Nylon Calculus Week 22 in Review: Proliferation of high-usage stars

This season has been one unlike any other, and it’s winding down. The articles now are concerned with a couple close playoff races, tanking performances, and awards talk. But, given the theme of this season, I think there’s still a lot of learn, and no one person can watch the entire season. We have all had different experiences with different bases of knowledge — it’s like we all saw the same movie but edited differently. There’s plenty more to talk about.

Devin Booker’s 70-point special

In perhaps the most unexpected event of the season, Devin Booker scored 70 points versus the Celtics without the aid of overtime. Yes, only five other players have scored that many in a single game: Wilt Chamberlain six separate times, and then one each from Elgin Baylor, David Thompson, David Robinson, and Kobe Bryant. What we witnessed was an event that doesn’t happen but every decade, roughly, with the exception of the era of Wilt Chamberlain’s scoring exploits. The question now is what we do with this information. It’s an incredible scoring feat, so how will we view his future and his potential?

Devin Booker’s inclusion in the 70-point club is going to boost his resume, as the rest of the group is all Hall of Fame caliber players and superstar scorers. Booker is far from that, and while his 21 points per game average seems nice, he’s the lead scorer on an awful team with mediocre efficiency. His future isn’t set in stone, but even the worst 60-point players were still All-Stars, like Klay Thompson and Tom Chambers. In fact, you have to go to 59 points to find the nearest average player: Purvis Short, who did this for the Warriors back in 1984 with his patented rainbow jump shot. He actually scored 57 in another game too, and he averaged 28 points per game in his best season — but he provided nothing else but scoring, and he usually wasn’t efficient.

Read More: Jeremy Lin’s return and his impact on the Brooklyn Nets

Muddling Booker’s performance is how his team catered to his scoring total. They force-fed him the ball in the later stages of the game, and intentionally fouled several times so he could get the ball back quicker. After calling a timeout twice to advance the ball and write-up a play for Booker in the final minute when the game’s result was already decided, Phoenix’s coach Earl Watson defended his actions stating, “It was about letting our kids be great.”

However, extreme scoring events often occur with extreme circumstances. The nature of Wilt Chamberlain’s game is dubious and controversial. David Robinson was gunning for the scoring title for the season, so the basketball played that day was not exactly pure; he scored seven points in the final minute of a meaningless game, which caused the opponent’s coach to say it made a mockery of the sport. The same scenario happened in 1978 when David Thompson scored 73. Even during Elgin Baylor’s game the media mentioned he was fed the ball quite generously in the closing minutes. Devin Booker’s game wasn’t an outlier in that regard; those games are usually forced. But be careful projecting his future — we have so few 70-point scorers that we definitely don’t have enough of a sample to make a definitive conclusion about what the scoring says about the player.

Swipe right

I’ve entertained a number of explanations for why homecourt advantage in the NBA has been declining for years, but here’s one I didn’t consider: Tinder has made it easier for NBA athletes to find women, and as such players are more efficient at allocating their time and they can get more rest. I would, however, emphasize that it’s one of many related causes, because teams are getting better at pampering their players in transit and finding them healthful sleep through modern medicine. But this does make sense.

Homecourt advantage varies based on how many games of rest a team has, how many time zones they travel, and what the altitude differences are. In fact, the league’s movement towards reducing the number of back-to-back games is greatly beneficial too since most back-to-back’s happen for away teams. But this shows how competitive the NBA is. Any way you can find that increases rest and health, pushing them away from the late-night club life, will have real benefits, no matter the method.

Bradley Beal: Has his time come?

For years, the assumption on Bradley Beal was that he was a star scorer in waiting, and despite several seasons of mediocre production, the perception of his potential hadn’t waned. It’s because — no surprise — he’s a scorer and was a lottery pick. But his numbers are up this season, and I’ve seen a lot of discussion about his breakout. This topic has a long history, and it’s partly tied to something that was not under Beal’s control: Randy Wittman’s control as head coach and his proclivity to design an offense heavily in favor of less efficient shot types (i.e. midrange shots.) Maybe I’ve missed the comments about the team, but perhaps we should not be surprised Beal and other guys like Otto Porter Jr. are having greatly improved seasons.

We can see the effects here when we look at how Beal has actually changed, and it’s all to do with his shooting — that’s it. His assist rate hasn’t changed, his rebounding is still low, and his defensive stats are still troubling. But, using Basketball-Reference’s stats, his rate of shots at the rim and beyond the 3-point line has gone up 16 percent, and his free throw rate has gone up too. He’s hitting an unusually high percentage of his long 2-pointers. That’s partly a function of how he’s changed his game, eliminating the great number of awful off-the-dribble midrange shots, but it’s a bit unsustainable too.

