Sports Tech Jumps Ahead
How Sparta Science is using deep levels of data to keep athletes performing at their peak while reducing injuries.
A blown ulnar collateral ligament—the UCL, in the elbow—is among the worst injuries a pitcher can suffer. And it costs teams millions of dollars. Sparta Science, a leader in the growing sports-tech industry, has figured out the answer. It’s in the legs, not the arm. Founder Phil Wagner believes he’s uncovered other secrets for minimizing athletic injuries while maximizing performance. So do his growing list of competitors. Game on!
PHIL WAGNER is a man who believes in controlling his physiological responses, not letting them control him. Each night, he uses a technique called coherence training to harmonize his heart rate and breathing to maximize the restorative delta sleep phase. Upon rising, he spends the first three hours of his morning under red lights to stimulate the mitochondria in his retinas to produce more energy. This regimen allows him to limit his sleep to four hours and still feel refreshed. Wagner also fasts 18 hours a day so his stomach won’t distract him when he’s too busy to eat. “The mental constructs around eating are actually pretty interesting,” he says of the condition most of us simply think of as “hungry.”
Still, when Wagner, who runs a sports-tech company called Sparta Science in Silicon Valley, glanced at his phone after a meeting one morning in 2010 and saw he had nine missed calls from his wife, the reaction of his animal brain was a pure fight-or-flight adrenaline surge. Those calls could only mean something was terribly wrong at the hospital, where his newborn son was being treated.
Three-month-old Mason had been looking jaundiced, so that morning Wagner’s wife had taken him to the pediatrician. There, Mason was given a diagnosis of something called biliary atresia. A medical school graduate, Wagner knew it was serious. “They call it the silent killer,” he says. Mason was the one infant in 20,000 born without a bile duct. He needed emergency surgery to create one, and spent the next few months in the ICU while doctors waited to see if he would need a liver transplant.
The waiting carried an extra charge of anxiety for Wagner: He was not only a new father; he was also a new entrepreneur. As any parent would be, he was reluctant to leave his sick child’s side, but he was also barely a year and a half into running Sparta. After a slow start, the company had finally begun drawing a regular clientele of elite athletes, who had heard about Wagner’s ability to improve their performance using a proprietary jump-testing analysis. This was no time for a hiatus— Wagner was the one person the company couldn’t function without. He was torn between his duties as a father and the demands of his startup. “I was really hoping they wouldn’t be mutually exclusive,” he says.
They weren’t. Today, Mason is a healthy 7-year- old, and Wagner is more than ever master of his interior states. On the rare days things aren’t going well at Sparta, all he has to remember is the crisis he and his family weathered seven years ago. “It puts stress into relative terms,” he says.
Profitable and with a staff of 15, Sparta is a leader in the rapidly emerging sports-science industry that is using machine learning and previously unheard of levels of data about athletes in action to help professional and college teams minimize injuries and maximize player potential. Sparta collects prodigious amounts of biometrics from a seemingly simple jumping exercise. But Wagner says that data is the key to truly understanding athletic performance.
For years, pro teams have employed quants armed with statistical analysis to draft, trade for, and pay players on the basis of their output on the field. Now they are taking that same approach to the athlete’s body, first to keep it productive—that is, off the bench—and second to increase output by helping it achieve its upper limit of fitness. Eventually, that technology will filter down to weekend warriors, too. Firms in this industry include P3 Applied Sports Science, in Santa Barbara, California; Fusionetics, in Milton, Georgia; and Kitman Labs, which, like Sparta, is a Silicon Valley company. That’s not coincidental. Data-centric sports-tech companies such as Sparta have become the latest recruits to Silicon Valley’s venture capital teams. Go ahead, call it moneyball for injuries—the team with fewest broken players wins.
WAGNER HAS ALWAYS had confidence in his ability to control things that others write off as whims of biology or luck. It’s at the heart of what Sparta does. Once a week, the 12,500 professional and college athletes who train at Sparta HQ, or whose teams pay from $20,000 to $200,000 per year to use its technology, step onto a square metal plate and perform a sequence of six vertical jumps, striving to get as high as they can. Data collected from sensors under the plate feeds into a machine-learning software algorithm, which analyzes the force generated during the jumps— how powerfully the athlete explodes off the ground and how evenly that force is distributed over time and between the left
“What about a 32-year-old Caucasian with a history of hamstring strain who has a .303 batting average? What does he need?” —Phil Wagner, Sparta Science
foot and the right foot. The algorithm then generates training recommendations for the athlete that may not seem connected to a given weakness. The magic is in the eight years of data that Sparta has now collected that can tailor workouts to correct specific deficiencies. A basketball player trying to develop a more explosive first step might be told he should add bench presses to his workout—because, somewhat counterintuitively, the data shows benching makes the legs more powerful. A soccer forward with an explosive first step but who lacks sustained power might hear she’s at risk for developing hip problems and ought to do more lunges to remedy the problem.
