Statcast at 10: From MLB’s secret project to inescapable part of modern baseball

The email landed in Cláudio Silva’s inbox on the evening of Dec. 6, 2011. One of the first things he noticed was the three letters in the subject line: MLB.

Silva was an NYU professor who specialized in data science and computer graphics. He had once worked at AT&T Labs and IBM Research. Those were initials he understood. But MLB? Silva grew up in Fortaleza, Brazil, a coastal city where baseball had little relevance. When he got his doctorate at the State University of New York at Stony Brook, he never bothered to learn the rules.

The email was written by Dirk Van Dall, who was working with Major League Baseball Advanced Media (MLBAM), the league’s digital arm. It was forwarded to Silva by Yann LeCun, another NYU professor and one of the world’s foremost experts on machine learning. Silva read the first few lines. It concerned a secret project in the works. “MLBAM is working with a vendor on technology to identify and track the position and path of all 18 players on the field,” Van Dall wrote. The problem, he continued, was that the resulting firehose of data would need to be compressed, coded and organized on the fly for use by broadcasters, analysts and coaches.

Van Dall didn’t mention the project could revolutionize the sport, transforming the way teams evaluate players or how fans watch games. Nor did he use the project’s eventual name: Statcast.

Leave a Comment