NBA Week in Review 14: All-Star Fever

Every year, I send in my All-Star ballot for the reserve selections, and every year I get ignored. Since I have no power or control over the All-Star Game, I’m instead going to complain on the internet about it with the choices I think should have been made. I’ll preserve the starting lineups even though one choice in particular was errant, but I’ll add an injury reserve for both conferences and those guys would theoretically take the spot the two Hall-of-Fame legends have stolen. And, naturally, I’ll tie in my selections with a few games from last week.

Western Conference

Guard, first spot: I think the player with the worst historical timing in the NBA is Chris Paul, who peaked when the league thought it was the right time to award Kobe Bryant in 2008 and when LeBron first peaked in 2009. He’s been in the shadows of great players despite being incredible himself, and it’s all the same this season. Nevertheless, he’s still a master at his position, and while we worried about how he’d age with his injury history, he’s so skill-based that he should be effective for a long time. For instance, his ability to pull-up for a midrange shot with great accuracy will buoy his value in pick-and-rolls even as he loses quickness. He’s still an otherworldly playmaker who’s had to score more with Griffin injured; his inclusion is unarguable.

Guard, second spot: The player for this slot, James Harden, is another obvious choice, but Houston’s collapse and his general indifference could lead to some skepticism. Yet, he still deserves a spot on the team easily because he’s an NBA offense unto himself. He’s the entire frame of the Rockets offense and they fall without him. Via NBAWOWY, the team scores 108.7 points per 100 possessions with him on the court, and 102.3 points without him (excluding minutes with their deep bench players KJ McDaniels, Montrezl Harrell, and Sam Dekker). Every season he’s been in Houston, he’s had a shot volume rate above 27, a rate of assists per 100 possessions above 7, and a TS% relative to the league average above 6 — so yes, he’s creating shots for himself, others, and he’s doing it efficiently[1. You can see more of those stats here.]. That’s a rare skill, and since 1967 he’s tied for first in the total numbers of qualifying seasons — and this year might be number five. He’s an easy selection, even if he’s a poor defender.

Table: one-man offenses since 1967, min. 1000 minutes played

Forward, first spot: Draymond Green was second last season in Defensive Player of the Year voting and probably deserved first, and he’s now a really useful offensive player too. His defense hasn’t slipped, and, as others have said, he isn’t a system player boosted by Stephen Curry and the rest of Golden State — he is the system. His value is hard to deny now, and he’s actually third in multi-year RAPM[2. This stat measures how well a team outscores its opponent when a player is on the court, adjusting for teammates, opponents, homecourt advantage, and a couple other factors. It uses multiple seasons of data weighed by recency for more useful — i.e. predictive and accurate — results.]. If you adjust for minutes played, he’s either first or second overall[3. Depending on what the assumption is on who he’s “replacing” on the theoretical team.]. Green’s not a big scorer, but there are more ways to provide value on the court.

Forward, second spot: One player has become a breathtaking, rim-shaking force on the court, and even the newer generation of advanced stats sees his high on-court value. DeMarcus Cousins is on a losing team, but that’s not his fault and when he plays Sacramento is formidable. The Kings are +11.6 points per 100 possessions better when he’s on the court through the games last Sunday, which is best on the team[4. There’s a tie between him and Casspi, but the tiebreaker is given to Cousins for being considerably better at basketball.]. He fills up the stat sheet so thoroughly and voluminously that he has few peers, and, yes, some of those stats are negatives, but he does a lot of valuable smaller things most fans don’t track, like drawing an offensive foul. Here’s an example where he gets whacked in the face by Jeremy Lin and sells the contact. He’s a huge dude, but he lets other referees know what other people have fouled him — and that’s very valuable.

Forward, third spot: For the last forward slot, I’m taking Anthony Davis. I suppose someone could mount an argument against his inclusion based on his team capsizing and some poor pure plus-minus stats. But he doesn’t run the front office or play all five positions on the court at once, and half-season plus-minus or basic on-off stats are really noisy and not entirely useful. He’s an excellent scorer for his size with enough versatility to flummox good defenses — which is more apparent when he has capable teammates on the court. With his range, his pick-and-roll and pick-and-pop ability, his ability to drive, and his low turnover rate, Davis, even in a down season is one of the best offensive players at power forward or center. And while his defense is overrated by his high rate of blocks, he’s still a plus there and that’s enough for an all-star selection.

