Amazon and Major League Baseball get tighter

In This Article:

Last week, Amazon Web Services and Major League Baseball issued a joint press release that was easy to gloss over at first as something pro forma: it said that the two have “extended their relationship” and that as part of the extension they would “leverage artificial intelligence and machine learning.” No one would be blamed for shrugging; MLB is one of many different entities that uses AWS, after all.

But what MLB and AWS are aiming to do could actually prove vital to pro baseball fandom. It is also yet another sign that Amazon has quietly become meaningfully entrenched with the major pro sports leagues.

MLB has been using AWS since 2015 when the two partnered up on Statcast across all 30 ballparks. (But MLB does not use AWS as its exclusive cloud storage provider.) Most fans may not even notice the AWS logo, but when detailed stats appear on-screen during a game broadcast to highlight, for example, the distance and velocity of a home run, a little box notes that Statcast is “powered by AWS.”

According to an AWS case study, here’s how that actually works. In all 30 MLB parks, a radar now sits behind home plate that pinpoints the position of the ball 2,000 times per second, and two advanced cameras above the third-base line pinpoint the positions of the players on the field 30 times per second. Attached to this data is a quick description, written by a human, of every single play on the field right after it happens.

All that data is transmitted and sent to the AWS cloud within 15 seconds after a play ends. That’s why a game broadcast can present Statcast data points only a minute after a big homer. An estimated 7 terabytes of Statcast data is generated per MLB game.

So, what’s new here? The partnership extension announcement touts AI and machine learning—what will that actually mean?

“We are exploring a range of use cases,” MLB CTO Jason Gaedtke tells Yahoo Finance. Those include: pitch sequence analysis that will take into account “new parameters reflecting game context”; using machine vision to recognize players and other objects within the video feed; and full automation of scorekeeping, game notes, and pitch classification.

In other words: more automation is coming to baseball stat-tracking.

MLB Statcast in action during a July 26, 2018 Texas Rangers vs Oakland Athletics game. (Courtesy MLB)
MLB Statcast in action during a July 26, 2018 Texas Rangers vs Oakland Athletics game. (Courtesy MLB)

Of course, some baseball purists may get concerned when they hear that. There are certainly widespread fears about the continued automation of every business. But these new capabilities sound like they could make watching baseball games on television (or on any screen) more interesting—a vital priority for MLB these days.