The Technology Road Map That Can Turn the New York Mets Into World Series Champions

By Joe Lemire3.16.2021

Al Bello/Getty Images

Not long after hedge-fund titan Steve Cohen purchased the the New York Mets for nearly $2.5 billion last fall, a person with inside knowledge of the club told the New York Post that its operating structure was “archaic.” Other assessments placed the organization’s standing closer to the middle of the pack, but either way, the Mets had clearly lost their first-mover advantage.

That advantage had taken hold around 2005, when the Mets quietly built their first proprietary analytics portal, making the franchise one of the sport’s early adopters of such technological infrastructure. Around 2009, the Mets began widely using TrackMan ball-tracking radars, once again placing them near the forefront of MLB’s next data evolution.

SOTI 2021: Join Us at State Of The Industry on March 24 & 25—It's Virtual and Free

But such advantages soon dissipated. Analytics growth stagnated and efforts to catch back up were stymied by ownership. The promise of Moneyball patriarch Sandy Alderson’s hiring as general manager after the 2010 season portended a robust analytics department, but that never materialized in either personnel or gadgetry to match the lofty expectations.

In 2011, the Mets’ director of baseball operations at the time, Adam Fisher, invited KinaTrax—a first-of-its-kind 3D motion capture technology for biomechanics analysis—for a trial. KinaTrax installed its multi-camera system at Citi Field and conducted the tests and captured the images the startup would use to sell its product for the better part of the decade. Mets ownership, however, balked at the six-figure price tag.

By the time Alderson left his post in 2018 due to health issues, the Mets employed just three full-time analysts, which by one measure placed them at the bottom of the National League. As technology has permeated all levels of the sport, the Mets often dabbled in the use of new devices but lacked a commitment to the full suite of offerings and a tech-first approach.

“They’ve fallen behind in recent years,” an industry source tells SportTechie, "but now they can start to set the pace into an expansive and exploratory area.”

Alderson returned when Cohen took over, now serving as president. A club spokesman declined to make him available for an interview, noting that the team’s plans to build out the analytics and technological infrastructure are still being formulated. So SportTechie has taken the liberty to put forth the Baseball Technology Roadmap that the Mets—or any club playing catch-up—should follow to keep pace with the Dodgers of the world. And just how wide is that chasm between the World Series champs and the distant hopefuls?

When it comes to organizational efficiency, consider the analogy put forth by Ryan Reinsel, a former Chicago Cubs minor league coach who is now chief product officer at the data management and visualization tool BaseballCloud: “Some teams have these jungle bridges that you have to tread lightly on and hopefully it doesn't break. And then you're seeing teams build these massive, cable superhighway bridges, where data and information can flow.”

Here’s how you can turn the former into the latter.

R&D Personnel

In a 2018 survey of big league clubs by The Athletic, the Mets were found to employ only three R&D analysts (which doubled to six by 2019), but the Dodgers and Yankees each had 20, still more than triple the Mets’ headcount. Adding more bodies/brains enables greater specialization of skills across key areas such as big league personnel decisions, player development, biomechanics, amateur scouting, and in-game strategy. There should be clear delineation between analysts and engineers, so the former can focus on understanding information and aren’t bogged down by time-consuming programming and database management.

Building out a robust department is critical. The average size in that 2018 analysis was 8.4 R&D analysts per club, so there’s no reason a cutting-edge 2021 department shouldn’t have at least 15. Even when so much of the data collection and analysis is automated, people still need to ask the questions that guide the work. Which step of a pitcher’s kinematic sequence portends to greater risk of injury? How much can a team compensate for a fielder’s poor range with better positioning? How would proposed changes to the baseball’s composition affect a power hitter’s production or a reliever’s wipeout slider?

