Counting Cards in Venture
What the MIT Blackjack Team Can Teach Us About Venture Capital
There’s a scene embedded in the mythology of the MIT blackjack team — later immortalized in the film 21 — that I sometimes think about as a venture investor. A group of mathematically gifted students walks into a Las Vegas casino and, night after night, wins. Not because they’re exceptional card players. Not because they have some supernatural ability to read a room. But because they had the intellectual honesty to admit that blackjack is, at its core, a game of randomness — and then systematically identified the few variables they could actually control.
The rest of the casino crowd was playing the game. The MIT team was playing the odds.
I think about this constantly when I think about how most venture capital is practiced today.
The Illusion of Control
Let’s start with an uncomfortable truth: venture capital has a lot of variables, and investors don’t control most of them.
Take the binary question of whether a company exits at all. Either it happens or it doesn’t. You can add value as a board member, make introductions, help with recruiting — but ultimately, you don’t control whether the stars align for an acquisition or a public offering. The exit price? Even further out of your hands. That’s a function of market conditions, comparable transactions, and the appetite of strategic buyers at a particular moment in time. The amount of capital a company raises over its lifetime — and the dilution that comes with it — is driven by the company’s business needs, the competitive environment, and the fundraising climate, not your preferences as an early investor.
Most VCs spend a disproportionate amount of energy trying to influence variables they fundamentally cannot move. They run complex models, execute standardized playbooks, try to “help” in ways that may add marginal value but don’t structurally change the distribution of outcomes. This isn’t cynicism — it’s math.
The MIT team understood this intuitively. Blackjack, in any given hand, is a losing game. The house has a built-in edge. You can play perfectly, memorize basic strategy, make every statistically correct decision — and still lose. I wrote about this in a piece on being right and still loosing. That’s not a flaw in the game. It’s the architecture.
So the team didn’t try to become better blackjack players. They tried to change the conditions under which they played.
Finding the High-Leverage Variables
Card counting is, at its essence, a discipline of attention. The team tracked which cards had already been played. As the deck depleted of low cards — cards that benefit the dealer — and grew rich with high cards — cards that benefit the player — the probabilities shifted. The game that had been systematically stacked against them became, in certain moments, systematically stacked in their favor. And in those moments, they bet big.
They didn’t try to control the shuffle. They didn’t try to control which cards came out. They waited until the randomness was working for them, and then they maximized their exposure.
In venture, there are analogous high-leverage variables — and they are fewer than most people think. After years of investing, I’ve come to believe the three that matter most are ownership, price of entry and n (the number of companies in a portfolio).
How much of the company do you own? What did you pay for it? How many companies are in your portfolio?
These are the variables that sit at the intersection of your control and outcome sensitivity. Unlike exit timing, exit price, dilution, or macro conditions, ownership percentage and entry valuation are things you actually negotiate. N is something that you deliberately execute. They’re decisions you make before the randomness of the startup journey plays out. And they have an outsized, compounding effect on your ultimate returns.
If you own 15% of a company that exits for $500M, your portfolio is in a fundamentally different position than if you own 3% of the same company at the same exit. The binary outcome — the exit itself — was the same. Everything downstream of your initial ownership decision was the same. But your returns are a factor of five apart.
This is why we structure our investments the way we do. We lead seed and series A rounds — typically $3 to $5 million — with the explicit goal of buying meaningful ownership at the earliest stage of product velocity. We’re not trying to be too clever about which companies will succeed. We’re trying to be disciplined about the conditions under which we invest.
The Humility Prerequisite
Here’s where most venture investors get it wrong, and where the MIT team got it exactly right: you have to accept the randomness before you can work with it.
The blackjack team didn’t sit down at the table believing they could will themselves to draw face cards. They accepted that any given hand was largely outside their control. That acceptance is what allowed them to be patient — to count, to wait, to hold back — until the moment when the deck was in their favor. And then to act decisively.
Venture capital requires the same humility, and it’s rarer than you’d think.
There’s a version of venture hubris that says: “We pick winners. Our judgment is the edge. We can identify the exceptional founders and back them before anyone else sees it.” And there’s something to that — pattern recognition, founder assessment, market intuition are all real skills. But they are not sufficient explanations for top fund returns. Macro conditions, timing, luck, the random variable of which competitor raises a massive round or which acquirer decides to go on a shopping spree — these are enormously consequential and almost entirely outside any investor’s control.
The investors who accept this tend to make structurally better decisions. They don’t overpay for hot deals. They don’t over-concentrate on a handful of bets because they’re “sure” about them. They build portfolios sized to let the underlying probability distributions express themselves.
Playing Long Enough to Win
The MIT team had one final insight that tied it all together: the math only works if you are able to stay in the game long enough.
On any given night, they could lose. The cards could run cold. The shuffle could go against them. But over a sufficient number of hands — across enough sessions, in enough casinos — the underlying probabilities would assert themselves. If you’ve shifted the odds in your favor by 2%, and you play a thousand hands, you will make money. Probably not on hand 17. But across the distribution.
This is portfolio construction in venture capital. The “n” in your portfolio isn’t just a number — it’s the mechanism by which you convert probabilistic edge into realized returns. If you have genuine ownership discipline and price-of-entry discipline, and you back a sufficient number of companies (we target 25 to 30 per fund), the power law mathematics of startup outcomes will work in your favor. You’ll have zeros. You’ll have singles. And you’ll have outliers that return the fund many times over.
Our historical data — and frankly, a lot of the academic research on venture returns — bears this out. Concentrated portfolios with high conviction but low ownership tend to underperform. Disciplined, systematically constructed portfolios with strong ownership targeting tend to generate consistent 4 to 6x fund returns.
This is also why firm-building matters as much as deal selection. In a recent post on building championship teams, I argued that genuine advantages in venture compound in exactly the same way: superior pattern recognition leads to better selection, which leads to better returns, which strengthens LP relationships, which opens better deals. The math only works if the system behind it is built to last. You have to be able to stay at the table.
Why Doesn’t Everyone Do This?
This is the question I get asked most often when I explain our strategy. If counting cards works, why does anyone ever play blackjack any other way?
The answer, in both cases, is probably ego.
Most blackjack players believe they have an intuitive edge. They trust their gut. They make “feel” decisions. Counting cards requires you to suppress all of that — to play a mechanical, patient, sometimes boring game — because you’ve accepted that your intuition is not the source of your edge. The math is.
Most venture investors believe they are exceptional judges of founders and markets. Some are. But very few are exceptional enough that their judgment overcomes poor ownership structures, high entry valuations, and under-diversified portfolios. The hubris of “I can pick winners” is, in many cases, the single biggest drag on venture returns.
I wrote about this directly in a post on vibe investing — the tendency to make investment decisions based on feel, momentum, and pattern-matching to hot narratives rather than disciplined process. The antidote, as I argued there, is showing your work: understanding how you arrived at a decision, not just that you arrived at the right one. Consistency leads to repeatability. Repeatability is the only path to the kind of compounded returns that actually build a great fund. The card counters knew this. They weren’t trying to win every hand. They were building a system that would win over time.
The MIT team was ultimately banned from Las Vegas because their strategy worked so well. It wasn’t glamorous. It wasn’t based on a mysterious gift. It was disciplined, humble, mathematically grounded, and relentlessly executed.
That’s the kind of venture firm I’m working to build.
Count the cards. Wait for the deck to tilt. Then bet as much as you possibly can.
Thanks for being a Perishable Knowledge subscriber.
If you are getting value from my blogs, I’d appreciate it if you share this post with someone who will enjoy it. (If you’re reading this email because someone sent it to you, please consider subscribing.)