Analytics Rumors

James is 10-for-25 (40 percent) in the playoffs on shots attempted when tied or trailing by two points or fewer in the final 24 seconds. Bryant was 7-for-28 (25 percent). James is 12-for-23 (52 percent) on playoff game-tying or go-ahead field goals with less than 10 seconds left. Bryant was 5-for-22 (23 percent). James has made five buzzer-beating game-winners in his playoff career. Bryant made one. This is not close.
As for what he’s focused on, Beal said it is being a consistent scorer “with efficiency.” He wants to increase his volume as a scorer without leaving his shooting percentages to suffer. He has been able to achieve that for the most part, this season averaging 30.5 points while holding a 52.0 effective field goal percentage. That is not easy to do, especially as a guard. But don’t let the efficiency talk lead you to thinking Beal is poring over the numbers, especially the advanced metrics. He also dropped a line on the show that may raise some eyebrows. “Honestly, I’m not an analytical guy. I say F the analytics, just go hoop,” he said.
Now that he has approached the hiatus with that same kind of mentality, studying ways to further understand his body and maximize what he can do, Millsap figures to be a major X-factor in Denver’s playoff run should one occur. With access to his own gym, leaning on a workout plan he created and used to spark the season’s early splash — he averaged 13.9 points through Denver’s first 18 games on 49.1/50/88.2 percent shooting splits — it will be fascinating to see whether Millsap can spring into the postseason with a similar energy level. The advanced data is clear: The Nuggets have been at their best over the past three seasons, on both ends of the floor, when Millsap is on the court and healthy. He could very well set the tone for how Denver approaches an unprecedented postseason tournament.
Battier is already the Miami Heat’s vice president of basketball development and he owes a lot of the success to his strong knowledge of analytics. Earlier this week, he credited his time with Houston Rockets general manager Daryl Morey and assistant Sam Hinkie for acquiring the skill. Battier played with the Rockets from 2006-11. “I was lucky to play for Daryl Morey and Sam Hinkie, who taught me how to look at the data,” Battier told The Athletic. “Analytics is like blackjack. When the dealer has a five showing, what do you do? You double down. Why? Because the book tells you that is the best play at the time and gives you the most chance to win the hand and win money.”
Battier was known his analytics knowledge as player when he helped the Heat win consecutive championships in 2012 and 2013. He would often lead discussions with teammates about player data and it was common to see him having long conversations in the locker room with writers who had an analytical background. “Analytics has a lot of the same aspects as [blackjack],” Battier said. “When you make Kobe Bryant go to his left hand, he’ll score 44 percent of the time. When he goes to his right hand, he’s gonna score 56 percent of the day. So you don’t need to be a math genius to understand that the defender should take Kobe to his left hand.”
Miami stands as one of the most analytics-friendly teams in the league, a skill that can in-part be attributed to Battier. So how did Battier get schooled in the world of analytics? Rockets general manager Daryl Morey deserves a share of the credit. “I was lucky to play for Daryl Morey and Sam Hinkie, who taught me how to look at the data,” Battier told The Athletic’s Kelly Iko and Mo Dakhil. “Analytics is like blackjack. When the dealer has a five showing, what do you do? You double down. Why? Because the book tells you that is the best play at the time and gives you the most chance to win the hand and win money.”
Beyond the Celtics, the top player isn’t Jordan either; Kobe Bryant edges out his mentor because his title teams were never as strongly favored as Jordan’s. For instance, the 2000-01 run, in which the Lakers set a record (at the time) with a 15-1 playoff mark, is even more astonishing in retrospect. Despite nearly full regular seasons from Bryant and Shaquille O’Neal, those Lakers had the point differential of a 51-31 team, and their playoff opponents were all superior: Portland (53-29 Pythagorean record), Sacramento (57-25), San Antonio (63-19), and Philadelphia (54-28). Maybe the 2000-01 Lakers simply underachieved in the regular season. But that one title was part of a trend: Bryant’s teams regularly advanced farther in the playoffs than expected. None of his titles were faits accompli—judging only by the quality of playoff opponents, his Lakers teams faced some of the toughest slates any champion has ever navigated.