Baseball Ops Data Platform

The Cleveland Indians pioneered the MLB’s first internal computer system in 2000 when they built DiamondView, a proprietary analytics platform with searchable databases. Now it’s believed that all 30 MLB clubs have their own versions—some whose names are famously public, including the Astros’ Ground Control, the Red Sox’ Beacon (née Carmine), the Cardinals’ Red Bird Dog, the Cubs’ Ivy, and the Pirates’ MITT—although assuredly with varying degrees of sophistication.

Such systems ingest vast quantities of data from dozens of disparate sources such as scouting reports, medical histories, and performance stats. Athlete management systems such as Fusion Sport’s Smartabase, Kinduct and Catapult AMS already collate biometrics, performance data, health records, and tracking data, but what MLB clubs’ baseball operations departments have built and maintain are wider in scope. They also ingest all available information on amateur draft prospects, international free agents, and every pro player with other organizations—with an eye toward trades and free agency.

The use of third-party analytics platforms has grown in popularity. In addition to incumbent players TruMedia and Sports Info Solutions, there are newer entities such as Zelus Analytics, which was founded by former Rays and Dodgers R&D boss Doug Fearing, and BaseballCloud, which spread like wildfire through college baseball. Those types of services can offer new perspectives or augment existing insights or provide validation (or rejection) of what clubs are doing internally.

John Curnutte, the VP of engineering at Zelus Analytics, has roots in the financial world—he helped build and maintain trading systems for Dimensional Fund Advisors, which has more than $600 billion in assets under management. (Cohen, of course, made his fortune through hedge funds Point72 Asset Management and, previously, S.A.C. Capital Advisors.) Curnutte says baseball and finance are “equal in terms of complexity and workload,” drawing on many different data providers. For the Mets, having an expert from Cohen’s endemic financial business audit the processes of his new baseball business could help identify inefficiencies, and contracting third-party analytics support would seem to be especially important while the team plays catch-up.

Infrastructure and Processing Power

The Hawk-Eye powered Statcast system generates 24 terabytes of data for each of the 2,430 scheduled regular-season MLB games every year, although 99% of that is raw video that doesn’t get saved for long. That data is processed on Google Cloud as part of a league partnership, but a club’s own data collection devices—bullpen pitch trackers, super slow-motion videos, biomechanics assessments, and minor league tracking systems— all add voluminous more bits and bytes for their proprietary algorithms to ingest. Analysts need to test their theories using these vast amounts of data.

Curnutte strongly endorses the use of a cloud provider. “First and foremost, it's not having to extend the time and energy managing your own infrastructure, but also being able to get access to highly scalable compute capacity,” he says, noting that Zelus uses Google Cloud. There’s great ebb and flow to baseball data, with huge intake after every game and periodic high-volume requests for, say, 10,000 iterations of a computationally complex model.

More franchises are finding use cases for deep learning neural networks, a more intricate application of AI using raw data and fewer human-entered parameters. (The Astros began using neural nets to govern their defensive overshifts nearly a decade ago.)

“If you start getting down into player-tracking data and some of the newer stuff coming out in the area of motion and kinematics, the size of data sets increase dramatically,” Curnutte says. So it’s necessary to have “data pipelines that can perform well in terms of ingesting that amount of information and data, but also keep them really well organized—and then accessible, in an easier way for running your machine learning that’s stacked on top of that.”

One of the most telling insights in the New York Post story about the Mets’ technology infrastructure was that it lacked the internal server capacity to support “30 players using their iPads simultaneously.” The Mets had built an internal app for players to access scouting reports, video clips and the like. Many tech vendors offer their cloud computing options, but enough work is done locally that every organization still needs significant processing power.

Everything needs to be accessible too.

“In the future, with the right infrastructure, you'd have it where technology can be accessed freely in real time from anywhere,” says BaseballCloud’s Reinsel. “If I throw that bullpen on a backfield, the major league coach should be able to see that live bullpen data on his phone and see what [the video] looks like whether he’s on a plane, if he’s in a car, if he's in the locker room.”

