SportVU and the Jazz

December 9th, 2013 | by Ben Dowsett
i (13)

Photo by Melissa Majchrzak/NBAE via Getty Images

While the general public would likely associate the name “Moneyball” with the 2011 film starring Brad Pitt, those who are analytically inclined would think of something entirely different. Spawned near the turn of the millennium, the term is used to identify the budget-friendly system with which Oakland A’s GM Billy Beane changed the way baseball talent is evaluated using advanced, computerized metrics. And while Moneyball itself was only one step along the road, it played a role in popularizing new ways to look at professional sports.

Fast forward just over a decade, and advanced metrics define baseball almost completely. They’ve also begun to play significant roles in most other major team sports, from American football to the highest levels of European soccer. And though many sports (hockey being a good example) still have their share of doubters as to their efficacy and application, the NBA is at a stage where nearly all such skeptics have given way to mounting evidence of their value.

A key player in this, both now and in the future of this stats revolution, is STATS LLC and their SportVU cameras. Many informed readers will already know the basics (a couple more detailed Grantland articles on the subject here and here for those who don’t), so just a quick overview:

Every NBA arena now has cameras that track player and ball movement on a level never before seen. This tracking can be converted into huge amounts of relevant data on everything from simple ideas (rebounding position, offensive spacing) to some truly revolutionary advances, one of which was discussed at length by Zach Lowe in the second link above, but which I think is so amazing that it’s worth mentioning here: a computerized “ghost” defender pioneered by the Toronto Raptors last season to show optimal defensive positioning for every conceivable situation.

It sounds far-fetched, but the Raptors converted the raw data into digitally viewable video diagrams. After conversion, they could break down any play to look like this, with each solid circle representing a real player:

See the transparent circles following the defensive players around the court? Those are the “ghosts.” Using insanely detailed data from hundreds of thousands of plays, the Raptors can show where the defender should be at every second of every defensive possession. They can alter this based on the defensive scheme they want to run, but far more importantly, they can factor, on an ever-changing basis (as more plays are logged each game), the tendencies of every single player into these optimal defensive positions. For instance, the program knows that giving up a long two-pointer to Kendrick Perkins is preferable to an open corner three for Kevin Durant; every variable in this vein is factored in using every play available from the massive database.

Any programming-inclined readers will recognize how truly awe-inspiring this sort of advance is. And even those within the league who are technically challenged can see the potential: with the right tweaks (many of which are likely in the works already), this sort of program will completely change the way basketball is coached and scouted at every level. I have other things to write about, but know that I’d happily go 3,000 words on this alone if Andy would let me – it’s simply amazing.

And while this particular level of uber-detailed tracking isn’t yet available publicly (and isn’t even being used by plenty of teams; the Jazz, among others, haven’t shown much inclination to delve that deeply), this season is the first where certain SportVU information is being made publicly available on NBA.com. So why shouldn’t we get in on all the fun? Let’s have a look at some Jazz trends, both good and bad, that the new cameras have given us access to this season:

• Gordon Hayward had, to the naked eye, been looking pretty tired lately. His offensive efficiency was slipping badly after a strong start to the year, and many including myself hypothesized that he was simply tiring out from the increased workload this year. And what do you know? Turns out Hayward has run a whopping 57.4 miles this season, tops in the NBA. His 2.6 miles per game puts him fourth league-wide, and it’s easy to see how this has had an effect on his overall game. I’m far from the first to point this out, but it provides the perfect example of the value of STATS; where we’d have only usage and minutes numbers to confirm the eye test in the past, there’s now concrete data to back this up (and as the revolution continues, SportVU cameras are projected to have a major role in assisting with injury tendencies and minutes limits as well). Corbin addressed the issue recently, saying it was largely due to injuries and now that the Jazz are getting everyone healthy he hopes to get Hayward back to more comfortable levels.

• Speaking of Hayward, he’s tied for ninth per game in the NBA in “hockey assists” – the pass that leads to an assist. This stat can be both relevant and misleading; secondary assists frequently are the passes that really open up a defense, but it’s impossible to know when this is the case and when it’s just random luck. That said, some other names in the top 10 for this category include John Wall, Chris Paul, Mike Conley and Tony Parker – at the very least, Hayward is in some excellent company.

