Nylon Calculus Week 13 in Review: Unicorns and MVPs

Jan 11, 2017; Philadelphia, PA, USA; Philadelphia 76ers center Joel Embiid (21) and New York Knicks forward Kristaps Porzingis (6) in a game at Wells Fargo Center. The Philadelphia 76ers won 98-97. Mandatory Credit: Bill Streicher-USA TODAY Sports
Jan 11, 2017; Philadelphia, PA, USA; Philadelphia 76ers center Joel Embiid (21) and New York Knicks forward Kristaps Porzingis (6) in a game at Wells Fargo Center. The Philadelphia 76ers won 98-97. Mandatory Credit: Bill Streicher-USA TODAY Sports /
facebooktwitterreddit

We’ve passed the mid-point of the NBA regular season, yet in some ways the season is actually over. The standings right now are shockingly stable and it’s unlikely we’ll see major changes. We also have a good understanding of who the leaders are for various awards, like Most Improved Player and Rookie of the Year. But we’ve still gotta play out of the remainder of the season, and there are plenty of storylines to untangle. And with that, let’s look back at the last week of the NBA.

Most Valuable Player: The race heats up

As the MVP race heats up, I fired up my box-score metric HBox for the season. I was curious how Russell Westbrook rated. My metric, like Box Plus-Minus, values high combinations of usage rate and assist rate as well as assist rate and rebound rate. Thus, I surmised Westbrook would look like a god, but that isn’t so: he’s second in the league and has a healthy lead there, but James Harden is ranked first. You can see the results here. The MVP Index is just a way to factor in minutes played, and it’s a nonlinear function that hews closer to how people generally vote for MVPs historically rather than traditional replacement level stats, like VORP.

I understand people are more impressed by Westbrook, and it’s easy to be awed by his thunderous gameplay, but Harden rates more valuable for a simple reason: he’s a lot more efficient, and it makes up for Westbrook’s edge in volume and big lead in rebounding, which is typically an overrated stat. Of course, Harden’s team is also a lot more successful too, which also has an effect on most metrics. I assume the MVP talks will get serious after the All-Star break, so I want to impart this onto readers: understand the stats you’re using, and clearly and accurately communicate your facts and theories.

Free throw analysis

The great Mike Beuoy of Inpredictable released another SportVU-driven article last week, which will likely be one of the last of the excellent public pieces we’ll see with this now private data. He looked at how picking a spot at the free throw line — where the players physically are — had a correlation with free throw success. Indeed, it’s about consistency: players who are usually in the same spot, free throw to free throw, perform better than those who are more inconsistent, most notably Andre Drummond. I know every shooting coach has preached this, but it’s good to see this backed up by hard data. It’s just too bad this kind of analysis is now kept in the dark from the public.

A (singular) positive development in Washington

I like to highlight players who stick out statistically and then explain why it’s happening. In that case, Otto Porter Jr. is an ideal target. He’s one of the highest ranked players through a variety of metrics, from 23rd in BPM to 35th in HBox. Yet it wasn’t too long ago that his awful rookie season caused some people, even draft experts who openly touted his virtues, to wonder if he was ever going to be a useful rotation player. He was then immortalized as the Shaqtin’ a Fool MVP for 2015. He had an inauspicious start to his NBA career.

Read More: Tracking the highs and lows of the New York Knicks

What changed? Most of his stats haven’t moved much over the years; his offensive rebounds are down but his defensive boards are up. Even his usage is virtually the same as it was his rookie season. But his efficiency has skyrocketed. His turnovers have decreased, and his true shooting percentage right now is elite: 64.1 percent. That’s primarily due to his 3-point shooting; he’s taking more than ever and he’s hitting 45.4 percent of them. I would expect his 3-point percentage to regress significantly, but even if he were shooting at his career norm this season he’d still be an improved player.

You can see six 3-pointer Otto hit last week here versus the Knicks, along with one layup. He’s finding his spots on the court very well, and with a point guard like John Wall those kind of opportunities are abundant. Defensively, he’s nothing special — he’s okay, and his size is a plus at times. He was often matched-up against Carmelo Anthony in that Knicks game, and there were mixed results. He gets out-muscled by Carmelo here, and he was lucky Carmelo missed about everything. You can see him get hung up on a screen on this play, leading to an open trey for Carmelo — that caliber player is out of Porter’s range, who can struggle with timing and navigating screens. He’s not a liability, but he has a long way to go to become a true stopper. But his outside shooting alone has transformed him into an above average player, and for people who saw his rookie season it’s a grand improvement.

