Nylon Calculus Week 11 in Review: All-Star arguments

facebooktwitterreddit

With the All-Star Game a month away, I thought this was a good opportunity to cast some imaginary votes. At the same time, I did a bit of wrangling to update Dredge, and now I’m able to update it more frequently without too much of a lag. I’d recommend reading the introductory article before diving in.

All the stats included in this article are through Sunday January 8th. And all opinions are, unfortunately, my own, and if you think I’ve being biased against a particular team or player, you are correct and it is part of my nefarious plan. With that, let’s look back at the last week in basketball while making All-Star Arguments.

Eastern Conference

Starting guards

There’s one lock here for a guard, and he should be an MVP candidate too: Kyle Lowry. He’s in a virtual tie on ESPN’s Wins metric (based on RPM and minutes played) and he’s near the top in Dredge too. He’s the driving force behind the Raptors, not DeMar DeRozan, who I’ll discuss in a few paragraphs, and the Raptors are roughly as impressive with the Cavaliers in terms of point differential.

Outside of Lowry, there’s a long list of other point guard candidates here. Kevin Pelton outlined the group here, and if you go by advanced stats, including Dredge, John Wall has a slight edge, at the very least. But are those stats missing anything? Let’s start with Isaiah Thomas, the point guard who has exploded in respect, partly because of a higher exposure level on the Celtics. His usage rate and his true shooting percentage have both increased significantly, which is quite rare for a player of his type at his age. It’s largely due to his Morey Index: he’s taking more 3-pointers and getting to the line more frequently. Consequently, only two other players — Kevin Durant and Stephen Curry — have added more points through scoring efficiency this season. While his offense has been spectacular, it is true that his defense is a borderline liability, and the effects are showing on the court.

Table: efficiency stats (points added relative to league average TS%)

PlayerTS%Shot usage%PtsOverAvg/ 100 Poss.
Kevin Durant64.925.2

3.26

Stephen Curry63.424.9

2.72

Isaiah Thomas61.430.4

2.58

James Harden61.028.9

2.39

Kyle Lowry63.821.9

2.37

Kawhi Leonard60.627.7

2.02

Giannis Antetokounmpo61.225.7

2.02

Damian Lillard59.627.5

1.71

LeBron James59.826.6

1.68

Rudy Gobert68.113.2

1.56

Kemba Walker has the next best case. Last season wasn’t a fluke. He’s truly a good shooter now, and he’s a well above-average scorer. He’s a steady hand leading the Hornets, where his usage rate is approaching 30 percent — it used to be rare for a player to hit a usage rate of 30 with a true shooting percentage above 60 percent, but Kemba, among others, is surprisingly close. I don’t think he quite has the case that John Wall has, because the Wizards would be so much worse without him, and that’s backed up by the numbers too. Wall is probably the better choice, and nothing about his stats is anomalous.

As for Kyrie Irving and DeMar DeRozan, I don’t think they have compelling starting cases. Irving’s defense pushes him out of contention, and the fact that Cleveland is controlled by LeBron, not him. I make this kind of argument every year about him, and I don’t think much has changed[1.]. Plus Wall has a minutes advantage over him (and almost everybody else.) As for DeRozan, my argument is the same even though his shots have increased in volume. I’ve rerun the same analysis I did over a year ago: Toronto lives and dies with Lowry, and without him DeMar DeRozan’s units, even when only looking at possessions with fellow starter Jonas Valanciunas, are atrocious — and don’t blame Jonas for this either, because there was a similar pattern with +/- star Amir Johnson when he played north of the border. DeMar’s empty calorie offense isn’t as valuable as it appears, and his defense is sneaky terrible.

Table: Lineup stats 2014-2017 with Jonas Valanciunas (via NBAWOWY.com)

On courtPossessionsOff. RatingTS%Def. ratingOpp TS%Net
Both6105.0111.555.7107.355.24.2
Lowry1549.0112.555.0109.254.23.3
DeRozan1378106.853.2114.255.8-7.4
Neither516.010153.2112.253.9-11.2

Starting frontcourt

There are two obvious answers here. LeBron James, perennial starting All-Star-caliber forward since the Grover Cleveland administration, is one answer, as he’s once again the primary force of a title contender. Then there’s Giannis Antetokounmpo, who’s also a legitimate MVP candidate. He’s on track to lead his team in points, rebounds, assists, blocks, and steals[2.], and there’s still a chance he could average two steals and two blocks a game — that kind of statistical combination is unprecedented. The non-box score stats are in his favor too, from +/- to the arcane — for instance, for a player his size with a high usage and assist rate, he has surprisingly few stolen turnovers. Giannis is every position simultaneously, and he takes steps with the width of Zeus himself. Plus last week he flushed a monumental game-winner at Madison Square Garden. At some point we’ll need to update his nickname to one of the Greek gods. I’d agree with the linked article: it’s Hermes, a trickster god with winged sandals.

