Author Archive

The Technology Behind the U.S. Women’s National Team’s World Cup Victory

The United States women’s national team went into the 2015 FIFA Women’s World Cup with a chip on their shoulder, trying to avenge a heartbreaking finals loss in 2011. But eagle-eyed viewers might have also noticed the chips the women wore under their shirts as well, as Will Carroll pointed out on Twitter.

The objects were Polar Global’s H7 heart rate sensors as suggested by this Wired article and confirmed by Polar Global. The USWNT is also listed as a client of Catapult, an Australian-based company that combines GPS and inertial measurement units (IMU) into a single sensor.

Strength and fitness coach Dawn Scott confirmed that her team uses heart rate sensors and GPS systems to monitor player performance. However, because the team does not have a formal relationship with either company, she could not discuss the specific devices she uses in detail. Nevertheless, she was still happy to answer general questions about how she and the rest of the American coaching staff used the devices.

The GPS system and heart rate monitor produce a wide range of metrics. Head coach Jill Ellis and her staff were mostly interested in measures of intensity, rather than total distance covered. Scott specifically discussed the percentage of high-speed running (running faster than 11 mph) and distance covered during high-speed running.

“For me they’re the main factors that then show how much a player’s involved in the high-intensity activities,” Scott said. “[That means] overlapping for your midfield player, making high-intensity runs into the box. For defenders, [it means] having to recover.”

But not every position calls for such high-intensity bursts. For those players, the coaching staff relies on meterage — a player’s average speed in meters per minute.

“So say a Lauren Holiday, who isn’t necessarily doing a lot of sprints when she’s in a holding midfield position, but she’s one of the ones who does the highest meterage, so for her, that is more of a marker of her work rate,” Scott said. “In one of the games where she was pushed into the attacking midfield role, she suddenly had a lot of max sprints.”

The games presented an additional set of challenges. Although this tournament marked the first time FIFA allowed players to wear monitoring devices on the pitch, FIFA retained the regulations prohibiting the use of technology on the sidelines. This prevented the coaching staff from using these systems to guide their in-game decision making.

“I don’t always see the purpose of real time [monitoring],” Scott said. “Sometimes in training we’ll take out the real time system, but for me that’s only if we want to get a certain physical output from a fitness point of view.”

Making matters more difficult, several of the stadiums in this summer’s World Cup were domed (like Montreal’s Olympic Stadium) or had large roofs overhanging the field (like Vancouver’s BC Place). This meant the team’s GPS-based systems were much less accurate during games.

“The interpretation of that data is crucial, especially when you’re giving that back to players and the coaches who are interested in that feedback,” Scott said.

Scott doesn’t rely on a single number to judge player performance, instead adjusting her expectations and the numbers she looks at based on the game plan for that particular match.

“It’s knowing your team, your opposition, it’s knowing your own players, and what their physical capabilities are as well,” Scott said. “Carli Lloyd’s numbers were very different in the first three games from the final three games when her role was very different.”

But unlike a coach for a club team, who can monitor their players’ workouts year round, Scott had the added challenge of making things as simple as possible for her players after their training session ended. That meant shelving the more complicated GPS monitors and giving each player a wrist-worn heart rate sensor to wear during training. To their credit, though, the players diligently stuck to the team’s training plan — and just as diligently sent the data back to Scott.

“The players were very good at giving us updates in terms of their heart rate loads,” Scott said. “And they also logged into an online training diary or physical monitoring system, where every single day they would log in, answer five questions about how they feel physically, and so I can then log in and see where a player’s physical state is.”

Scott traveled across the country, working with coaches for every National Women’s Soccer League (NWSL) team to come up with a plan that kept the national squad healthy without hindering their club’s chances of winning. Scott was quick to praise her NWSL counterparts for their cooperation.

“The clubs were given guidelines in terms of when we want to train, when we want the players to have a day off, and also ideally how long the training session should be with the player,” Scott said “And to be fair to the clubs, in that crucial period in the leadup to the World Cup, they stuck to the programs we sent.”

Off the field, Scott is working towards a doctorate from the University of Western Sydney. Unsurprisingly, Scott’s research focuses on the physical demands and training loads of elite female athletes, with a focus on soccer players. Scott’s research relies on the hundreds of hours of game data she has collected from USWNT athletes since 2012.

“The main focus is going to be to develop a training model, so looking at what are the physical demands of women’s football,” Scott said. “And then with that, we look at what is the training intervention is to prepare the players physically for those demands,” Scott said.

But not all Scott’s methods are quite so high-tech. During March’s Algarve Cup in Portugal, players complained about stiff necks and poor sleep. So before this summer’s Women’s World Cup, each player was given an allowance to buy their own pillow to take on the road with them. The result, Scott said, the team was better rest and improved performance.

“When I first suggested it, people looked at me like I’d gone mad,” Scott said. “But the players appreciated it, because it just meant something they had at every single hotel.”

(Header image via GoToVan)

Review: Motus Global’s mThrow

When Motus Global’s sleeve was announced last spring, it was supposed to save baseball, stemming the flood of Tommy John surgeries plaguing the majors. Now, the device that teams have been using to study their pitchers’ mechanics since last fall is available to the public. The mThrow has been on sale through the Motus website since March, and began shipping in early May. Eager to see what the device had to offer, I plunked down the $150 (plus $20 for an additional compression sleeve) and waited anxiously.

The box that the mThrow comes in is taken up mostly by the compression sleeve. The actual IMU — the sensor that actually tracks the arm’s motion — is a tiny blue thing, about the size and shape of a circus peanut*. The IMU charges by induction, so all the user has to do is plug in the charging station, place the sensor on top of the station, and wait about an hour.

* – But slightly better-tasting.

