Sifting through SportVU: Using Bucks’ data to find the strengths and weaknesses of a team under construction

John Henson and Milwaukee's other bigs are protecting the rim at an elite rate (photo credit: Brian Babineau, NBAE via Getty Images) .
John Henson and Milwaukee’s bigs are protecting the rim at an elite rate (photo credit: Brian Babineau, NBAE via Getty Images).

Two months ago, we took our first glance at the new SportVU player tracking data provided by NBA.com/Stats. Using an extremely small sample size of six games, we saw that rookie Nate Wolters was passing well, limiting his turnovers and struggling to score on the drive. Zaza Pachulia was getting a star-level number of touches at the elbow. Gary Neal was making jumpers at an unsustainable clip. Ekpe Udoh wasn’t rebounding.

As one could imagine, some of those trends persist while others do not. Here’s a riveting sneak peek: Udoh still doesn’t rebound.

If you need a refresher on SportVU, this is the primer from our preceding piece:  “Per its website, SportVU’s data revolve around four essential measures: speed, distance, player separation and ball possession. These are collected by six cameras — mounted to each arena’s catwalks — that track player and ball movements 25 times every second, according to NBA.com. In short, the system collects and synthesizes endless advanced statistical measures. The Bucks are no strangers to SportVU; as early as 2011, they were one of the first 10 NBA teams to implement the system. Now, all 30 teams utilize the service following a summer agreement between STATS and the NBA. The degree to which teams actually implement the data into game preparation or roster evaluation is largely unknown.”

(Note: An arbitrary filter of at least 10 games played was used to weed out extraneous players. The data do not include Tuesday night’s game against the Warriors. For some players, the database wasn’t updated to reflect Saturday night’s game against the Suns, either.)

Revisiting Old Trends

1. Wolters – hockey assists per game

NBA/SportVU definition: “Quantity of passes made by a player to the player who earned an assist on a made shot. Assister must make a pass within [two] seconds and [one] dribble for passer to earn a secondary assist.”

  • Through Milwaukee’s first six games: 2.5 (1st in the NBA [t])
  • Through Milwaukee’s first 33 games: 1.2 (30th in the NBA [t])

Wolters’ dip in secondary assists — or hockey assists — per game is largely a result of fewer minutes. While he was averaging 32.1 minutes through Milwaukee’s first six games, when fellow point guards Brandon Knight and Luke Ridnour were injured, his average is now at the 20-minute mark (which doesn’t account for some questionable DNP-CDs). The current category leader, Chris Paul, is averaging 2.2 hockey assists in 34.9 minutes per game.

Interestingly, Knight is tied with Wolters at 1.2 hockey assists per game.

2. Wolters — touches per game and time of possession 

NBA/SportVU “touches per game” definition: “The number of times a player touches and possesses the ball”

  • Through six games: 76.2 (19th) 
  • Through 33 games: 50.0 (101st)

NBA/SportVU “time of possession” definition: “The total amount of time a player possesses the ball per game in minutes.”

  • Through six games: 6.0 minutes (12th [t]) 
  • Through 33 games: 4.0 minutes (43rd)

Obviously, the substantial decreases in time of possession and touches per game reflect the point guard’s minute reduction.

Unsurprisingly, Knight now leads the Bucks in touches per game (77.2) and time of possession (5.9 minutes) — ranking among the league’s top 25 in both categories. By comparison, John Wall leads the NBA with 99.4 touches per game and a time of possession of 8.2 minutes.

Caron Butler and Ersan Ilyasova are the only other Milwaukee players to rank within the top 100 in touches per game. Butler is somehow averaging 51.8 touches in 27.2 minutes per game. In other words, he is getting the ball nearly twice each minute on the court. Since he is shooting 36.4 percent from the floor and only averaging 2.0 assists per game, some of those touches would be better served elsewhere.

3. Wolters — shooting percentage on drives

NBA/SportVU definition: “The field goal percentage of a player on his drives to the basket.”

  • Through six games: 21.4 percent
  • Through 33 games: 43.5 percent

While shooting 43.5 percent on drives isn’t an ideal marker — ranking well outside of the NBA’s top 100 — it nearly doubles Wolters’ earlier percentage. He is now surrounded by a few of his teammates within this category: Neal (42.1 percent), Ilyasova (45.5 percent) and Giannis Antetokounmpo (47.1 percent). 

Two prominent offensive contributors, O.J. Mayo and Khris Middleton, are struggling with this aspect of their game. They are shooting 36.0 and 32.4 percent on drives, respectively. With a lack of playmakers, the Bucks have to count on Mayo and Middleton to create a decent amount of their own shots, so this is a troubling trend. More often than not, they have settled for long jump shots — Mayo and Middleton are both averaging less than three drives per game. By comparison, LeBron James and James Harden drive over six times per game.

