Nicolas Darvas: the trader whose edge was being unable to watch
In 1957 a working ballroom dancer was somewhere between Saigon, Tokyo and Calcutta with a dance troupe, and he was also running a stock portfolio in New York. He could not see a chart. He could not call his broker on a whim. What he had was a week-old copy of Barron’s and one cable a day listing a handful of closing prices. Every trader reading this has more market information on their phone right now than Nicolas Darvas had in eighteen months — and that is exactly the problem this profile is about.
KEY TAKEAWAYS
- The stop went on at the same instant as the buy. Not after the trade was uncomfortable — at the moment of entry, as a resting order with the broker, cabled from wherever he was dancing.
- The box sets the position size. Risking $100 on a $10,000 account buys $2,500 of a stock with a 4% box and only $500 of one with a 20% box. The chart decides, not your enthusiasm.
- Distance was the edge. At one chance in ten thousand of overriding your plan per look, checking weekly leaves 99.7% of six-month positions intact and checking every five minutes leaves 0.53%.
- The famous number is contested and we print both. $2,000,000 on the cover; $216,000 ascertainable according to the New York Attorney General in December 1960. A court blocked the probe in January 1961 and investigators admitted they had not traced every account.
- Even the small number is remarkable. $216,000 from his stated $36,000 stake is 10.5% compounded monthly for eighteen months. Nobody writes a book with that on the cover, which is the whole lesson about trading stories.
Who was Nicolas Darvas?
Nicolas Darvas was born in Hungary in 1920 and trained as an economist at the University of Budapest. In June 1943, aged 23, he left the country on a forged exit visa carrying fifty pounds sterling, and made his way to Istanbul. He later found his half-sister Julia; the two became a professional dance act that worked Europe and then the United States, and by the mid-1950s they were touring internationally.
He came to the market by accident. As he tells it in the first chapter of his book, a Toronto nightclub offered to pay him in shares of a small Canadian mining company called Brilund rather than cash — 6,000 shares at 50 cents, about $3,000. He could not make the engagement, felt bad about it, and bought the shares anyway. He then forgot about them. Some weeks later he noticed the stock near $1.90 and sold. That is the entire origin story, and it is worth being blunt about what it teaches: nothing. It was luck, he says so himself, and the years of Canadian penny-stock losses that followed are the reason the book gives two separate chapters to crises.
What is unusual is what he did next. During his off hours he read something like 200 books on markets and speculators, sometimes eight hours a day. Two he reread almost weekly: Gerald M. Loeb’s The Battle for Investment Survival (1935) and Humphrey B. Neill’s Tape Reading and Market Tactics (1931). Both are books about losing less, not about picking better. That is a detail worth noticing before any of the numbers.
What was the box, and what did it actually do?
The box theory is one sentence long. A stock moving upward does not move smoothly; it pauses inside a range, then jumps to a higher range and pauses again. Darvas drew those ranges as boxes stacked on top of one another. While price stayed inside a box, he did nothing at all. When it broke out of the top of a box, he bought — and in the same instruction, he placed a stop-loss order just below.
That is a structure stop, which is one of the four legitimate places a stop can go. We cover the other three in stop-loss placement, and the box itself is nothing more than the support and resistance zone you already know, drawn as a rectangle and given a rule.
The part people skip is the consequence. If your stop sits under the box, then the box has decided how much you can lose per share before you have chosen anything. And if the amount you are willing to lose is fixed — the 1–2% rule — then the box has also decided your position size.
Read that chart the right way round. It is not saying a deep box is a bad trade. It is saying that a deep box is a small trade. Most beginners do this backwards: they choose how much to buy first, then look for somewhere to hide a stop that fits the position. Darvas’ sequence — structure, then stop, then size — is the sequence taught in position sizing, and it is the reason the method survives even if the fortune does not.
Why did being 8,000 miles away help?