Pairing the sensor is simple, too, taking just a few taps of the smartphone app. The hardest part of setting the thing up is probably wedging the sensor into its little pocket in the compression sleeve, and then pulling the sleeve on so that the sensor rests over the infamous ulnar collateral ligament. In fact, the design might be overly simplified. In an effort to make the sensor more water-resistant, there are no lights on the sensor to tell the user of the charge level. The only way to check is to pair the sensor with the smartphone app; if the app doesn’t recognize the sensor, it probably needs to be re-charged.

The app is currently available only for the iPhone; an updated version was approved this week. The software now computes five metrics from the sensor data: pitch count; maximum arm speed, a rotational velocity measured in revolutions per minute; arm slot at release; maximum shoulder rotation relative to initial position; and, of course, torque on the UCL. These are then combined into three headline numbers: performance, a measure of mechanical consistency; workload, currently an additive function of elbow torque; and a “throw meter,” an energy bar that drains from blue to orange as the workload increases and consistency decreases.

I ran some preliminary testing of the mThrow, connecting it to an iPhone 4S and throwing 17 fastballs, 17 changeups, and 17 curveballs to the best of my extremely limited ability; all but six throws were recorded. Even if there’s no difference between their speeds and movement, you can still see a difference between my initial warmup tosses (the first dozen, with much lower arm speeds), fastballs (about 13-25), curveballs (26-43, with much lower torque values), and changeups (44 onward, with decreased arm speeds).

image2
This simple relationship was confirmed with a second test using a HitTrax system, which can track speed and late break of pitches as they cross the plate. My subject was a 45-year-old with some collegiate pitching experience who threw ten fastballs and ten curveballs. By comparing the HitTrax velocity report (right) to the mThrow statistics (left), we can see the correlation between the decrease in arm speed and the decrease in velocity as the subject switched from fastball to curves.

donfrancesco
Lastly, I brought the sleeve to a local high school (Blackstone Valley Tech, Upton, MA) to get some insights from active players. Assistant coach John Burke, pitcher Nick Laren, shortstop Joe Corsi, and catcher Jack Lynch took turns throwing an assortment of pitches from a number of release points, seeing how their throwing motions stacked up. The session supported some beliefs — for instance, that the quick motion Lynch uses to throw out would-be base stealers puts more torque on the elbow than a standard pitching delivery. But others were surprisingly contradicted: despite everyone’s belief that sidearm throws put less stress on the elbow than an over-the-top delivery, the app didn’t seem to report a relationship between arm slot and torque.

Chief technology officer Ben Hansen says the mThrow is still in its infancy, and says that the device’s official consumer launch is not scheduled until later this summer. The app currently relies on data compiled from Motus Global’s work with MLB prospects at last fall’s instructional league to generate its workload number, but Hansen and his team are working to produce more meaningful metrics from a more complete data set.

“We’re just capturing as much data as we can to see what’s normal,” Hansen said. “We also have controlled studies going on at every level. We have [NCAA] D1, D3, high school, and Little League players wearing it religiously.”

At this early stage, the app seems to be designed more for Motus’s professional clients than for public users. Maybe the best example of this is the tagging feature, which allows users to tag individual throws as belonging to bullpen sessions, long toss, or game action, and to further break throws down by pitch type. But at the moment, the tags are unavailable to the user after selection, getting passed on to Motus Global with the sensor data but not visible on any of the trend screens. Hansen confirmed that the tags were being used in the company’s research for their MLB clients, however.

“Every week we send reports broken down by tags where we compared each pitcher to the league averages for that pitch type,” Hansen said. “The teams love using the tags and breaking things down into the different pitch types.”

It’s a tantalizing view of an exciting feature that could still be a couple years away. And it’s not just super dorks like me who would find those analytics useful. The key to a good changeup is matching the same arm speed used to throw a fastball, so it’s easy to see coaches like Burke using the arm speed metric to give feedback to young pitchers just learning to throw the pitch. But without any way to divide pitches into different categories, this sort of feedback isn’t possible yet.

“We are looking into a web portal to give users more in-depth analytics,” Hansen said. “But right now we’re focused on getting the analytics right before we move on to other platforms.”

It’s probably still too early to judge the mThrow fairly, and I’m almost definitely not the right person to do it (sabermetrically-inclined tech geeks who can’t pitch are not Motus’s target market). And it’s true that more research could produce findings that actually help young pitchers stay on the field and off the operating table. But as currently constructed, the mThrow raises more questions than it answers, and left me wanting more. Like a top pitching prospect, the technology needs some time to mature before it can make a meaningful contribution.


KinaTrax Gives Rays In-Game Markerless Motion Capture Data

In an effort to keep their pitchers healthy, the Tampa Bay Rays have enlisted the services of markerless motion capture company KinaTrax. As Jeff Passan of Yahoo! Sports reported Monday, the Rays are the first team to partner with the Philadelphia-based company.

When asked about the technology Tuesday, KinaTrax founder Michael Eckstein was reluctant to reveal much of the technology that drove his company’s system. Images from “multiple cameras” positioned throughout the ballpark (an earlier test used eight) are stitched together to create an unobstructed, 360-degree view of the pitcher. Eckstein compared his system to the commercially-available Microsoft Kinect, which uses infrared and sonar tracking to capture a user’s position for video gaming or other applications.

“The Kinect has a focal length of 8 to 14 feet, and captures 30 frames per second,” Eckstein said. “The challenge is, how do you scale that up to an MLB stadium, where you have to capture 275 to 300 frames per second from 350 feet away?”

Once the data is collected and uploaded to cloud storage, “proprietary algorithms” are then used to identify the position of body landmarks like joints and calculate the distances, angles, velocities, and accelerations between the various body segments. In an earlier talk at the 2013 SABR Conference in Philadelphia, Eckstein claimed that the positions measured by the system were accurate to within 1.5 centimeters.

