Did you know that Spurs PG Tony Parker led the NBA in drives this postseason? Or that Tim Duncan actually has a higher Rim Protection Percentage than Serge Ibaka? How about that LeBron James, had he chosen to take the shot at the end of the Eastern Conference Finals Game 5 against the Pacers instead of passing it out to Chris Bosh for a corner three, had a 70% chance of making it and tying the game (a full 27% above the NBA average)? Brian Kopp did.
People tend to think the biggest change to ever happen to the league was the adaptation of the 3-point line in 1979 or the surge in national interest from the Larry Bird vs. Magic Johnson rivalry in the 80's or the emergence of Michael Jordan and responsibly-sized shorts in the 90's, but to me they're all wrong.
To me the biggest change to ever hit the NBA probably happened three years ago when five teams decided to install SportVU player tracking cameras in their arenas so STATS LLC Senior VP Brian Kopp and his team of analytics geeks could quantify the game in a ways few ever even imagined.
A little background: SportVU (pronounced sportview) is a six-camera system for NBA games where statisticians can can collect player and ball data at a rate of 25 times per second. The ability for cameras to track players without them wearing special equipment was derived from a technology originally intended for the Israeli military, before scientist Miky Tamir used it to start SportVU as a way to track soccer matches. STATS later purchased SportVU in 2008. In the 2011-2012 NBA season SportVU worked with five teams. In 2012-2013 that number rose to 15. This 2013-2014 season was the first to feature SportVU cameras in all 29 NBA arenas.
Right now on NBA.com you can look up fun facts like exactly how many miles Dwyane Wade ran per game this postseason (2.3) and who leads the playoffs in pull-up shot attempts (surprise surprise Russell Westbrook), but the thing that makes this sports tech so revolutionary is its gargantuan potential. These cameras have the potential to change everything about pro sports as we know it from staff halftime routines to diet suggestions based on player exertion levels to defensive assignments based on GHOST DEFENDER PROJECTIONS.
This shit is intricate and insanely deep, so what better than for us to chop it up with STATS' SVP as we sit mere hours away from the start of the NBA Finals?
COMPLEX: Are there any general trends you saw in the data pertaining to the Spurs and Heat that you think would be interesting to laymen who may not know a lot about the SportVU system?
Brian Kopp: Good question. I don't always get the chance to dive in personally into the actual output but there were a couple stats that came out after the Spurs/Thunder series. The Spurs just pass the ball so well. I know that sounds like a very obvious statement but consider that the Spurs averaged an extra 60 or 70 passes per game than the Thunder did this season. They have more players that average 40 touches per game than I think any other team in the NBA, so they spread the ball around a lot. Now not every team that passes the ball a lot is efficient, there are some teams who pass the ball around a lot because there's nobody who can score. But for these guys passing the ball a lot is part of their game and it creates better opportunities.
That's not necessarily what we saw with the Heat. There was one stat that sounds simple but it's based on very complicated algorithms where it's a drive to the basket. In the last series Paul George in Game 4 was complaining about the Heat shooting more free throws saying, "You can't say we don't attack the rim more than the Heat does." Well actually the Heat did a LOT more of that than the Pacers did. And when LeBron drives to the basket and gets to the rim he's just so effective. He shoots over 70% on drives to the basket even when there are one or two players defending the rim. Doesn't matter if both players are 6'10" or taller. He's just really good at finishing the ball in the half court. We can also look at those drives and ask how they start, whether it be an isolation or a pick-and-roll action to start the drive, and LeBron 70% of his drives come from isolation. He doesn't need a screen to get him going.
Those are the things that stand out. The Spurs are really good at spreading you out and passing the ball, and with the Heat the data supports what your eyes see: when LeBron puts his head down and goes to the rim he's a tremendous finisher and hard to stop.
C: What stats do you guys work with that you feel should be common knowledge and right along points and assists in the box score?
BK: One of the most flawed statistics that we use in basketball is assists. There's just no context to it. It's not capturing whether it was a good pass or not, it's not capturing what was the role of the passer vs. the shooter, so we're starting to look at things like Free Throw Assists where a guy got fouled. You should still get credit for that. What we Assist Chances or Assist Opportunities where he made the pass and it would have counted as an assist had me made the shot. It's similar to Rebounds vs. Rebound Chances. What kind of passes are you getting to your teammates? Are you getting them shots at the rim and open threes or are you giving them long two-pointers with 2 seconds left on the shot clock?
There are guys like John Wall who has a lot of assists but a lower conversion on those assists because his teammates aren't always converting. This may be a bit old from earlier in the year but I remember when Chris Paul passed to one of his teammates and they had a shot within 5 feet of the basket, they shot 90%. If you think that's a lot of Blake Griffin alley-oops and DeAndre Jordan alley-oops, but you gotta remember the league average is around 60%.
There are always some that surprise you, too. The Field Goal % when you're the passer, one of the guys who always ranked high the last couple years is Carmelo Anthony. His teammates convert at a high percentage when he passes it to them. You can argue he doesn't do it enough, but when he does his guys do make the basket.
C: Any other insights on this year's Finals?
BK: I think this match up is going to come down to health. What's Tony Parker's ankle gonna be like? He led the league in drives to the basket and so that was a big part of their offense-- him getting to the rim. Not just for himself but for the team. You wonder if somebody else can fill that role. Dwyane Wade's been playing very well-- can he keep that up? Bosh has played a lot of minutes this year, LeBron has played a lot of minutes this year...they're freak athletes but when you go The Finals four years in a row those are long seasons.