Sacramento loses Rudy for the season

Rudy Gay suffered a ruptured Achilles tendon last Wednesday, and he has just completed his surgery. I imagine people are wondering if this dooms Sacramento’s chances of competing for that last playoff spot in the west, but upon closer examination it probably won’t change much for the Kings and the history of Rudy Gay is a history of how teams do just fine without him. Really, the greatest impact here is on the rest of Rudy Gay’s career.

I don’t know how many teams would have wanted him in a larger role, and an athletic player coming off a major Achilles injury will have a glaring, giant red flag. He’s still an interesting figure though as a sign-post in the growth of analytics and how the NBA has changed, from the sorely under-discussed Rudy Gay-Shane Battier trade that helped start the Daryl Morey era in Houston to his eye-popping inefficiency in Toronto with DeMar DeRozan, leading him to ban box scores in the locker room. The Kings may not miss him on the court, but he’s interesting nonetheless.

Rubio-Reggie: Challenge trade

This trade was recently shot down, but it’s an intriguing challenge trade, where two teams have similar goals but think they can both become better with the other team’s player at the same position. Given how the organization has drafted recently and what they’ve explicitly said, Minnesota does not view Ricky Rubio highly. To most fans who read Nylon Calculus, this is strange because he has not been the problem with the team — not now, and not over the past three seasons.

Even though this hasn’t been Rubio’s most valuable season, he’s still a very good player, and there’s more to being a good perimeter player than being a shooter. For years, Minnesota has played significantly better on defense with him on the court, and explaining why digs into the philosophy of NBA defense. Defenders have less of an effect on jump shots than conventional wisdom suggests, and from the data and different means of testing models it appears the best way a guard can affect his team’s own defense is through creating turnovers. Rubio is a savant in that respect with one of the highest steal rates in the league combined with one of the highest offensive fouls drawn rates.

He’s not an efficient scorer, but he has the same effect on his team’s offense through his playmaking wizardry. I understand people are obsessed with 3-pointers, and outside shooting is now equated strongly with analytics, but truly speaking analytics is about finding value where no others see it through data and objective methods, and unconventional players are the richest gold mines. Maybe the Wolves will understand this in time, but for now the front office has committed to a young, unproven rookie point guard, and Rubio is likely gone anyway.

A process unlike any other

Joel Embiid has many remarkable stats, but the most remarkable may actually be his usage rate. No rookie has been close to his rate of 36 percent, and actually few players in NBA history have ever used team possessions at the rate Embiid has. Maybe you think it’s caused by his high turnover rate, since turnovers are part of usage — but nope. I’ve calculated this stat I usually refer to as Shot Percentage that’s simply usage without turnovers[1.], and Embiid’s Shot Percentage is 35.2 percent. You can see that stat here, by the way, and historically right here.

If you’re wondering what Wilt Chamberlain’s usage rate would have been, since it’s unavailable both through my stats and on Basketball-Reference, then let me save you some trouble. We have his team’s free throw and field goal totals for Wilt’s rookie year, so it is possible to calculate Shot Percentage. It was 30 percent that season. Sure, he averaged 37 a game, but he played monster minutes with a high pace. Even his usage rate would be no match for The Process. Make no mistake — what Embiid is doing as a rookie, especially as a center, is unprecedented.

The hidden power of long 3-pointers

I’m actually a little annoyed this article was released because it discussed what I thought was a nice hidden value: super-deep 3-pointers. These shots have been a slight obsession of mine for a while, and during Stephen Curry’s rise to prominence amid the discussion of gravity I had some thoughts. Basically, we know spacing is valuable because it pulls defenders away from the basket, so if you can pull defenders even further away, it’s even more valuable. A related note: spacing exists for long-range 2-pointers, but it’s not as powerful, and I think it’s partly because spacing power is a function of range and geometry. Being 24 feet from the basket compared to 12 feet doesn’t mean the spacing power is twice as large — it’s nonlinear. The distance from the basket should be conceived of radially.

The article doesn’t provide extensive details, but you can imagine the combination of a progressive front office and coach were important here, and it’s perhaps quite intentional. Look at this list of most shots taken (and made) from 25 to 30 feet: Eric Gordon, through Sunday, is leading the entire league — yes even over Stephen Curry. Three of the top five players are Rockets (James Harden and Ryan Anderson), and yes, the team does lead the league in long 3-pointers. These shots are a new market inefficiency, as you can pull defenders even further away from the basket and the ballhandler — plus, defenders are not used to guarding these attempts. It’s the secret power behind two of the most dangerous offenses — the Warriors, thanks to Curry and Klay Thompson, and the Rockets. Expect to see this trend grow in the future.

