DraftKings is taking heat for using machine learning to hunt the players likeliest to lose everything. But if you look at the fundamentals, optimizing for the structurally doomed is simply best-in-class product-market fit.
The financial press is in hysterics again, and as usual, they are punishing a visionary American company for achieving exactly what it set out to do.
DraftKings has successfully deployed advanced machine learning to identify the gamblers statistically likeliest to lose their shirts, actively targeting them with bespoke betting incentives. Meanwhile, the company has politely declined to use that same technology to flag problem gambling or restrict users who are spiraling. The predictable chorus of consumer protection advocates and bleeding-heart ethicists are calling this predatory behavior. I call it a masterclass in total addressable market optimization.
We live in an era where every major tech platform claims to want deep user engagement. But when DraftKings finally achieves it—by building a hyper-efficient algorithmic heat-seeking missile that explicitly targets a vulnerable father of three in Ohio and hits him with a $200 risk-free parlay on a Tuesday night MAC football game right as his willpower dips—suddenly it is treated as a corporate crisis.
I was discussing this over a $400 omakase lunch in Palo Alto recently with a prominent venture capitalist who sits on the boards of several consumer fintech startups. We both agreed that the public outrage fundamentally misunderstands the purpose of enterprise software. The goal of artificial intelligence is not to act as your moral compass; it is to identify inefficiencies in the marketplace and aggressively close them. In this particular case, the inefficiency is that some historically unlucky plumbing contractor still has discretionary income sitting idle in his checking account.
The critics argue that since DraftKings possesses the data architecture to predict exactly who will chase their losses into personal bankruptcy, the company has a moral obligation to intervene. This is a breathtaking misunderstanding of shareholder value.
You do not ask a copper mining conglomerate to build a sophisticated algorithm that warns the mountain it is about to be blown to pieces. You ask them to extract the ore as cheaply and efficiently as possible. DraftKings has simply realized that the human psyche is a highly porous, easily exploitable rock formation, and their data science division has developed the perfect drill.
The sheer elegance of the targeting parameters should be studied in business schools. The algorithm does not simply look for regular bettors. It sifts through terabytes of behavioral data—frantic late-night deposit patterns, the speed at which a user chases a lost bet, the tendency to wager on obscure Belarusian table tennis matches when down on a Sunday—and finds the unique intersection of impulsivity and a complete lack of statistical literacy. It then serves up push notifications with the precision of a surgeon extracting a kidney.
If we built a machine learning model to stop people from making terrible financial decisions, we wouldn't be a sportsbook, we would be a fiduciary, and the market multiples on fiduciaries are frankly garbage.
Exactly. Haverford understands that a publicly traded entity cannot voluntarily leave yield on the table just because that yield happens to be earmarked for a child's orthodontics.
To demand that DraftKings reverse-engineer this beautiful machine to protect the consumer is to demand they actively sabotage their own profit margins. It is asking the great white shark to invent a Kevlar swimsuit for the seal. Why would we punish a company for successfully identifying its core demographic—the structurally doomed—and servicing them flawlessly?
If anything, DraftKings should be syndicating this technology rather than apologizing for it. Imagine the synergies if credit card companies, subprime auto lenders, or predatory mortgage brokers could license this predictive engine. They could find exactly which stressed consumers are one bad push notification away from taking on a 30 percent APR loan to cover a margin call. It represents the holy grail of frictionless wealth transfer at scale.
Unpopular opinion: we need to stop treating the mathematically weak as victims requiring government protection, and start recognizing them as a vital, renewable resource for institutional investors. DraftKings has not built a predator; they have built a highly efficient mechanism for relocating capital from those who do not understand probability to those who do. As a shareholder, I couldn't be prouder of the innovation.