However, they don’t take up a large share of his offense, and he’s actually had an improvement in an area of his game I don’t think many people are mentioning: his free-throw percentage is at a career high, and he’s taking more of them too. His percentage from the charity stripe may regress, but if he can keep drawing free throws and shooting more often from behind the 3-point line, he can remain an efficient scorer.

Table: Bradley Beal’s shot distribution

source: b-ref

In a weird way, the fact that his production hasn’t changed much over his first four seasons bodes well for him. Basically, he’s built up latent potential, as his performance as a younger player indicates a stronger player than we’ve seen until this season. NBA players grow in fits and spurts, not smoothly. I still don’t expect him to transform into a star player, because of his defense, his injury history, and his limited ball-handling, but he should be a more potent offensive weapon from now on.

Ricky Rubio es mágico

Here’s the Unseld of the week, which is a marvelous backwards-toss, across half-the-court assist from Ricky Rubio. I know he’s loved substantially more by some subsets of the NBA-sphere, but I hope fans are appreciating him. That pass is a direct feed to the joy of basketball, and it’s the type of high-value play that can get lost when statistics lose their resolution as it was boiled down to just an assist.

The art of the charge

As someone who discusses charges and offensive fouls drawn in general as much as anyone, I had to comment on a recent FiveThirtyEight article about charges and Anthony Tolliver. The article uses a charge rate stat, which I believe is just charges/(charges + blocking fouls), where blocking fouls can be coded as “shooting block fouls” from what I’ve surmised. But for some reason, the stat provided only goes back to 2015, and available data actually goes back to 2011. In the sample FiveThirtyEight provides, Anthony Tolliver has an unparalleled rate. However, I found that three other players had nearly equal rates: Gerald “Crash” Wallace, Kurt Thomas, and Shane Battier, a player whose defensive skill and overall value helped inspire me to investigate non-traditional defensive metrics.

I should also say that it’s unfortunate we don’t have the data available for the prime seasons of Jason Collins, who was the maestro at drawing these stats. Lastly, I need to mention that we’re overvaluing the charge and ignoring all offensive-fouls drawn that aren’t classified as charges. This is unfortunate because my research, and others, have found that these other offensive-fouls are more valuable, and they’re available going back to 2006 (plus just 2000, for some reason.) If there’s a real interest in this stat, I can provide it somewhere, though I question whether or not it’s an improvement on anything — that’s for some testing I may attempt later on. For now, we can marvel at Tolliver’s performance here, but he wasn’t without his peers.

Defensive tools

Thanks to the internet, there are an immense number of defensive statistics to apply in the Defensive Player of the Year discussion, which appears to be concentrating on Draymond Green and Rudy Gobert. You really don’t even need advanced stats to reach that conclusion, but you can get into the weeds when you try to figure a, say, top 5. And please don’t rely on raw on-off stats for this — that’s not a reliable statistic. It doesn’t adjust for teammates, opponents, and it’s prone to wild fluctuations. You generally need multiple seasons even for an adjusted plus/minus to have stable results, and even with more advanced (and better) techniques like RPM, you can still get weird numbers.

For instance, if one were to rank players by DRPM*minutes played, Rudy Gobert is first by a decent distance, then there’s Draymond Green, a bigger gap, and then you get Anthony Davis, Robert Covington, and Gorgui Dieng. I don’t suppose people are arguing for those latter two players. These raw plus-minus stats are even worse. For instance, there’s one candidate I haven’t even mentioned, and it’s because his team has been defending much better with him off the court — but the reasons have more to do with bad luck, as opponents have converted unusually high percentages when he’s on the court, than anything else. Be warned.

This sounds crazy but don’t eat an entire bag of sugar a day

In case you missed it, there was a wonderful, long-form article about the role of peanut-butter and jelly sandwiches in the NBA. It’s a story about comfort, simple food versus elegant food, athletic performance, and the current pushback against sugar. I just had to point out Dwight Howard’s sugar addiction: he apparently used to consume the sugar equivalent of two dozen candy bars, and this was discovered when he complained of tingling in his leg while on the Los Angeles Lakers. I’ll go with a conservative estimate here where I’ll assume a candy bar has on average 25 grams of sugar. That translates to 600 grams of sugar, which is, yes, 1.3 lbs, for those who aren’t comfortable with metric units. I can’t fathom that. This isn’t even a new story, but I’m surprised I don’t remember seeing this before; it certainly explains how he broke down so young.