To counter the “he’s on a terrible team and doesn’t deserve the all-star game” argument, let’s use some basic, conventional logic: when you have worse teammates, you get more offensive attention and it’s easier for other teams to deal with you. Teams love it when Omer Asik shares the court with Davis because they can ignore him for large segments of the play. For instance, when Davis isolates here and tries to drive, Brook Lopez, who was “defending’ Asik, comes up with his huge frame and blocks the lane. Plus, Alonzo Gee was crowding the paint with another guy because he was slowly moving under the basket. It’s that sort of floor-crunching that can kill a player’s effectiveness, and it’s more common on poor teams.

Wild card, first spot: After contemplating for a while, I think the right pick here is DeAndre Jordan. I’m been skeptical of his value for a while, but he’s definitely not a bad player. And as weird it this would sound to most fans, he might be better on offense than defense — he is, you know, shooting 71.5% from the floor. The metric PT-PM, for instance, utilizes rim protection and still has his offense rated higher. Actually, his value is best expressed in terms of defensive attention. Since he’s such an active dunk and lob threat near the rim, defenses have to be paranoid any time he’s near the rim and, say, Chris Paul has an angle. That warps opposing defenses, and it sometimes frees a teammate. That’s valuable. He’s still not a great defensive player, but he’s good enough there that box score-agnostic stats like RAPM see him as a plus. That’s just enough for the All-Star Game.

Wild card, second spot: While I don’t think Golden State should have three selections just because they have an argument for best (first half) team of all-time, Klay Thompson has performed well enough for a good argument. I don’t think his defense is as good as his reputation suggests — Golden State’s defense hinges on Iguodala, Draymond, and (when playing) Bogut — but with his elite shooting and his now well-rounded game he’s a large positive on offense. He’s good enough now that he’s risen above his 3-and-D label, and it’s the combination of Curry and Klay that stymies opponents on the perimeter; you can’t leave one open. If Curry didn’t exist, we’d be discussing Thompson’s chances at breaking the three-point record. Oops.

Kobe’s stolen/injury replacement spot: The west is usually crowded, and unfortunately the Kobe farewell tour has claimed a victim: Damian Lillard is no worse than he was last season as an All-Star, but there’s a chance he won’t be invited. He’s taking on a larger role in scoring and his shooting efficiency has been fine; he’s still a pretty good playmaker, but his limitations there keep him from progressing further. His defense certainly takes a chunk out of his offensive value, but the All-Star Game isn’t about only selecting guys who aren’t significant negatives on defense, or else we wouldn’t invite James Harden or, yes, (old man) LeBron James. Plus, as I mentioned with Klay, we’d be more impressed with his shooting if Curry hadn’t broken the league.

Apologies to LaMarcus Aldridge, who lost some stats for the Greater Good; Gordon Hayward, who really deserves a strong All-Star argument one of these seasons but is usually stacked behind a number of stars in the west; Dirk Nowitzki, whose shooting will always make him a threat; and Kevin Garnett, who will definitely not get the Kobe Bryant retirement treatment despite an incredible career. But the most interesting argument would be for Ricky Rubio, who annually rates well by advanced stats and has been largely healthy this season. By traditional stats he looks mortal, but the deeper you go the more amazing he seems[5. Here’s one straight-forward stat where he excels: he creates the most turnovers out of anyone from 1997 and on.]. This may or may not be correct, but the Wolves are roughly 1000 points per 100 possessions better when he’s on the court versus Zach LaVine at point guard. But I doubt he’ll ever be an All-Star given the staggering depth right now and his scoring woes.

Eastern Conference

Guard, first spot: Jimmy Butler isn’t a selection I need to argue too heavily. The biggest mark against him is that his “pure” adjusted plus-minus stats aren’t elite — but they’re not bad either. He’s a solid scorer for the Bulls with great all-around skills, from outside shooting, to improved passing, to a low turnover rate. He’s not a defensive specialist anymore so he’s not highly valuable there, but he’s still pretty good. The SportVU metric PT-PM puts him in the top 30 and with his high minutes load, he’s produced a lot for Chicago. Overall, partly because of the weakness at guard in the east, he gets the selection here easily.