Among the teams using Sparta are the NBA’s Cleveland Cavaliers, MLS’s San Jose Earthquakes, MLB’s Colorado Rockies, and the NFL’s Atlanta Falcons, who reached the Super Bowl in February. Colleges such as Auburn and Penn are also on board. Gabe Bauer, head strength and conditioning coach for the Rockies, first came across Sparta six years ago, when the team was looking for some sort of technological solution that would allow players to track their workouts without toting around a clipboard and a pen. “We weren’t even looking at the injuryprevention side of things originally,” he says. Like every coach, Bauer balances competing imperatives: On the one hand, everescalating salaries mean that the cost of injuries, in the form of lost playing time, is rising too. On the other hand, there are so many new training technologies vying for the time and money of people like him, he can’t afford to waste either on a product
that doesn’t have a clear ROI. By Sparta’s calculations, it saved teams like the Rockies $12.1 million each last year by keeping players off the disabled list. And that’s before you factor in the performance gains, which are harder to attribute purely to fitness training but, Bauer says, are just as palpable. Sparta coaches raised some eyebrows on the Rockies training staff when they recommended that infielder Daniel Descalso stop lifting weights because his scan results showed it wasn’t helping his force profile. But the Rockies went along with it. Last season, Descalso’s slugging percentage, a measure of batting power, jumped 100 points. “The results we’re getting back, honestly, they’re amazing,” Bauer says.
Wagner got into the athletic-performance field the way many people do: He was a jock who wasn’t quite talented or lucky enough to make a career out of it. As an undergraduate at the University of California, Davis, he played defensive back on the football team but endured a series of shoulder surgeries, wrist injuries, and concussions. “I love playing, so I was willing to endure a lot of pain,” he says.
After graduation, Wagner, who studied exercise physiology, landed a job as a strength coach at UC Berkeley, an hour away. The football program there was full of top-notch athletes, including future Ravens quarterback Kyle Boller, but the team was terrible, losing most of its games. When the coaching staff was fired after his second season, Wagner was left with the sense that he had succeeded, but he had no means of proving it: “I thought there’s got to be some way to validate that these guys got better.”
Ready to give it another try as an athlete, Wagner moved to New Zealand to play for a semipro rugby team. It didn’t go well. In the first minute of his first game, Wagner took a shot that knocked him out and separated his quadriceps tendon from the bone. “When I came to, the muscle had basically rolled up like a window shade,” he says. “That was the one where I was like, all right, I’m out of here.”
From that point on, Wagner redefined his mission, from staying healthy enough to play sports to finding a way to help people like him stay healthy enough to play sports. If his fellow coaches didn’t know how to do that, he thought, perhaps doctors did. He was accepted to the Thomas Jefferson University medical school, in Philadelphia. After two years, he walked away, thinking he had learned all he needed as a coach, but then the dean of USC’s med school, where he was contributing to research, offered him a transfer slot. It was too good to pass up. Wagner agreed to finish his MD.
It was at USC that he first encountered the sometimes curious data being collected by force plates, which measure biomechanical movement. Some athletes, he noticed, produced an incongruously small “impulse,” a function of force multiplied by time, for no obvious reason. Driving down I-5 to San Diego one day, Wagner got stuck in an epic traffic jam. With nothing better to do, he pulled out a notepad and started noodling on the force-plate data he’d seen. “It was like a lightning rod,” he says. “For three or four hours, the pen didn’t leave the paper.”
That epiphany was the basis of the methodology Sparta uses today. A jump consists of three phases: the load, in which the athlete crouches down and gathers energy through eccentric muscle contraction; the explosion, or transition from eccentric to concentric muscle action; and the drive, or extension and follow-through. What Wagner discovered is that different types of athletes generate different amounts of force in each phase of the jump. An athlete who relies on quickness, like the defensive back he had once been, demonstrates high explosion, while those who perform elongated actions, like a pitcher’s stride off the rubber, show a lot of drive. While some degree of imbalance is normal and even useful, Wagner’s hunch—which data collected in hundreds of thousands of Sparta scans has now validated many times over—was that athletes who become deficient in some aspect of force production are the ones who get hurt.
