surprising how often we combine the two official types of rebounds — offensive and defensive — into total rebounds. That’s a loss of information, as offensive rebounds are quite distinct and more valuable individually. Westbrook, for example, is one of the most ferocious offensive rebounders at his position in NBA history, but we can slice rebounds into more categories as well. I’ve done this before, but I thought it deserved an update: what happens when you separate rebounds by free throw versus field goal?

Looking at 2017 first, there’s something that stands out about free throw rebounds, and the real surprise is the gigantic margin: Russell Westbrook has 53 defensive free throw rebounds as of Sunday, while the next closest player, DeAndre Jordan, has 39. You can see the tweet and clip that inspired me below. His frontcourt players are often boxing out for him on easy free throw rebounds. I understand it’s important to get the ball in his hands early, but in this case I think it’d make sense for his teammates to pass him the ball while he’s near, say, halfcourt. Wouldn’t that be faster?

I question the accuracy of that tweet, by the way. Westbrook has 54 total free throw rebounds, and he had 37 last year. Perhaps that meant at this point during the season, but in any case it’s still quite a leap statistically.

How does Westbrook compare historically? I can grab these stats going back to 1997, and for a fairer comparison I’ll look at things on a per minute basis. Westbrook is 15th in that time frame with the best rate since 2013. You’ll notice some repeats in the table below, like Marcus Camby and Drew Gooden, and some rebounders who have been labeled overrated by their stats alone, like Carlos Boozer.

Table: top FT DRB rates, 1997-2017, min. 1000 MP

The flipside to that defensive free throw rebound is the offensive board, which are rarer and tougher to collect due to positioning. They’re significantly more valuable, consequently. Referring to the table below, you’ll again see a lot of repeat names, which is a good sign because it’s a mark of discrimination. Westbrook, by the way, is ranked much lower, as he’s only grabbed one offensive free throw rebound. That makes sense, though, because the offense is only allowed two players at the edge of the paint for free throw rebounding, and their frontcourt guys usually take those stations.

Table: top FT ORB rates, 1997-2017, min. 1000 MP

There’s another way to analyze free throw rebounding: proportions of defensive boards that come from missed free throws. Westbrook’s proportion is 14.4 percent, which ranks 289th since 1997 among qualified seasons (1000 minutes played.) It’s not the highest this season either, as Jae Crowder is tops with 17.5 percent, and that’s got nothing on Jalen Rose’s mark of 25.4 percent in 2005. Of course, some of this is influenced by how many opportunities players have, which is controlled by the playing environment. Free throws have been trending downwards for a few years slightly, so it’s better to compare Westbrook to his peers from this season, where his stats are truly anomalous.

Next: A quick look at the Pelicans' stagnant offense and situational play

I realize that Russell Westbrook is a menace to opposing teams no matter what the stats exactly say, and that all stats should be taken with a grain of salt, but it’s important we all understand that not all rebounds are the same, among other truths. Some rebounds are tougher than others, and they should appropriately receive more credit. It’s best when Westbrook grabs a defensive board because he can blitz the defense and lead his team to an easier opportunity, but I question the utility with free throw rebounds, where defenses are more prepared and he could just as easily stay close to the halfcourt line and await a pass from someone else before attacking. Soon we’re going to get serious with MVP discussions, and we should all understand what the stats are indicating.