Nylon Calculus Week 4 in Review: Clippers and the Hanging Tech

The NBA season is in high-gear, and its unpredictability is, strangely, welcome. No one expected either the Los Angeles Clippers or the Los Angeles Lakers to be this good this season, and the NBA MVP race is shaping up to be one for the ages. The rise of the 3-pointer is no coincidence either because it decreases predictability, all other things equal. This is healthy for the future viability of the league because the excitement factor goes up, and it’s a fun challenge for people who try to analyze the league.

With that, let’s try to analyze the state of the league over the past week.

The new kings of Los Angeles

Coming into the NBA season, people imagined the Golden State Warriors, flanked by the near 7-foot, shooting phenom Kevin Durant, would crush and roll over the league with ease. Instead it’s the Los Angeles Clippers who are outscoring teams by almost 15 points a game with only two losses, a month into the season. They’re second in offensive efficiency and defensive efficiency, per Basketball-Reference, which is a combination you rarely see outside of title winners. They were a forgotten team in the Western Conference, dismissed partly because they’ve had no exciting changes, and analytic-based projections saw them taking a major step backwards. Now, even with some major regression to the mean, they’d still be a true contender.

Read More: Celtics, Embiid, and truer shooting percentage

The Clippers starting lineup of Chris Paul, J.J. Redick, Blake Griffin, and DeAndre Jordan, along with a fifth Beatle like Matt Barnes or, this year, Luc Richard Mbah a Moute, has been consistently fantastic over the years. It’s no different this season, as every starter has a net rating around 25 while no one else on the team is close. In fact, the starting lineup last season was stellar too with a rating of 19.4; they just had so many missed games, mostly from Blake Griffin, that they didn’t play too many games together. People were cautious in projecting the lineups with Mbah a Moute because, historically, his abysmal offense has outweighed his top-notch defense.

We should be thankful he’s finally found a home, however, after years of toil in basketball’s Siberia triumvirate of Sacramento, Minnesota, and Philadelphia, because he appears to be a genuinely nice and thoughtful NBA player. He also found Joel “the Process” Embiid, so we all owe him a lot. I’m just worried his suspiciously high shooting percentages will plummet and he’ll lose his luster.

There’s a team-level anomalous stat too, however: opponents are shooting 28 percent from 16 feet to the 3-point line against the Clippers, which is about eleven percentage points under the average and six from the nearest team. That kind of magnitude can have a significant effect on a team’s performance level. For instance, if opponents had shot the league average percentage on those shots, and after accounting for offensive rebounds, it would translate to an additional 2.2 points per game for their opponents. And that’s ignoring the real effect it would have on their offense: the Clippers, like every team, shoots much better off defensive rebounds compared to made field goals, which itself would nearly result in another loss of two points per game.

The assumption here is about the role of luck in defense. 3-point percentage is not controlled well by defenses, but that’s only partly true for long two-pointers. However, it’s an illustration of the great consequences shooting stats can have on both ends of the courts, and in any case, even if the Clippers are defending long jumpers well, I would not expect opponents to continue shooting that poorly. People can cite Chris Paul’s lasik surgery as the cause here in their unbelievable start, but the evidence there is murky; his shooting percentages could be unsustainable too. Feedback loops are crucial for elite teams, and it’s unclear at this point just how good the Clippers truly are.

ESPN’s RPM Has Been Released

It’s that time: Real Plus-Minus is now available on ESPN, and we all free to criticize and cherrypick to support arguments we have already decided on as we so desire. Chris Paul is the leader right now, buoyed by a ridiculous +3.3 mark on defense for a small point guard — only DeAndre Jordan and Steven Adams outrank him overall on D. The rest of the leaders are mostly the usual suspects. Jimmy Butler is higher than expected, however, ranking third overall in RPM. Then there are a couple of surprising kids near the top too, like Giannis Antetokounmpo and Kristaps Porzingis. And, naturally, I have to mention the player ranked last: Evan Turner, who was heavily criticized by Nylon Calculus-type folks and has been a major disappointment in Portland.