Instead of looking at the changes as shackles that are holding the team back, Lindsey is looking at ways the Jazz can take advantage of the situation. Part of that effort has fallen heavily on the Jazz analytics team, led by coordinator of analytics Cory Jez, whose group of consultants are working on draft modeling informed by statistical work and analysis.
“By definition, if you’re in management or you’re in scouting, a good percentage of your job is remote work, is video work, it’s database work, whether it be the scouting database or the statistical database,” Lindsey said. “We’ve had an opportunity to shift those percentages and using technology to a greater degree versus the in-person scouting and hopefully we can tip a few of the odds in our favor relative to the draft and free agency moving forward.”
The Houston Rockets and general manager Daryl Morey have become pioneers of the NBA’s analytics movement in recent years, which is why they’ve employed coaches like Mike D’Antoni, have played essentially without a true big man and shoot more 3-pointers than just about any other team in league history. With that being said, former Rockets head coach and Hall of Fame power forward Kevin McHale isn’t a fan of analytics and even took a bit of a jab at his old team during a radio interview this week. “You guys are caught up a little bit in Houston on numbers. Numbers do not win games. The one number that wins the game is if you have one more point than the other team. That’s the biggest number. But numbers don’t win games,” McHale said on ESPN 97.5’s The Usual Suspects.
In fact, since the NBA moved the 3-point line back in 1997, no one has come close to topping Jordan’s mark of midrange makes in a season. Most midrange buckets in a season (since 1997-98) 1997-98 Michael Jordan: 671. 2005-06 Dirk Nowitzki: 564. 2003-04 Kevin Garnett: 540. 2000-01 Glenn Robinson: 506. 2005-06 Kobe Bryant: 502. 2008-09 Dirk Nowitzki: 501.
Even prior to the 2011 lockout dip, defenses were catching up, in large part because of the way Tom Thibodeau-inspired “strong side overload” defenses were able to effectively shrink the floor before Golden State and Steph Curry definitively proved “jump-shooting teams” could win at a high level — and as has been noted in this space many, many times before, describing those Warriors as somehow more reliant on jump-shooting than previous top-level teams is largely ahistorical — fully energizing the pace-and-space era in which we find ourselves.
At a high level of abstraction, the basic story is familiar — more pick-and-rolls, fewer isos and postups. But this analysis merely scratches the surface. Areas for future exploration getting deeper into each play type include: Where are all these additional 3-pointers coming from? Which play types more often result in 3-pointers than previous? If there has been a rise in spot up 3-pointers, why has there been so little change in the efficiency of spot up plays?
Storyline: Coronavirus
“T.EA.M. is about where the love of the game can take you,” Musante said. “It’s about meeting students where they are. T.E.A.M. is an equalizer, ensuring that young folks who are often underestimated have the opportunities to explore their passions, develop skills and build networks to unlock their potential, right now.” But with all the work put in, all the recommendations, all the networking, McQueen still sees the hole in the system — which is why it matters a great deal when this topic comes to the forefront. In the NBA currently, there are at most a handful of African-American video coordinators, including Chris Kent with the Bulls, Ryan Frazier with the Suns and Ryan Lumpkin with the Wizards. The others on McQueen’s list have moved up the ranks, but the pipeline hasn’t been replenished diversely. As data and analytics have become more popular, McQueen has watched the talent become even less diverse at the entry levels.