Guard, second spot: John Wall hasn’t had his best season, but I try not to penalize players too heavily when they have a long track record of playing at an All-Star level, especially when the issues are mainly related to shooting percentages. Wall still has his exquisite passing, and the Wizards without him would be adrift on offense. He’s pretty good on the other end of the court too, sometimes, and he’s enough of a leading playmaker and overall force that I can put him here above a long line of point guards. Just appreciate all his nifty passes. With his own speed and how fast he can hurl the ball it’s tough for defenses to recover, and he can quickly find guys at the rim like with this play. Wall’s in, even as the Wizards suffer.

Forward, first spot: The Atlanta Hawks have been disappointing and worse than most people expected, but that’s not because of Paul Millsap. There are multiple explanations why Millsap continues to be underrated, but the most effective one is that he plays for the Hawks and he’s a jack-of-all-trades player who’s good across a number of categories. People often overrate guys who have one prominent skill, usually scoring, but are deficient in many other areas. You don’t think about a scorer’s help defense, rebounds, and steals because you’re thinking about his scoring. Millsap is the opposite, and people don’t realize how good he is because his value sneaks up on you. But he’s an easy pick, even if his short-term plus-minus stats are mediocre.

Forward, second spot: I don’t think I have to explain that Chris Bosh rates well across a variety of advanced stats. In some ways he’s like Anthony Davis, but understated. He’s a well-rounded big with range whose efficiency is boosted by a low turnover rate, and he’s a definite plus on defense. He’s currently on track for another season with at least 300 midrange jumpers attempted with a conversion rate over 45% — that would be his sixth such season and his fourth in a row. In fact, since 1997 there are only 66 seasons that qualify for big men[6. Dirk Nowitzki has 13 seasons, by the way.]. People used to concentrate on his deficits, like his low raw rebounding numbers, but he’s highly skilled and he’s made multiple deserving All-Star appearances — this will be another.

Forward, third spot: Bosh, in fact, is a great lead-in for my next all-star selection, and it’s one with a more subtle explanation: Kevin Love. When the big three first struggled in Miami, Chris Bosh was often singled out because his traditional, basic stats declined across the board and he was shipped over from a perennial doormat that never seriously challenged the conference. Since LeBron James and Dwyane Wade used a large number of shot attempts, few were left for Bosh, but he was valuable in other ways. His pick-and-roll defense was superb, which was a vital to their defensive style. It meant he was away from the basket on defense often, and on offense he was usually spotting up and stretching the floor — thus, his rebounding totals were depressed. But he had a valuable role, and defense and other vagaries of teammate interaction are tough to detect accurately.

Kevin Love is having the same issues with the gap between his conventional stats and his impact. People are aggressively negative about his defense, and this reminds me of Bosh’s critics with respect to his rebounding and his “softness.” There’s some confirmation bias here too because people have never really accepted a guy of his size and athleticism being a star, so they select examples where he fails more often than when he succeeds. While Love’s defense in space and against pick-and-roll’s is poor, the effects aren’t as bad as people imagine and he has positives elsewhere[7. If you think he’s an unplayable liability on defense, consider that the Timberwolves with Pekovic at center we an above average defense, and the Cavaliers were too even before their injured guys came back.]. As Kevin Pelton explained, he rebounds very well and doesn’t foul, and when he’s on the court Cleveland does, indeed, rebound better and foul less often. But it’s his unique offensive value that sets him apart. People complain that he’s “only” been used to space the floor, but that’s incredibly valuable since he’s an actual threat and unlike a lot of stretch 4’s he’s a real scorer too, and you can’t just hide any perimeter guy on him. His ability to post-up and score efficiently with a high rate of fouls drawn erases a few options opposing defenses have. He’s a great passer too, and on a team with scorer and shooters everywhere that skill is magnified. LeBron and Kyrie Irving are slashers, and they can get to the basket more easily with him on the floor. Over the past couple seasons the Cavaliers have played significantly better with him on the court, even after adjusting for teammates and opponents.

The notion that Love is not valuable or that he should be traded immediately because they, and he specifically, did not play well in a game against the Warriors — people exclude their first game against Golden State — is extremely narrow-minded and short-sighted. You don’t load up your team to match-up against a single opponent, especially one in another conference, because the league changes too quickly and there’s no guarantee of facing each other. Plus, it’s not like anyone has an answer for the Warriors — so what’s your replacement for him? Draymond Green? Finally, it’s interesting that the one time the representation argument (great/conference leading teams should have two selections at least) is not invoked is because of him. I hate that argument, but let’s rephrase it intelligently: the Cavs were very good for the first part of the season despite missing Kyrie Irving, Iman Shumpert for a long stretch, LeBron lost his jump shot, and Timofey Mozgov has been ineffective. Where is that value coming from? Love’s a smart pick here.