Read More: Dominant or not, the Golden State Warriors are good for basketball

Since RPM is available now for the current seasons, I should cover a few basic things about the metric because it’s so often misunderstood. RPM is a mix of plus-minus numbers and countable, discrete stats, from points and rebounds to play-by-play actions like blocks that go to the defense (aka Russell’s) to a handful of non-standard adjustments and stats like height and a “luck” adjustment on free throws. If a player shoots a free throw, people on the court don’t get credited based on whether or not it goes it; it’s based on the expected result. For example, if you’re only on the court for DeAndre Jordan’s misses and you’re his teammate, you’ll look worse in plus/minus than you should. However, this adjustment is only made for free throws that can’t be rebounded, because possessions afterwards are directly affected by how the free throw sequence ended —  e.g. it’s easier to run a fast break on a missed free throw than a made one.

The “pure” plus-minus model and an SPM-only version are mixed and weighed together to produce the best out-of-sample results. That means that the coefficients aren’t arbitrarily weighted, and it’s tested to make sure the metric actually means something — you’re closer to the reality of player value when it tracks future wins. There’s no public source for all the nuts and bolts of RPM, but that’s what I’ve learned from the creator himself over the years. You can use it at your own discretion; just be wary that RPM is only a ranking of RPM itself and the Platonian ideal of how good a player “truly” is will probably never be found and is a myth itself.

The post-up game is dead, long live the post-up game!

After years of hearing about the death of a basic basketball play type, there’s been a bit of a resurgence thanks to a plethora of young and skilled big men, from Karl-Anthony Towns to Joel Embiid to Jusuf Nurkic (seen below spinning around a hapless defender.) I know people are crafting narratives already, but the post-up game isn’t about to envelop the league and some tales of its importance from past decades are a bit overblown anyway.

Last season, via the NBA’s play type statistics, around 7.5 percent of all plays were post-ups, while this year it’s close to 7.1 percent. Efficiency is down a little too. While some people wax poetic over the sad loss of posting up — this concerns mostly big men posting up, by the way —  I think it’s a great trend, overall, because you’re moving the ball away from slow behemoths and into the hands of more skilled players and moving the ball around the perimeter. Of course, handing the ball to a skilled big man can reward greatly, but now our skilled big men are spotting up from behind the arc — and perhaps this is all for the best. Long live the post-up game.

We are all Porzingis

Coming into the season, my expectations for Kristaps Porzingis were muted based on one principle: he appeared to vastly outperform expectations, so I assumed a sophomore slump was likely. But after a career-high game of 35 points, we should all prepare for a world where a 7-foot-3 unicorn is running around New York, shooting 3-pointers, driving, dishing, and dunking on everyone. And I’m afraid to use this comparison but the most similar player I can think of at his size is more myth than man: young Arvydas Sabonis. Porzingis will never be the same level passer Sabonis was, but I think we have to appreciate the player for how unique he is — and let’s hope we see more lineups with Carmelo Anthony at power forward where he’s best and Porzingis at center.

The empty calorie scorer

Harrison Barnes, Andrew Wiggins, and DeMar DeRozan — that trio creates a gulf of vehement basketball arguments, and few players have a larger chasm of their perceived value between conventional and numbers-based analysts. DeRozan, thanks to his 31.7 points per game figure, has received serious MVP talk. Andrew Wiggins was ranked 31st before the season by SLAM online, and his increased scoring output has garnered him even more attention. People are already chipping away at the summer criticisms of the Harrison Barnes contract, due to his (you guessed it) scoring numbers. Their supporters will claim, ha! They are as good as we all already knew, and this season will be a vindication. But the problem the analytics crowd had with these players wasn’t their scoring; it was everything else.

Harrison Barnes, for example, has nearly doubled his points per game average, but his BPM has actually fallen precipitously from -0.2 last season to -3.0 currently. If you think that BPM does not value these focal point scorers well, or their defense, then you can point to RPM, which sees Barnes and Andrew Wiggins as slightly below average and DeMar DeRozan, contrary to his eye-popping scoring numbers, as just a little above average. These players, despite their athleticism and size, don’t rebound, block shots, or even steal at a high rate. Even their assist rates are not noteworthy. Thus, advanced metrics do not see them as particularly great players. They’ll pile on the points to gain the attention, but their deficits in nearly every other area of the game washes out those benefits. They’re empty calorie scorers.