It is probably no surprise that capturing such detailed visual information hundreds of times per second is a costly process. Eckstein estimates that a typical game could produce up to 1.4 terabytes of data. The data is owned by the teams — since it identifies each pitcher and is thus considered medical information, even KinaTrax can’t access it without permission once it’s collected. For teams unable to work with the raw data, KinaTrax can also develop reports on key metrics; Eckstein said in his 2013 presentation his system was capable of generating these reports overnight.

“Some teams have the ability and the staff to say, ‘We want these kinds of reports, and these kinds of analytics,’ and then we can go out and produce them,” Eckstein said. “And then if teams have very qualified staff, they’ll get the raw data to work with themselves.”

Although KinaTrax worked with the Mets in 2013 to develop their system, Tampa Bay is the first major-league team to install the system and collect game data. And while it’s too early to draw any conclusions from the data collected by the system, Eckstein is happy with KinaTrax’s early performance.

“We’ve successfully recorded thousands of pitches, and the system is working as expected,” he said.

According to Eckstein, KinaTrax had discussed possible arrangements with 17 MLB teams between the Winter Meetings, Cactus League, and Grapefruit League before finally coming to an agreement with the Rays. Eckstein was excited about working with the Rays, praising their front office acumen and even the symmetrical shape of Tropicana Field (which made camera installation easier).

“The Rays are among those top major-league teams that understand what we’re doing and have an understanding of big data,” Eckstein said. “We couldn’t ask for a better team for our pilot.”

Teams have already proposed a number of different uses for the system. For major league pitchers, teams could use the system to demonstrate “best practices,” and highlight the subtle changes in mechanics that could separate a great outing from a poor one. But Eckstein also discussed the possibility of installing the cameras in minor-league parks, allowing teams to better teach proper mechanics to young arms while also developing “longitudinal patient records” of changes to a pitcher’s kinematics over time.

“All of the teams we’re speaking to want them in their major league stadiums,” Eckstein said. “But the really innovative teams tell me, ‘Where we will get the most benefit out of this is with our Single-A or Double-A teams.'”

Once installed, the system can also be adjusted to capture mechanics in bullpen sessions, and could be modified to track hitters’ swing mechanics. For now, though, KinaTrax is primarily focused on the action on the pitcher’s mound.

“There’s a consensus among teams about this anecdotal evidence of pitchers who are great in their bullpens but then lose it on the mound,” Eckstein said. “But truth be told, it’s the in-game information that managers, coaches, and scouts are after.”

Before founding KinaTrax, Eckstein worked in the technology sector for 25 years, helping companies figure out how to use technology to develop competitive advantages. A baseball fan, Eckstein found himself at a lunch with a Phillies senior executive in 2012, and the conversation turned to Roy Halladay’s early-season struggles.

“He said, ‘Wouldn’t it be great if we had a way to measure his mechanics and see what he’s doing wrong?'” Eckstein said. “And I said, ‘Oh, this will be easy. We’ll go to Microsoft and they’ll come up with something.'”

It wasn’t that easy, of course. The leap from the existing technology to in-game motion capture from hundreds of feet away required the development of an entirely new technology platform, which became the basis for KinaTrax.

Before Monday, KinaTrax first announced itself at the 2013 SABR Conference in Philadelphia, where Eckstein gave a talk and brief demonstration on his system. At the time, KinaTrax had persuaded the Mets to let them test their camera system in Citi Field. The eight-camera test was successful, but no actual game data were recorded.

Now that the word is out on KinaTrax, Eckstein plans to return to the Winter Meetings and put his newly-tested product before the decision-makers in MLB front offices.

“We’re going to have serious discussions with teams about agreements for the 2016 season,” he said.

But go on the company’s website and you’re greeted not by a picture of a Major Leaguer or of Tropicana Field but by a youth baseball pitcher. This is not just a nice image: Eckstein said KinaTrax is planning to scale its system down for college, high school, and even youth-level teams.

“Clearly the arm motion is very different for an eight or 12-year-old versus a major league pitcher,” Eckstein said. “But we feel that with the nuggets we’ve learned, and with cameras that don’t have to capture 275 to 300 frames per second and don’t have to be 350 feet away, we can bring the price of the system down to that level.”


Blast Motion Sensor Augments Metrics with Adaptive Video

The first thing to keep in mind about Blast Motion’s sensor is that it’s not just designed for baseball. Yes, like the Diamond Kinetics SwingTracker, you can attach the sensor to the end of a bat to track swing speed and direction. And like the Zepp sensor, the Blast sensor can also be used to track a golf swing. But Blast’s approach revolved around designing a high-quality, general purpose sensor, and then building specific applications for baseball, basketball, golf, action sports, and athletic performance around it.

“We didn’t approach this as trying to design a swing sensor or a specific sport product,” senior director of marketing Donovan Prostrollo said. “What we designed it to do was to be a natural motion capture product, and then we applied that to different sports, so it doesn’t pigeonhole our product.”

At the heart of the Blast sensor are inertial measurement units (IMUs), the combination of accelerometers, gyroscopes, and magnetometers that have become ubiquitous in devices like smartphones and tablets. But Blast has made two improvements to make the device more accurate. First, Blast Motion uses multiple IMU chips (although they wouldn’t disclose how many) to capture a wider range of movements. Second, the Blast sensor was also designed to use what founder Mike Bentley referred to as “tactical-grade” technology, a combination of more precise sensors, more processing power, and on-the-fly calibration that improves the device’s accuracy and consistency from one movement to the next.

But despite the intense technological focus, both Bentley and Prostrollo stressed the importance of keeping their outputs simple for the end user.

“You’ll find other solutions out there really overwhelm users with numbers, which is both good and bad, because if users don’t know which number to focus on, you’re not really helping them, you’re actually potentially making it worse,” Prostrollo said.