Ball Tracking

The baseball is tracked with remarkable accuracy anywhere it travels in a big league game, thanks to Statcast. Many minor league ballparks have TrackMan radars installed, and a few also have Hawk-Eye.

Bullpens and batting cages, however, need their own coverage as players and coaches increasingly want every training activity to be captured and quantified for the sake of pitch design, progress of player development, injury risk flags, and signs of fatigue. Rapsodo, HitTrax, portable TrackMan units, Diamond Kinetics PitchTracker, and YakkerTech fill this domain. For the ease of data compatibility as well as accuracy of measurement, having a uniform solution is preferred—or at least a way of reconciling, standardizing and syncing the information.

YakkerTech is the newest entry into the space and the fastest growing; as a camera-based system, YakkerTech can measure metrics such as spin rate directly and has the potential for biomechanics analysis. “We see the industry starting to really turn towards the optical-based, data-capturing solutions, and it's just going to be even more bullish on that front, from markerless motion capture to ball flight,” Reinsel says.

Edgertronic cameras have become the predominant high-speed option in baseball, particularly for studying pitching granularly. Rapdoso has added its own competitor, and other super slow-motion cameras less known in baseball can also fill the need.

Biomechanics

Few areas of baseball are gaining as much interest as the analysis of player biomechanics, both for preventing injuries and maximizing performance. The area has great potential, Reinsel says, but a lot of organizations are applying the wrong process. In this realm, the right person needs to be brought on board before the technology upgrades are purchased. “That biomechanist should be hired first, instead of being hired last,” Reinsel notes. “Let's know what we’re setting up before we have the technology dictate what we’re going to do.”

But when it comes to those technology options . . .

The OnBaseU coaching certification program focuses on three distinct but interrelated areas—strength and conditioning, medical and biomechanics, and hitting and pitching coaching—to make sure a player has the best fitness and skill set to reach his potential.

Though the Hawk-Eye system does offer player limb and body motion tracking, its highest frame-rate solution is not standard in Statcast. Some clubs can choose to install the higher-grade versions themselves while others are using other enterprise options such as KinaTrax and Simi Motion. Other systems, more commonly found in bullpens and labs, include Dari Motion, Vicon’s Capture.U, and Qualisys.

More accessible options that require only one smartphone (ProPlayAI, Mustard) or two (Uplift Capture) are gaining momentum for training and scouting purposes as the aforementioned lab and ballpark systems are too expensive to place in every minor league facility and on spring training back fields. When only an iPhone is required, coaches and scouts will be able to collect helpful data whether they are at a big league ballpark, high school field, or Dominican academy.

Wearable options from K-Motion and 4D Motion remain popular for now but may have less use in the coming years as the camera-based systems continue improving their accuracy. No matter how data is collected—by phone, sensor, or multi-camera triangulation—companies like Reboot Motion provide advanced analysis.

Player Fitness, Health and Wellness

As the dissection of in-game performance statistics continues to proliferate throughout the sport, gains will become increasingly marginal. Keeping players healthy and available to play, however, remains fertile ground for giant leaps.

Sport science is rapidly becoming a priority among baseball organizations, and more tools than ever are available for monitoring player workload. The GPS or LPS devices so common in cardio-heavy sports such as soccer, football, hockey and basketball are not quite as directly applicable to baseball, although a half-dozen MLB organizations have signed on with Catapult Sports, which has shown value in tracking hitters’ and pitchers’ rotational accelerations as well.

The best throwing tracker available is the Motus sleeve, which Driveline Baseball acquired in Feb. 2020. Motus quantifies throw counts as well as stress and torque on the elbow. For more general load monitoring, sleep tracking, and recovery assessing, the Whoop strap is popular among MLB players and is approved to wear in games. Oura provides similar metrics, although a ring is less practical (and, in fact, currently prohibited) for the particular demands of baseball.

