Big data meets big dunks on the cutting ‘edge’ of sports

Sloan Sports Analytics Conference

Walking into the Boston Convention and Exhibition Center, I find myself holding the door for none other than Mark Cuban. The affable owner of the Dallas Mavericks doesn’t have time for a photo. We’re both running late.

Michael LewisWe’re here for the Sloan Sports Analytics Conference, the sports geek’s version of Coachella, a gathering of who’s who at the intersection of sports and technology. What started seven years ago with a few stats geeks meeting in spare MIT classrooms, inspired by works like Michael Lewis’ 2003 sabermetrics showpiece, Moneyball, has blossomed into a mainstream extravaganza presented by ESPN, attracting 800 enterprising students looking to schmooze their way into their dream job or simply shake the hand of NBA coach Stan Van Gundy or famed New York Times political prognosticator Nate Silver. They’re joined by 1900 attendees that include 90 sports teams from the NBA, NFL, MLB, NHL, EPL and MLS. If you’re a sports enthusiast with a predilection for numbers and technology, this is the place to be.

I’m hoping to get a glimmer into the future of sport and I’m not disappointed. Gone are the days of the jocks versus the dweebs. We’ve since graduated high school and the image of the obsolete coach laughing off the Excel-toting poindexter is a tired cliche. Long since accepted, geeks like San Francisco 49ers COO Paraag Maraathe and Houston Rockets general manager Daryl Morey – an MIT alum and chairperson of the conference –have infiltrated the upper echelons of the game, ushering in an information-driven revolution that is reshaping the landscapes of our beloved pastimes. This is the evolution of sport.

At Sloan, sport is treated like you would expect MIT to treat any field of study: as a true science. “Sport is this great laboratory,” Lewis says during the weekend’s kickoff panel, Revenge of the Nerds. It’s about hypothesis and experimentation. You have an idea and you test it.

nate-silver“It’s a good criteria for measuring success,” agrees Silver, who started out playing poker. One of the event’s recurring topics, gambling is the purest manifestation of this ideal. Evaluate, theorize, and commit. “Gambling is a really great way to test that belief,” says Morey while moderating a panel called True Performance & the Science of Randomness.

In sport, as in life, there are no guarantees. We deal in uncertainties and so we owe it to ourselves to understand the kinds of odds we’re dealing with. “Poker players are the best at analyzing probabilities on the fly,” Silver says during the same panel. And if you believe in something, then put your money where your mouth is. Many of those present have done just that. More often than not, they’ve won, tipping the scale in their favor through thoughtful preparation and an instinct for the next big thing. Jeff Ma, the MIT blackjack player who famously “brought down the house,” is a speaker. So is Erik Seidel, winner of eight World Series of Poker bracelets along with Haralabos “Bob” Voulgaris, a professional gambler on the NBA, or “sharp.”

In gambling parlance, it’s all about finding “edge,” any sort of advantage you can back up with hard numbers and facts. It’s only when you’ve got an edge that can you make a wager. Otherwise, you really are just gambling, which is not what those at the Sloan conference do.

sloan sports conference revenge of the nerds

This persistent search for edge is the essence of why we’re here. Biometric analysis provides athletes with the physiological awareness to train better, run faster, last longer, and race smarter. Technological advances allow us to see the invisible, interpret the inexplicable, and measure the unquantifiable. Big Data leads to more informed decisions, innovative coaching techniques, and new kinds of team strategies. This never-ending quest for edge is, as the immortal Charlie Sheen would say, about one thing: “Winning!”

Take basketball, which doesn’t naturally lend itself to analytics. Sure, there’s a ton of data we already measure, like shooting percentages, rebounds, blocks, but these metrics paint a low resolution picture. Stats like points per game or steals, while useful, only tell us the destination while completely ignoring the journey. That might work for game like baseball, but basketball is fluid, dynamic. For proper scientific evaluation, we require context. What that really means is we need the technology to capture it.

As such, the star of the Sloan conference wasn’t a brilliant statistician or a writer or progressive GM or a doctor-cum-trainer. It was a visual tracking system called SportVu, one that brings to the game of basketball a brand new perspective.