That's something we actually track -- in addition to just speed and distance -- we actually track individual player loads and outputs. That's more for trainers so we can start to see different trends with people and players on both sides. If this thing goes 6-7 games I think that's what it's going to boil down to: who has enough left in the tank?
C: Curious to see if Tim Duncan's age shows up in the data. I know you didn't have the cameras set up 10 years ago to compare but even from last year to this year I wonder if there was a significant dip.
I'll tell you this story. Two years ago when the Spurs lost to the Thunder in the Conference Finals they won the first two games and then they lost the next four. The Spurs and Thunder have been clients with us for a couple years they were some of the first teams (along with Houston, Dallas, Golden State, Boston, Milwaukee, and Minnesota) and so we had data for them and...Duncan fell off a cliff. He was a downward spiral in terms of his output; his body started to break down. The Spurs know that and have taken some actions to rest him throughout the season but I'm sure their training staff are doing things none of us see to make sure they have something left in the tank. I'm sure they're working on spreading the load but even in that Game 6 this year against the Thunder he had more minutes than he had in all but three games this season. Definitely looked like he was tired. It seems like he's stronger than he was the past couple years, though, because of the steps the Spurs are taking to make sure he has something left.
C: You'd mentioned some of the teams who were early adaptors, are there any specific coaches or players who have been more accepting of this technology or ones who have been more hesitant?
BK: Yeah, absolutely, but I'm much more happy to name the teams who were open versus the teams who are hesitant. It's mostly the teams you'd expect that are highly analytical, but the thing is a lot of the guys from those teams have gone on and now the tree is growing. Sam Hinkie is the GM of the Sixers now and was one of our first clients in Houston. Dennis Lindsey is over in Utah and he worked with us in San Antonio.
One of the most impressive teams in using the data is the Washington Wizards. They have one of the smartest guys working on it, a group that really received it and integrated it throughout the organization, its been really impressive to work with them. Really using the data in unique and interesting ways.
C: Maybe that's why they surprised some people in the playoffs this year, too.
BK: Obviously they got some nice young guys but they use the match ups from the data very well. One of the most effective pick-and-roll match ups is Wall and Gortat or Wall and Nenê. Those two paired up are some of the most effective and we can measure that stuff and they're taking actions based on it. It's interesting to see.
C: Do you ever foresee a future where fans can look up the deep data accessible to your team in real-time during games?
BK: Yeah, the NBA wants to do that. The NBA is already using our stuff on NBA.com and in their previews. Stuff like "who guarded who best on each of the teams" is a new algorithm we're working on and we showed the NBA and they told us to write some stories. So if you look now my team actually writes those and they just put it online with some pictures. They want more and more of this stuff to get out into the public domain and we're all for that. The goal here isn't to sell some analytics stuff to teams, honestly the goal is to become part of the language of basketball.
So when people talk about basketball they're not just talking about rebounds but Rebound Chances. Not just talking about assists but are talking about Assist Opportunities. That's what we want people to be talking about.
What I always said to my team is you shouldn't have to explain it for 30 seconds. It should be something that when you say it or explain it briefly that people get it. So something like Protecting the Rim you don't have to think about that. There's some more detailed stuff where we have to explain it, but we want people to visualize these ideas right away.
C: Have to ask this because I'm also a football fan -- is the NFL late on this? Are they more hesitant to it? Is there just more red tape to go through to get this partnership going? Because I've never understood, especially with today's technology, why the most popular league in America still relies on referees eyeballing first downs. There should be a sensor in the ball and sensors along the sidelines telling me the exact point where the runner was down. When can we expect to see this stuff?
BK: When we acquired SportVU in 2008 they had a soccer technology, so my job was to take it to other sports. The first one we worked on was football with the NFL. We went to Lambeau, set up the cameras, built out the system to show them it works, and it all broke down because they're the NFL and working out a business arrangement with them is very difficult. They didn't allow us to go and work individual with teams like the NBA, it's an all-or-nothing proposition with the NFL. They're gonna control everything that happens. I know the NFL has been looking at solutions like putting RF tags in the should pads, and the good news is they have places on the uniform to put things like tags.
The goal here isn't to sell some analytics stuff to teams, honestly the goal is to become part of the language of basketball.
Talking about sensors in the ball, I actually have a ball in my office that has as chip inside. You can tell. We gotta figure out how to put something in the ball without altering the performance of that ball. Even if you get that worked out there's going to be testing and the first time a kicker misses a field goal it's gonna be the chip's fault -- there's gonna be issues they have to work through. The good news is they've been testing technologies the last couple years and those companies have approached us. No matter how you collect the data, you still gotta make sense of it. Even if you have a chip in your shoulder pad, that just tells you where you are, it doesn't put it in context of if you're catching the ball, if you're breaking a tackle, what kind of formation, etc. There's a lot of other contextual information and that's what we do everyday.
I do think it's coming sooner than later in the NFL, but I don't think you're gonna see as much transparency with it as you do with the NBA. They're not going to be as open with making it public like the NBA did. If you think about even MLB they've been doing PITCHf/x and a couple other things for a while in baseball and while they release that data to the public, they released the raw data. That's great for academic researchers and stuff, but if you look at the baseball box score it hasn't changed. You look at the NBA and the box score they're putting player tracking data into the box score. They're integrating this and putting it into the public. Other sports aren't. Other sports are still trying to figure out which technology they should use and they're a couple years behind.
