TrueHoop: Stat Geekery
December, 11, 2013
Brian Spurlock/USA TODAY SportsRoy Hibbert isn't the only reason the Pacers have the NBA's best defense.
A tricky thing about basketball is that it's tough to know what's happening on defense. So tough that credit and blame are almost impossible to hand out from afar.
Back in the days of isolation basketball, maybe you could say, with some confidence, that Mark Jackson just scored on John Starks and that's that.
But nowadays, by the time Kevin Durant gets to the rim, the primary defender was supposed to force him to the baseline instead of the middle, but he got to the middle anyway because he's Kevin Durant. The big man was supposed to meet Durant as he arrived near the hoop, but that big man has also been drilled to close out on the wide-open 3-point shooter he has left to be here. So he's a half-step farther away, and all that together created a tiny seam, which is all Durant needs.
You could probably watch a play like that and figure out good ways to blame all five defenders, or their coach, for the Durant bucket.
It's tricky stuff. And yet we can't ignore it -- indeed it really is half the game.
Offense is easy, by comparison. So many little things have long been tracked on offense -- who shot the ball, who passed it to them before they shot it and whether it was a 3 or a 2 have always been fundamental to recording a game. That stuff has always been in highlights and box scores. It's public, searchable and well-known. In the last decade, our understanding of all that has only grown with many new measures.
It's not hard to get a sense, at a glance, who can score.
On defense, though, wow. It used to be that notoriously noisy adjusted plus/minus was the go-to measure, but that's not readily publicly available anymore. There are SportVU cameras in the sky at every arena this season, but it takes a dozen hours of Zach Lowe or Kirk Goldsberry sifting to glean anything conclusive from them. Haralabos Voulgaris has long been tracking this stuff, but his database is private. In other words, it's tricky even to find out the most basic things such as which players were on the darned court when the other team scored most efficiently.
Which means making an evidence-based case that one player or another is awesome at defense is tough -- or nearly impossible this early in the season, when the sample sizes are small.
But we're not entirely without tools. And we do have lineup data, and the fact is there are combinations of players against whom it is crazy tough to score. Whether or not those players are the cause of the other team's bad offense, it's too soon to say. But if I were looking for players who are making it happen on defense, here are some names for the early season short list.
The resurrection of the pioneering NBA Israeli's game has been told as one of stroking 3s and attacking the rim.
But something is certainly happening on defense, too, which may overshadow all of that.
With Casspi on the floor, the Rockets have given up 94.8 points per 100 possessions, which is almost as good as the league-leading Pacers. When he's on the bench, the team has given up 104.1 points per 100 possessions, which is pedestrian.
The defensive bottom line is that the Rockets have gotten 9.3 points worse on D when Casspi checks out. The number could be thick with early-season noise, but it's eye-opening nonetheless.
Looking at two-man combinations, you can see that almost any Rocket with Casspi is effective. With Terrence Jones and Casspi in, the Rockets only give up 85.8 points per 100 possessions. With Patrick Beverley: a stingy 90.6. Seven of the top 12 Rockets defensive combinations feature Casspi. Dwight Howard appears in that list only once ... with Casspi. Meanwhile, there aren't many Rocket lineups that perform well on D without Casspi.
It's possible his defensive qualities are overstated by these stats. But I don't think it's possible he's bad on defense.
I'd also suggest it's a long shot the plus/minus obsessed Rockets are eager to sit him. Casspi is also helping the team on offense. Terrence Jones and Chandler Parsons have been similarly effective. Which makes you wonder, as Omer Asik trade rumors heat up ... does it really make sense to trade for a shooting forward such as Ryan Anderson? Maybe so, but if playing Anderson means limiting minutes for Casspi, Jones or Parsons, it's tough to imagine the Rockets getting more effective in the process.
The Pistons' rookie hasn't gotten much attention this season, and rookie guards almost never have good defensive statistics.
But a quarter into the season, Caldwell-Pope looks like an exception.
The list of the NBA's top three-man defensive units so far this season are largely Pacers, as we'll discuss. At the time of this writing, nine of the top 25 are from Indiana, in fact. Which means players on 29 rosters are competing for the 16 remaining spots. So when I tell you that Caldwell-Pope is on the list five times himself, with a grab bag of Pistons ... well, something is up.
Worth noting: The Pistons, generally, aren't even good at D, ranking 20th in the league.