“At the end of the day, [the athletes] would love the technology to just completely disappear,” Bentley added. “And that’s one of the goals of Blast is how do we make the device disappear.”

When compared to other bat sensor apps, Blast lacks the three-dimensional rendering of the swing. Instead, the Blast app revolves around video, typically captured by setting the device on a tripod and automatically clipped so that only the events of interest are included. Prostrollo argued that the focus on video gave Blast an edge in capturing the entire movement, not just key metrics.

“We decided from the beginning to capture video and do it natively as part of the app so it’s really integrated into our DNA,” Prostrollo said. “And the cool thing about that is when you pair video and you compare the level of consistency out of our product it really does an amazing job.”

And Blast recently announced an adaptive slow-motion feature that adjusts the playback speed around the event.

“Basically, we know exactly when the impact occurred, when the swing started, and when the swing ended, and based on that we can speed up and slow down the video,” Prostrollo said. “We can also take the metrics and overlay them on top to get this dynamic fill, so it’s not just a metric in isolation.”

Blast verified the accuracy of its metrics using motion capture systems. As an example, Prostrollo said the system was within 1 mph of the motion capture system “85 percent of the time” and Bentley claimed that the Blast system “outperforms our optical system when you talk about rotational velocity” as verified by higher-end devices more commonly used to test aeronautical and military-grade IMUs.

Bentley and Prostrollo stressed not only the device’s accuracy but also the device’s consistency, so that identical swings or jumps would produce identical sensor readings. They attributed this consistency to improvements in their manufacturing process, and claimed it made a big difference to the professional athletes they collaborated with.

“The challenge is pro athletes absolutely can recognize that day one, the amateur athletes won’t necessarily realize that a product’s not as accurate as they want until it’s too late: they’ve purchased it, they’ve gone out, they’ve tried it, and they wonder why their swing speed varies by 6 mph when it’s all the same,” Prostrollo said.

Despite being a relatively new company, the founders of Blast Motion have been in the inertial sensor business for a quarter of a century. Before entering the sports world, their focus included military and medical products.

“This is not the first sensor we’ve ever manufactured,” Bentley said. “When we originally designed the sensors, it wasn’t for a single application. We wanted to be able to use the sensor and cross-pollinate across all applications.”

As Blast Motion began adapting its offerings for new markets, it worked with coaches, professionals, and other subject matter experts to design useful applications. But Bentley said there was a lot of overlap between the biomechanical elements underlying the different sports. Even more surprising, he said, was the overlap between social circles across different sports.

“What’s pretty unique about when you do get into the inertial world of working with different professional teams, how many baseball players work with professional golfers, and how many golfers play with hockey players,” Bentley said. “So the world is pretty small, and when you get a pretty exciting product, the word travels pretty fast in those worlds.”

The company currently works with a number of action sports ambassadors including Mike “Hucker” Clark and Greg Lutzka, as well as some NBA and MLB players they declined to name, citing confidentiality. And Blast Motion is working closely with bat manufacturer Easton as it gears up to release the Easton Power Sensor this summer. Little information is currently available about the project, but judging from the screenshots in the iTunes App Store, the interface at least will be very similar to Blast’s Baseball Replay app.

Looking to the future, Prostrollo said the biggest change would be not on the technological side but rather on the adoption side, as wearable sensors like Blast become more and more ubiquitous among both amateurs and pros.

“We’re at the point now where the average consumer has access to this technology, it’s no longer the pro athlete,” Prostrollo said. “What you’re going to see is a whole new generation of athletes leveraging the data and the technology, having a history to go back on, and really be able to do something very meaningful and different with that.”


Kitman Labs’ Profiler Helps Keep Athletes on the Field

Sports analytics has moved on from the days when an ambitious amateur could fire up Excel or a relational database and make earth-shattering discoveries. Modern front offices must incorporate not only on-field performance but also medical histories, training results, biomechanics data, and a host of other sources into their decision making. To help teams better manage and access that mountain of data, Dublin-based Kitman Labs has developed the Profiler, a system that combines disparate data sources into analytics describing player health and injury risk.

Chief product officer Stephen Smith described Kitman Labs’ offering as “the operating system for sports teams.” The strength of the system is in its ability to combine data from multiple areas — including medical, biomechanical, and on-field sources — to produce more holistic analytics that can better inform team decision making regardingl athlete training and injury prevention. Smith said he was first inspired to create the system while working as an athletic trainer for Leinster, an Irish rugby team.

“One of the biggest challenges I had as a practitioner was that all the fitness data was held in one area, all the medical data was held in one area, and all the performance analytics were held in another area,” Smith said. “That just made it very hard to understand what any of the information actually meant.”

An example of the power of Profiler is demonstrated through a software application that allows users to collect markerless, three-dimensional biomechanical data from an off-the-shelf Microsoft Kinect. The software can calculate select joint angles from an athlete a few feet away — even during rapid dynamic movements, such as running, kicking, or throwing a pitch. And although Smith insisted that the Kinect software was “probably five percent of what we actually do,” he was enthusiastic about its ability to make motion-capture based analysis more accessible.

“Biomechanical information that you would garner in a normal professional sports environment would take you hours to actually get because the downtime is huge, and the cost of that is pretty difficult, and you just can’t access that day to day because it takes too long,” Smith said. “The software that we’ve created jumps professional sports teams into the next generation of real-time technology.”

When interviewed, Smith refused to name the specific organizations that have partnered with his companies.

“We definitely don’t like to speak about our clients because a lot of the information we’re housing, as you can imagine, is very sensitive data on very high-profile athletes,” he said.

But some of his clients have been less tight-lipped about their relationship. In March, The Los Angeles Dodgers announced that they would be partnering with Kitman Labs in their farm system, declaring themselves “the first American sports team” to sign with the company. Across the pond, Kitman Labs works with British Premier League squad Everton, along with a number of Irish rugby teams. Other organizations, including the San Francisco Giants, have also tested this system.