For the curious, the LD50 of sugar is 30 grams per kilogram, at least for rats (LD50 refers to the amount of a substance in grams that would kill 50% of the subjects per kilogram of their own weight). Dwight Howard is listed at 265 lbs, which is about 120 kg. So no, he wasn’t close to his LD50 target of nearly 3600 grams of sugar, but I’m still concerned.

But there is a bit of a sugar problem in the NBA (and the world); this wasn’t the first weird case I’ve seen. Nikola Jokic used to drink three liters of Coca-cola a day, which is “only” 324 grams of sugar. Caron “Tough Juice” Butler used to drink two liters of Mountain Dew a day, which is about 250 grams. Derrick Rose had a notorious sweet tooth too, stating that he literally ate two pounds of candy a day sometimes. These athletes are, or were, young, and they’re able to process a shocking amount of food. But they’re multi-million dollar machines who should not be consuming an amount of sugar that would be downright dangerous for anyone. You won’t get that by eating a pb-and-j before a game, but please don’t eat two pounds of candy.

How shot usage has changed over the years

This season has been an incredible one for individual statistics, from Russell Westbrook’s path to the triple-double average to the astonishingly high number of 50-plus point games. There are a number of explanations, mostly tied to the offensive renaissance and the greater reliance on more efficient shot types. But there’s another trend that’s been gnawing at me, and it’s one that seriously affects our accounting of who the best players are and how they compare historically. Stars are dominating the ball now more than ever, and it might truly be unprecedented.

You can quantify how much a player is dominating his team’s offense through usage rate, the common formula that shows the proportion of field goals, free throws, and turnovers a player is using compared to the rest of his team. But that’s only available for seasons going back to 1978, when individual turnovers were first tracked, so we don’t have a complete picture of the pattern of usage rates historically. Consequently, you usually only see studies going back to 1978, which doesn’t show the full picture.

The solution here is to use my shot usage% stat, which is simply the proportion of field goals and free throws a player is taking relative to his own team (like the traditional usage percentage, this equates a free throw to a field goal with the coefficient of 0.44.). With that, you can calculate this stat for the entirety of the shot-clock era. Let’s start with something simple: how does the standard deviation change year-by-year? Standard deviation should show how shot usage can become concentrated at the top because it’ll lead to a greater spread and, therefore, a higher standard deviation. You can see a graph below.

During the early periods of the NBA, the standard deviation fluctuated greatly, and it was usually under 4.5 percent. However, by the mid-70’s, it stabilized a little more around 5.0 and has only oscillated a bit since then. But this season is different; the spread is significantly higher. Over 5.7 percent of the players right now with at least 1000 minutes have a shot usage rate of 30 percent or higher — that used to happen with half the frequency or never at all. In fact, there are four players with a shot usage rate of 35 percent or higher, which translates to 1.5 percent of the league, and it’s only ever been above one percent twice — in 1962 when only Wilt Chamberlain qualified due to his whole averaging-50-points-per-game-thing and in 2006, the year of Kobe Bryant, where Tracy McGrady and Allen Iverson also made the cut.

But we can do a little better with what’s known as a kernel density plot, which can show the distribution with more precision. It’s more illuminating to to see these for every season, but that’s a few dozen images. The solution here is the GIF, and I hope the timing is correct — you can see how the distribution of shot usage flexes and darts as the years go on by. The right tail is probably what people are most concerned about, and indeed 2017 is again an outlier. But the left tail of the graph is important too. If we have more ball-hogs now, does that mean we have a higher proportion of super-low usage players too? It appears that’s not the case, and that was actually most common in the 1990’s.

Anecdotally, people have noticed that there have been a strange number of big games this year, but we have some evidence to back things up now: the “usage inequality” of the league is higher than it’s ever been by a significant amount. This leads to more high scoring games and more players who do very well by all-in-one metrics.

Next: Nylon Calculus -- Golden State's offense before and after All-Star Weekend

We’ll see if this trend continues — perhaps it’s a one-year blip or the end of the season games will regress things further. But this might be here to stay, and we’ll probably see a number of articles on why this is happening, from the data-ball movement to the effects of hand-checking. Russell Westbrook is not an anomaly but simply one of the trend’s best examples.