Recent history indicates taking a foreign-trained player in the top 10 picks is a bigger risk than a U.S. player. Using Win Shares, a statistic that provides a single number to evaluate a player, only 7.7% of foreign-trained top 10 picks develop into All-Star-caliber NBA players vs. 20.7% of U.S. trained players. Foreign-trained players, however, slightly out-perform their U.S. competition in the rest of the first round, 12.9% vs. 9.5%. Though not talked about much leading into the draft, Bitadze was going to be a first-round pick for someone. Several teams, including the San Antonio Spurs, had him high on their draft boards. After failing to trade up into the lottery, the Pacers took him one spot before the Spurs picked.
The 76ers have a 99.0 defensive rating when Simmons is on the bench, which would rank No. 1 in the league by a mile, and a 107.2 rating when he’s on the floor. A few weeks ago, I asked Philly coach Brett Brown what he thought of this statistical discrepancy, which has only grown more extreme in the time since. “I would put that completely into the weird-stat, I don’t care basket,” Brown said. “Ben Simmons is an all-league defender. Slice it up any way you want. Just look at what he does, and the versatility with which he does it. Guard a five man. Guard a point guard. Go guard the best player. He’s 6-foot-10. He’s a stud of an athlete. Just look at the size of him. So that metric you just said — I’ve never even heard it, to be honest with you — but I dismiss it. I aggressively dismiss it.”
“He’s a free safety at heart,” a league scout told CBS Sports. “We tend to think of great defenders as guys who can lock in on one player and shut him down. Ben can do that, we’ve seen it, but he likes to move around and interrupt different things. I don’t know if he’s the kind of guy you can just say: ‘Go get their best scorer and take him out.’ Although nobody can really do that anymore.” If we have even slightly overrated Simmons’ defense, we have likely done so on account of, first and foremost, the coverage Joel Embiid supplies. Philly’s fully healthy starting lineup, for instance, has a suffocating 97.1 defensive rating over 240 total minutes, per NBA.com. But you take Embiid off the floor, and the defensive rating for all lineups that include Simmons falls to 109.5, per Cleaning the Glass.
According to Cleaning the Glass, the Celtics have outperformed their location-effective field goal percentage 56.8 percent to 53.5 percent over their last seven games after being neutral in that category all season. Even with regression seemingly inevitable, Boston’s best trios have been silly during this streak. That’s one of the reasons why Stevens likes this team even though he believes health issues have kept him from giving it a proper evaluation. “I don’t think we’ve had enough time to truly evaluate the group as a whole, especially as you move into a week like this, right, where you have the trade deadline coming and everything else,” Stevens said. “But we feel good about our team when we haven’t had everybody and we feel better about our team when we have.”
So with all that rigmarole out of the way, let’s go to the leaderboard. We’ll look at points created per 36 minutes to account for differences in playing time (along with a minimum of 750 total minutes played). This distinction is important because the Bucks’ frequent blowouts mean Giannis doesn’t play nearly as much as other stars. Among qualified players, Harden ranks first in minutes per game while Giannis ranks 68th; the Rockets star enjoys 21 percent more playing time in which to collect stats than his Bucks counterpart. And, voila, that adjustment moves Giannis to first place in our humble little stat.
Sabonis is second in screen assist points but also contributes to double-digit points with his passing, so he ranks no. 1 even as LeBron holds a healthy lead for the actually recognized assists crown. The Pacers big man won’t enter the MVP discussion, and nor should he, given the absurd talent and statistics of the top quartet. But he’s the fulcrum the Indiana offense rests on, and he adds value in a comprehensive way beyond the basic box score. One wonders whether he’ll maintain this level of productivity once Victor Oladipo returns from injury.
By almost any measure, the Bucks are the best team in the NBA, yet they still crush opponents when Antetokounmpo is on the bench, posting a net rating of plus-7.9. That would still rank No. 1 overall in the NBA. Make no mistake, that’s an impressive mark for Milwaukee’s title chances. And, no, it’s not Giannis’ fault his team is awesome when he sits. But in a weird way it could hurt Antetokounmpo’s MVP case — and it’s vastly different from what’s happening in Lakerland. Even though they have Anthony Davis, the Lakers without James are a negative team. They post a net rating of -1.6, which would rank 18th in the NBA — in between the Detroit Pistons and the Chicago Bulls. Folks, the Pistons and Bulls aren’t very good.