Skill Development

Bat sensors made by Blast Motion and Diamond Kinetics are increasingly seen in elite youth ball, and as of 2018, they are approved for in-game use at all levels of the minor leagues. As a result, some players will have a decade of longitudinal swing data by the time they reach the majors. The sensors aren’t permitted in MLB games—ideally that ban is overturned soon—but they can be used in batting practice, and optical bat-tracking solutions are in development.

In addition to the pitch insights provided by the aforementioned ball-tracking technologies (Rapsodo, YakkerTech, et al.) and high-speed cameras, BaseballCloud’s BallR tool, which Reinsel designed, presents advanced metrics such as spin, seam orientation, and axis with digestible visualizations. BallR can even recommend ideal fingertip placement for throwing certain pitches.

“We can measure your hand size and relate that to how you hold the ball,” Reinsel says. "And then, in the future, you're going to be able to say, ‘Hey, I want at least 16 inches of vertical break on my curveball,’ and then be suggested, ‘Here's the best grip based off of your hand size, and based off of your biomechanics and your arm slot, that’s going to allow you to get that right spin direction to create 16 inches of vertical break.’ ”

Former Mets pitching coach Dan Warthen was so well known for instructing his starters to throw a high-velocity slider that the pitch became known as “the Warthen Slider” (even if some think it better resembles a cutter). Having that type of BallR insight, plus ball-tracking data on pitch mixes and tunneling, could help future prospects make more informed decisions about which pitches to add to their arsenal—and how those high-velocity pitches pair with big breaking curveballs, for instance.

Other baseball-specific tools, such as virtual reality hitting environments, are becoming commonplace in the sport. WIN Reality and Monsterful’s RibeeVR can help hitters visualize that night’s starting pitcher and work on getting their timing right. More and more big league clubs are minimizing the importance of traditional batting practice in favor of more competitive, game-like situations. VR can help fill that void, as can some smart pitching machines.

Vision training and decision making are also skills that can be improved. Most clubs, including the Mets, contract both a performance vision expert and a mental skills coach. Apps such as GameSense, uHIT and Vizual Edge as well as tools like Sports Vision Trainer and a Dynavision board can also help on these fronts.

Communication

No matter how good your technologies, models or theories might be, they have to have real value for the humans who are playing the game, managing the game, and making roster decisions. Critical to this endeavor is consistent messaging throughout the organization. Turning data into actionable information is only helpful when it’s communicated effectively.

Coaches remain the most important conduit in this organizational approach, and a growing number of clubs employ liaisons—either a former player or an analyst—who can translate advanced data into actionable recommendations. Selective dissemination of this info is also important, with those liaisons making judgment calls based on their personal relationships with players.

Furthermore, each organization needs to instill the same vernacular, the same principles, the same KPIs, and the same philosophies in A ball as they do in the big leagues and the scouting war room. That will make advancement through the farm system more attainable for players when the expectations are clear, and it will make the transitions smoother because the terminology and ideas will be the same at each level. It also can help coaches, scouts, executives, and medical practitioners communicate more effectively across different levels and departments of the organization.

No ballplayer, now matter how strong his interest in analytics, wants to be overwhelmed with numbers—especially in the midst of the action. Most people, in baseball and otherwise, say they are better visual learners, which means there should be heavy emphasis on video—to this end, Synergy Sports Technology accounts are essential—and converting tabular charts and graphs into colorful visualizations and even immersive environments. BaseballCloud does this effectively in the public domain, and many clubs have their own proprietary platforms for doing the same.

And it may not be long before that those visualizations could even reach, say, a bullpen where a reliever might be wearing Google Glass or some other AR headset showing hot and cold zones over the plate for the batter he is about to face in the game. “Adding more context within the environment is going to allow us to train and develop players at a better level and a better rate,” Reinsel says, “and I think we will be able to do that inside of VR and AR.”

Question? Comment? Story idea? Let us know at talkback@sporttechie.com

‍