MLB Advanced Media (MLBAM) wanted a new way to capture and analyze every play using data-collection and analysis tools. It needed a platform that could quickly ingest data from ballparks across North America, provide enough compute power for real-time analytics, produce results in seconds, and then be shut down during the off season.  It turned to AWS to power its revolutionary Player Tracking System, which is transforming the sport by revealing new, richly detailed information about the nuances and athleticism of the game—information that’s generating new levels of excitement among fans, broadcasters, and teams.


It was a legend-making play for fans of baseball—a sport built on legends going back 150 years. In the third inning of the winner-takes-all Game Seven of the 2014 World Series, the San Francisco Giants and Kansas City Royals were tied at two. The Royals’ Eric Hosmer hit the ball hard, driving it toward centerfield. If the ball cleared the infield, the hit could have sparked a rally.

But Giants second baseman Joe Panik made an amazing dive to snatch the ball, resulting in two outs—including Hosmer, who was thrown out at first base after a diving attempt to beat Panik’s throw. A possible Royals rally fizzled, and the Giants went on to win the game—and the World Series—by a single run.

Panik’s play fueled plenty of talk on social media and in bars and broadcast booths. But more details about the play emerged from a system hosted in the cloud—a new big-data solution called the Player Tracking System, which MLB Advanced Media (MLBAM) created using Amazon Web Services (AWS).

The solution, which revealed Hosmer could have made it safely to first base by running through the base instead of diving, captures and analyzes the subtle complexities of every play in games. Launched into full production at all 30 MLB ballparks for Opening Day of the 2015 season, the Player Tracking System is generating new excitement with data delivered within seconds after the action occurs, including information sent to broadcast companies under the brand name “Statcast.”

Joe Inzerillo, executive vice president and chief technology officer of MLBAM, says AWS was key to making Statcast a reality.

“Consumer behavior is changing. It’s going online, it’s going mobile, and this kind of technology is crucial for the game to evolve,” he says. “One of the most exciting things we’ve worked on is Statcast powered by AWS. For the first time, we can measure things we’ve never been able to measure before.”

  • AWS Services Used

  • The Benefit of AWS

    • AWS can handle data streams from fluctuating game schedules across the country
    • Can ingest, analyze and store 17+ petabytes of data per season
    • MLBAM can scale down during off days and in the off season
    • Delivers new ways fans, broadcasters, and clubs can analyze plays and players
    • Data can be used for broadcasts, MLB apps

  • About MLBAM

    MLBAM is the digital services division of Major League Baseball. The company operates the official website for the league and the 30 Major League Baseball club websites via MLB.com, which offers news, standings, statistics, and schedules, as well as live audio and video broadcasts for subscribers. MLBAM also owns and operates MLB Radio and BaseballChannel.TV, and either runs or owns numerous other websites such as Minor League Baseball, YES Network, SportsNet New York, and the World Championship Sports Network.

Statcast data overlaid on video of the Panik-Hosmer play

Data from the Player Tracking System (Statcast) overlaid on video of the Panik-Hosmer play. The red section on the right shows that if Hosmer had maintained his speed instead of diving to the bag, he would have been safe by about a foot.

Data plays a huge role in baseball, with fat volumes of statistics cataloging the game’s arc over the seasons. This information, however, is historical and static. MLBAM wanted to change its approach to statistics by capturing and analyzing data in real time to reveal greater subtleties about the sport.

MLBAM considered an on-premises IT solution, but ultimately ruled it out. “We looked at using compute capabilities in all the stadiums,” says Dirk Van Dall, MLBAM’s vice president for multimedia technology development. “But distributing the data efficiently and from so many locations would have involved a lot of time and investment in expensive IT resources that would sit idle for about half the year.”

The AWS cloud offered an ideal alternative that could support as many as 15 games on a single day—and some days with just one or two.

“AWS provides nationwide coverage for reasonable round-trip times for sending data between the game sites and the cloud, and multiple services that we used for building Statcast,” Van Dall says. “It provides great scalability, so we can burst when we need it the most, manage just one, two, or many games on a single day, and then shut down the resources during the off season.”

The workflow begins with two data-acquisition systems at the stadiums that provide coordinate information. A Doppler radar system sits behind home plate, sampling the ball position 2,000 times a second. Two stereoscopic imaging devices, usually positioned above the third-base line, sample the positions of players on the field 30 times a second. Data from these systems is augmented by brief written descriptions of each play entered by personnel on the field after the action is over.

Ten to 15 seconds after a play is completed, the data is transmitted over private networks at the stadiums, aggregated, and then sent to the AWS cloud using AWS Direct Connect, which provides a dedicated network connection for rapid data delivery. MLBAM uses Amazon Elastic Compute Cloud (Amazon EC2) for the compute power behind the solution. The coordinate data from each play is stored in Amazon Simple Storage Service (Amazon S3), which will expand to hold the vast amount of information generated through the solution. MLBAM anticipates that an average of 7 TB of data will be generated per game. With 2,430 games in a season, that’s about 17 petabytes of data each season.”

MLBAM uses Amazon ElastiCache to temporarily store game information in memory caches instead of on hard drives, which enables fast retrieval of the data for analysis tasks. Amazon DynamoDB powers queries and supports the fast data retrieval required, while Amazon CloudFront delivers a scalable solution to serve up the APIs.

AWS Lambda, a serverless computing service that runs code in response to events, supports analysis of data feeds in the solution’s metrics engine. “Lambda is really clever. It’s where we take the raw data, do some cleaning up and error detection, then create the metrics that bring more insights into plays—the throws, the player’s acceleration rate, the top running speeds,” Van Dall says. “We’re accessing a truly big data mine, and have yet to scratch the surface."

The analysis happens within milliseconds after the data is received, a key for broadcasters to take delivery of the raw metrics and video within 12 seconds after a play is complete.


The Statcast architecture powered by AWS. Click on the image to enlarge.

Speed, scalability, and the ability to capture, analyze, and deliver large quantities of data in different ways are central to MLBAM’s efforts to innovate for the benefit of everyone who loves the game, especially fans who now have reliable metrics for those arguments about who, for example, runs the bases most efficiently or has the fastest reaction times to fielding line drives.

“We’re giving fans empirical information to power that conversation, which is a huge part of what sports is all about,” says Inzerillo.

Broadcasters also have new information to use for on-air analysis, further enhancing viewer engagement, while clubs have new data and tools to analyze and coach players.

“We believe the Player Tracking System powered by AWS will deliver new and more exciting information to apps and devices, and that will appeal to a younger generation of fans, who are used to video games and who have a lot of expectations about the viewing experience,” Van Dall says. “It delivers a new level of excitement to baseball.”


To learn more about how AWS can help you capture and analyze big data, visit our Big Data details page.


Joe Inzerillo describes how AWS empowers MLBAM to innovate.


Joe Inzerillo discusses how AWS helps MLBAM get products to market quickly.