When Kitman Labs signs a new client, the two first collaborate to determine which data and which metrics are most important to the organization, and what sources of information the organization already collects. Kitman’s sports scientists then work with the coaches and training staff to demonstrate how to use the system and understand the analytics the system produces.

“We have a very experienced team of sport scientists who all understand the individual nature of each sporting discipline that we work with, and the uniqueness of each club, team, and athlete,” Smith said. “Those sports scientists will actually be on the ground with teams for a number of days actually helping them to get up to speed.”

Smith says he understands that his company offers an appealing solution to clubs looking to maximize the return on their sizable investments in player salaries, not to mention strength and conditioning, coaching, and other aspects.

“I presume that [general managers] want tools to ensure that they can get the best value from their athletes,” Smith said. “I think the clubs just love the idea of being able to try and maximize on that by being sure they can keep the athletes on the field.”

But the growth of biomechanics data has led to rumblings in some quarters. Some have expressed concerns that medical data which suggests an injury risk could be used against athletes during negotiations. (An example can be seen in the controversy surrounding the Houston Astros’ dealings with top overall pick Brady Aiken last summer.) Despite this, Smith said he hasn’t seen any pushback from athletes on teams using this product, and insists that the system was designed primarily with athlete wellbeing in mind.

“One of the largest driving forces for us in doing this is that we want to protect athlete welfare, we want to improve the standard of care that is given to athletes worldwide,” Smith said. “It’s there purely for the team to use that information to empower their decision making, and that way they can ensure the athlete makes it onto the field in the best possible shape.”

Kitman Labs was born out of Smith’s postgraduate research into injury risk factors, as well as his professional experience as an athletic trainer. The company was founded in October 2012, with its first product offering coming online in June 2013. By early 2014, Kitman Labs had signed their first partnerships with soccer and rugby teams, and were looking to expand into the American market.

“We kind of expected that the market over here would be pretty far ahead of what was going on in Europe,” Smith said. “But when we came over, we realized that it didn’t look like there was anybody else trying to do something like what we were doing.”

The company opened its first American office in Menlo Park, California, in September 2014. Since the MLB season was just wrapping up, Kitman Labs initially focused on expanding into baseball to coincide with teams’ buying cycles. Kitman Labs is now looking to expand into other sports, developing new applications in both professional and collegiate sports leagues.

“We’ve had early success with baseball in the U.S., but we’re actively working with NBA teams, NFL teams, and we’re actually now branching into the NHL as well,” Jeff Eckenhoff, a member of the Business Development team, said. “We’re pretty sport agnostic.”

And with the expansion into new sports comes an expansion of staff, as Kitman Labs looks to add sports scientists and engineers that can help them adapt their solutions for new clients. Smith said his company is actively looking to fill eight vacancies.

“We need industry experts from basketball, football, and baseball to come and be part our team, and to help us solve the largest problems for each of these sports so that we can truly help these teams to uncover the sources of injuries,” Smith said.

Still, Smith insisted that his company’s expansion would not come at the expense of Kitman’s current offerings.

“We don’t want to be a company that walks into a market and grabs a huge collection of customers and then walks away with their checks in our back pocket,” Smith said. “We want to change the face of sports science and sports medicine and we’re going to do that by incredible focus and by being extremely diligent.”


HitTrax System Makes Batting Practice Perfect

Professional baseball is a grind, with daily games and countless hours of batting practice for hitters. But younger hitters working in a batting cage lack the feedback of seeing how that last hit would have traveled on the diamond. To help hitters get that experience, the company InMotion has developed the HitTrax system, capable of tracking batted ball speed, launch angle, and a number of other parameters that tell hitters how far each ball would have traveled during an actual game.

The system consists of separate hardware and software components. The hardware, encased in the rectangular white box seen above, consists of three near-infrared cameras and two near-infrared LED arrays that better illuminate the ball. Like other motion-capture systems, multiple cameras track the ball as it crosses the camera volume. The location of the ball in each camera’s field of vision, combined with the known distances between each camera, are combined to measure the position of the ball in three dimensions.

The box containing the cameras is positioned inside the cage, a few feet behind the batter and just in front of the plate, in a fixed position for both right- and left-handed batters. You typically wouldn’t want to stand by the box when someone is in the cage, of course, but the hardware is still well-protected from foul balls: the LED arrays are behind bulletproof glass, and the front of the box is “made from the same material as hockey boards,” according to Tom Stepsis, InMotion Systems’s director of marketing.

The tracking data is then fed into a physics engine to project the distance each hit would travel in the real world. But in addition to distance and trajectory, HitTrax also estimates whether each batted ball would result in a hit or an out. The fielders’ ability has been programmed to match the hitter’s, so high school hitters will face high school fielders, whereas more skilled fielders and deeper fences await older hitters.

InMotion, based in Northborough, Mass., claims that the speeds reported by the HitTrax system are accurate to within one mile per hour, as compared with conventional radar guns. Stepsis also claimed the distances reported were accurate to within five percent of the actual distance, as measured manually with a tape measure. The system does not track the ball’s spin (which has been shown to have an important impact on the distance a fly ball travels) but instead makes its calculation based on the first few feet of trajectory captured by the cameras.

The HitTrax software is controlled by a touchscreen, where the user can enter personal information, change settings, or switch between training and game mode. In training mode, the system can produce detailed spray charts, strike zone “hot and cold” zones, and trajectory data such as launch angle and exit velocity. Reports and leaderboards are available online so players can track their performance and get a sense of how a change to their swing mechanics might translate to in-game performance.

