Nylon Calculus Week 9 in Review: A gift of stats

As much of the world celebrated Christmas or various other winter holidays, the NBA rolled on and had its annual featured day of national games on Christmas day itself. We had battles between Los Angeles teams, some young stars on display in an Oklahoma City vs. Minnesota game, a New England battle of the media with Boston vs. New York, and a rematch of the finals.

Our year will soon end, but the NBA will not — there’s still a lot to look forward to.

Pure Boogie

Last week, DeMarcus Cousins had the quintessential DeMarcus Cousins game. He scored 55 points, he turned the ball over five times, and received five personal fouls. He also drew 17 free throws and blocked three shots. He received two technical fouls, got ejected, and then was immediately rescinded of one and came back.

The game ended with a narrow victory over the Portland Trail Blazers and Cousins had a bizarre post-game rant whose audio was cut short. It was everything you could hope from a Cousins game, and more, in perhaps the strangest 55 point game ever.

You can see all his field goals here — he was relentless.

I’d say a center hitting 55 points with a season long usage rate in the high 30s is a throwback performance, but Cousins also launched several 3-pointers and hit five of them, flummoxing Matt Femrite’s theory that he was perhaps the best candidate to break the 50 point game without a 3-pointer drought. (Don’t worry; Anthony Davis broke that with a game back in October where he also had seven steals and four blocks.)

Cousins is an oddly entertaining player, but some of his worst traits were also on display, like his incessant taunting directed at Mason Plumlee. I don’t know what his future will be, or where, but I wish his ferocious on-court behavior could be properly contained. I know franchises loath to let their big names go, but sometimes an organization fails and a star would be better off on another team. Boogie’s antics should be on a bigger stage, hopefully for the better and not the worse.

Warriors-Cavaliers: A Return

I suppose I should comment on the Golden State Warriors-Cleveland Cavaliers rematch, as everyone has. Maybe every single angle has been attacked and analyzed, but we can think of the game like this for the Warriors: they lost an away game by one point, therefore the game proved they were the superior team.

Listen: What did we learn on Christmas?

I know that’s not a popular view, but it wasn’t the worst result for Golden State, especially if you think Kevin Durant should have gone to the free throw line at the end. They didn’t obliterate Cleveland, like I suppose they had hoped when they picked up the former MVP over the summer, but few people expected that, didn’t they? And with Stephen Curry largely silent for the game, I’d favor the Warriors in their next rematch, but this is still a top-notch rivalry.

BPM and Assists, Part the Thousand

A high assist rate is wholly important for basketball-reference’s BPM, and there’s no better illustration than Thon Maker’s page. Despite high shooting percentages, rebound rates, blocks, a usage rate near 20, and a PER of 22.7, he has a severely negative BPM (-6.9).

That’s a consequence of having zero assists.

It’s nothing to worry about now, of course, because he’s recorded 45 minutes as of last week, but it’s greatly affecting some of his advanced metrics. BPM is an improvement on previous metrics, including basketball-reference’s outdated Win Shares model, but it’s not perfect either, and you can see the cracks at the extremes.

The Greatest Show Up North

In a season where the Warriors are dominating headlines and crushing (most) opponents, it’s a bit odd to see Toronto fighting them for top spot in offensive efficiency, since, you know, the Warriors added Durant and Toronto’s biggest addition (who has actually played) was Jakob Poeltl. Plainly speaking, Toronto is doing so well because of the inspired play of Kyle Lowry, good depth, and some surprisingly good minutes from young players like Lucas “Bebe” Nogueira. Unlike the Warriors, who bludgeon opponents with high shooting percentages, the Raptors offer balance by shooting well, crashing the offensive glass, keeping turnovers low, and getting to the foul line frequently.

I suppose people may cite DeMar DeRozan, midrange artist, as a primary contributor, but I am not a supporter and the evidence doesn’t lean in his favor. The team scores much better when he’s off the court — a blistering 121.4 points per 100 possession rate — even though he’s often with the starters. This is no anomaly either; the team lives and dies with Lowry, not DeRozan. Thus, it’s no surprise that advanced stats don’t see DeRozan as a star at all, not matter how many points he scores a game.