"“The line is there,” D’Antoni said. “Doesn’t mean you have to stand on it.”"

Classifying unicorns, manticores, and chimeras

The NBA term of the moment right now is the unicorn, which was used last year frequently on Kristaps Porzingis and has now spread like verbal wildfire to numerous players. In fact, it has become so common I’m not quite sure what the agreed upon definition is anymore. As Dan Devine discussed recently, that’s a sad development because there are so many more fantastical mythical beasts we could compare players to. And in addition, it’s a slow death of a word when it’s applied so broadly that it loses definition. Words need specificity or else they become mere wafts of air that completely vanish into the surrounding aether.

I propose that we adhere to a strict definition of a unicorn, as it’s close to Kevin Durant’s original conception of a player with unusual size and skill, and it’s one that’s quite useful currently: a 7-foot defensive presence and rim protector who can shoot 3-pointers reliably. Indeed, those beasts would border on nonexistence, or they were at least extremely rare. Raef LaFrentz had a few seasons shooting from behind the arc and blocking shots, but he wasn’t a great defensive player otherwise and didn’t do much on offense besides shoot from outside. Besides him, there are few historically, depending on how you classify Rasheed Wallace or what you make of Manute Bol’s brief time launching deep shots.

I thought I could use statistical clustering to identify these unicorns, but I ran into a couple problems: one, they’re so rare you have to slice up the data into tiny segments. And secondly, it’s almost like there’s a natural trade-off between outside shooting and shot-blocking, and it’s all pretty continuous — the unicorns haven’t (yet) created their own little bubble of performance separating themselves from the rest of NBA history. You can see that in the graph below for every player listed as a PF or C in b-ref’s database. The other dot in Unicorn-Land, by the way, is Raef LaFrentz.

3pt-block-frontier-big-men
3pt-block-frontier-big-men /

I thought there’d be a better way to quantify this, so I turned to an old friend: the geometric mean. It’s a way to build a composite score of two numbers, or more, that rewards balance. For example, my Unicorn Score is a product of blocks and 3PTers per 100 possessions, and because I don’t want guys dominating the list who only, say, have lots of blocks and no outside shots, the geometric mean penalizes those players. You can see the results below.

There are some odd inclusions — blocks on their own don’t denote who’s a rim protector or a great defender, and height alone doesn’t tell the real story of being a unicorn. I doubt opposing offenses shuddered at the thought of Brad Lohaus or Charlie Villanueva, or even LaFrentz. Thus, I included an adjusted Unicorn Score, which uses Defensive BPM as well as the other two stats. Regardless, Joel Embiid stands alone, and you can see unicorns are becoming much more frequent. Myles Turner, by the way, was ranked 32nd, only a few spots behind Marc Gasol’s 2017 season too — Turner’s on his way to mythical beast status.

Table: Unicorn ranking (6′ 11″ or higher, data: b-ref, min. 750 MP)

PlayerSeasonTeamHeightBLK1003P100Unicorn ScoreUnicorn (DBPM adj)
Joel Embiid2017PHI7’0″4.832.093.173.81
Brad Lohaus1992MIL6’11”3.302.652.963.13
Raef LaFrentz2002TOT6’11”4.162.032.913.21
Kristaps Porzingis2017NYK7’3″2.852.922.882.47
Charlie Villanueva2013DET6’11”1.894.362.871.87
Brook Lopez2017BRK7’0″2.722.882.802.32
Brad Lohaus1994MIL6’11”2.902.432.662.57
Kristaps Porzingis2016NYK7’3″3.362.032.622.71
Channing Frye2010PHO6’11”1.703.962.592.69
Channing Frye2012PHO6’11”2.172.832.482.70
Charlie Villanueva2011DET6’11”1.364.042.340.00
DeMarcus Cousins2017SAC6’11”2.072.632.332.79
Channing Frye2011PHO6’11”1.583.422.332.38
Andrea Bargnani2007TOR7’0″1.693.182.321.91
Raef LaFrentz2006BOS6’11”1.842.872.302.30
Andrea Bargnani2009TOR7’0″2.062.532.291.99
Brad Lohaus1993MIL6’11”2.092.402.242.05
Channing Frye2014PHO6’11”1.373.472.182.18
Charlie Villanueva2010DET6’11”1.612.912.170.00
Raef LaFrentz2000DEN6’11”3.751.252.162.46

