Amazon, MLB add machine-learned stats to cloud deal

Amazon.com Inc will provide a new set of real-time statistics and graphics on live baseball games later this season, expanding its cloud computing deal with U.S. Major League Baseball, the two organisations told Reuters.

The agreement, which makes use of Amazon Web Services’ artificial intelligence and machine learning capabilities, follows a similar deal with the National Football League in November. Neither side disclosed financial details.

Amazon and MLB are hoping the new stats will give fans insights as they follow games on TV and online. A new logo and branding will show off Amazon’s machine learning technology to a wider audience.

One stat under development is a real-time pitcher heat map that changes to reflect specific situations such as who a pitcher is facing, what stadium they are at, the time of day, and whether a team is still in the playoff hunt, to show fans where the next ball may be pitched.

“All these multiple variables that basically we can’t keep in our heads and are not easy to calculate manually, we can now feed those into these large, cloud-based machine-learning systems and see what emerges,” said MLB Chief Technology Officer Jason Gaedtke.

The global cloud infrastructure market is forecast to be worth nearly $82 billion in 2018, according to research firm Canalys. AWS accounted for 32 percent of the market in the first quarter of 2018, followed by Microsoft Azure with 16 percent and Google Cloud Platform with 7 percent, according to Canalys.

“Sports leagues make good reference customers for cloud providers,” said Blair Hanley Frank, principal analyst at ISG, a technology research and advisory firm.

“They’re large, high profile enterprises with complex needs and interests in developing new digital experiences for consumers that translate well to cloud usage.”