Wild card, first spot: Again, I don’t subscribe to the notion that every good team “deserves” one all-star, like we vote by constitutional representation, but there’s one player who I saw near the top of multiple player metric leaderboards: Jae Crowder. After digging into the numbers, including my own, I’d say the driving forces are his solid defense, led by his high steal rate, for a good defensive ball-club, coupled with his efficient, three-point heavy scoring — essentially, he’s one of the better 3-and-D guys out there. One could be wary in being bold enough to pick him for an All-Star Game, but he’s a key component in transforming Boston into a surprisingly competent team, along with Isaiah Thomas. This isn’t a short-term fluke either because multi-season RAPM ranks him highly. He has Kawhi-like hands of death, wrestling the ball away from Nikola Jokic here and poking the ball away from Danilo Gallinari on a drive. He’s a versatile defender too, playing a lot of small ball power forward and switching and trapping a lot, like this play where he trapped Ramon Sessions against he baseline and forced a deflection. He gets the annual Andre Iguodala selection, as a well-rounded defensive wing whose value exceeds a few more obvious names.

Wild card, second spot: There’s one player left who ranks highly on my stats and a few others, and he’s not an unreasonable All-Star case. Kemba Walker has had a mini-breakout season for the Hornets, who most people left for dead over the off-season. While Nicolas Batum has helped the team turnaround, Kemba has been the leading player on offense and his defense, I believe, has been decent too. He’s had little to no growth the past few seasons, but that actually means a player is more likely to improve — sometimes it’s just one or two below average seasons that mask a sharp age curve and suddenly a player erupts, or you may think of it like a lagged improvement that delivers everything at once. Whatever case it may be, Walker’s efficiency is good enough that he’s not hurting his team. This was a tough call, but looking at every available stat and how each player controls his own team, I think this is the safest selection.

Bonus/injury replacement spot: Isaiah Thomas has the best combination of scoring and playmaking out of anyone left. I shouldn’t need to defend his worth anymore because the Celtics have clearly performed well with him over the past year, and there was a similar pattern on his two previous teams. His defense limits his value, but he’s not truly a matador. Also, don’t count out his size in the post-season — few people feared J.J. Barea but everyone had trouble stopping him in the 2011 playoffs.

Apologies to DeMar DeRozan, who’s been driving and getting to the foul line like a mad man, and a few others including Andre Drummond and Pau Gasol. The former has crazy rebounding numbers, but raw rebounding stats aren’t useful on their own and I don’t believe he’s a difference maker on defense yet. The latter can reliably score and rebound, and he also consistently doesn’t box out and lets the other team score on him. Reggie Jackson is somewhere around the conversation, but I thought Kemba was better. But the guy I’d consider for the next spot is probably Al Horford, who’s quietly been very effective for a team that’s once again being ignored.

Yet Another NBA Metric: Play-by-Play Model Coming to a Tableau Near You

For reference below, I’ve included the top players ranked by a metric I’ve been working on. It’s similar to the model I explain here, except that I’m not including SportVU data — but there are a large number of other variables I’ve tested and considered that virtually no other public metrics use. I still need to error-check and do more out-of-sample testing, and the version below has no interaction stats (i.e. no assists*rebounds), but the results are useful and unique enough that I thought they’d provide some value. This metric will eventually be posted here. The leaders below are mostly known stars, but Caldwell-Pope is an interesting inclusion. However, BPM, basketball-reference’s metric, rated him 28th by their value score at the time I pulled this data. And, naturally, I have to mention near All-Atar Zaza Pachulia: he’s actually rated well by advanced stats the past few seasons, although I’d hesitate to select him over, say, Draymond Green because I am a sane person.

Table: play-by-play statistical plus/minus model leaderboard, 2016

By the way, Stephen Curry has the best single season SPM out of anyone since 1997, and that’s even with regressing his stats to the mean based on his low minutes because it’s only been half a season. The all-star game isn’t good enough to reward his brilliance.