The Problem
In article #001, I wrote about why I cap my leverage at 1.5x.
At the time, that number came from instinct and one personal scare — my account had drifted to 2.25x leverage with 77% margin usage, and one night I woke up with a single thought: “Another 10% drawdown and I get a margin call.”
But instinct needs evidence. So this month I did something simple: I pulled out the most famous blowups of the last three decades, one file at a time.
I only wanted to check whether my discipline held up. After reading them all, I turned on every light in the house.
Six Files
File 01 · LTCM (1998) — The smartest team on Earth, 25x leverage
Long-Term Capital Management had two Nobel laureates in economics on its board (Scholes and Merton) and the former bond chief of Salomon Brothers at the helm. Their models were nearly mathematically perfect — “the market makes small mistakes, but it doesn’t stay wrong for long.”
In 1995 and 1996, they returned over 40% annually. Then, from January to September 1998, they lost more than 90% of their capital. Balance-sheet leverage: 25-to-1. Notional exposure: over a trillion dollars. It took the Federal Reserve convening 14 banks to inject $3.6 billion to stop LTCM from dragging Wall Street down with it.
File 02 · XIV (2018) — The “safe income” trade that went to zero in a day
XIV was an exchange-traded note that shorted the VIX. The logic was seductive in its simplicity: volatility trends down over time, so stand on the selling side and collect rent every year.
The logic was right for five straight years. On February 5, 2018, the VIX went from 18.44 to 37.32 in a single day — more than double. XIV fell 84% after hours, then closed down 93% the next day as Credit Suisse announced liquidation. The bank lost over $500 million on the product.
One individual investor watched three years of savings — $4 million — evaporate overnight.
File 03 · Crude Oil Treasure (2020) — Ordinary people can blow through their accounts too
On April 20, 2020, the May WTI contract settled at −$37.63 a barrel — the first negative oil price in human history.
Bank of China’s “Crude Oil Treasure” product failed to roll its positions in time. The result: investors didn’t just lose their principal — they owed the bank money. This wasn’t a game for professionals. Participants were ordinary office workers buying what was advertised as a “100% margin, no leverage” product. It still blew through their accounts.
File 04 · Archegos (2021) — The leverage nobody could see
Bill Hwang, a legendary trader from the Tiger cubs lineage, built positions across multiple investment banks using total return swaps. Every bank thought he was a small client. Nobody saw the whole picture.
In March 2021, two trading days erased his $20 billion fortune. Credit Suisse lost $5.5 billion. Nomura lost $2.9 billion. No matter how well you hide leverage, a blowup is always public.
File 05 · The AI Stock God (2026) — Four times leverage, 36 hours
This one is last month’s news. Leopold Aschenbrenner, 24, former OpenAI researcher, ran a hedge fund with eight employees and +439% returns in the first half of the year. Then he applied four-times leverage to his AI thesis.
In July 2026, he lost 67% in a month. Over the following 36 hours, he sold roughly $16 billion in equities to Citadel in a single block trade — not because he chose to sell, but because he had to. “We let you down this month,” he wrote to investors. The Bank for International Settlements cited the case in a systemic risk warning.
File 06 · A contemporary specimen (2026)
I recently researched a Chinese-language options education platform that publishes its real account data on its website. I actually respect the transparency.
But I ran the numbers on their leverage: portfolio value ÷ net equity = 3.93x. They had borrowed $461,000 from their broker and paid $89,000 in margin interest over four years — 48% of their total deposits. Their worst month in 2026 was −71%. The account fell from $1.3 million to $157,428 — −87.9%.
And he’s still selling courses on how to earn stable income.
What I Took Away
Put all six files side by side, and the pattern is counterintuitive:
| Common thread | What it means |
|---|---|
| All were brilliant | Nobel laureates, Wall Street veterans, AI prodigies. Intelligence does not prevent blowups |
| All had won first | None were beginners. Every one made real money, then added leverage |
| None got the direction wrong | LTCM’s model was right. The VIX mean-reversion thesis was right |
| All were forced out | Not one exited by choice. Every one was liquidated at the worst possible price |
| All went to zero | Not drawdowns — permanent capital loss |
The first four aren’t fatal. The fifth is.
And of the five, I think the most dangerous is the second: they all won first.
Because success is the best leverage salesman.
You win once, confidence rises. You win twice, you start to think your capital is too small. You win three times, and you think: “If this is so stable, why not use someone else’s money and earn more?”
LTCM increased leverage after two consecutive years of 40%+ returns. The AI stock god used 4x leverage after +439%. The platform I studied earned 410% in 2025 and lost 79% the next year.
Adding leverage after winning is the most natural-feeling, most lethal move in investing.
One more thing deserves its own paragraph:
They didn’t lose on judgment. They ran out of time.
LTCM’s mean-reversion model wasn’t wrong — the market just stayed irrational longer than their margin lasted. XIV’s “volatility trends down” thesis wasn’t wrong — February 5, 2018 just didn’t belong to “the long run.”
The real mechanism of high leverage isn’t that it makes you wrong. It’s that it strips away your options at the worst possible moment.
At 1x leverage, you can wait out a crash. At 4x, when the crash comes, you can only sell — and you sell when everyone else is selling, at prices set by panic.
That’s what “forced to make the right decision at the wrong time” actually looks like.
The Insight
Leverage cannot buy you returns. It can only buy you speed. And the price of speed is your margin for error.
Investing is a survival game, not a speed game. The ones who last aren’t the fastest — they’re the ones who never get knocked out.
Here’s how I’ve turned this into non-negotiable rules:
- Leverage cap at 1.5x, hard stop at 2.0x. The math is simple: at 1.5x I can survive a 20% drawdown; at 3.9x, a 25% decline wipes me out.
- Never use margin loans for directional trades. Interest is a certainty; trading returns are not. Using certain costs to chase uncertain gains means starting 6% in the hole every year.
- In a losing streak, cut leverage — never average down. The platform in File 06 kept high leverage through six consecutive losing months. That’s not trading. That’s acceleration.
- Before every trade, ask: if the underlying drops 20% tomorrow, do I get liquidated?
Closing
I filed these six cases in my knowledge base under a single tag: cautionary tales.
Not because they scare me. Because I know — I haven’t won enough yet, which means my most dangerous moment hasn’t arrived.
The truly dangerous moment isn’t now. It’s the future day when my options income engine has run smoothly for a dozen months, when the equity curve is so clean that risk feels theoretical. On that day, someone — possibly me — will say: “If it’s this stable, why not add just a little leverage?”
These six files are my answer for that day.
Trading Insights #009 · To be continued