But game mode, Stepsis said, was entirely separate. Here, hitters can compete in home run derbies and on teams in simulated games. The system also includes fun features, like power boosts, to affect trajectories.

Despite its name, the HitTrax system is also capable of tracking pitchers. The system tracks the horizontal and vertical break of the ball, the “end speed” as the pitch crosses the plate, and where in the strike zone the pitch was located. Because the cameras are fixed in front of home plate, however, more in-depth statistics like release point, starting speed, and a more detailed trajectory of the ball’s path to home plate, are not available.

Prior to founding InMotion, the company’s founders had decades of experience with motion tracking technologies and a passion for baseball. It took InMotion “a solid two years” to develop the HitTrax system to the point where it was ready to be sold. Stepsis said that, because the product was so unlike other available offerings, the initial marketing focused on showing potential customers how to use the system.

“When we introduced this, part of the hurdle was explaining what it was to people,” Stepsis said. “And seeing is believing, so we did a lot of demos. And then once people saw it, word of mouth started to spread, and things really took off.”

The system is now in facilities across North America, along with some high schools, colleges, and the occasional private residence. For those in publicly-accessible facilities, the price for a session can vary widely.

“There are some places that charge over $100, there are some places that just put this in a coin-op [batting cage] and just charge double, so instead of $1 for 20 balls, it’s $2,” Stepsis said.

InMotion has gotten positive feedback from players, coaches, and facility owners as a training tool, but Stepsis said some users were also using it for tryouts or scouting purposes.

“Some of our customers who own facilities are also MLB scouts, and they love it,” Stepsis said. “They feel like the data we’re providing them just paints this elaborate picture of what the player’s like.”

As InMotion grows and HitTrax becomes more popular, Stepsis hopes that his company will be able to give players and coaches instant access to the type of data that will allow them to monitor their progress and quantify the effect of any changes in their swings.

“We’re not coaches. We just want to be data providers,” Stepsis said. “It’s all about making the indoor training environment more engaging and more beneficial.”


An Analog Approach to Enjoying Baseball

It’s Baseball Week on TechGraphs. Our writers have been describing tools they use to keep up with the baseball season. Bryan Cole’s is below.

Look, I get it. Technology makes baseball better. There’s no question. I’m not going to sit here and pretend that being able to flip between any MLB game happening live anywhere around the world*, with a little pull-up menu that instantly that shows how your fantasy teams are doing isn’t amazing. It is.

* – Certain blackout restrictions apply.

If you wanted to follow the 1912 World Series, here’s what you did: you went down to the newspaper office and you stood outside and you watched an electronic scoreboard with mechanical players that operators updated every time they got a telegraph from the stadium. This sounds like an “uphill-both-ways-in-the-snow” style exaggeration, but this really happened. Some people paid as much as 50 cents — the same price as a bleacher seat at the actual Series! — to watch these scoreboards.

Still, baseball is one big nostalgia trip for me: listening to the game on the radio, scoring the game on paper, saving ticket stubs to commemorate the games you went to. But there are still ways to bring some of these parts of the experience into the 21st century.

Radio Broadcasts

We’ll get the easy one out of the way first. An MLB.tv subscription of course includes the home and away audio broadcasts of all games, and you can get an audio-only subscription for just $20 all season. If you speak Spanish (or want to learn the extremely hard way), those broadcasts are available too. If you want still more people talking about baseball, there are a number of baseball-centric podcasts* that go incredibly deep on virtually every aspect of the game.

* – I’ve been looking for a daily baseball podcast on the level of The Basketball Jones for a couple years now, but I still haven’t found anything quite like that.

Keeping Score

Baseball Reference is a thing of wonder. Your dad can start reminiscing about this time he saw Yaz play a doubleheader against the Senators when he was in elementary school and boom, you can tell him who the winning pitchers were before he gets to the part where their car overheated in traffic on the way home.

Before then, the only way to get those details was to keep score with paper and pencil (or a pen, if you felt confident). There are a number of different scorekeeping guides and books, with varying levels of complexity. If you want to just mark down whether a player reached base safely or not, that’s fine. If you want to track balls and strikes, cool. If you want to try to indicate where the ball was hit on that tiny little diamond they give you, go for it.

My only advice is to get one that’s wider than it is tall. Most of the books available in sporting goods stores are designed for Little League coaches, so they have something like 16 lineup slots and only nine innings. But if you happen to be scoring an extra-inning game, your choices are either (a) stop keeping score at the most interesting part, or (b) copy all of the lineup information over again only to have the lead-off hitter hit a walk-off homer in the bottom of the 10th.

Obviously you don’t have to do this to enjoy a baseball game. At the professional level, you don’t even have to do this if you want to see how your favorite player is doing, since it’s usually a couple of smartphone clicks away. But it does give you something tangible to remind you of the game you went to and that, yes, Dad, Tim Wakefield did give up six homers in that game.

Paper Tickets

The ticket stub is a built-in souvenir, a reminder of the specific game you went to (so you can look it up later on Baseball-Reference). And recent tickets — the ones with photos on the front — tell you even more: a generic shot of the stadium or fans cheering tells you that team probably wasn’t very good. But teams have stopped mailing out those admittedly easy-to-lose pieces of paper, instead sending a PDF fans can print at home. And that’s great, and they’re actually really convenient if you’re meeting up with people, but no one’s going to pay for a super-sized PDF printout to hang on their wall.

For once, technology actually offers ways to counter this. Apple’s Passbook can be used in a number of MLB parks and lets you hang on to past tickets, meaning you can actually take your collection with you. Then again, when you think about the phone you were using ten years ago, you realize these digital tickets might not be with you for as long as you’d think.

