NBA Week 11 in Review: Winter Winds

The season is coming close to the halfway mark, and we’ve covered a lot of basketball over the season so far. The West is stratified, and the East is still a morass excluding Cleveland at the top and a few on the bottom. We’re now firmly entrenched in the Stephen Curry era where LeBron James has moved down a tier — although nothing is truly over until the season and the playoffs end. But it’s truly a different era. The Spurs are solidly Kawhi Leonard’s team, as it was last season. Russell Westbrook has been Oklahoma City’s best player as they crush opponents. And New York is buzzing about Kristaps Porzingis, not Carmelo Anthony. This is a new world, and there’s still a lot to learn, so let’s head into a discussion of a few topics from last week and beyond.

The Two Empires: the Spurs and Warriors

Right now San Antonio and Golden State are devouring the rest of the league, blowing out opponents with regularity. In fact, using Basketball Reference’s SRS team metric, no season has ever had two teams rated so highly — the Lakers and Bucks in 1971 were at +11.65 and +11.70, respectively, whereas the Warriors and Spurs are sporting ratings of +12.19 and +10.80. You’d expect some regression to the mean the rest of the season — remember when the Pacers were on pace to have the greatest defense ever? — but what’s surprising is that the Spurs have the better rating. Generally, point differential is a better indicator of future performance than pure wins, but point differential has its issues too and it can skew based on a few factors.

When people think of how strong a team is, they think of how much damage the team can do with its best players, not how well they outscore another team through the full 48 minutes. For example, the Spurs outscore their opponents by a significant margin regardless of how many starters are on the court. They actually have a larger point differential with zero starters in the game than all five — that’s the sign of great depth. Defining “bench players” as the five non-starters with the most minutes and “deep bench” as everyone else, the only situation where San Antonio didn’t do well was with five bench players. They even played well with the deep bench guys.

The Warriors, by contrast, have a more traditional pattern. They don’t do well without at least three starters in the game, and they don’t do well either if there’s two or more deep bench guys on the court. That’s normal, but it means during garbage time their huge leads are chipped away minute by minute. San Antonio, however, can increase its lead during that time, inflating its point differential. For instance, using Nylon Calculus’s tool, you can see that the Spurs have the second best rating during the 4th quarter while the Warriors are far down the list with a negative rating. That doesn’t mean they’re “choking” in the 4th quarter, of course; you can look at their stellar overtime rating if you want some hard numbers on how they perform in the clutch. But it’s something to consider as the grand San Antonio-Golden State regular season game drawers nearer, and if we see them in the playoffs. A standard simple point differential, or net rating, is a good reference point, but there’s a lot of information it’s hiding.

Curry’s Trick Pass

For a lighter basketball note, Stephen Curry is known for his shooting, of course, but he’s an underrated passer and he has a great chance at taking two MVPs in a row — part of that impact is with his skills as a playmaker and creative passer. The “purity” obsession with the point guard position is mostly gobbledygook, but Curry’s a real point guard in the traditional sense, not just a shooter. He’s actually like an improved version of Steve Nash. For instance, he had a sensational pass in the below Vine where he threw a no-look alley-oop over his back, which should make the basket worth at least two more points. The problem with guarding him is that even if you contain his shooting — nigh impossible — he can kill your defense with spectacular passes.

Gold Found in Denver

There’s one NBA rookie who’s received little national attention compared to his performance level, and it’s a bit unfair. Like Karl-Anthony Towns and Kristaps “the Savior” Porzingis, it’s another skilled center: Nikola Jokic, a Serbian drafted in the second round in 2014. He was a darling of advanced stats projections and won the Adriatic league MVP the year after he was drafted. Like a prototypical international man of mystery, he’s multi-faceted with good range, as he’s been occasionally shooting from behind the three-point line.