Dan Feldman and Rob Mahoney have both dug into this phenomenon recently. The gist is that the Pistons started the season terribly on defense, when Caldwell-Pope never played. They got a little better all in all, and then Chauncey Billups -- who has been terrible on defense at this age -- got hurt. So Caldwell-Pope earned his minutes by replacing a bad defender and while joining a lineup that was finding its feet.
He's also, to the naked eye, a wiry and active defender who gets around screens far more effectively than Billups or Rodney Stuckey.
Caldwell-Pope has played close to 500 minutes, during which time the Pistons have given up a stingy 96.9 points per 100 possessions.
When he has sat, Pistons are allowing 108.4. The difference is 11.5, at least some of which, you'd think, has to do with the fact that this rookie guard is living up to his predraft reputation as a committed defender.
It's a closely guarded secret that the Bobcats are good at something, but today their defense is fourth best in the league, just after the Bulls and just ahead of the Heat and Thunder. But line up the NBA's best defensive player combinations in terms of points allowed per possession, and Kidd-Gilchrist's long and noticeable name is all over the place. There are three four-man Bobcats lineups with MKG that play better defense than the best four-man combination of Indiana Pacers. If you rank the whole league's best two-man defensive combinations, the top five pairs are all Pacers -- except for Kidd-Gilchrist and Gerald Henderson, who are third in the whole NBA in that ranking.
Kidd-Gilchrist, who is out with a broken finger at the moment, has played nearly 500 minutes this season, during which time the Bobcats have basically been the Pacers, with a 94.8 points per 100 possessions. When he's on the bench, they give up more than 100.
This is fascinating. Durant is famous as a scorer and was not long ago derided for sub-par defense. Jackson is a guy who can create his own shot. But they can, evidently, make you feel them on defense.
When opponents have the ball, Durant and Jackson have been, by the numbers, a top-10 NBA defensive duo. And it's not a simple case of the Thunder being great at defense. It's worth considering it might be something about this combination. One of the best five-man defensive units in the NBA (minimum 50 minutes played) is Durant and Jackson with Serge Ibaka, Thabo Sefolosha and Kendrick Perkins. That lineup is one of the Thunder's most used and has an incredible defensive rating of 78.3. At the moment, if you substitute Westbrook in for Jackson, you have one of the Thunder's most familiar lineups, and one that gives up 103.3 points. The Westbrook lineup faces the best opponents and would be expected to perform a little worse. But 25 points per 100 possessions is a massive difference.
It's also noteworthy that lots of Thunder players have great defensive ratings when they're on the floor. Jackson, though, is the standout for whom, thus far, sitting has led the team to play much worse defense. Could be a fluke. Worth keeping an eye on.
Related: Put defense and offense together, and Durant and Jackson are, at the moment, literally the best-performing duo in the whole NBA.
The other Pacers
We know Roy Hibbert is really good at defense. We know his Pacers have been one of the best defenses ever thus far. When Kevin Pelton (Insider) wrote about this the other day, he pointed out that the Pacers were giving up fewer than 94 points per 100 possessions in a league that averages 106. No other team is close. So the Pacers are killing it.
And as I just dug through NBA.com/stats looking at player combinations, there's no arguing Hibbert is the dominant reason. In fact, if you take every two-player combination in the league, from every team, the best combination out of all of those thousands, in terms of holding opponents to the fewest points per possession, is the Pacers' Roy Hibbert and David West.
In and of itself, that does not prove they are the two best defenders. Far from it. But it would be just about impossible for them to be so high on the list while being lousy at defense. And that they belong there is affirmed by this: The second best combination out of the whole league? Hibbert and Paul George. Fourth best is Hibbert and George Hill. Amazingly, Pacers account for nine of the league's dozen most effective two-player defensive combinations, and Hibbert is part of most of 'em.
Just as it's impossible to argue Hibbert is anything but great on defense, it's also impossible to argue that he's the only reason the Pacers are good. The Pacers' center is only playing 30 minutes a game, and the Pacers are good on defense all night.
This is not a question of the starting five carrying everybody. None of the Pacers' five-man lineups, in fact, are in the league's 10 most effective defensively. It really is a team effort.
When Hibbert is on the bench, the Pacers give up 98.7 points per 100 possessions, which would still be a top-10 NBA defense.