Ridnour, who is shooting just 39.4 percent from the floor, is actually excelling on the drive. He is shooting 50.0 percent on 42 total drives.

4. Pachulia — elbow touches per game

NBA/SportVU definition: “Touches that originate within the 5 foot radius nearing the edge of the lane and free throw line, inside the 3-point line.”

  • Through six games: 10.0 (7th)  
  • Through 33 games: 9.3 (6th)

Pachulia, who is recovering from a fractured right foot, has only played in 16 games this season, so this is still a relatively small sample size. Only Marc Gasol (18.1), Blake Griffin (12.9), Kevin Love (11.8), Pau Gasol (10.7) and LaMarcus Aldridge (9.7) are averaging more touches at the elbow than Pachulia.

Ilyasova is second on the Bucks in elbow touches per game (5.5), which makes sense with his immunity to the post and ability to stretch the floor (more on this later).

5. Pachulia — opponent field goals made at the rim per game

NBA/SportVU definition: “Field Goals Made by an opposing player or team at the rim while being defended.”

  • Through six games: 3.4
  • Through 33 games: 2.8

Pachulia improved slightly at protecting the rim prior to getting injured, though he’s still allowing opponents to shoot 52.4 percent at the rim.

6. Udoh — rebounding chances

NBA/SportVU definition: “Measures the number of rebounds a player recovers compared to the number of rebounding chances.”

  • Through six games: 48.1 percent (318th out of 397 eligible players)
  • Through 33 games: 45.2 percent (385th out of 403 eligible players)

Udoh still can’t rebound.

As a team, the Bucks continue to struggle on the boards. According to Basketball Reference, they rank dead last in defensive rebounding percentage (71.0) and 19th in offensive rebounding percentage (24.8).

Middleton leads the Bucks in percentage of rebounds per chance (there still has to be a better way to word this) at 62.8 percent, followed by Henson at 62.3 percent.

Ilyasova, who is usually quite active on the glass, is converting on just 53.3 percent of his rebounding chances this season. However, that may be a result of his lingering ankle issues.

7. Gary Neal — catch-and-shoot opportunities

NBA.com/SportVU definition: “The percentage of [catch-and-shoot] field goals that a player makes.”

  • Through six games: 66.7 percent (2nd [t]) 
  • Through 33 games: 45.5 percent (63rd [t])

Regression to the mean is cruel, but Neal is still shooting 51.4 percent on catch-and-shoot threes, which ranks in the league’s top 25.

(Nevertheless, the Bucks are aggressively shopping him around the league, according to ESPN’s Marc Stein.)

While Middleton is struggling on the drive, he remains an extremely consistent spot-up shooter. He is shooting 46.0 percent on catch-and-shoot field goals and 46.1 percent on catch-and-shoot threes. That combination gives him an impressive 62.3 catch-and-shoot effective field goal percentage.

New Trends

1. Bucks’ frontcourt — opponent field goal percentage at the rim 

NBA/SportVU definition: “Field goal percentage of an opposing player’s or team’s shots at the rim while being defended.”

  • Henson: 41.7 percent 
  • Sanders: 38.3 percent
  • Raduljica: 35.0 percent

Henson ranks among the NBA’s elite at rim protection. He’s allowing opponents to shoot just 41.7 percent at the rim, a rate that is similar to the likes of Roy Hibbert (40.9) and Omer Asik (40.7) and better than Andrew Bogut (43.9) and Dwight Howard (48.3).

Equally impressive is Larry Sanders (small sample size alert), who’s allowing opponents to shoot just 38.3 percent at the rim — and still thinks he can improve.

Now, the key question is this: Can Henson and Sanders — two big men who reject, alter and prevent shots at the rim — play next to each other on offense? Injuries have delayed the answer, but the Bucks need to find out sooner rather than later.

Contrary to the eye test, Miroslav Raduljica has been extremely effective at protecting the rim, as well. Despite his slow foot speed and inability to consistently alter or block shots, he’s limiting opponents to (relatively small sample size alert) 35.0 percent at the rim, which is the best rate among Milwaukee’s frontcourt players. (SportVU data are quite kind to Raduljica, as highlighted in Alex Boeder’s latest piece on Bucks.com.)

Although his physique differs a bit from Raduljica’s, Antetokounmpo is using his extreme length to limit opponents to 41.7 percent at the rim on 3.0 attempts per game. Udoh and Ilyasova are allowing opponents to shoot 46.2 percent and 50.0 percent at the rim, respectively.