As for the third slot, there are three schools of thought: Jimmy Butler, who’s played exceptionally well; a real frontcourt player you could actually play inside; and Joel Embiid, so we can hook him up. There are a large number of non-Jimmy Butler candidates who all have cases that are tough to separate, but I should address Joel “the Process” Embiid himself. It’s nice to have a truly post-modern center in the league and while he’s one of my favorite players already, some of his stats are troubling. He has a mountain of turnovers without an appropriate number of assists, and his team is terrible. Factor in his low minutes total, and he doesn’t have the merits of an objective candidacy — vote for him for other reasons. Still, across a broad swath of advanced metrics, he’s well above average, like his +1.35 mark in RPM and his Dredge is +1.32 (when I don’t regress to -2 based on his minutes.)

The best non-Butler candidates are Kevin Love, Paul Millsap, and Al Horford. But I’m afraid none have a strong enough case against Butler. Kevin Love has had his best season yet in Cleveland, but he’s not the kind of primary offensive fulcrum that Butler is, and his defense is overrated by box-score stats. Al Horford is quietly an overrated player, and his inability to rebound or box-out has hurt Boston — he’s still very good, of course, but not the kind of team-leading talent Jimmy has become. Finally, Paul Millsap, who’s on the trade blocks, has quietly aged — his blocks, his rate of shots at the rim, and some of his percentages are on the decline. As a result, he’s not a starting All-Star caliber player. Thus, I’d have to stick with Jimmy Butler, which is okay because then Giannis is the center. Let chaos reign.

Bench guards

I’ve already covered both Kemba Walker and Isaiah Thomas in the starting guards section. Needless to say, I think these are pretty drama-free picks, except for the extreme Boston/Brad Stevens critics. I’ll save some words here.

Bench frontcourt

There are two locks here I feel should be fairly uncontroversial: Kevin Love and Paul Millsap. Based on their current production and their history, I’m fairly certain they belong in the game. Millsap’s case is built on his versatility, and Love’s scoring and efficiency has been at their highest points since his 2014 All-Star season. Both have solid cases, and as an added bonus they’re actually frontcourt players, and not just big wings.

As for the last frontcourt choice, there are a number of choices, from high-end role players with good stats from Jae Crowder, Nic Batum, Otto Porter Jr., Patrick Patterson, and Kentavious Caldwell-Pope to bigger names like Paul George, Dwight Howard, Al Horford, Kristaps Porzingis, and Andre Drummond. The choices here are statistically close, and it comes down to philosophy and preference. While I’d love to see Porzingis or Joel Embiid at the game, those two guys are a bit overrated by their countable stats because they’re still learning NBA defense, which is especially true in the case of The Process. (Horford got the better of Joel, ultimately, in a game last week too.) Some of those role players have shockingly good overall stats, but even I can’t get myself to vote for Otto Porter. It comes down to Al Horford and Paul George, based on their long-term value (i.e. they’re proven to be this good or better.) But I should skip ahead to the wild card section because I may not have to make a choice in case they’re both ahead of another guard.

Wild cart slots

Besides Kyrie Irving, I don’t think another guard has a strong All-Star case. Jeff Teague has been pretty good, but nothing spectacular, and the movement off-season has shown George Hill is arguably the better player. I’ve already discussed DeMar DeRozan; I think he’s quite good and a useful piece (in many cases) on offense, but he’s woefully overrated. I wouldn’t want him over Paul George, for instance, who’s a much better defensive player even in a down season. Thus, I’m going with Al Horford (with Paul George taking a frontcourt bench spot) here and Kyrie Irving taking the last slot, with apologies to more esoteric choices like Nicolas Batum.