With his background, Wagner didn’t have to be told that teams would pay a lot for any product that reduced the risk of injuries, even fractionally. For reasons that aren’t fully understood, injury rates in most sports have risen over the
years, even as sports medicine has become more sophisticated. (Better diagnosis is thought to be part of it; longer seasons and heavier training regimens factor in too.) A single common injury, rupture of the anterior cruciate ligament, costs NFL teams an estimated $63 million annually. In baseball, where the average player salary tops $4 million, more than one in four pitchers will tear their ulnar collateral ligament, which requires Tommy John surgery to repair—replacing the damaged ligament with another one from the patient’s own body or one from a cadaver. Typical recovery time exceeds the length of a baseball season. The epidemic of UCL ruptures has spawned a cottage industry of self-styled experts and quasiscientific hucksters, such as the pitching guru who tells his clients to crawl on the floor like babies, claiming there’s something in the primal motion that’s therapeutic.
After securing a loan with the help of the California Small Business Association, Wagner signed a lease on a 9,000-squarefoot warehouse in Menlo Park, bought a pair of $20,000 force plates, and declared Sparta open for business. That was in 2009, right in the middle of the recession. “We had three months of zero customers,” he recalls. “Having no business experience, I’m thinking, ‘Oh, shit. What did I just do?’ ”
What Wagner did have was contacts in sports, especially in the Bay Area. He started cold calling agents to see if any of their clients might want to try out this force-plate analytics thing. He managed to attract a couple of baseball players and a quarterback from Stanford. Word of mouth spread. Second baseman Chase Utley, then with the Phillies, became an early client. Two years later, NBA guard Jeremy Lin, who had grown up in nearby Palo Alto, started working out there. Although Wagner was mindful of the massive PR value names like Utley and Lin carried, he says he never took VIPs on for free. If Sparta’s methodology worked, there was no need. “We hope to consistently provide every team and individual at least 10 times in immediate return value on their performance and injury risk,” he says.
AS SPARTA WAS GAINING notice, it was also gaining competitors. Around the time Sparta was spinning up, a few hundred miles south, in Santa Barbara, another jock-turned- doc, Marcus Elliott, was opening a clinic that used high-tech equipment to analyze athletes’ movements and assess their injury risk. While Sparta has the appropriately Silicon Valley feel of a startup in a garage, Elliott’s P3 Applied Sports Science has a posher vibe consistent with its clientele of Los Angeles celebrity athletes. Elliott, a Harvard Med grad who looks like the doctor in a soap opera, spends his vacations doing things like paddleboarding with San Antonio Spurs star Kawhi Leonard. P3 uses force-plate analysis, but it supplements it with a constellation of high-speed cameras that capture the motion of every joint and limb and render it all in 3-D graphics.
Across the country, near Atlanta, there is Fusionetics, founded by Micheal Clark, a physiotherapist who spent many years as head trainer for the Phoenix Suns. During Clark’s time there, the Suns were consistently one of the healthiest teams in the NBA despite its roster of older-than-average players. Clark’s system, based on a theory of “movement efficiency,” involves identifying and treating range- of-motion limitations that produce compensatory injuries. Among the tools he uses to diagnose is a system called Optojump, which measures how and where athletes interact with the ground in jumps. Kobe Bryant has said he never would have lasted 20 seasons in the NBA without his Fusionetics treatments.
And right down the road in Menlo Park is the headquarters of Kitman Labs, founded by Stephen Smith, the former rehab coach for Ireland’s Leinster Rugby. Kitman offers yet another tech solution to the movement-and-injuries problem: It tests athletes daily using a 3-D-motion-sensing system of the type found in video game consoles. The measurements are supposed to detect small changes in range of motion that signal positive or negative responses to training. Billy Beane, the innovative Oakland A’s general manager and subject of Michael Lewis’s
Moneyball, was so impressed by Kitman he signed on as an adviser. And those are just Sparta’s head-to-head competitors. There are also a slew of sport-specific firms using motion capture, wearable sensors, and other methods to help pitchers, tennis players, or golfers optimize their motions.
Next to 3-D motion capture, Sparta’s technology can seem a little, well, spartan. Why be content with looking at an athlete through the soles of his feet when you could be measuring his whole body? Wagner’s retort: All of the others are gathering data they don’t know how to interpret. They’re compiling haystacks instead of finding needles, and then drawing conclusions no more scientific than the baby-crawling pitching guru’s. “That’s a key problem in sports science right now—it’s more about marketing,” he says. Wagner points across the gym to where Daniel Descalso is flying down an AstroTurf strip used for sprint workouts. “We could gather force data on this sprint and everyone would write stories and talk about how cool it is. The problem is, it has no scientific validity.”