Developed by a company called (what else?) STATS, SportVu collects the location data of every player and the ball 25 times per second throughout an entire game. Described as “the holy grail of the NBA,” it achieves this using a series of strategically placed high definition cameras around the stadium. A few years ago, only four teams signed up. Now, the roster includes half the NBA, leaving you to wonder what the other 15 teams are waiting for.


With the resulting “x, y, z” data, you can essentially “see” everything that goes down in a game – like how often a player sprints, the speed at which he accelerates, and when he begins to fatigue. You can assess not only the number of rebounds but also the amount of rebound chances and even the quality of rebounds – if you’re rebounding in traffic or surrounded by teammates. Golden State Warrior baller David Lee might get plenty of boards, but the ones he racks up are relatively easy. According to SportsVU, he struggles when he has to reach.

Sometimes the data confirms what we already know. For example, Lebron James is indeed a certified beast. Or, when players refer to Kevin Love of the Minnesota Timberwolves as a guy who “goes hard,” it’s because he really is doing the most running on the floor.

Speaking of running, while wing players like James and Dwayne Wade are usually the fastest, centers actually sprint the most out of any position. So we can stop calling them lazy when fail to execute the pick-and-roll. In fact, setting up the pick might appear easy, but SportsVU shows that it actually requires tremendous acceleration.

We know that teams average 1.0 points every possession, but 1.2 points when they drive. When James Harden drives, that average increases to 1.45. We’re also privy to the subtleties by showing how many 3-point opportunities he creates for his Houston Rockets teammates when he does charge the basket. Is it then any surprise that Morey’s team drives to the basket more than any other in the NBA? In a few years “the way people play [basketball] will be completely different,” says Morey, during Revenge of the Nerds. These are early days, all over again. SportsVU offers a fresh perspective, one that half the teams have yet to acquire and the other half has barely begun to utilize.

New dimensions of information require pioneer interpreters and the sensei-in-chief of SportVu metrics is Kirk Goldsberry, a visiting scholar at Harvard, who applies his studies in geography – which includes examining access to nutritious foods in urban areas and investigating real-time traffic maps – to the basketball court. His paper, CourtVision: New Visual and Spacial Analytics for the NBA was the research award runner-up at Sloan 2012.

The Dwight Effect A New Ensemble of Interior Defense Analytics for the NBA overall nba shootingGoldsberry downloads the data into million-celled spreadsheets to create visualizations like the one above, which maps out Rajon Rondo’s shots and reveals some of his idiosyncrasies. The Boston Celtics point guard is a poor perimeter shooter, but it he appears to have a sweet spot. He’d be wise to position himself as such when launching from downtown.

But conventional stats almost always favor offense, so this year, Goldsberry teamed up with Eric Weiss to present a research paper titled The Dwight Effect: A New Ensemble of Interior Defense Analytics for the NBA. They want to answer questions like, what happens to the opposing team’s offense when Dwight Howard is in the paint?

The Dwight Effect A New Ensemble of Interior Defense Analytics for the NBA figure 3

Using spatial analysis, Goldsberry and Weiss take into account how close a defender is to a player and his distance from the basket. This produces invaluable metrics like “invisible blocks,” as they call it. With Lee in the on D, teams score almost at will, as per the above chart. And while Larry Sanders’ aptitude for preventing close shots won’t boost your Yahoo Fantasy stats (yet), coaches now have a statistical basis for appreciating defensive effectiveness of the Milwaukee Bucks power forward.

This kind of powerful analysis will inevitably seep into other sports. During Football Analytics panel, former Kansas City Chiefs GM Scott Pioli laments the limited value of the NFL Combine. “I don’t care about Manti T’eo’s forty time,” he says, pointing out that “in-game speed,” which requires vision, awareness, and intelligence on top of pure athleticism, is all that matters. Pioli suggests that T’eo is a great player whose poor Combine results don’t do his on-field capabilities justice. Soon, thanks to visual tracking systems like SportVU, we’ll know definitively.