According to an NBA spokesperson, beginning with the 2017-18 season, the league added a stats auditor to watch games at the replay center in Secaucus, New Jersey. Using a DVR-style device to review plays during breaks in action, this individual can serve as a fallback option in the case of a misawarded number in the box score. All decisions are made by the in-arena stats crews, the NBA official stressed, but the auditor can work in conjunction with them and recommend certain changes as the game speeds by. The more eyes, the official said, the better.
Overall usage of zone defenses — schemes that call for defenders to guard a specific area of space rather than an opposing player — is up 50 percent from last season, according to Synergy Sports data. The increase per possession is 10.8 percent, according to data provided to N.B.A. teams by Second Spectrum. A leaguewide average of 2.3 zone possessions per game remains modest, but the rise Carlisle referenced is more tangibly reflected by the number of teams regarded as regular zone practitioners.
With Giannis off the floor, the Bucks’ are posting a net rating of plus-6.8 points per 100 possessions, which would rank fifth in the NBA overall. Their defensive efficiency of 103.3 when Antetokounmpo sits would be No. 1 in the league overall. The fact that Milwaukee thrives when Giannis is resting means Milwaukee can seek out more opportunities to save his legs. It also gives the Bucks a rare edge across the league.
On the season, Young is shooting 48.9% on non-restricted area paint shots and is fifth in the league with 2.9 makes per game. He is shooting 47% on floaters this season on the most attempts in the league (116), per Second Spectrum tracking. It’s a give and take: Analytically inclined defensive schemes encourage you to take them, but Young is proficient at making them. As Garnett says, it’s about discretion. “Just because that play is always going to be available for you,” Garnett says, “doesn’t always mean it’s the right one all the time.”
After having the best interior defense a season ago, the Bucks have simply improved to the best interior defense of the modern era through 36 games. As good as Lopez was last year, he has only improved this year, leading the league in Block%, along with a career best 1.3% Steal rate while also easily posting the best combination of Contest % and FG% Allowed against those contests among big minute players.
The ensuing discussion lead all the usual places about why the 3-pointer is bad, but there was also a somewhat new argument to me: people don’t like the three because they don’t like watching all the missed jumpers that result from the increase in 3s. Which is an odd reaction, because it’s not true. As I’ve talked about in exhaustive (and exhausting) detail, the shots replaced by the additional 3-point attempts have been long two pointers, and primarily assisted long 2s. Catch-and-shoot 19 footers and the like. And those shots have been replaced because despite being worth a point less, the NBA as a whole doesn’t make them much more frequently than it does the slightly longer 3-pointer, and this trend has held fairly steady over time
Storyline: Three-Point Line Debate
In 2003-04, the NBA was 4.2 percent more accurate on 2-point jumpers (defined as shots from 15 feet and out) than it was on 3-pointers. So far this season, that gap is 5.0 percent. Meanwhile, over that time, the proportion of total shot attempts from 15 feet and out has risen all the way from 46.1 percent of shot attempts in 2003-04 to 47.8 percent this season, with this number actually peaking in the 2007-08 season at 49.1 percent. Now, the increase in pace of play across this time period – there are about 20 more shots per game now than in 2003-04, means that the total number of jumpers taken per game has risen from around 73.5 per game in 2003-04 to just over 85 per game this year.
But there’s a catch, and it’s one that will almost certainly be a factor come postseason: The Sixers make a decent percentage of the threes they take, but they aren’t really taking that many. They rank in the bottom five in the share of their shots that come from behind the arc, and teams, essentially daring Simmons and other Philly players to shoot, have often used zone against them. In fact, the 191 possessions the Sixers’ offense has played against a zone defense this season are almost 60 more than the next closest team, according to Second Spectrum — a pretty firm indication that teams don’t fear Philadelphia’s ability to hurt them from the outside. That’s what made the win over the Bucks so intriguing: With how the Sixers defend, they can blow great teams out when they shoot well from outside and take care of the ball for a change.