The fundamental language of baseball is one of tradition, of grizzled scouts and outdated equipment and 60-year-old men wearing uniforms and suboptimal strategies because That’s How It’s Always Been Done. It’s ridiculous, sure, but if you squint (or if your vision is going), you can convince yourself that Ted Williams and Babe Ruth and Sandy Koufax could actually still survive in this game, giving it a link to the past none of the other sports really enjoy. In a few minutes, the nostalgia will pass, and I’ll be back to looking at StatCast data while watching two games at once.

Until then, get off my lawn. I just found the tickets from that road trip I took to see Pedro Martinez pitch against the Braves in Shea Stadium.

(Photo by Scott A. Thornbloom/U.S. Navy)

Wilson, SportIQ Team Up to Produce “Pro-Quality” Smart Basketball

Look close at Wilson’s Wx “connected basketball,” and it’s hard to tell what’s so different about it until you spot the Bluetooth logo by the inflation valve.

“That is not something we usually deal with,” Wilson’s Vice President of Innovation Bob Thurman chuckled.

The ball was presented, along with an accompanying mobile app, at last month’s Sloan Sports Analytics Conference in Boston. Wilson developed the basketball in partnership with SportIQ, a Finnish company whose player tracking solution combines wearable sensors with synchronized video to help coaches analyze their teams’ performance.

When asked, SportIQ CEO Harri Hohteri (pictured above) was reluctant to talk about the “secret sauce” behind his company’s basketball. But he was quick to differentiate it from other smart sensor basketballs like 94Fifty’s.

“[The 94Fifty ball] was designed around shooting mechanics as a training tool,” Hohteri said. “But the first thing for us is the consumer side of things. We wanted to develop a professional-quality basketball.”

Hohteri, who played four professional seasons in Finland’s Korisliiga, insisted the feel of the basketball was of the utmost importance to players. “I can’t tell the difference between this and a game ball,” he said.

The quality of the basketball is further underscored by SportIQ’s partnership with the Korisliiga. For the third straight season, Finnish players are wearing the company’s sensors (and using its basketball) in league games. The data is used to automatically tag events in a synchronized video that Hohteri says coaches are using to track the efficiency of their offensive sets. And because it relies on sensors, the system doesn’t need the extensive camera setup used by STATS’ SportVU tracking system.

“It’s about doing things more efficiently,” Hohteri said. “That’s the whole idea. We can do the whole thing in real time with less manpower than teams are using now.”

But for those of us who aren’t running a professional basketball league, Wilson’s connected basketball is launching this year. The demo at the conference included a smartphone app (projected onto a television) that showed players their accuracy from various distances on the floor. A machine learning algorithm in the app detects makes and misses without the need for an additional sensor attached to the net, unique among smart basketball systems. Each distance stripe was color-coded, according to the percentage of shots made from anywhere in that arc.

SportIQ’s partnership with Wilson started in August 2012, when SportIQ first began its relationship with Finnish basketball. Because Wilson is the official basketball of the league, Hohteri approached Wilson’s innovations department about developing a smart basketball.

“At the same time, our business director was asking us for a way to measure makes and misses in the driveway to keep kids in the game,” Thurman said. “So we agreed that we would help engage them on tracking the basketball, and they would help us with this make/miss aspect.”

Thurman hopes the partnership between their companies will combine SportIQ’s intelligence with Wilson’s broad user base to “gamify” practice and inspire the next generation of basketball players.

“We want to activate 12-15 year old kids, to get them off the video games, and get them back in the park, to be more active,” Thurman said.


OptiShot2 Allows Golfers to Play Year-Round

Winter is finally starting to release its grip on the U.S., but temperatures are still below average and there are several feet of snow on the ground across the northeast. However, golfers can still work on their game without trudging through the snow thanks to simulators like the OptiShot2.

OptiShot is a small company based out of Traverse City, Michigan (2015 snowfall: 89 inches). Their golf simulator allows users to play a full round indoors using every club in their bag*. Even in the summer, the simulator can still come in handy: since there’s no walking or hunting for balls, users can play a full 18 holes in under an hour.

* – While you can use real golf balls with the OptiShot, you might want to consider using their foam practice balls — or no balls — if you’re actually playing inside.

The heart of the OptiShot system is a row of sixteen infrared sensors embedded in a practice mat. The sensors track the club tens of thousands of times per second, sending data via a USB cable to a laptop. There, the OptiShot software computes the speed of the club head, club face angle, and swing path. Of course, the system also computes the trajectory of the shot, projecting it onto three-dimensional replicas of famous courses from around the world. For added realism, the software also includes penalties that limit the range of shots out of the bunker and rough.

But OptiShot CEO Russell Edens says the company’s user base is what sets the OptiShot apart.

“We have the largest user base on the planet,” Edens said. “You can play against other golfers around the world online, and we hold events for people to play against each other.”

To keep its customers happy, OptiShot is constantly digitizing new courses, releasing the Riviera Country Club simulation (pictured above) last month. The device ships with 15 courses, replicas of the courses that host the U.S. Open, PGA Championship, and Ryder Cup; additional course simulations are available for purchase online. OptiShot’s course designers rely on high-resolution photos and detailed three-dimensional data “that maps every contour, dip, and bump” of the famous courses, Edens said.

OptiShot trained its system using launch monitors (like TrackMan) that track a shot’s speed and trajectory, and Edens is proud of the system’s accuracy.

“Frankly, the accuracy of OptiShot for the price point is shocking,” Edens said. “When it come to sidespin, we are often better than most simulators on the market.”

Edens said the touring and teaching professionals OptiShot has worked with to design and test the system were impressed by the system’s ability to capture their game.

“The reaction is nearly always the same,” Edens said. “They hit a shot, nod their head and say, ‘Yup, that’s my shot.'”