2. Wolters — catch-and-shoot opportunities

NBA.com/SportVU definition: “The percentage of [catch-and-shoot] field goals that a player makes.”

  • Wolters: 17.6 percent

Earlier in the season, efficiently scoring on drives appeared to be Wolters’ NBA weakness. Now, it’s jump shooting.

At this point, he’s extremely tentative to shoot wide-open jumpers and often defers to his teammate, even when the shot clock nears expiration. Considering he made 34.2 percent of 462 three-point attempts in college, it’s probably a confidence thing at this point.

Not-so-shockingly, former Bucks forward Luc Mbah a Moute has the worst catch-and-shoot percentage — 8.3 percent — among players with at least one catch-and-shoot field goal.

3. Henson and Ilyasova — close touches per game

NBA/SportVU definition: “Touches that originate within 12 feet of the basket, excluding drives.”

  • Henson: 5.6 (16th [t])
  • Ilyasova: 1.6 (117th [t])

At times, statistics shed light on something previously undetected. Other times, they reaffirm what is already quite apparent. The latter scenario is the case with the Bucks and close touches.

John Henson and Raduljica are the only two players on Milwaukee’s roster who can post up with any consistent success — and the Bucks are ensuring they get their looks in the right spot. Henson, who features a smooth lefty hook, gets 5.6 close touches per game, while Raduljica, who creates easy opportunities with his massive frame, is averaging 2.0 close touches in just 10.5 minutes per game.

Ilyasova is different story. While he gets plenty of looks at the elbow, the Bucks don’t bother posting him up at the block with any regularity. That’s a reasonable strategy, given Ilyasova’s preference to fadeaway over post defenders rather than go up strong. Tellingly, Antetokoumpo — who struggles to physically match up with any NBA player — averages more close touches per game (1.7) than Ilyasova.

4. Luke Ridnour — free throw assists per game 

NBA/SportVU definition: “Quantity of passes made by a player to a recipient who was fouled, missed the shot if shooting, and made at least 1 free throw. Recipient must be fouled within 2 seconds and 1 dribble for passer to earn a FTA assist.”

  • Ridnour: 0.7 (12th [t])

This metric isn’t very significant — every NBA player is averaging between 0.0 and 1.1 free throw assists per game — but it’s pretty interesting. It seems unfair that players aren’t credited with a standard assist even when they directly setup teammates for two foul shots. Thus, tacking on free throw assists to players’ standard assist numbers may lead to a more accurate — though still not perfect — gauge of passing skill. It’s probably no coincidence that Minnesota’s Ricky Rubio, a renowned passer, leads the league in free throw assists per game. Wolters and Knight are both averaging 0.4 per game.

Categories: Stats and Stuff

Twitter: @pdschmitt1

17 Comments

  1. Unrelated to SportVU but related to Bucks’ stats, is there a tool/website where one can look up the total +/- stats through the season and compare players? I am curious to see how everyone compares on the team.

      • Thanks, Ross. That site definitely has the easiest interface for finding +/-. Unfortunately, it doesn’t update regularly. It takes some navigating, but +/- information is available through Basketball Reference and NBA.com/Stats, too.

  2. If the Bucks have any doubt how valuable John Henson is … look at how unbelievably bad the team is without him.

    At least when Henson is on the floor they have some inside post up presence, points, block shots, rebounds and shooting efficiency. A player shooting 52-53% every night is something that is unheard of for the Bucks the past 4-5 games. Or more accurately, the entire season.

    Even with Larry Sanders back he isn’t even remotely the impact player Henson is.

    • I agree — even though they were pretty bad with him, too. I think the main concern now is whether Henson is truly a power forward. He plays a lot like a center, which might complicate future plans.

  3. I like Nate, and hope he makes it with the Bucks, but don’t understand all the attention he seems to get. I’m still shocked that he didn’t develop his outside shot more in the offseason, especially after summer league. It does seem like he has the form to be a good shooter, but it could take awhile. My hunch is that Nate would do better in the D-League gaining experience — and then perhaps bring him up later in the season.
    As it is, it seems that Nate is an unready rookie, Luke is a declining veteran, and Brandon is a shooting guard. Couldn’t we trade a couple of our current players for a more promising alternative at the point position? He doesn’t have to be great, but someone with a little flair and a little savvy, who has a little creativity and a little leadership, and can maybe get our offense running a little more smoothly and energetically.
    Otherwise, it seems hard to make much of the statistics for our guys, because everything seems so disconnected and confused and clunky and clogged. It doesn’t seem like we’re putting our guys in positions to succeed. It’s tough on the fans, too, at least for me.

