In a memorable episode of The Simpsons, the cunning Mr. Burns recruits actual Major League Baseball players for his company’s softball team to win a wager. However, just before the championship game, he substitutes the eight-time National League All-Star Darryl Strawberry with Homer Simpson, illustrating that sometimes even the best analytical decisions can lead to unexpected outcomes.
“You’re a left-hander, and so is the pitcher. If I send up a right-handed batter, it’s called playing the percentages,” Mr. Burns explains to Strawberry. “It’s what smart managers do to win ballgames.”
Baseball, particularly at high levels, relies heavily on analytics, with teams employing numerous data engineers to delve into intricate statistics that steer managerial decisions. Yet, like Mr. Burns, there’s a risk of oversimplifying clever analysis into absurdity.
The Oakland Ballers, a pioneering independent team in the Pioneer League, has taken the concept of statistical strategy a step further: they allowed an AI to manage a game.
Founded by edtech innovator Paul Freedman, the Ballers emerged as a response to the Oakland Athletics’ controversial relocation—an event perceived as a critical loss for local fans. Surprisingly, the Ballers, affectionately abbreviated as the Oakland B’s, have cultivated a passionate community of supporters who rallied against the A’s exit. In just their second season, they clinched Oakland’s first baseball championship since 1989.
According to Freedman, “The Oakland Ballers uniquely know what it’s like to function like a major league team in a minor league environment. Our creativity and experimental nature allow us to explore advancements that major leagues would approach with caution.”
Minor league teams frequently play the role of testing grounds for cutting-edge technologies before they’re adopted by the majors—think instant replay systems and automated ball-strike techniques. The Ballers have embraced this spirit, especially with Freedman’s tech background, while adding elements of playful experimentation that traditional teams might shy away from.
Last season, the Ballers collaborated with Fan Controlled Sports, enabling fans to dictate managerial decisions in a thrilling game. The Ballers lost, partly due to fans favoring whimsical decisions over strategic ones, like having a pitcher pinch hit.
This time around, upon securing a postseason spot, they partnered with AI company Distillery to implement AI software designed to manage a baseball game in real time.
“Baseball provides an ideal setting for a trial like this because it is fundamentally data-driven, and decisions can be made rapidly after each pitch,” Freedman noted.
Distillery trained OpenAI’s ChatGPT using over a century’s worth of baseball statistics to mimic the decision-making process of the Ballers’ human manager, Aaron Miles.
“The AI’s role was to replicate the decisions our human coach would make, blending strategic creativity from Miles with analytical capabilities to identify patterns,” Freedman explained. “Our human ingenuity remains safe; AI acts as an optimization tool rather than a replacement.”
The AI-managed game unfolded smoothly, with the AI making identical choices regarding pitching substitutions, batting orders, and pinch-hitting that Miles would have chosen. The only deviation came when Miles needed to substitute the starting catcher with a backup due to illness.
Even as AI took temporary managerial control, Miles displayed a light-hearted approach to the experiment, humorously extending the AI tablet for a handshake before the game. However, the AI initiative elicited mixed feelings among Oakland fans, many of whom perceived tech companies like OpenAI as prioritizing technological advancement over reliable, tested products. This sentiment mirrored their feelings about the numerous sports teams leaving the city.
“It seems like the Ballers are catering to Bay Area tech enthusiasts rather than traditional baseball fans,” lamented one commenter. “It’s over for Oakland.”
This unexpected backlash was not anticipated by the Ballers, and Freedman has no immediate plans to replicate the AI experience. Yet, it highlights a broader cultural conflict in baseball and technology. “It’s disheartening to hear our fans express dissatisfaction,” Freedman acknowledged. “However, it’s crucial that we engage in dialogue about the benefits and drawbacks of these new technologies while there’s still time to reflect.”