• Both Derrick Favors and Enes Kanter have immense skill and potential as rebounders, but SportVU helps us differentiate between their rebounding “styles”, if you will. I wrote about Kanter’s high level of proficiency on contested rebounds a couple weeks ago, and this hasn’t changed. He currently sits second of qualified players, grabbing 56.0% of his available contested rebounds, one of only 12 players league-wide over 50%. Favors, on the other hand, grabs only 42% of his contested boards, but is in the league’s top 25 in uncontested rebounds, with 5.5 per game. Once again, the tracking data confirms what the trained eye had already suspected: Kanter excels at grabbing tough, in-traffic rebounds while Favors is clearly the better box-out man. SCH editor Andy Larsen did an excellent breakdown of Kanter’s declining rebounding last season for SLC Dunk that addressed this box-out issue among others; with Favors effectively matching Kanter’s offensive rebounds per-36-minutes (an area considered Kanter’s forte), this could explain some of the reasoning behind the latter’s recent flip-flops in and out of the starting lineup.

• And while Favors has impressed in the above category and several others, one area of concern is something most would consider a given – his rim protection. Long-armed and athletic, the big man has flashed top-10 defensive potential and was widely expected to make a leap toward that territory this year…but some of the data seems to indicate that he hasn’t done so. Of players defending five or more rim attempts per game, Favors is just 42nd in opponent FG% at the rim, allowing opponents to shoot 52.4%. This is worse than guys like Jermaine O’Neal and Mason Plumlee, and compared with the figures of true elite big defenders like Hibbert (38.2% allowed), Bogut (43.0%), and Asik (43.2%), it certainly doesn’t look promising. Still a somewhat small sample size for this sort of data, but something the Jazz would like to see Favors improve on.

• Finally, to end on a positive note: Marvin Williams can shoot the ball! Who knew? Maybe Ty Corbin did, because his newfound embrace of small-ball has freed up Marvin for all sorts of buttery goodness like this:


SportVU tells us that Williams has an effective field-goal percentage (field-goal percentage weighted to include the value of three-pointers) of 61.0% on catch-and-shoot looks like those, good for 12th in the NBA of qualified players. Guys right around him on this list include Klay Thompson, Kevin Martin, and Dirk Nowitzki – Marvin is in the company of some true sharpshooters. Some major credit to Corbin here for, at least temporarily (and before Marvin’s recent injury, which won’t keep him out much longer), revitalizing the career of a much-maligned former lottery pick with some smart systematical changes. I’ve said it before and I’ll say it again: what looks like a miserable season has actually included a lot of these sort of real positive steps for a developing franchise. Looking forward to more fun on the horizon in Jazzland.

Ben Dowsett

Ben Dowsett

Ben Dowsett is a life-long Jazz fan and general sports fanatic based in Salt Lake City. He also writes for Nylon Calculus (Hardwood Paroxysm/Fansided Network), and can be heard on the airwaves for the SCH podcast and appearances with ESPN AM 700. With a strong background in both statistics and on-court fundemantals, he writes primarily as an in-depth strategic analyst. He can be found on Twitter at @Ben_Dowsett.
Ben Dowsett
Ben Dowsett

Latest posts by Ben Dowsett (see all)

  • Utah’s Early Off-Ball Issues
    Analytics
    9
    November 24th, 2014

    Utah’s Early Off-Ball Issues

    As the league reaches the end of November, initial excitement over a new season begins to give way to more of a “daily grind”...Read More

  • To Board Or Not To Board
    Analytics
    2
    October 16th, 2014

    To Board Or Not To Board

    Like giddy teenagers on a fourth date, we’re still just getting to know the 2014-15 version of the Utah Jazz. With a new...Read More

One Comment

  1. cw says:

    I don’t want to be too negative. Gordon Hayward has the ability to be a real useful starter for a good team. But…. if you go back to the start of preseason, some 30 games ago, Hayward has had a streak of about 5 games in, I think early november, and then one good game since. So he’s had about 6 out of 30 decent games. I watch him and see physical limitations. He has a hard time getting to the hoop. He gets into the lane but his momentum dies before he gets to the basket. He also has some clunky “moves” that sometime result in baskets but often result in misses. He also shoots a lot of shots the are WAY off as well as just simply missing. He has always been a streak shooter–with months long streaks, both good and bad–since his freshman year in college. His basketball IQ seems surprisingly low. WHy doesn’t he take the guards that frequently guard him to the post? Why hasn’t he developed a pump fake? The last thing is his temperament. He has been shying away from shooting open shots. I don’t think he has the testosterone levels to take command of games and a team the way some people (his agent) are expecting. I think he is a complementary player. At this point I wouldn’t really bet on any of this changing. It is possible, but not likely. Which is to say, I don’t think what we are seeing form him is the result of running 2.6 miles three times a week.

Leave a Reply