The romantic version of this story says Darvas succeeded despite being on the road. The mechanical version is more useful: he succeeded partly because of it.
Consider what his setup physically prevented. His stop was a resting order sitting with a broker in New York; once cabled, it executed on price whether he approved or not. His information arrived once a day, as a short list of closing prices, and once a week as a printed magazine that was already stale. There was no intraday chart to stare at. There was no button to press at two in the morning. He could form an opinion, but he could rarely act on one.
You can put a number on what that is worth, and the number does not depend on knowing anything about Darvas. Suppose you have a plan and a stop, and suppose that on any single occasion you look at the price there is some small probability p that you override the plan — you close early out of boredom, you widen the stop, you take a small profit that was supposed to run. Over N looks, the chance you reach your planned exit with the plan intact is (1 − p)N. That is all.
Set p at one in ten thousand — an absurdly disciplined trader, wrong about their own plan once in every 10,000 glances. Hold a position for six months. Checking a weekly magazine, 99.7% of those positions survive. Checking a daily cable, 98.7%. Checking hourly on a phone in a market that never closes, 64.6%. Checking every five minutes, 0.53%.
Move p to a still-flattering one in a thousand and the daily trader holds 87.8% of the time while the five-minute trader is at 0.00%. We are not measuring p and we are not claiming to know yours. The point is structural: whatever p is, it is being raised to a power, and the exponent is a lifestyle choice. Darvas’ exponent was 130. Yours, if a chart app is open on a second monitor, is in the tens of thousands.
PRACTICE CORNER
Darvas’ rule was that the exit existed before the entry did, as a resting order he could not easily take back. That part is not history — every major exchange lets you attach a stop at the moment you open a position, and lets you set a price alert instead of watching. Do the two together on your next trade: mark the range, put the stop under the low of the range as you enter, then set an alert and close the app. You have just cut your own exponent from thousands to one.
Referral links — they never change our assessment. Education only; most retail traders lose money.
Did he really make $2,000,000?
This is the section the rest of the internet leaves out, so here is the record.
Time profiled Darvas in May 1959, while he was still dancing. The book followed in 1960 and sold enormously. Then, in December 1960, Time reported that the New York Attorney General had “thrown the book” at him: his story was “unqualifiedly false”, and the state could find ascertainable profits of only $216,000. It was the first action taken under a broadened state law against misrepresentation in giving investment advice. Darvas called it a “cynically irresponsible action, book burning by publicity”.
On 13 January 1961 Time reported the sequel: a court blocked the investigation, ruling it an “unwarranted invasion of the free press”. The same report noted that state investigators had admitted they were unable to track down all of the dancer’s brokerage accounts.
So neither figure was ever established. $2,000,000 is his claim. $216,000 is what one set of investigators could verify before they were stopped, from accounts they admitted were incomplete — a floor, not a measurement. Anyone who tells you confidently which one is true is telling you about themselves.
What we can do is take both figures seriously and see what each implies. Darvas states his starting stake as $36,000 and his run as eighteen months.
Here is the part worth sitting with. The prosecution’s number — the deflating, debunking, you-were-lied-to number — is a return of 10.5% a month compounded for a year and a half. That is a career-defining stretch for a professional fund manager. It is also, on the cover of a book, completely unsellable. A cover carrying the smaller figure does not get written, does not get published, and is not being quoted seventy years later.
That asymmetry has not gone anywhere. It is why the screenshots in your feed show the account that went up forty times and never the one that went up six. Six is the good outcome. Six is what a very good decade looks like. If the only records you ever see are the fifty-fold ones, your sense of what normal success looks like is being set by a selection effect, and you will size your positions accordingly.
What does this method cost you?
Every profile in this section has to answer that question, and Darvas’ answer is unusually clean, because he said it himself late in life.
“I keep out in a bear market and leave such exceptional stocks to those who don’t mind risking their money against the market trend.”Nicolas Darvas, You Can Still Make It in the Market (1977), p. 126.