Because Sparta has been measuring one thing and one thing only in the same way for almost nine years, Wagner says, its data is all “clean,” or usable. That mountain of clean data yields the kind of rigorous insights that not only impress coaches but also stand up to peer review. Sparta has turned its data into six scientific papers (four published, two pending) on injury risk factors, more than any of its competitors. The crown jewel is one published in collaboration with researchers at the famed Steadman Clinic in Vail, Colorado, showing a correlation between force-plate patterns and UCL tears, the injury that costs MLB teams more than any other. The startling conclusion: Pitchers who demonstrate the most drive relative to the other parts of their jumps blow out their elbows more often.
How is it possible to deduce anything about a pitcher’s elbow on the basis of how he jumps? “How could you not?” says Wagner. “There’s a big rubber on the mound for a reason. People forget that all movement is initiated through the ground. Even in swimming, races are won or lost in starts and turns, which are essentially a jump.”
When it comes to the intersection of sports and data, Julie Allegro, founder and general partner of Fyrfly Venture Partners, is an unusually discerning critic. At her firm, she focuses on machine-learning startups with large, proprietary data sets. But she also grew up in the sports world: Her father, Jim Allegro, was a founding executive at ESPN. “So everybody sends me all their sports startups,” she says.
Last July, a friend who is a professor at Stanford business school told Allegro about a local sports-tech startup she ought to know about. After bootstrapping for seven years, it was looking to raise venture capital for the first time. Allegro agreed to meet Wagner and liked what she heard. “I know a lot of companies claim to have proprietary data. But the volume of data [Sparta] has—that’s a real competitive advantage,” she says. “They’re so far ahead of the competition, and with machine learning the way it is now, that’s just accelerating.” With Fyrfly leading the round, Sparta set out to raise $1 million in seed capital; when it closed the round after a month, it had $2.7 million.
Wagner wanted the money to hire more engineers. Sparta’s database was growing so big, he says, its machine-learning platform was having trouble keeping up. “It’s like patching a roof—you can patch it only so long.” The startup’s Silicon Valley location is a mixed blessing: Talent abounds, but competition makes it expensive. Just down the road in Menlo Park is Facebook, and Google and Apple aren’t far away either. “The good thing is I can offer tickets to Warriors games,” Wagner says, the NBA team being a client. With seven engineers now operating in-house, a few yards away from the training area where famous athletes bench and squat, Sparta is able to mine insights that are not only counterintuitive but wholly unexpected. For instance, it has found that white American athletes benefit more from single-leg squats, while Dominicans do better with bilateral ones. No one knows why. But that’s the beauty of following the data: It works even if you don’t know why. “We’re just starting to scratch the surface with ethnicity,” he says. “We want to continue to get more granular with prescriptions.” The end goal, Wagner says, is what’s often referred to in health care as “n of 1 medicine”: completely individualized treatment. “What about a 32-year- old Caucasian with a history of hamstring strain who has a .303 batting average? What does he need?”
Unlike some of its rivals, which have rushed out versions of their products for the corporate and consumer markets, Sparta is moving cautiously. It recently started working with U.S. Special Operations Command, a natural extension from elite sports. Special forces soldiers, Wagner says, aren’t much different from NFL players, with one big exception: “The consequences are much greater. Instead of not making the Super Bowl or getting another contract, you die.”
In another two or three years, Wagner thinks Sparta will be ready to introduce medical and consumer products. “Ultimately, our mission is to make sure if there’s a desire by any individual to be active, there should be no physical obstacle to that,” he says. With a sizable nest egg for the first time, he says, “we’ve got our handcuffs off.” Athletes, after all, belong on the field—although, for Phil Wagner, life on the disabled list has worked out pretty well.
“They’re so far ahead of the competition, and with machine learning the way it is now, that’s just accelerating.” —Julie Allegro, Fyrfly Venture Partners
THE DRIVE Pitchers rely on the extension phase of their motion to generate speed. And that’s where the problems arise.
PUSHOFF Pitching replicates jumping. Reaching back replicates a jump’s “load,” and pushing off the rubber replicates the “explosion,” the transition from eccentric to concentric muscle action. THE ELBOW The varus torque a pitcher generates in his UCL while throwing a fastball is about 32 Newton meters. That much torque is at the limit of what the elbow can withstand.
Phil Wagner sits under red lights in the morning to stimulate the mitochondria in his eyes, a process he says gives him more energy.