The big question is what will happen to their offensive rating the rest of the season. That rating is nearly six points higher than it was last season, when they had largely the same roster, and correspondingly their 3-point percentage is significantly higher this season too. Some regression downwards is definitely expected. They received another surge from a low turnover rate, and that one is trickier to diagnose. I actually don’t know what to expect from this season, except from a simplistic Bayesian perspective where I expect them to regress to their 2015 and 2016 performance levels.

Sometimes mysteries exist, and there’s no easy explanation.

Golden State Point Barrage

For Nylon Calculus and similar thinkers, team points per game is an evil, archaic statistic, like something from phrenology. But it’s still a conversation starter, and it’s what we see at the end of the game. Since I became an obsessive fan during the NBA’s dead-ball era (1999-2004), this one is unbelievable for me: the Warriors are averaging over 117 points per game, and it’s the most since the early 90s.

Read More: Visualizing lineup balance and understanding team fit

Last year the same was true — they averaged 114.9, and only one other post-2000 team had averaged 111 or more. (Surprise, surprise: the Don Nelson Warriors were that one team.]. This year, we may see several high-scoring squats, from the Houston Rockets to the Toronto Raptors and the Cavaliers. But the Warriors, adhering to their franchise’s tradition of gaudy scoring statistics, are in a league of their own.

Best (Theoretical) Offensive Lineup Ever

I listened to some discussion recently on building the best possible offensive lineup ever, and I can’t help but build my own team. For a plausible constraint, I’m filling out the NBA positions with realistic options, instead of just playing five 3-point shooting bombing point guards.

• For center, I’ll actually go with Karl Malone, as he was part of some of the best offensive teams in modern NBA history, he’s a great pick-and-roll finisher, he can crash the glass, and he can take advantage of smaller players so opposing defenses can’t just switch every time. Plus, you need someone to set screens.
• For the forwards, I’d go with Durant and Larry Bird. Both are Hall-of-Fame shooters and scorers, sure, but they’re also accustomed to playing next to other high scorers where they can cut and spot-up and Bird’s supernatural passing ability can be utilized greatly here.
• For a shooting guard, I’d go with Ray Allen, who can come off screens and draw a ton of attention for other players. He’s also another plus ball-handler.
• For point guard, I was torn between Curry and Steve Nash, who are remarkably similar but offer different legendary specialties: Curry with his shooting and range and Nash with his passing. If I had to choose, I’d go with Curry, because his range and ability to hit tough shots are so unique it’d draw defensive attention even on a team with these players.

This is all for fun, of course, but it’s a useful thought experiment for people who write about the league because it forces you to tackle questions like, what does an ideal offense look like? And which skills will compliment these scorers? You can throw out a lineup of Magic Johnson, Michael Jordan, LeBron James, Charles Barkley, and Shaquille O’Neal that looks good on paper, but how does it actually function? We’ve seen from recent super-teams that fit does indeed matter. The important question is how much influence does fit have, and that’s still up for debate even in 2016.

Tough Talk About Marcus Smart

Marcus Smart was a draft prospect with some intriguing traits who transitioned that into a promising NBA rookie season where he was well above average on defense for his age but struggled shooting. But his development has stalled, and overall his stats have been getting worse each season. It’s beyond simple scoring stats, too, as his RPM has seen a discouraging decline for a young player from +2.22 in 2015 to -0.01 last season to -0.64 now. However, if you follow discussion about Boston, you see many positive articles, like about his leadership and his growth as a point guard. But there’s been some ongoing and long debate among Boston fans about his future and his offensive development.

The problem isn’t Smart’s defense — his defensive stats have held up, and he’s still the tenacious defender that came into the league. But his shooting stats have cratered, and his chances at becoming a league-average shooter are becoming dimmer and dimmer. He hit 34 percent of his 3 3-pointers in his rookie season, but he’s been sub-30 percent shooter since then. Given that his career free throw percentage is in the low 70s and he couldn’t even surpass 30 percent from the college 3-points line, it’s likely his shooting stats during his rookie season were an aberration. He’s also a high volume outside shooter, which has caused people to question his shot selection.

