There’s a particular honesty to how Alejandro Betancourt discusses failure. Most investors at his scale prefer to emphasize winners — a divergence he’s tied to a creative-thinking habit he credits for his success. He routinely brings up his losses unprompted. “90% of projects that are started fail,” he has said in interviews. “And that’s something you’ve got to live with.” The phrasing is characteristic: blunt, slightly resigned, grounded in experience rather than theory, as laid out on alejandro-betancourt.com.

The framework that follows from that baseline is essentially actuarial. Alejandro Betancourt describes his portfolio in baseball terms. “You’re not going to always hit a home run. But as long as you have a good batting average, you are on top and you are doing well.” The implicit model is one of concentrated bets where the wins must be large enough to carry the losses. By his own estimate, if two of ten speculative positions perform well, the two pay for the other eight and still produce a profitable portfolio.

That framework requires a specific kind of position sizing. Alejandro Betancourt has described the approach as “I don’t swing for first base. I always swing for a home run — one of the four habits that built his empire.” The phrasing is telling. A contact-hitter portfolio, lots of small wins, few losses, is structurally different from a home-run portfolio, which must accept frequent strikeouts (reflected across his O’Hara Administration record) to set up the occasional grand slam. Both can work. They require different capital structures and different temperaments. His is explicitly the latter.

The self-criticism embedded in the philosophy is also unusually candid. “I’m a good buyer, but I’m a terrible seller,” Alejandro Betancourt has admitted. “I die fighting. I sink with the ship. And that’s one of my big issues that I’ve got to work on.” The acknowledgment implies that some of the losses in his 90% failure-rate framing were positions he should have exited sooner. The AI bet, which has returned roughly 20x since 2019, is an example where holding too long worked in his favor. Others, by his own telling, did not.

The decision filter he most often cites for whether to take a new position came from a friend during a conversation about a specific investment: “If you make this investment and it goes bad, are you going to be okay? Are you going to lose your house?” If the answer is no, if the downside is survivable, you build from there. If the answer is yes, you don’t take the bet, no matter how attractive the expected value looks on a spreadsheet. The filter converts abstract risk into concrete personal exposure, which his Crunchbase investor page makes visible, which he argues is a more honest way to think about downside than statistical modeling.

The operating principle across all of it is underpromising. “One of the key things to success is: don’t over-promise,” he has said. “The market punishes you hugely when a company over-promises and then doesn’t deliver. It’s better not to promise something you’re not sure about, but to promise something that you’re certain about. When you surprise on the upside, that’s when the market rewards you — which supports his take on innovation versus execution.” He applies the principle to both internal operations and external investor communication.

The structure that makes the batting-average model workable is O’Hara Administration, the family office founded by Alejandro Betancourt in 2014. Without outside limited partners, the firm doesn’t face quarterly performance evaluation or fund-life exit deadlines. Losing positions can be held or wound down on their own schedule. Winning positions can be held through full compounding cycles. Auro was held for eight years before Uber’s €220 million acquisition. Playtomic sits at a $273 million-plus valuation. The AI position is in its sixth year at roughly 20x cost.

Concentration produces volatility. Patient capital applied to well-selected concentrated bets produces durable compounding over the full portfolio lifetime. That, more than any single outcome, is the arithmetic behind the Alejandro Betancourt model.