Why Being Wrong 90% of the Time Still Makes You Rich

Cover image

Every self-made American fortune looks, from the outside, like a story about being right. The real ones are built on a payoff structure that makes being wrong most of the time completely irrelevant.

Let's get straight to it. Bookmark this one. You will want the exact structure when you look at your own bets.

Here is the part that took me years and real money to accept. A real edge, by itself, makes almost no one rich. Renaissance was right on barely 50.75 percent of its trades. That edge alone is a rounding error. It became more than 100 billion dollars because of what was built around it: a structure that turned a razor-thin advantage into the greatest fortune in market history.

That structure is not a secret. It is a handful of mathematical facts, and every self-made American billionaire who started from nothing runs the exact same machine, whether they can name it or not. Consider where they started:

Warren Buffett had a paper route.
Jeff Bezos started Amazon after quitting his job.
Peter Thiel arrived in the country as an immigrant.
Ed Thorp was a math professor with no capital and a formula.

None of them started with money, or with information you cannot get. They started with a structure.

By the end of this article you will understand:

The five mathematical forces that convert a small, real edge into generational wealth.
Why each one is non-negotiable.
Why skipping any single one is how people with genuine talent still end up broke.

Note: This is deliberately long and every part builds on the one before it. If you want a motivational quote, this is not for you.

The Whole Machine, In One Screenshot

Before the detail, here is the entire structure. This is the part to save.

Find a real edge, even a tiny one. That is the prior problem, and it is the hard one.
Make it convex. Cap the downside at losing 1x, leave the upside uncapped.
Own it, don't rent it. Take the edge through ownership, not wages, so the upside is actually yours and it compounds.
Place enough bets to catch the power-law winner, because one bet returns more than all the others combined.
Size to survive. Never risk ruin. The penalty for betting too big is permanent.
Then hold, long enough for compounding's exponent to do the work the rate alone never can.

Everything below is why each line is there, and what happens the moment you drop one.

Part 1: Convexity, or Why Being Wrong 90 Percent of the Time Still Wins

A convex bet loses a little when it fails and wins enormously when it works. The engine is that asymmetry, not accuracy.

In his 2015 letter to Amazon shareholders, Jeff Bezos wrote the clearest sentence ever put on paper about how self-made fortunes are actually built: "Given a ten percent chance of a 100 times payoff, you should take that bet every time. But you're still going to be wrong nine times out of ten."

Read that as math, because that is what it is. Take the bet a hundred times:

Ninety times it fails. You lose your 1x stake each time.
Ten times it hits. You make 100x each time.
Net result: you were wrong on 90 percent of your bets and you multiplied your money many times over.

Accuracy was never the point. The shape of the payoff was the point.

Bezos went further, and this is the line people miss. He compared it to baseball:

In baseball the outcome is capped. Swing as hard as you want, the most you can score on one hit is four runs.
In business it is not. "Every once in a while, when you step up to the plate, you can score 1,000 runs." As he put it, "this long-tailed distribution of returns is why it's important to be bold. Big winners pay for so many experiments."

That is the whole of Part 1. A self-made fortune is not built by being right. It is built by structuring bets so that:

Losses are capped at what you risked.
Wins are not capped at all.

The founder who understands this does not fear the string of failures. Amazon could only ever lose 1x on the Fire Phone. It could make what AWS made. The failures are the price of standing in front of the one bet that pays for everything.

The mistake that keeps talented people poor is the opposite shape: bets with capped upside and uncapped downside. A salaried expert selling a fixed number of hours has exactly that structure, and no amount of skill fixes it. More on that in Part 3.

Part 2: The Power Law, or Why One Bet Returns More Than All the Others Combined

Wealth is not spread on a bell curve. It follows a power law, where a single outlier is the entire result and everything else is noise.

Convexity tells you to structure bets so the upside is unbounded. The power law tells you why that upside is where all the money actually is. It is the most counterintuitive fact in this article.

Peter Thiel put it plainly in Zero to One: "We don't live in a normal world; we live under a power law." The difference:

On a bell curve (a "normal" distribution), outcomes cluster around an average and extremes are rare.
Under a power law, a tiny handful of results are so large they dwarf the sum of everything else.

Thiel calls the practical version "the biggest secret in venture capital": the best single investment in a successful fund equals or outperforms the entire rest of the fund combined.

This is not a Silicon Valley opinion. It is measured. Hendrik Bessembinder's 2018 study in the Journal of Financial Economics looked at every US common stock since 1926, roughly 25,000 companies. The findings stop people cold:

The best-performing 4 percent of companies (about 1,092 firms) account for all of the net wealth the US stock market created in 90 years.
The other 96 percent, collectively, merely matched one-month Treasury bills.
More than half of all stocks had a negative lifetime return.
Five companies alone (ExxonMobil, Apple, Microsoft, GE, IBM) produced 10 percent of all the wealth ever created in US equities.