Of course, George, who has been discussed as a candidate as both MVP and a first-team all-NBA defense, is a big part of that. Even though he's the epicenter of the Pacers' offense -- in a role where many players would catch their breath on defense -- George expends serious energy guarding some of the league's finest scorers. Despite those challenges, he's still a mainstay among the Pacers' best defensive combinations. When George sits, opponents score a little better than when Hibbert sits.
But you know who else has been on the floor for long minutes of great defense for the Pacers? Almost everybody. David West, C.J. Watson, George Hill, Orlando Johnson, Lance Stephenson, Luis Scola -- these are not the Pacers' most famous defenders. I have named eight Pacers in this article. Put any three of those players together on the court, and Pacers are playing good defense.
When any or all of them are on the court, the Pacers as a team average better defensive performance than the Spurs, who are the league's second-best defensive team.
It's almost impossible to find any combination of Pacers players that is bad on defense. It's amazing. (3-point specialist Chris Copeland might be the one exception. He has not been great on defense, the statistics say, but he is also new to the team and has averaged less than four minutes a game, so it's hard to know what the future holds for him.)
Clearly, coach Frank Vogel knows something.
December, 9, 2013
December, 3, 2013
It's hard to find a productive NBA big man. But the Celtics, Thunder and Nets pulled it off, with unheralded rookies Vitor Faverani, Steven Adams and Mason Plumlee playing well. David Thorpe on big men most of the league missed.
November, 19, 2013
Clinical psychologist, professor, author, and former Suns advanced stats consultant Stephen Ilardi says naturally occurring testosterone has a massive effect on the brain. A few days ago, Ilardi told TrueHoop TV why it is that fans of the winning team riot.
October, 30, 2013
By Ethan Sherwood Strauss
Layne Murdoch/NBAE/Getty Images
James Harden helped his career tremendously by insisting on being a leading man.
When discussing the James Harden trade, the main focus is normally on Oklahoma City. It’s about whether they lost big, whether there’s still time for them to emerge victorious from a now-infamous deal. Even in a league that celebrates brilliant individualism, transactions are viewed through a team lens. A “good contract” is rarely good for the player, for instance.
We’ve been through the implications in Oklahoma, so I’d rather parse what the trade meant for Houston’s rising superstar. Not only did Harden make an extra $24 million by spurning Sam Presti’s last, best offer, but he also received a giant boost in status.
While I’m certain his 2012 Sixth Man award was gratifying, that plaudit is likely trumped by this year’s All-Star and All-NBA selections. Over the course of a season Harden surged from No. 26 in #NBArank to No. 4. In the annual survey of NBA general managers, Harden surpassed Kobe Bryant as the league’s best shooting guard.
What’s funny about Harden’s reputational ascendance is that it’s difficult to prove he got any better between last season and this season. His Win Shares per 48 minutes declined, and his moderate boost in PER can be ascribed to more shots taken. Under greater defensive scrutiny, his true shooting dipped a whole six percentage points. On many defensive possessions, the closest thing to Harden moving toward an assignment was his beard’s slow growth in that general direction.
This isn’t at all to say Harden had a bad season; he was a free throw machine whose overall success was instrumental in wooing Dwight Howard from Los Angeles. It’s just that the extra praise had more to do with Harden’s increased floor time and opportunity than his improvement as a player. Before, Harden was underrated, his super-efficient production obscured by far fewer minutes and touches than a star of his caliber usually receives.
Once a player is locked into a certain kind of role, it’s hard for NBA observers to envision that player in a different role. Daryl Morey wisely didn’t fret over whether Harden could handle being “the man” in an offense, instead trusting that a former bench player would continue to be himself as a starter. Harden was so efficient as a super sub that he could afford to be a little less efficient in a bigger role.
Another All-NBA team, another All-Star Game, and Harden will have equaled stylistic twin Manu Ginobili’s individual awards resume. Not only does Manu serve as an apt player comparison for Harden, but he also offers a glimpse into where Harden’s career might have been headed.
Per minute, Manu’s production has been on par with Kobe’s -- the Argentine leads in win shares rate and trails slightly in PER. Would Manu be considered Kobe’s equal had he averaged 37 minutes over his career? This is unknowable, in part because an oft-used Manu very well could have been an oft-injured Manu. But Harden gives a good indication as to what a season of Manu would have looked like on another roster. Such a season might not have been on the level of late-aughts Kobe, but would have almost certainly resulted in more fame than Manu’s been accustomed to.