Watching Jokic play, he’s got good hands and nimble footwork, meaning you have to keep track of him near the rim because he’ll take advantage of any mistakes. He’s a good low-post scorer, and teams should consider guarding him on the three-point line too. He’s an excellent passer for a big man, as he recently had 9 assists in a game — you can watch them here. In the first video, he has a nifty no-look pass during a semi-fast break opportunity, and you can see an assist he has often in the second one where he hits a guy cutting to the rim with a pass from the high post.  The last video has a strange one: he goes into the hook shot motion and at the last second passes it — I assume it’s a pass — to Gallinari. Towns and Porzingis are getting all the attention, but we have three skilled, well-rounded rookie centers all deserving of an award.

Raising the Young Pups

Last week, Minnesota suffered an ugly loss. They were playing Denver, who scored 78 for the game including 12 in the fourth quarter, and they still lost — 74 total points and nine in the last period. They attempted only eleven shots from behind the arc, well under half of the league average, and that’s not far from their norm. They’re last in the league in proportion of three-pointers to all field goals taken and first in long two-pointers. The Timberwolves are a rebuilding project, and while they’re taken the steps to surround their young talent with veterans they do not have a modern offense. It’s arguable this is damaging long-term, and it’ll be important to note if this changes in the future.

Patty Mills Versus Tony Parker: the Bias of History

Tony Parker has rebounded after a disappointing 2015, while Mills has done the same after an injury sidelined him last season. Parker has all the accolades and the All-Star appearances, but this is a legitimate question: who’s the better player? Most fans, I would think, would lean heavily toward Parker, and the more conventional and mainstream the analyst — think halftime commentator on a major network — the most absurd the question seems. Parker has what I would call the bias of history: we know what’s happened in the past, not the future, so we tend to overrate older players with storied backgrounds over unheralded younger players.

Right now, one could sift through various advanced stats and look at the numbers. Parker has the edge in a few box score metrics and stats, but the more “advanced” the stat the better Mills looks — he’s ahead in BPM and RPM. But the point here isn’t somehow proving one is better than the other; the point is that it’s hard for a lot of people to even accept this as an argument. Tony Parker has been part of a great team for a long time, which colors our perception of him, and how good he’s been in the past doesn’t make him even greater now[1. With predictive statistics, this isn’t necessarily true because the best predictions typically use the results of the past, but that’s not this argument, e.g. Michael Jordan isn’t better than most NBA guards out there now just because of his past and his famous name.]. We can accept that many point guards now are better than him, maybe even including Patty Mills. It’s an important step in analyzing basketball more objectively; we have to be aware of our cognitive biases.

Larry Nance Jr.: How High?

Growing up watching basketball, one question was constantly springing in my mind, and it was something that had few answers given: how high are these players jumping during the game? You can look at various angles and make estimates and there are measurements taken at the combine, but the former is inaccurate and the latter isn’t during a real game. But with new data, some of these miscellaneous questions can be answered.

There was an excellent feature released recently about Larry Nance Jr. where the Lakers were flabbergasted by his athleticism. He reportedly had a vertical of 44 inches, which is in the upper echelon of NBA athletes. How does this translate to actual NBA games? Since SportVU data (at least whats publicly available) only includes height from the ground for the ball, I need a good example where Nance is reaching straight up with the ball in his hand. I have a decent one below including a Youtube video and a graph from the SportVU data. The purple dots are where Nance meets the ball. At the apex with Nance, the ball is at 10 feet and 10.1 inches. That translates into a vertical around a mere two feet, but there are a few explanations here. Players rarely use their max vertical jumps during games; Nance caught the ball far from the rim. But this data is imperfect, and there are often some hiccups with the ball’s height. Also, I’m not entirely sure if the ball’s height is measured from the center of the ball or the bottom, as using the bottom of the ball as a reference point would allow detection of something like dribbling easy and intuitive to spot — that would put his catch at somewhere near 11′ 2″. It’ll be interesting to use this method on dunks that are like, say, this missed one from Shannon Brown. You just gotta catch lightning in a bottle once, and we have the tools to study it.