Sit with two facts side by side. The typical stock is a loser. The market's entire gain comes from a sliver of extreme winners. That changes what "a good bet" even means:

The winner is not the average of many outcomes.
The winner is the outlier.

If your strategy is not built to have exposure to the tail, you have, mathematically, built a strategy to match Treasury bills while feeling busy.

This is why Thiel's rules are so violent. He argues you should:

Only make bets that could, by themselves, return everything.
Concentrate, because rule one eliminates almost every available bet.

Diversification, in his framing, is a hedge against ignorance: you place many small bets because you don't know which one is good. The self-made fortune comes from the opposite discipline. Do enough work to become non-ignorant about a small number of bets that can hit the tail, then concentrate on them. Spreading yourself across a hundred safe, capped, average bets does not reduce risk. It guarantees you never touch the 4 percent that mattered.

Part 3: Ownership, or the Multiplier That Wages Mathematically Cannot Have

A wage is a linear function with a ceiling. Equity is a convex function that compounds. This single difference is why every name on the list built through ownership and not salary.

Here is a fact that sounds like trivia and is actually the hinge of the whole thing. Warren Buffett, one of the richest people who has ever lived, has drawn a salary of 50,000 dollars a year from Berkshire Hathaway for decades. His fortune has essentially nothing to do with that salary. Ninety-nine percent of his net worth is tied to his ownership of Berkshire stock.

This is not a quirk. It is the mathematics of two different kinds of income, and the gap between them cannot be closed by effort.

A wage:

Is a linear function with a hard ceiling. Income equals hours times rate.
Does not compound. Last year's paycheck does not earn this year's paycheck.
Resets to zero every January.
Is capped, because there are only so many hours and only so much any market pays per hour.

Equity:

Is ownership of an outcome, and outcomes are power-law distributed and unbounded (see Part 2).
Compounds. The value it gains this year becomes the base that gains value next year.
Works like multiplication, where a salary is only addition.

This is why the answer to "how did they get so rich from nothing" is never "they were paid a lot." It is always the same structural move: they converted their edge into ownership of a compounding thing.

Bezos owned Amazon.
Thiel owned slices of PayPal, Facebook, and Palantir.
Buffett owns Berkshire.

The salary, where it exists at all, is a rounding error. The whole fortune is the equity, doing the one thing a wage can never do: compounding on itself against a power-law upside. Your edge is worth almost nothing while you rent it out by the hour. It becomes a fortune only when you own the thing it makes.

Part 4: Compounding and Time, or Why the Exponent Does the Work

Compounding is multiplication repeated. The rate gets the attention. The number of times it repeats is what actually builds the fortune. Time is the exponent.

Now the force that turns ownership of a compounding asset into a number with ten zeros. It is the one people understand least, because human intuition is built for straight lines and compounding does not move in a straight line.

Morgan Housel made this concrete in The Psychology of Money with a fact about Buffett that reorganizes how you think about wealth:

Of Buffett's roughly 84.5 billion dollars in net worth at the time, 84.2 billion came after his 50th birthday.
About 81.5 billion came after he turned 65, the age most people retire.
He did not become a billionaire until he was 56.

As Housel put it, his skill is investing, but his secret is time. Then Housel ran the counterfactual that proves it. Buffett has been compounding since he was around ten years old, roughly eighty years. Suppose instead:

He started investing seriously at 30, not as a child.
He earned the exact same phenomenal returns.
He retired comfortably at 60, a normal-length career.

His net worth would be about 11.9 million dollars. Not 84 billion. 11.9 million. That is 99.9 percent less, and the only variable that changed was how long the compounding ran.

Here is the math underneath it:

Each period multiplies your wealth by (1 + r).
Over n periods, wealth scales as (1 + r) to the power of n.
The rate r sits in the base. Time n sits in the exponent.
An exponent does not add to your result. It detonates it.

This is why Buffett, compounding at about 22 percent a year, is far richer than Jim Simons, who compounded Renaissance at an almost supernatural 66 percent, three times Buffett's rate:

Simons ran his machine for a few decades and withdrew along the way.
Buffett let his run for eighty years.
The higher rate lost to the longer exponent.

Rate is what everyone chases. Time is what actually pays. The practical consequence is brutal and freeing at once. You do not need a spectacular edge. You need a real one, structured for ownership, that you can keep running for a very long time without interruption. The most expensive thing you can do to a fortune is stop the exponent early. That is exactly what a blow-up does. Which is Part 5.

Part 5: Survival, or Why One Ruinous Loss Erases Everything

Because wealth compounds multiplicatively, a single total loss is absorbing. You cannot compound upward from zero. The whole game is to never be forced to stop.

Everything above depends on one condition the math makes non-negotiable: you have to still be in the game. Multiplicative growth punishes ruin in a way additive thinking completely hides.