Eastern conference starters:
Kyle Lowry
John Wall
Jimmy Butler
LeBron James
Giannis Antetokounmpo

Bench:
Isaiah Thomas
Kemba Walker
Paul George
Paul Millsap
Kevin Love

Wild cards:
Charlie
Kyrie Irving
Al Horford

Western Conference

Starting guards

Shockingly, only one player has an air-tight statistical case, and it’s not Russell Westbrook. James Harden does exceptionally well across a variety of metrics, and Houston is punching above their expected level thanks to his inter-positional shenanigans. His defense isn’t quite as bad as it’s been before, and it’s showing on the court.

Westbrook’s Dredge is abnormally low. There are a few reasons, and some are a bit complicated, but I’ll summarize them. Firstly, Westbrook is crushing BPM; he has the highest mark ever with that metric. BPM uses interaction stats, and the AST*TRB variable, as well as AST*USG, are propelling him to unprecedented heights. However, I transformed a few interaction stats to snuff out outliers in Dredge. I may have been too aggressive, and Westbrook is a historic outlier.

The truth is somewhere in between Dredge and BPM for Westbrook. There are also a few subtleties to analyze. Russell’s efficiency is average, and based on how you value efficiency it can hurt him. There are also non-standard stats, like stolen turnovers. Basically, turnovers that end in steals are about twice as damaging as non-steal turnovers, and Westbrook has a very high proportion of those with one of the highest turnover rates ever. Additionally, Andrew Johnson’s PT-PM, which has done quite well at predicting future wins, is just as pessimistic as Dredge is about Westbrook.

Meanwhile, Chris Paul has few peers in his ability to avoid turnovers while scoring and assisting at an elite rate. Paul also has a better rating on RPM, though I could entertain an argument it’s because of Doc Rivers’ obsession with bench units and non-linear returns. I think I’ll go with Paul here, and I can explain it without any fancy stats: I truly believe he’d be more valuable on a better team than Westbrook, while Westbrook is more fully realized on a mediocre one.

Starting frontcourt

Both Kawhi Leonard and Kevin Durant are locks, so instead of wasting words describing why these titans should start, here’s Joel Embiid rumbling to the rim and dunking.

The last slot for a starter has a great many candidates, and many of them are traditional centers. I’ll fast-forward to the two best candidates: Draymond Green and DeMarcus Cousins have the best overall cases, and metrics actually have them neck-and-neck. However, you can carefully examine the evidence, and come to a reasonable conclusion: countable stat metrics (especially box-score based) are most likely overrating Cousins while undervaluing Draymond because the latter is much better at team defense and other small factors that don’t show up easily in basic stats than the former player. You can see this when you move to +/- stats, and then add in the bonus of Draymond elevating a superior team to all-time historic levels. He deserves to be a starter, and he’s on track to be in the Defensive Player of the Year conversation again.

Bench guards

The selections here are easy. Russell Westbrook has to be chosen, and hey, if he’s coming off the bench, he’ll have a chip on his shoulder and will try to throw down a 40/15/15 game at All-Star weekend. Then you have Stephen Curry “slumping” with a 63.4 true shooting percentage, scoring 25 points per game. Yes, I’ll take slump Curry over most players, please.

Bench frontcourt

I think four guys stand out, and it’s up to everyone else to beat their cases: DeMarcus Cousins, Rudy Gobert, Anthony Davis, and Marc Gasol. Since there are three spots here, and I’m confident one of them will take a wild card spot, I’ll quickly pick one guy to throw to that wild card discussion. Cousins is a game-changing monster, Gobert is a defensive beast with some sky-high percentages finishing at the rim, and Marc Gasol is an all-around maestro. Gasol is actually being underrated by most metrics because they don’t consider his team’s magical and consistent ability to win close games, so I’ll use that as a tie-breaker over poor Anthony Davis and his wreck of a team.

Wild cart slots

This happens every year in the west: the wild card competition is fierce and chaotic. Here’s my short-list: Damian Lillard, DeAndre Jordan, Anthony Davis, Karl-Anthony Towns, Klay Thompson, LaMarcus Aldridge, Eric Bledsoe, Mike Conley, and Gordon Hayward. Apologies to valuable role players, like Trevor Ariza and George Hill. Then there’s Patrick Beverley, who Dredge thinks is one of the most valuable players in the league based on his defensive profile and low mistake game; PT-PM loves him too. And I wish I had space to argue for Nikola Jokic, who plays like the next Marc Gasol and should have a solid all-star case with more minutes and experience next season.

Anthony Davis should be the first wild card pick. It’s not his fault the rest of the Pelicans roster is an irradiated wasteland. He’s had some monster games, and it’s without much help. Somehow he’s propelled Tim Frazier to the assist-pairings leaderboard. But now it gets tricky.

