Sensors Are So 2016: Here's What's Next in Advanced Sports Analytics and Technology



  • Creating a dream team has always meant finding the perfect mix of talent and personality—but as Big Data takes a leading role in value assessment, will draft picks become nothing more than an algorithm?

Who’s better at predicting a winning team—the general manager or an algorithm?

In the excitement leading up to draft day, a general manager’s first priority is quieting the noise. It comes from a thousand directions at once—advisors, the press, a sea of social media followers shouting out picks from across the web and so on. There is a constant deluge of expectations and doubts as everyone waits to see what the general manager has up his sleeve. And yet, amid the hysteria, the GM must possess the mind of a monk—close his eyes, take a deep breath and synthesize all that he knows, all that he doesn’t and add to that one very human ingredient: instinct. And then he must open his eyes and construct the dream.

Or does he? Because while the draft day madness ensues, somewhere far away from the noise, tucked in the dark recesses of a titanium box, is a server humming along without a care in the world. It is geeking out in its own robotic way, churning through player data and conjuring a very different version of the dream.

The players that the computer likes aren’t in the headlines. They don’t have huge social media followings. In fact, some of them are the last on anyone’s list—if they made the list at all. Because we can’t see what the computer sees. We haven’t run the data under a microscope. And even if we did, we wouldn’t know what it means: that this handful of nobodies is actually the somebodies who, when properly matched together as a team, will become a juggernaut to rival all others.

Imagine “Moneyball” but powered by overachieving robots

In his book, “Big Data Baseball,” author Travis Sawchik offers the 2013 Pirates season as a perfect example. After the longest losing streak in North American pro sports history, team morale—and not to mention the salaries—was at an all time low. The Pirates had come to a collective breaking point. Something had to change. Big Data had started to nudge its way into the arena, but no one in the club had really looked to it as an answer.

And then it happened. Feeding the data through systems like PitchF/X and TrackMan, the Pirates were able to translate every piece of game data into a series of visual insights—future plays, winning lineups, etc.—all at the click of a button. Most importantly though, the platforms showed the Pirates how to work with what they had, rather than spending more time and resources on finding the perfect fix.

That’s why Big Data is here to stay

Not because it works, but because it’s relatively affordable. Relative, that is, to an industry that’s had to pay talent scout salaries for the past 50 years. Today, few teams have money or resources for the kind of speculation that sports managers used to do. Everyone had to get lean and mean. And is there anything leaner or meaner than a closet stacked with a blinking tower of hard drives?

But what kind of data are they looking at anyway? Where is it coming from?

Today, sensors are everywhere. Name the sport. For pro football, they’re embedded in the players’ shoulder pads, transmitting speed, direction and the impact of a tackle. For boxers, it’s not just about tracking how hard a punch is landed, but the fancy footwork, the strategy and all of the nuances that trainers have tracked on paper for decades, but are nearly impossible to catch from the ringside. Or take the NBA, with its own GPS wearables, tracking everything from heart rate to player fatigue to lower risk of injuries. Even Rafael Nadal is using a tricked out racket with microsensors that not only tells us exactly where on the racket that killer slice is coming from, but his heart rate as he crushes it over the net.

How do we make this data valuable?

St. Lawrence University statistics professor Michael Schuckers is attempting to answer that question—at least for the NHL. Teaming with St. Lawrence women's hockey coach Chris Wells, the two developed Total Hockey Rating (THoR), a two-way player rating that measures every on-ice action event for every player, giving it a value according to its likelihood of a potential goal. THoR also translates each player’s stats into an overall value to the team. That said, Schuckers admits there’s still a ways to go. Recording passes and real-time puck positioning are just a few improvements on a growing to-do list.

As these more complex analytics become available, we’ll be able to hone in and measure player performance at a microscopic level. Just like high-frequency trading disrupted the financial world by handing financial analysts the keys to a machine, player performance tracking and analytics will drastically transform the sports industry—from the bodies on the field to the decisions in the team boardroom.

Opening up data to the right fans

As they were entering the 2014 draft, the Sacramento Kings faced the same challenges of every other team: more data than they knew what to do with. But then owner Vivek Ranadivé and his staff realized, maybe they didn’t have to be the ones handling the data. What about the fans? Somewhere out there, there had to be at least a few Kings fans who knew how to translate the data into something meaningful. With that idea in mind, the team launched Draft 3.0, inviting the brightest analytical minds to help them decode the data. Inevitably, they tapped nine fans from around the country—from college students, to IT Consultants, to a Quant Fund Manager—to help them make draft decisions.

With great power comes great responsibility

Whether your sports data supercomputer is a humming server tower or a roundtable of super nerds, it’s not the information you put in, but the wisdom you pull out. And as teams rely more and more on Big Data to improve their performance, it makes you wonder if clubs will begin to invest as much in player tracking and analytics as they do on the players themselves. And if teams have access to that much valuable data, think about the data storage implications, the computer power needed to process it and, of course, the cybersecurity risks that go along with it.

But can we get back to humanity for a second?

What about personality, plain and simple? Let’s not forget that these athletes are the furthest things from robots, full of their own dreams, drives and imperfections. How does a computer know if two players will get along—let alone an entire team? It’s a good question, but one that lends itself to an unexpected, if not slightly scary, answer—that a computer might know us better than we know ourselves.

Yes, on the surface personalities may clash, but the one thing that trumps the ego, especially on the playing field, is winning. And players know when they have a recipe that works, they’re willing to put personal feelings aside (or at least out of reach) for the good of the team and the love of the game. And isn’t that really what sportsmanship is all about?

The fact is, no team is a perfect fit. No ego is too big or too small. But maybe there is a way to take all of that imperfectness and make something spectacular. Maybe it takes a computer to lead us to our better selves. 

Big Data is revolutionizing the game, but it can also revolutionize your business.