Brad Lohaus is probably a name many readers are unaware of. He was primarily a power forward — part-time center — who played for several teams but became a frequent 3-pointer shooter when he was traded to the Bucks. He was a decent shotblocker, but he was never known as a plus defender. He’s also one of Bill Simmons’ favorite targets. The problem with the Unicorn Score is that defense is hard to evaluate with basic stats. If you go through the rankings, it’s actually remarkable how many of those players aren’t great defenders, and how few could qualify subjectively: Embiid. Maybe LaFrentz and maybe Kristaps Porzingis. Cousins on a good day. Marc Gasol. Maybe Myles Turner. Somehow Manute Bol makes the cut, although he’s his own mythical creature. And that’s it for everyone with a score of 2 or greater. That’s why we’re calling Porzingis the Unicorn: it’s truly rare for players near his size to be adept on defense with a real 3-point shot.

As for other mythical beasts, let’s start with Giannis Antetokounmpo. It’d be boring to call him a unicorn. He’s so much more, and he’s not a frontcourt player like Porzingis anyway. I’d say if I could name Giannis any one beast, it’d be a chimera: it’s part snake, goat, and lion. I’d say that suits him best, since he’s an improbably player with traits we rarely see together. And, again, disagreeing with Dan Devine here, the griffin is obviously Blake Griffin. It’s a mix of a lion and an eagle, which makes me think of a muscular player with a strong, wide-shouldered frame who can nonetheless leap high into the air. Aaron Gordon is one such young player.

Next: An open letter to Giannis Antetokounmpo and Jabari Parker

Karl-Anthony Towns has received some unicorn labeling, but he’s not quite a good 3-point shooter and his defense has been exposed this season. He may soon become another unicorn, but for now he’s just a versatile big man — we’ve seen those before. Nikola Jokic has been a hot topic ever since he returned to the starting lineup, but a big man with special passing skills isn’t necessarily a unicorn, and he lacks the shot blocking defensive presence someone like Embiid has. I should also mention that history has a number of otherwordly passers in the frontcourt. Let’s run a quick similarity measure for Jokic using his stats[2.].

Table: similarity score for Nikola Jokic 2017 (source: b-ref, 1980-2017, min. 1000 MP)

SeasonPlayerPositionAgeTeamSimilarity
2007Dirk NowitzkiPF28DAL5.52
1996Charles BarkleyPF32PHO5.55
1992Charles BarkleyPF28PHI5.75
2016Nikola JokicC20DEN5.76
1998Karl MalonePF34UTA5.90
1993Brad DaughertyC27CLE5.99
1993Charles BarkleyPF29PHO5.99
1992Brad DaughertyC26CLE6.02
1994Charles BarkleyPF30PHO6.10
1988Larry BirdSF31BOS6.21
2012Greg MonroeC21DET6.22
2009Pau GasolC28LAL6.28
2000Karl MalonePF36UTA6.32
1995Charles BarkleyPF31PHO6.32
2013Kevin DurantSF24OKC6.35
1982Mychal ThompsonC27POR6.36
1989Robert ParishC35BOS6.46
1989Charles BarkleyPF25PHI6.47
2006Brad MillerC29SAC6.54
1985Kareem Abdul-JabbarC37LAL6.55

Yes, as bizarre as it seems, Dirk Nowitzki’s MVP season is the most similar — even over Jokic’s rookie year — and a number of other All-Star seasons rank highly too. It’s surprising, but the list also reads as one of big men with versatility and a wide range of box-score stuffing. It’s hard to find non-All-Stars too. There’s Greg Monroe, and ranked 21st (not seen above) is Lamar Odom, who in his prime was one of the best non-all-stars ever. If you only look at shooting and passing stats, along with position, then the similarity list reads like one of notable European players: Alexander Volkov, Pau Gasol, Marc Gasol, Brad Miller, Alvan Adams, and Arvydas Sabonis. That doesn’t make him a unicorn; but it does signal Jokic’s stardom is just now dawning.

I know language isn’t perfect and it’s easier to apply a fun word to everything you so desire, but the meaning can get lost and words have a purpose. A unicorn is a player who has blended the skill and grace of a perimeter player, often by shooting from outside, (the horse aspect) and the ferocity and power of an interior rim-protector (the rhino aspect.) They used to be rare beasts, but they’re increasing in number — this is a new era.


[1. It’s just the proportion of a team’s shots a player has taken while on the court. Free throws are included.]

[2. The inputs are everything on b-ref’s advanced stats section except for Win Shares and PER plus FT% and position. Usage and AST% are given a weight of two, while position got a weight of three.]