Soccer is another sport desperately in need of a technological boost, and I’m not referring to goal line technology. At the Soccer Analytics panel, when asked if they could name a single statistical metric, the panelists were at a loss. “There is not one metric for soccer,” concedes Chris Anderson, a Cornell professor and semi-professional player who authored the soccer analytics tome, The Numbers Game: Why Everything You Know About Football Is Wrong.

the hidden foundation of field vision on english premier league soccer figure 2In search of further insight in the beautiful game, I make my way towards The Hidden Foundation of Field Vision in English Premier League Soccer Players, which received rave reviews on the Twittersphere (it’s a repeat from the day before). As a rare American “football” fan, I find myself wading against the current as a crowd of attendees flood the opposite way. What’s all the fuss about? Betting Analytics. I call a last second audible and follow the money.

Congratulating myself on a clutch play, I bunker in for one of the more entertaining panels of the conference. If the majority of the two days was a mashup of Charlie Rose and 60 Minutes, what I was about to watch was the Sloan edition of Jersey Shore.

In one corner, we had Voulgaris, the sharp, whose vicious tongue and outsized ego befit a guy that puts food on the table by being more right than you. On the other end is Matthew Holt a director at Cantor Gaming, the largest bookmaker in the U.S. If I told you to imagine America’s number one bookie, you wouldn’t be very far off. Overweight with a slight lisp, Holt may as well have greased his hair back with mayonnaise.

The tension is immediate and Voulgaris isn’t pulling punches. “Some of the dumbest people I know are bookmakers,” he says, coming in swinging. He adds that “bookmaking is a parasitic business” for “avoiding bets” and preying on the weak. “They really don’t know what the real numbers are.” By now, people are hollering from the stands.

Bob has a reason to be testy. He’s had a terrible year, as Holt points out. There’s much less “edge” these days, shared among an increasing number of sharps. To his credit, Holt takes Bob’s condescending jabs like a gentleman, despite his appearance, and admits that “it’s been an awkward season so far.” Moderator Ma interjects to lighten the mood. “As awkward as this panel?”

sloan sports conference

At last year’s conference, Holt had flaunted Cantor’s willingness to take a $1 million bet from anyone. Predictably, it was all hot air. In fact, bookmakers set artificial limits on sharps like Voulgaris to reduce their risk exposure. After all, it’s a zero sum game. Voulgaris reveals his maximum allowed bet to be only $5000 and that it’s “harder and harder to get money down.” “What am I going to do?” he wonders out loud. All the edge in the world is pointless if you can’t apply it.

That’s a thread that connects everyone here at Sloan. Sport is a great laboratory but only if you get the chance. “Opportunity is not always fair in sports,” says Maraath, during the opening panel. And even when those opportunities come, there may not be enough data points to see the probabilities through. There’s only 16 games in a football season. Most coaches only get a few years when they would really need 20, says Alec Scheiner, President of the Cleveland Browns. Soccer is even more depressing. The average top flight manager only gets about a season and a half before he’s fired and 80-percent of them will never manage a top team again, says Prozone’s Blake Wooster. Under these conditions, experimentation isn’t brave, it’s suicidal.

jeffery maEven as technologies like SportVU give us better understanding of the journey, as humans, we still only seem to care about the destination. For fans, it’s understandable that it’s a tough sell. “We haven’t won in 10 years, but we’re doing everything right!” jokes Ma during True Performance & the Science of Randomness. “How do you build a culture that appreciates the process?” wonders Morey at the same panel. How do we balance our short term need for results and our long term need to innovate?

Ultimately, there’s nothing wrong with being right. At the very least, you can sleep soundly at night knowing you did everything you could. When asked by Ma why he does what he does, given the struggle to find a fair shake. “Beating someone,” Bob responds reflexively, but after a pause, adds, “It’s about being right about something.”

That’s really why we’re here, what all the data and formulas and charts are for. The numbers don’t lie. Is there a signal to all this noise? While the robots compute, we, as humans, seek. More than being right and gaining edge, win or lose, it’s about truth. We’re not there yet, but seven years into the Sloan sports conference, we’ve never been closer. The panel ends, and a mob of fans swarm Voulgaris like some kind of rockstar. Being right still has its benefits.

Get our Top Stories delivered to your inbox: