Moneyball in the Age of AI
"Moneyball: The Art of Winning an Unfair Game" by Michael Lewis
My first thought reading this book was, “Wow, the American public really used to care a lot about baseball.”
The amount of technical baseball knowledge that is needed to get through this book really surprised me given its popularity, clocking in 1.7 million copies sold at last count in 2016, and a 2011 movie starring Brad Pitt.
But lucky for me, I have a lifelong passion for baseball, and lucky for you, this article isn’t about that.
To briefly summarize the plot, Michael Lewis brings us into the dugout of the Oakland A’s, whose management shook up the game by applying data-driven insights to assemble a surprisingly winning team. By applying cold, hard analysis and an f-you attitude to baseball’s establishment, the As’ general manager Billy Beane used the League’s smallest budget to win game after game and churn out overlooked players, turning them into stars.
Much of the discussion centered around how the romance of baseball clouded the judgement of talent scouts and led to an inefficient market for players. By contrast, Beane and his cohort used unusual techniques to uncover what, exactly, led to winning games.
Amidst the multi-million dollar contracts and flashy homeruns, their data crunching revealed a metric that was completely underappreciated by conventional thinking: on-base percentage.
They didn’t care how a player looked as long as he got on base (to paraphrase: “He’s not selling jeans”). More players on-base meant more positioning that set up the team to bring more players past home plate. More players passing home plate equates to more runs, more wins, more happy fans.
Crucially, they recognized overlooked players who had the sagacity to read pitchers well enough to make it to the plate via walks – a free pass to first base when the pitcher fails to throw enough pitches into the acceptable strike zone. It requires maturity, prudence, and patience in a hitter. It’s not as sexy as a home run, but hey, it keeps the lights on.
That’s enough baseball for now.
What I want to highlight is the way that we are about to step into a universe where everything has the potential to become Moneyball. As in, insights gleaned from easily accessible and super duper AI tools have the potential to disrupt just about any field.
I wrote about the effect of AI on the US stock market and its elucidating and revealing effects on my own life, but one thing I’ve noticed is the repeated concern that we still don’t really understand how, exactly, AI will justify the insane valuations that it currently claims.
Even as recently as in August of this year, the Economist wrote about the flailing support investors were showing and the surprisingly few American businesses that use AI to produce goods and services (5.1%, apparently).
Hardly a home run.
In my opinion, the Cloud computing cycle didn’t do all that much more than bring stuff that was offline, online. That’s not true innovation, that’s just superior accessibility and organization.
But AI’s edge lies in its ability to surface new insights that we currently aren’t aware of, such as my own experience I detailed in this prior piece.
I think Moneyball is a useful heuristic to consider AI’s potential. Specifically, crunching data enabled Beane to appreciate new metrics that unlocked enormous value for his team.
Here’s an example. Reading this reminded me of a podcast episode between Dr. Brené Brown and researcher Dr S. Craig Watkins, in which they discussed AI and race. Dr. Watkins recounts:
“So there was this study that some colleagues of mine and others did at MIT where they were looking at medical images. And they essentially trained a model to be able to try to predict if the model could identify the race and ethnicity of the persons whose medical image this belonged to, if you strip the image of any explicit markers of race or ethnicity. And the model was still able to predict with high degree of accuracy, the race of the patient who was imaged in this particular screening…
The story here is that these models are picking up on something, some feature, some signal that even humans can’t detect, human experts in this space can’t detect. And what it suggests is that our models are likely understanding and identifying race in ways that we aren’t even aware…”
Naturally, this has enormous application in domains where race can be used to make potentially detrimental life-impacting decisions, such as in hiring and creditworthiness assessment, many of which are already unfair games. These are serious risks that must be understood and addressed.
But this lodged in my mind and stuck the landing because it was an example of algorithms picking up on the paranatural, things that exist but we just don’t see and understand yet.
And that is terrifically exciting to me.
But wait – Moneyball certainly has its critics. As writer Benjamin Charles Germain Lee wrote in Jacobin,
“The mechanization, quantification, and financialization of everything undoubtedly renders life less warm and dimensional, cramming it into spreadsheets while also fine-tuning the machines of exploitation and inequality. Maybe such stories are actually cautionary tales, in which case we’re cheering at our own submission to the tyranny of numbers with dollar signs in front of them.”
To be fair, Billy Beane wasn’t playing Moneyball because he wanted to take the romance out of the game, but because he had to be thrifty with his talent budget. But just as Beane used data to challenge baseball’s traditions, we must harness AI to uncover hidden value while ensuring the metrics we prioritize actually lead to wins in the game of life.
I think many of us are deeply disturbed by the way that we fear AI can make many of the worst aspects of society, well, worse. Michael Lewis has become a cautionary tale himself, as his success-sniffing led him to the vaunt the rise of Sam Bankman-Fried, and we know how that ended.
Whatever Moneyball Lewis had mastered could use a software update.
I am 100% all for rendering life warmer and more dimensional. That’s why I choose a position akin to sitting in the bleachers like an Oakland A’s fan in late baseball season of the early 2000s, cheering with surprise and delight that my team is winning.
I look forward to the ways in which applications of AI can partner with us to truly move the needle in the domains where we desperately need new efficiencies – in health, in climate science and engineering, even in app-based dating, for crying out loud.
In fact, dating is a great example of an area where people who struggled to find a like-minded person in their community suddenly had the chance to connect to many more potential partners, to enormous early success. But the financialization of the dating industry has led to its dire enshittification.
I could use a hidden metric or two to link me to my soulmate. Now that would be some real value creation.
And as for the Oakland A’s, the story has recently taken a turn for the less romantic. They’re leaving Oakland for Las Vegas.