Read as an admission rather than as advice, this is the honest limit of the whole method: it needs a rising market to be in.
The box method buys strength. It requires stocks making new highs, which requires a market in which stocks make new highs. Darvas’ famous run was 1957–58, and he was explicit that he stayed out otherwise. He also never sold short:
“I have never done it myself because psychologically I am not cut out for short selling. But I think markets have now changed their character so much that all experienced investors should seriously consider it.”Nicolas Darvas, You Can Still Make It in the Market (1977), p. 89.
So the method has no answer at all for a falling market except to be absent from it. In crypto, where a drawdown of 70–80% from a cycle high is ordinary rather than exceptional, “be absent” is a strategy with a very long duty cycle, and most people are not psychologically equipped to sit it out.
The second cost is arithmetic and it applies even in a good market. A stop under the box gets hit by false breakouts, and false breakouts are common — see false breakouts and trend exhaustion. Each failed attempt costs you the full planned risk. Three failed attempts at 1% each is 3% of the account gone before the trade that works even begins, and you re-enter the eventual winner higher than you first tried to buy it. That is not a flaw in the method; it is the price of the method, and it is charged in advance whether the fourth attempt works or not.
Where this stops working
The 24/7 market is not the 1958 market. Darvas’ daily cable arrived after a session that had closed. Crypto has no close, so “check once a day” is a decision you have to make and defend, not a condition the market imposes on you.
A resting stop is not a guaranteed price. His stop orders filled at whatever the market offered. Yours will too. In a fast move a stop becomes a market order at the next available price, which in thin crypto pairs can be well below the level you chose.
Do not read “look less” as “plan less”. The whole reason Darvas could ignore the price was that the exit was already placed. Ignoring the price without a stop in the market is not discipline; it is absence.
The arithmetic in this article is a model, not a measurement. We do not know your p, and it is almost certainly not constant — it rises when you are down, when you are bored, and when you are leveraged. The model shows the shape of the effect, not its size for you.
His figures are his figures. The $36,000 stake, the 6,000 Brilund shares, the eighteen-month run: all come from a book written to sell. We have used them because they are what the record contains, and we have labelled every calculation built on them as conditional on his own account being accurate.
FAQ
Who was Nicolas Darvas?
A Hungarian-born ballroom dancer (1920–1977) who taught himself to trade while touring, and wrote “How I Made $2,000,000 in the Stock Market” (1960). He fled Hungary in June 1943 at 23 with a forged exit visa and fifty pounds, danced professionally with his half-sister Julia, and read roughly 200 market books in his off hours before he had a method.
What is the Darvas box theory?
Darvas treated a stock’s price as a series of stacked boxes. While price stayed inside a box he did nothing. When it broke out of the top of the box he bought, and at the same moment placed a stop-loss order just under the breakout. As higher boxes formed, the stop moved up with them. The stop never moved down.
Did Nicolas Darvas really make $2,000,000?
It is unresolved and the record says so. In December 1960 Time reported that the New York Attorney General called his story “unqualifiedly false” and could find ascertainable profits of only $216,000. In January 1961 Time reported that a court blocked the investigation as an “unwarranted invasion of the free press”, and that state investigators admitted they had not traced all of his brokerage accounts. Neither figure was ever proved.
Does the box method still work in crypto?
The mechanic transfers — a range, a breakout, a stop under the range — because it is just a structure stop. What does not transfer is the market. Darvas bought only rising stocks in a rising market and said plainly that he stayed out of bear markets. In a range-bound or falling market the method produces a run of small stopped-out losses, which is exactly what it is designed to do.
Why did trading from abroad help him?
Because it removed the option to interfere. His stop was a resting order sitting with a New York broker; his information was a daily cable and a week-old Barron’s. He could not close a position on a feeling at two in the morning, because there was no button to press. Distance did not improve his judgement — it reduced the number of occasions on which his judgement could be used.