I actually suspect that Harden will receive more renown than Manu ever did while never being quite as good as Manu was. This is amusing to consider when, just a year ago, Harden was being chided for ignoring legacy considerations. The framing was that, by leaving a contender, the league’s best sixth man was consigning himself to a kind of ringless obscurity.
Instead, it's the opposite. Even after getting bounced from the playoffs by his old team, Harden is more famous and more highly thought of than ever before. Sports pundits reward players for winning titles, but we’re also quite reductive in assigning credit for those titles. Kobe fans wear “5 rings” shirts as though the fact Kobe won rings matters more than anything about the Lakers. A “Whose team is it?!” culture means Kevin Durant would have gotten more than the lion’s share of lauding for hypothetical Thunder championships. Harden would have been an extra in the film about KD's career. Leaving a contender was Harden’s chance to salvage a legacy, not destroy it.
Much as the media derides players for their selfishness, we’re also the ones who handsomely reward selfishness. When faced with accepting a below-market deal and a continued bench role, Harden sided with his own interests, faced some criticism, and put himself in the express lane to lavish praise. The lesson is that “being the man” isn't some silly fixation of the selfish athlete. The role has real, positive consequences on how you'll be considered and remembered.
October, 20, 2013
By Kevin Arnovitz
Kevin Lee/NBAE/Getty Images
Projecting what Kobe Bryant can contribute in 2013-14 is guesswork, but the numbers aren't optimistic.
Our #NBArank project set off a firestorm this past week when Kobe Bryant was unveiled as our panel’s choice at No. 25.
The panel was asked to rate each player from 0 to 10, with the following guidelines: "Rate the overall level of play you PREDICT for each player for the upcoming NBA season. This includes both the quality and the quantity of his expected contributions, combined in one overall rating."
In other words, how much can we expect this player to contribute during the upcoming NBA season?
With that in mind, many voters on the #NBArank panel projected that, coming off surgery for a torn Achilles tendon, Bryant’s production in 2013-14 will fall off dramatically. We don’t know when Bryant will return to the court and, when he does, we have no idea how the injury will affect his performance.
Those questions inspired the team at Princeton Sports Analysts to investigate. PSA is a collective of Princeton undergrads who study advanced analytics, stats and the economics of sports. To get a better sense of how Bryant’s injury might affect his output this season, the gang at PSA turned to a paper published in March titled, “Performance Outcomes After Repair of Complete Achilles Tendon Ruptures in National Basketball Association Players.”
The study looked at 18 players who suffered the injury, and the findings were discouraging, as summarized by PSA:
Of those 18 players, 7 were never able to return to NBA action, 3 returned for just one season, and the remaining 8 would go on to play 2 or more seasons. And of those players that returned, their performance suffered drastically, especially in their first season. In their study of the 11 players that returned to the NBA, the players' PER (player efficiency rating), decreased by an average of 4.57 points. In the second, it decreased by 4.38 points.
... If you decreased his PER by the average reduction of 4.57 ... you’d find that Kobe would’ve ranked 49th in the league last year, some 24 spots higher than where ESPN has him in their NBA Rank. Kobe is an animal, but the stats indicate that the anger towards his NBA Rank of 25 is far from justified.
That’s a steep drop, and if we want a case study in the impact of an Achilles injury, Elton Brand provides a helpful example. In 2005-06, Brand ranked 6th in the league with a PER of 26.67. The following year, he dropped to 23.16, which was still good for 14th overall. That following summer, Brand suffered a ruptured Achilles tendon, missed all but eight games in 2007-08 and has never recorded a PER better than 18.5 since.
Bryant is a different caliber of ballplayer, the ultimate outlier -- in conditioning, preparation and intensity. It’s reasonable to assume Kobe will apply a unique level of focus into his rehabilitation so that, when the next study of NBA Achilles injuries is released, he’ll reside at the far end of the production axis on the scatterplot graph.
The PSA team also notes that age, something that’s been cited as working against the 35-year-old Bryant, wasn’t determined to be a factor in recovery, so talk of the challenges an older player faces coming back from an Achilles tear doesn’t conform to the data.
October, 16, 2013
#NBArank calls Kobe Bryant the NBA's 25th best player. J.A. Adande says that's not going to sit well with Kobe.