And One More

A few years ago, an English professor from the University of Toronto, Ian Lancashire, was studying the patterns and word choice of Agatha Christie’s writing when he made an odd discovery. After his computer program analyzed her 73rd book, Elephants Can Remember, he found that total vocabulary (i.e. different words used) fell by 20 percent compared to her other novels and her use of indefinite words — like anything or something — increased dramatically.

At the time the book was written, Agatha Christie was 81 years-old; she would go on to publish only three more novels. That 73rd book was heavily criticized for its plot and shoddy writing, but textual analysis revealed a deeper truth: she was likely suffering from Alzheimer’s. Looking back, friends stated she complained of concentration issues and had fits of anger. Her writing was never the same either, and the plot of Elephants Can Remember is illuminated in retrospect — it’s about a female novelist coping with memory loss trying to help solve a crime from the past.

When you do in-depth, thorough analysis, you don’t know what you’re going to find; but that doesn’t mean it’s unnecessary. Just like in various branches of science, discovery is often unexpected but it can be incredibly valuable. Thus, we experiment with new basketball stats and dredge lightly used data not for weirdly specific numbers for arbitrary questions but for the unfathomed riches in areas not yet mapped.

For another journey into basketball numbers, there’s one commonly cited stat that hasn’t been rigorously studied: And-1’s. Every NBA fan knows what they are, and they occur fairly frequently. Looking at mere totals tells you something, but it’s like only looking at total points — you’re missing a lot of information.

The vast majority of and-1 fouls are near the basket, typically layups or dunks from players driving hard to the rim. Naturally, I wondered how those shot types compared and correlated with and-1 attempts. The leaders for this season are listed below. Note that the and-1’s included are only those for dunks and layups — plus I have FG% on layups and dunks excluding and-1’s. Andrew Wiggins is the current leader; his explosive scoring punch near the rim leads to a lot of plays where the defense had no chance at stopping him without a hard foul. Quick, rangy guards are common on the leaderboard, as are athletic big men with a tendency to dunk hard. Brook Lopez and both Gasol brothers usually rate well too, but that’s probably because of their length and great touch.

Table: Layups and dunks compared to and-1’s, 2016, min. 100 FGA

Most of the leaders for 2015 aren’t too surprising[1. Given the reasonably small samples at work here, it’s not surprising there is some degree of year-to-year variation in the names, but certain players such as James, Eric Bledsoe and Blake Griffin tend to show up year after year.] — they’re mainly long, athletic players — but one player at the top is, strangely, 76er Robert Covington. I don’t know how many people have noted his particular skill here, but it’s a good sign — a sign of athleticism and concentration when finishing around the rim. Lance Stephenson had one of the worst offensive seasons ever, at least with his shooting, but he somehow earned one distinction in Charlotte[2. For what it’s worth, both those players have seen substantial drops in their and-1 rate, ranking near the bottom.]. For all his issues, he increased his (historically low) inefficiency with his ability to draw fouls and finish.

Table: Layups and dunks compared to and-1’s, 2015, min. 100 FGA

Lance Stephenson’s FG% was actually being inflated due to those freebies, which is scary because he shot 37.6% overall, and that leads to one important consequence of this little exercise: it doesn’t make sense to include field goals made when fouled because they are unlike other field goals. For example, if a player is hacked and throws up a shot, there is no chance at being credited with a missed shot as long as the foul is called. If a player only attempted shots when being fouled, his FG% would be 100.

You can see how the numbers change when you adjust the FG% for the players who draw the most And-1’s; it’s fairly dramatic. LeBron’s legendary percentages are boosted by his ability to draw contact and finish — which is still valuable, of course, but it puts his finishing skills in a different perspective. I didn’t set out to adjust a traditional stat like FG%, but after sifting through the data and considering the outcome, I think it’s a worthwhile endeavor. It would be a more accurate representation of who the best shooters are and how shots are made. Which is why basketball analysis should always consider exploration — the key to discovery.