To summarize the all-in-one metric stats, I made the table below. I know this is metric soup, but it’s not too much to handle. Remember that BPM is box-score only, and VORP is a minutes weighted version. Dredge is like BPM but with miscellaneous stats you can find in play-by-play logs, like offensive fouls drawn. Then MVP Index is minutes weighted stat of Dredge, but it’s nonlinear and gives more credit to high-end contributions. PT-PM is like Dredge but with SportVU stats. (It also uses raw +/-.) Finally, RPM is similar to Dredge but mixes in +/- numbers and other factors.

Table: 2017 metrics (Dredge and PT-PM are unregressed)

PlayerBPMVORPDredgeMVP IndexPT-PMRPMWinsMetric mean
Gordon Hayward41.73.682.03.83.333.743.69
Mike Conley5.11.63.962.12.71.522.473.32
Damian Lillard3.51.72.851.02.31.693.592.59
LaMarcus Aldridge2.31.22.010.33.41.613.212.32
DeAndre Jordan2.51.42.010.42.81.483.382.21
Eric Bledsoe3.61.72.851.0-0.11.633.812.01
Karl-Anthony Towns2.81.61.480.30.91.163.231.58
Klay Thompson-10.3-0.060.42.81.373.430.78

Klay Thompson has the lowest rating: he’s a defender without defensive stats, as his impact isn’t felt in rebounds or steals. Plus he had a slow start. Towns would have been fun, but the Wolves have been horrible on defense and he’s a major culprit; he has a lot to learn. Bledsoe is hidden in Phoenix; he’s been quite good but his team has been awful, and they’ve been better when he’s off the court, killing him in metrics like RPM and PT-PM. I would not expect that trend to continue, but ignoring still doesn’t put him over the top of, say, Mike Conley. DeAndre’s largely been himself, but his offense is pretty limited and he’s not a Defensive Player of the Year-caliber guy. He’s still very good, but the west is tough. LaMarcus Aldridge is under-appreciated, but there’s nothing that particularly stands out.

I’d love to make the case for Damian Lillard as a Portland fan, but alas, they have been terrible and it’s hard to take him above the last two players: Mike Conley and Gordon Hayward. Conley is having a potential career season after signing the richest contract in NBA history, and his stats are up in several areas. The Grizzlies should not be winning this many games either, and it’s largely thanks to him and Marc Gasol. However, his 3PT% is flukishly high, and with how many he’s taking it’s having a real effect on his overall numbers. He’s overrated on defense and underrated on offense, and he has one of the better careers out there without an all-star appearance, but I think I have to slight him and award someone else.

Gordon Hayward was my darkhose All-Star pick before the season, like Isaiah Thomas was the year before, but I do think he has a legitimate argument. Utah’s an elite team when their core is healthy: when he and Gobert are on the court, in nearly 900 minutes, they’re outscoring teams by 8.3 points per 100 possessions. And with George Hill and Hayward, the figure (in only 155 minutes) that figure is a preposterous 27.5. Every metric loves him, and he’s an all-around offensive plus who’s a good defender too. His basic per game stats are slightly depressed playing in the slow Utah environment, but he reads like a real All-Star — take that, Paul George, who should also be an All-Star[3.].

Western conference starters:
Chris Paul
James Harden
Kawhi Leonard
Kevin Durant
Draymond Green

Bench:
Russell Westbrook
Stephen Curry
Marc Gasol
Rudy Gobert
DeMarcus Cousins

Wild cards:
Anthony Davis
Gordon Hayward

And if Tony Allen gets to vote for Luke Babbitt and Mo Williams, I get to pretend vote for every spot. It’s in the rules.

[1. Irving’s age curve is like a highway in Nebraska. His BPM starting from his rookie season: 3.3, 3.3, 3.2, 3.3, 1.6, and 3.4.]

[2. You’ll see this historical fact cited frequently in the coming months, so I’ll explain a couple things few others will. One, this accomplishment is usually done with totals, not per game numbers. The latter would include a Tracy McGrady season, as long as you ignore guys with fewer than 34 games played for the 2003 Magic. Also, the NBA has maliciously ignored the existence of the ABA, so you’ll rarely hear about the exploits of Julius Erving in 1976.]

[3. He may also deserve a Mokeski, though Kevin Love is having a resurgent season.]