The Next Moneyball: Artificial Intelligence in Baseball

A rundown of how recent (and not so recent) technology has changed the game

Reading Time: 3 minutes

When Billy Beane decided to employ a recent Harvard graduate to use advanced statistical analysis to build a championship baseball team, he changed the game forever. While Beane’s famous early 2000s team never won a World Series, multiple 100-win seasons and a new record for the longest winning streak got the attention of teams across the MLB, all while on one of the league’s lowest payrolls.

Most people know Beane’s story as it was popularized in the book—and later in the movie—Moneyball. The machine learning techniques that were used to algorithmically determine a player’s value were light-years ahead of the archaic methods that had been used in baseball up to that point. Previously, player analysis had been based entirely on stats that had a large factor of luck behind them, with an example being the RBI (Runs Batted In). While the RBI has historically been used to measure the value of a hitter, it largely depends on whether the player is at-bat with runners on base. For example, a home run with two runners on base gives a player three RBIs, but a solo homer only gives a player one. There’s no clear difference in what the player actually hit, so a stat that gives the player in the first scenario more credit is inherently flawed. Instead, Beane used stats like Fielding Independent Pitching (FIP) to get the best team possible. FIP is a metric that only uses stats that a pitcher has perfect control over (walks, strikeouts, and home runs) to take any element of luck out of the equation. However, stats like FIP were just the first step.

While Beane was the first to really integrate Artificial Intelligence (AI) into sports, the techniques he used were not actually all that advanced. Today, sports are on the cutting edge of AI development. A prime example is MLB’s StatCast AI, which uses Amazon Web Services (AWS) to calculate real-time statistics such as launch angle, pitch speed, and even the likelihood of a catch being made. Data from StatCast has opened the game up to a whole new level of statistical analysis, with players like Chicago Cubs star Kris Bryant checking data in between innings to make sure they’re at the top of their game.

Nowadays, every team in the MLB has its own statistical analysis department, and they use data from StatCast and other sources in ways that would have been unimaginable 15 years ago. One example of this is the prevalence of the shift in today’s game. While teams have long employed the shift (moving an extra player one side of the field) against left-handed hitters, today’s data allows for individualized shifts for each hitter on an opposing squad. If you watched any of the Boston Red Sox’s games last year, you’d have noticed the Red Sox outfielders looking at little cards in their back pockets in between at-bats. These cards tell players where to position themselves for each hitter, using data straight from the Red Sox data department.

Artificial Intelligence and data analysis are certainly revolutionizing the way the game is being played at the professional level, but they are also having a major effect on college and even high school baseball. This is through technologies like those provided by Rapsodo. Through the use of high-speed cameras and machine learning techniques, Rapsodo’s products can give real-time metrics like spin-rate on pitches, as well as simulate the exact distance a ball hit inside a batting cage would travel on a real field. In many ways, this is even more impressive than StatCast, because it allows people who aren’t getting paid millions of dollars access to baseball AI. In fact, Stuyvesant’s own baseball team has been able to use a Rapsodo machine.

Computer science and sports have long been thought of as polar opposites. However, machine learning and data analysis are making athletes who choose to listen to the data better on a daily basis. This piece of advice that used to be ridiculous is now perfectly plausible: if you’re looking for a career in sports, majoring in CS is the way to go. From potentially adding robot umpires to expanding the StatCast system, AI in baseball is here to stay.