    • Nate gets a lot of attention because he’s really different. He’s a non-foreign white point guard. Not too many guys like that. I think a lot of people are intrigued by Wolters because he’s probably one of the most average-person guys in NBA history. I imagine him playing lots of Call of Duty in his off time. It’s also incredibly hard to compare him to anyone that’s played in the NBA. I think most people would assume by looking at him that he’s a not very athletic sharp-shooter. He’s never really been a catch and shoot guy, especially at South Dakota State. He has incredible ball handling skills and can change speeds very quickly. Google “Nate Wolters” “Deceptively Quick” and you’ll find plenty of results. His jump shot has never been NBA-caliber. He scored with his quickness, getting to the line, and using his crazy-good floater. He was a career .342 3% shooter out of college, so it’s not like he was a Brady Heslip or anything. When he was still in college, I remember plenty of doubts about him ever playing in the NBA because of his defense.

      As far as my comparison goes: I know it’s been discussed here before, but I like to think of him as another Bryce Drew. Sure, Nate kind of stunk in the tourney and didn’t hit a memorable buzzer beater, but there’s some obvious comparisons… He was another midwestern white American point guard, extremely normal, good passer, scored a lot, spent all 4 years in college for a small school, and came from the same conference (when Valpo was in the Mid-Conn/Summit). He spent 6 years in the NBA as a backup PG, and I would imagine Nate has roughly the same trajectory. Disclosure, I obviously went to the school, but I still think it’s a good comparison.

      • It seems like you know a lot more about Nate than I do. I heartily hope he has a really good career with the Bucks.
        One of my concerns is that it seemed like Ish Smith was about three times faster and/or quicker in summer league, and more natural as a point guard. Both are poor shooters, but I think both have the form to be at least good shooters.
        I do wonder if the best thing for Nate might be D-League for at least a little while. From what I’ve read of Nate, he seems like a really good guy, and I wish him well professionally and personally.

      • Bryce Drew was a terrible backup NBA PG, by every metric. His career PER was 9. He was short and slow. He played 78 minutes…in a SEASON…TWICE.

        Get beyond this ridiculous comparison, valpo kid.

        And I urge you to look a bit deeper than university school size and skin color to compare NBA players.

        • That implies there’s some sort of necessity in comparing players. There’s not. It was just a dorky aside that’s not a conversation about a trade that’s never going to be made or how much upside a potential draft pick has, which every basketball conversation seems to gravitate towards anyways.

    • I really like your post,…the offense does seem very clunky and often the spacing is horrible. Many of the half court possessions are herky-jerky,with little or no fluidity!! Seems that are best chance to score is in transition,without that fast break points…we will struggle to win ANY game.

      • Thanks, dipsydoo toyou, for the very much appreciated positivity. If we could somehow get better direction and distribution from the point position — to go along with Larry, John, Giannis, Brandon, Khris, Miroslav, maybe Nate, and any veterans with a good attitude — the second half of this season could actually be really fun and exciting and hopeful.
        This would give me something substantial to enjoy this season, and there’s a good chance our record would still be down enough to get a really good draft pick, perhaps a top three. I’d be delighted with a top pick, just don’t want to put an extravagant amount of hope on one very young guy out of college.
        (I’m still not completely giving up on Brandon at the point, but it’d be nice to have another option for this season; if Brandon isn’t suited for point, I’m hopeful for him at the two, that he might be a really good starter, and maybe even a star.)

        • P.S. If we package Gary Neal with another one of our vets and/or one of our second-round picks, maybe we could get a point guard plus that gift card to a sporting goods store.

    • Thanks for the post. I wanted to revisit each trend I previously covered, so that’s the only reason why Wolters was discussed in length.

      I shy away from player comparisons for him. I don’t think there’s an apt one.

  4. Pingback: Daybreak Doppler: What To Do This Weekend? Go Panthers? | PocketDoppler.com

  5. I don’t have my head wrapped around all these uber-nuanced and ‘advanced’ metrics yet.

    But the one that jumps out is #1, the Opponent FG% at the Rim. In this era of ball, where it’s all about getting to the rack and shooting the 3, this seems like a key defensive metric.

    My second favorite is Udoh still can’t rebound. :)

    • You’re exactly right about rim protection. Reputation is also key, because preventing opponents from attacking the basket is even more important than blocking or altering shots. Larry Sanders definitely gained that reputation last season — it will be interesting to see if or when Henson gets there. If you’re interested, Kirk Goldsberry dubbed this as the “Dwight Effect” at the Sloan Conference.