Yet some of the game’s greatest big men push back on the notion that there’s no longer a place for low-post scoring. “You ever heard of the term cooking the books? That’s what [big-man analytics] are to me,” Hall of Famer Shaquille O’Neal says. “‘Oh, our company is doing great. We’re doing this. And we’re doing that,’ when actually it’s a bulls— company. “My contention is that if a big man comes in and wants to dominate, he can dominate — and easily. Because we have shifted away from physicality, teams don’t know how to play [that style] anymore.
“We need to be able to do everything on the basketball court,” Embiid says. “That’s what I’ve been trying to do, to kind of change the narrative around big men.” Brown insists he hasn’t abandoned using Embiid in the post and claims his goal is for him to have at least 15 touches a game from there. That’s a fine strategy for the Sixers, but the 29 other teams do not have a Joel Embiid at their disposal. “The numbers don’t support their skill set,” one NBA general manager says about big men who live primarily in the post. “They’re DOA — dead on arrival.”
Kuzma can provide some degree of playmaking along with his scoring chops. He’s the only player on the Lakers aside from their stars who’s a dual threat, but his assists are way down from last year — 1.7 assists per 100 possessions vs. 3.6 last season — as Davis and Rajon Rondo have dominated the ball in the second-unit lineups that Kuzma often plays in. His ability to score without occupying the ball is a tremendous asset for a superstar-driven team, but the Lakers have taken that to an unproductive extreme as Kuzma can go several minutes without a touch. That makes it very difficult for a scorer to establish a rhythm.
On Saturday night, Harden dropped 60 points on the Atlanta Hawks in less than three quarters of play. It was another virtuoso performance by the world’s greatest offensive basketball player. Through 19 games, Harden is averaging an incredible 38.9 points per game and, barring injury, he’s on pace to win his third consecutive scoring title, something only MJ and Kevin Durant have done in the 3-point era. But the most stunning thing about Harden isn’t his numbers — it’s his style. He’s a rarity in pro basketball, regularly inventing new fundamentals. We haven’t seen scoring numbers this big since a 23-year-old Jordan put up 37.1 PPG. Before that, the only comparison was Wilt Chamberlain’s prime in the early 1960s. And Harden thrives much like Chamberlain did — in the kinds of isolated one-on-one matchups that were supposed to be dead by now.
Harden’s isolation volume and efficiency are both tops in the league. Over 50% of Harden’s points stem from isos. None of the league’s 11 other 25 PPG scorers are even approaching 30%. While most other teams in the NBA frown upon hero ball as an inefficient strategy reserved only for necessary moments, the Rockets have reached the opposite conclusion simply because they have Harden. He is an offense unto himself, and the numbers are startling.
Many observers, including myself from time to time, have derided the monotonous aesthetic of Rockets basketball. Critics argue there’s too much dribbling and not enough passing. The whole thing is gimmicky, predictable and tiresome. Rockets general manager Daryl Morey sees some hypocrisy there. “Nobody was mad when Kareem was getting it dozens of times per game,” Morey told ESPN.
Despite awesome scoring totals, Harden’s percentages from the field aren’t as dominant as other great scorers. He converts his shots at average rates, but that’s deceptive because he hedges on efficiency by only shooting in the best spots. And he achieves a massive subsidy by getting to the line more than anyone else in the NBA. Harden is the savviest foul-hunting guard this league has ever seen. He’s led the NBA in made free throws in each of the past five seasons, but he’s taking it to new heights this season. The dude is going to the line 14.4 times per game, and over 12 of his nearly 39 points per night are coming at the charity stripe.