Sure, it’s better than a pull-up midrange shot, but his True Shooting Percentage on 3-pointers is 41.9 percent while it’s 49.1 percent on all other shots. With his build and skillset, his scoring strengths are likely slashing and posting-up. What’s most concerning, however, is that he’s shooting 44 percent at the rim this season. If he can’t even finish at the rim his offense could seriously jeopardize his role going forward — he has some real issues that need to be addressed.

Updated Dredge

Over the last summer, I finished an overall player metric, Dredge, that uses a wealth of stats, everything from traditional assists and rebounds to offensive fouls drawn and goaltending violations. I haven’t yet released the metric’s results for this season partly because I wanted a bigger sample of games and partly because I needed to get things setup for this season. But now I feel like it’s the time to unleash it, so merry Christmas, NBA.

You can view the complete results here, but I pulled out a few interesting leaderboards. In the below table, you can see the leaders so far for this season. Durant is on top by a fair margin, which is a modest surprise. The metric has usually loved him, thanks to his combination of high-volume efficiency, versatility, size, and 3-point volume. Lowry is the more shocking inclusion; but Dredge has usually favored him thanks to some of his defensive stats that don’t get picked up well by conventional stats. Also note that the Raptors somehow have the best offense, well, ever so far.

The MVP Index is on nonlinear scale of 0 to 100, where a ton of minutes playing at an average level does little for the index but a few minutes at a really high playing level can actually move the needle. And since Dredge’s ratings are regressed, you’ll see some unimpressive MVP Index numbers right now, but it does appear we have a nice race so far with Durant leading. The rest of the usual suspects are listed in the top 20 leaderboard as well, but there’s at least one name that really stands out.

Patrick Beverley somehow has the best Dredge on the Rockets. It’s not entirely a pure anomaly, because advanced plus/minus stats have usually adored him, and his BPM is +3.4 right now. But he’s drawn an impressive 24 offensive fouls, which is weighed highly by the metric. Trevor Ariza is also rated highly. Dredge stands in stark contrast to many other stats out there giving credit in proportionally larger doses to James Harden. I suppose it could be because Harden doesn’t show up well in non-box score stats, generally, and I weigh “spacing” highly, especially for big men — i.e. Houston’s bigger 3-point bombers are getting more of the credit on offense than in other models.

There is one player missing from the above table, though, and he’s arguably leading the MVP race: Russell Westbrook, who was rated at a mere +1.7. He was still the highest rated player on the Oklahoma City Thunder, but his numbers aren’t leading the league like they are in BPM, which actually shares a few of the same interaction stats (e.g. AST%*TRB%). But the Thunder are one of the “luckiest” teams defensively so far this season, and the adjustment hurts his rating. Also, if you peruse his teammates’ BPM ratings, you’ll note a lot of extremes, and, unlike BPM, I regress everyone heavily toward -2. Thus, as the season progresses, Westbrook’s Dredge rating should increase substantially more.

Read More: The killer crossover James Harden uses to break defenders

And finally, I may actually change how I value efficiency in the next version because my metric seems to favor high-volume, high efficiency scorers unreasonably more than high-volume, average efficiency scorers, like Westbrook. I should also note that Westbrook’s stats are unprecedented and will break a few models, including BPM, while plus/minus stats, which are at least partially neutral to counting stats, like RPM don’t see him as an “undisputed” best player.

If you’re curious about the lowest rated players, there’s a table below. Glenn Robinson III, son of former NBA player and number one pick Glen Robinson, owns the dubious distinction of worst rating. Curiously, Tim Hardaway Jr., another son of an NBA player, is right behind him. The bottom of these rating systems is usually filled with young players. Sometimes these players are really young and are just getting experience, but it’s not encouraging to see Dario Saric, a 22-year-old who played professionally overseas, and Buddy Hield, a 23-year-old partly drafted to his ability to contribute immediately, ranked so low. Jeff Green is the longest tenured veteran on the list, but his proclivity to low ratings in advanced stats is notorious.

I would caution against using Dredge as an end-all, be-all metric, like I would with any metric. Consult other stats. Read up on weaknesses and blind spots and statistical assumptions. But referencing something like Dredge is still better than completely ignoring advanced stats, or even just all-in-one metrics — you can learn a lot about basketball value and test out theories. And, hey, my season win predictions were mainly based on Dredge ratings, and they haven’t tanked. There’s something in the madness.