ASMI’s Glenn Fleisig talks about Tommy John Surgeries, Wearable Sensors at Sloan Conference

For such a small piece of tissue, the ulnar collateral ligament can do a lot of damage when it tears. According to Jon Roegele, 90 professional baseball players underwent Tommy John surgery to repair a damaged UCL in the last calendar year, costing teams tens of millions of dollars and months of lost service time. Young prospects, grizzled veterans, pitchers, catchers, outfielders, flamethrowers and junkballers: all went under the knife and emerged with the infamous long, curved scar.

The high stakes go a long way to explain why Glenn Fleisig, Ph.D., research director at the American Sports Medicine Institute, was invited to speak on the epidemic at last weekend’s Sloan Sports Analytics Conference. Fleisig — who also serves as chairman of USA Baseball’s Medical and Safety Committee, an advisor for Little League Baseball, and on Major League Baseball’s Elbow Task Force — presented “Analytics of the Tommy John Injury Epidemic” Saturday morning to discuss the ASMI’s research, dispel some common myths, and talk about the “Pitch Smart” guidelines he helped develop.

Even in the few months since the Pitch Smart guidelines were released, Fleisig said he was “blown away” by the amount of positive feedback he’s received.

“For 20 years, Dr. [James] Andrews and myself have been giving guidelines that have been pretty well received in general by those in medicine,” he said in an interview. “Since the Pitch Smart guidelines came out, we’ve seen the same kind of response from coaches and parents across baseball.”

But Fleisig still devoted a significant amount of time to clearing up misconceptions about the surgery’s success rate (only two-thirds of MLB pitchers who underwent Tommy John surgery made it back to the majors and stayed there) and effect on pitchers (pitchers do not see a bump in performance immediately after the surgery).

The highlight of Fleisig’s talk was a discussion of the risk factors associated with UCL injuries. Fleisig reported the results of a recent study into the biomechanics of 80 minor league pitchers — 40 of whom had returned from Tommy John surgery, and 40 healthy controls. The paper looked at a number of potential risk factors (including amount of arm abduction, stride length, and the infamous “inverted W“), but found no difference between the mechanics of the previously injured and healthy pitchers. Fleisig admitted, however, that the study was not predictive: by looking at the biomechanics of only those pitchers who successfully returned to pitching, the study overlooked any mechanical flaws that had since been corrected, as well as any biomechanical issues in those pitchers who were unable to make a successful comeback.

A better predictor of UCL injuries, Fleisig said, was overuse. A recent ten-year study of 476 adolescent baseball pitchers found that pitchers who threw over 100 innings in a year were three times more likely to be injured; those who regularly threw over 80 pitches in a year more than four times; those who pitched for at least eight consecutive months five times; and those who regularly pitched fatigued 36 times more likely to suffer a UCL injury.

“I hear people say that throwing isn’t a natural motion,” Fleisig said during his presentation. “Throwing is natural. Pitching at max effort, 100 or more times in a game, that’s what’s not natural.”

Fleisig believes overuse of young pitchers is largely responsible for the rise in professional pitchers getting Tommy John surgery, even as organizations are placing increasingly strict limits on prized prospects like Dylan Bundy and Joba Chamberlain.

“A lot of the time, the damage has already been done by the time they get to pro ball,” Fleisig said. “When you pitch, you produce lots of microscopic tears that can heal with rest, but if you pitch too much, then [the UCL] is like a frayed rope, and there’s not much you can do.”

Fleisig’s group has calculated that a 90-mph fastball puts as much as 100 newton meters (N-m) of torque on the elbow. Fortunately, this force is instantaneous or UCLs would be tearing on virtually every pitch. But the ligament still absorbs a significant part of that force; as muscles tire, the amount absorbed by the UCL can increase, leading to the dramatic increase of UCL injuries in pitchers who regularly throw when fatigued. Fleisig said that strength and conditioning can help reduce the risk of injury, but stressed the importance of the entire kinetic chain, from legs to trunk to arm muscles.

“I wish I could say there was one key to avoiding injuries — and that would make a good article — but that’s not reality,” Fleisig said. “The secret isn’t the legs, the secret isn’t the arms, the secret isn’t the trunk, it’s the whole body.”

In addition to his work with ASMI and the sport’s governing bodies, Fleisig also serves as a consultant for Motus Global, a movement analysis company whose mThrow sleeve for pitchers will be available to the public this spring. The headline number produced by the system will be a measurement of workload, derived from the sum of the torque put on the elbow across all throws a player makes, including pitching in game situations, warmups, and drills such as long toss. The sleeve (and the metrics it provides) are still in the early stages, but Fleisig is confident that it represents an improvement in teams’ understanding of workload.

“We don’t know what the right exact formula [for workload] is,” Fleisig said. “For instance, if you throw two throws at 50 N-m, is that the same amount of danger to your elbow as one throw at 100 N-m? Probably not. But before the Motus sleeve, workload was just added up by people counting how many throws someone did.”

Fleisig admits, however, that users will rely too heavily on the system’s workload calculation, especially at the amateur level where coaches, trainers, and players may lack the rapport to establish when a pitcher is truly fatigued.

“The potential customers for the Motus sleeve are not only pro pitchers but also high school kids,” Fleisig said. “I would hope that the pro teams would click through to see the extra data the sleeve produces, whereas I think the summarized information might be the best for many of the amateurs.”

A number of companies aim to improve further on Motus Global’s innovation, and are working on systems that can automatically capture full-body kinematic data without markers. Fleisig does not consult on any of these projects, but has seen some of them in action. Most, including the system Sportvision presented at Saberseminar last summer, require manual digitization, where a human observer annotates the location of the pitcher’s joints at every frame. But despite the demand for automated systems, Fleisig stressed that these technologies are only a tool in the arsenal of coaches, trainers, and pitchers to prevent injuries and improve performance.

“Biomechanics experts and technology are not the keys to optimizing pitcher safety and performance,” he said. “The key will be the expertise of coaches to use this new information.”