Watch what a total loss does to a compounding sequence:

You go up 50 percent, then 80 percent, then 200 percent. A beautiful run.
Then one bet takes you to zero. Not down 90 percent. To zero.
Everything before it is annihilated.
You can never come back, because any rate of return multiplied by zero is still zero.

There is no percentage gain that recovers from a wipeout, because compounding grows what you have, and you have nothing. A total loss is an absorbing barrier. It ends the exponent from Part 4 permanently.

This is why the long-run growth of wealth is governed by the geometric mean of your returns, not the arithmetic mean, and the geometric mean is savagely sensitive to catastrophic losses in a way the average is not. A strategy with a great average return and one path to zero has a true long-run growth rate of exactly zero, because on a long enough timeline that path gets taken. Survival is not a virtue you add for safety. It is a term in the equation.

The tool built precisely for this is the Kelly criterion:

John Kelly, a physicist at Bell Labs, published it in 1956.
It grew directly out of Claude Shannon's information theory, and Shannon himself refereed the paper.
It sizes each bet to maximize the expected logarithm of your capital, which is exactly the growth rate of multiplicatively compounding money.
Bet too small and you leave growth on the table. Bet too big and you march toward the absorbing barrier. There is one growth-optimal fraction, and it depends on your edge.

The man who proved it works was Ed Thorp, another self-made-from-nothing story built on math, not capital:

He used Kelly to size his bets while counting cards at blackjack, and got banned from casinos.
He turned the same principle on Wall Street with Princeton/Newport Partners.
The fund compounded at roughly 19 percent a year for nearly two decades, with only three losing months in its entire history, and returns uncorrelated to the market.
He never blew up. That was the whole achievement.

Thorp's most-repeated advice is the line that separates survivors from legends-who-are-now-broke: bet fractional Kelly, less than the formula says, because the penalty for overbetting is far worse than the cost of underbetting. Err toward survival, always. The self-made fortune is not the person who bet the most. It is the person who bet big enough to matter and small enough to still be standing when the power-law winner finally arrived.

Part 6: The Machine, End to End

Every piece is necessary, and the fortune is the pieces running together, patiently, for a very long time.

Step back and look at the full structure. The parts connect into one clean pipeline:

Start with a real edge. The hard prior problem, the one my last two articles were about. On its own, worth almost nothing.
Make it convex. Failure costs 1x, success is uncapped.
Own it. So the unbounded upside is yours and it compounds instead of resetting every year.
Spread it. Enough independent bets to catch the tail, because one winner is worth more than everything else combined.
Size to survive. Because one ruinous loss is an absorbing barrier and no rate recovers from zero.
Let it run. For decades, because time sits in the exponent and the exponent builds the number.

That is the machine. It is the same machine whether it is:

Renaissance turning a 50.75 percent edge into 100 billion dollars across millions of tiny, well-sized, uncorrelated bets.
Thiel's fund concentrating on the few companies that could return everything.
Bezos running experiment after capped-downside experiment until AWS scored a thousand runs.
Buffett owning one compounding asset for eighty years and refusing to stop.

Different arenas. Identical structure. And this is the answer to the question the article started with. How do people get rich from zero? Not by being right:

Renaissance was wrong 49 percent of the time.
Bezos was wrong nine times out of ten.
Bessembinder's data says 96 percent of companies went nowhere.

Being right is not the input. The structure is the input: convex bets, owned rather than rented, spread wide enough to catch the tail, sized to survive, compounded long enough for the exponent to take over. Run that structure patiently and a small real edge becomes a fortune. Skip any single piece and genuine talent still ends at zero.

The Summary

The self-made American fortune is not a lucky pick and it is not raw genius. It is a repeatable mathematical structure, five forces stacked together:

Convexity. Losses capped, wins uncapped, so being wrong most of the time still wins.
The power law. One outlier is the entire result, so build to touch the tail, not to match the average.
Ownership. Equity compounds while a wage only resets, so the fortune is always in what you own.
Compounding. Time lives in the exponent, so a long uninterrupted run beats a higher rate.
Survival. A single wipeout is permanent, so size to never be forced to stop.

Buffett, Bezos, Thiel, Thorp, Simons. Different arenas, one machine. The edge is the ticket in. The structure is what turns the ticket into the fortune.

Here is the question I want you to sit with. If the fortune comes from the structure and not from being right, then the person with a mediocre edge who runs the full machine for thirty years beats the brilliant one who nails the call, rents it out by the hour, over-bets, and quits early.

So which are you actually building: a great bet, or the machine that turns any real edge into wealth and simply refuses to stop?

Your answer tells you whether you think like someone chasing one big win, or like someone constructing a thing that compounds for decades.

Drop your answer in the comments. There is no wrong answer. But there are very revealing ones.