Those three scorers are right at that empty calorie boundary line as seen in the graph above, and for those curious the two scorers next to Barnes are Devin Booker and Klay Thompson, whose defensive BPM has sank to unholy levels as the whole team struggles on defense. The outlier in the top middle portion is, of course, Chris Paul.

Foul miscellanea

The great sin of the box score is not simply that a complex game is summarized by a few items, but that we are led to believe each item is itself homogeneous and can therefore be added together. Personal fouls are an ideal example. Somehow, fouls that lead to nothing are lumped in with shooting fouls and even technical fouls. They’re not all the same, and they all have differing levels of (mostly negative) value. The sheer breadth of types is fascinating, however, so I thought I’d take a dive into the play-by-play data and pull out a few examples.

When I derived a play-by-play metric, I found that loose ball fouls were actually slightly positive, especially when just looking at defense. And this makes sense: going for a loose ball is positive indicator for a player, and since we don’t have the hustle stat of loose ball recovered going back to the 90’s this is our closest proxy for said recovery. You can see the leaders below; it’s a mix of rough-and-tumble guys like Danny Fortson and Dennis Rodman with some notable big names like Charles Barkley. Oddly, 1998 is over-represented at the top, and I don’t think it’s a major error. That season had the most loose ball fouls, according to my data, but it wasn’t outrageously different from any other season.

Table: foul leaders, source: stats.NBA.com pbp

Clear path fouls, on the other hand, are most likely not positive indicators, and at the very least they’re particularly harmful fouls depending on the specific rules of the era. The attempt itself to stop a fast break is a noble one, and it can be quite valuable, but the penalty is severe now: two free throws and possession. With that said, it’s a rare foul, and only five players have had at least four in a single season. Allen Iverson owns the dubious distinction of owning the most in a single season, but he has an excuse at least: it was his rookie season.

Table: foul leaders, source: stats.NBA.com pbp

There’s a rarer foul type still: the “Non-Unsportsmanlike” technical foul. Broadly speaking, many technical fouls are under this umbrella term, but they’re not specifically coded as such, so there’s perhaps little meaning in this category. But for those curious, the most in a season is a mere two by the infamous Darius Miles. A related foul is the “taunting” one, where only three players have had three in a season: the mercurial Steve Francis, Eddy Curry, and the finger-wagging Dikembe Mutombo. Also, by coincidence, another foul type had only three leaders with three in a season: the enjoyable rim-hanging technical. Two leaders there aren’t surprising — Amare Stoudemire and Tyson Chandler, two frequent high-fliers — but I was delighted to learn Chris Webber was the third name. Most people sadly know him for his floor-bound, injury-ridden days, but he was spry in his youth — or perhaps he was just afraid of letting go of the rim.

Related Story: The Golden State Warriors are embracing super-villainy

Here’s one I’m not sure how to judge yet: defensive three-seconds. As you can see below, the leaders are an impressive set of Hall of Famers, with the exception of Brendan Haywood, who was good defensively that season. This is a weak proxy, at least, of rim protection, because it’s showing which players are camping out in the paint. For some players, it’s also related to a tendency to stay inside to the team’s detriment and not defend away from the rim (Shaq and Haywood.) I’m also flummoxed by Dikembe Mutombo’s gigantic first-place lead over the nearest guy. I know this data isn’t perfect, and perhaps there’s an error here. In any case, he’s still a brilliant defender and that was one of his best seasons, so I’m not quite sure what to do with this stat yet.

Table: foul leaders, source: stats.NBA.com pbp

Splitting fouls into their constituent parts, much as physicists do with atoms to quarks, is fascinating and more of the game is told this way. But there’s a lot more you can do with this data once you combine it with other stats and scale it differently with rates — and we’ll see that in another Week in Review.