Nicolas Darvas
The Darvas box: buy the break of the box top, but only on real volume
Nicolas Darvas was a professional dancer who traded by cable while touring the world, with nothing but closing prices and volume arriving days late. The method he built around that constraint — boxes stacked on boxes, a stop always at the current floor — turned a small account into the two million dollars in his book's title. The constraint is the reason the method is teachable: everything it needs fits in two numbers a day.
- Style
- Breakout
- Approach
- Mechanical
- Difficulty
- Intermediate
- Horizon
- Position (weeks to months)
- Holding period
- Weeks to months
- Time needed
- 20 minutes a day
- Markets
- Single stocks · Growth names · Liquid ETFs
- Source
- Nicolas Darvas, How I Made $2,000,000 in the Stock Market (1960) — Nicolas Darvas
The rule set
- Identify a box: a range the stock has been contained in for several weeks
- Buy when price breaks above the box top on volume well above its recent average
- Place the stop just below the box floor
- As the stock builds a new, higher box, raise the stop to that box's floor
- Sell when price falls back through the floor of its current box
What makes it distinctive
- Runs on closing price and volume alone — Darvas traded it from overseas on cabled quotes that arrived late
- The box is a visible risk boundary: the floor tells you where you are wrong before you buy
- Volume confirmation is the part most retellings drop, and it is the part that filters the false breaks
When it works
Strong bull markets with leading growth names that consolidate and then break out in steps — the staircase the boxes are drawn around.
When it fails
Choppy markets never form clean boxes, and defining the box is where the discretion hides. False breaks on light volume are common, and a wide box means a wide stop.
How a decision moves through it
Input
Daily bars and daily volume
Closing prices and share volume — the two numbers a 1950s cable could carry. Nothing intraday, nothing fundamental enters the mechanical rules.
Measure
The box: a 20-day high-low channel
Darvas drew the box by hand off the recent highs and lows; a 20-day Donchian channel is the same boundary written as a formula. The top and floor sit at actual extremes, so they hold still while the stock consolidates.
Measure
Volume against its own 50-day average
One ratio: today's volume divided by its 50-day mean. The method never asks whether volume is large in the abstract — only whether it is large for this stock.
Decide
Break the box top on expanding volume
The bar's high clears the previous bar's box top and volume runs at least 1.5 times its 50-day average. Both halves are required; a quiet break is what a false break looks like.
Size & protect
Stop below the floor, capped at 8%
The floor of the current box is where the thesis dies. The shipped version adds a hard 8% stop for the boxes wide enough that the floor alone would risk too much.
Act
Buy the break, sell the broken floor
Buy when the top gives way; exit when the low falls through the current box's floor. As new boxes stack higher, the floor — and the exit — rise with them.
A box is a stock deciding, drawn as a rectangle
Watch a strong stock for a few months and it rarely rises in a straight line. It runs, then it stops and trades sideways in a band — a few weeks where every rally stalls near the same high and every dip finds buyers near the same low. Darvas drew a rectangle around that band and called it a box. The top is where the sellers have been winning; the floor is where the buyers have.
The box is not a prediction. It is a description of where the stock has recently been willing to trade — and the break of it is the stock leaving that description behind.
Darvas drew his boxes by eye from cabled closing prices. The product's version writes the same boundary as a formula — the highest high and lowest low of the last 20 days — which loses a little of his judgement about where a box starts and ends, and gains something his version never had: two people running it get the same answer.
The half everyone forgets is the volume
A break of the box top, on its own, is just a 20-day high — and a page already exists in this library for the system that buys every one of those. What made Darvas's version different is that he would not take the break unless volume expanded with it. A genuine breakout in a stock under accumulation comes with a burst of activity; a drift through the level on thin trade is the signature of a break about to fail.
The shipped rule makes his instinct a number: volume on the breakout bar must run at least 1.5 times its 50-day average. That single condition is the difference between this system and a plain channel breakout, and it is the condition most modern retellings of the box method quietly leave out.
Boxes stack, and the stop climbs the stairs behind the stock
One box is a trade; the method is the staircase. After the break, a strong stock runs, stalls, and builds a new box at a higher level. When it does, the stop moves up to the new floor. The position is never given more room than the current box, and the exit is never a decision — it is wherever the floor of the newest box happens to be.
That is the whole system: a boundary, a volume gate, and a stop that only ever rises. Darvas ran it from hotel rooms on another continent, reading quotes that were days old, and the method never asked him for anything the cable could not deliver. The pages that follow take each part in turn — the exact rules, what the box geometry does to risk, which markets form boxes worth trading, and what the method looks like when the market stops cooperating.
Five ways into this system
- The box rules: how a box forms, when the break counts, and where the exit sitsDarvas drew boxes by eye; the mechanical version draws them with a 20-day channel and gates the break with a volume test. This page states both, and where the formula differs from the hand.6 min read
- The box height is the risk: sizing when the stop is a level, not a percentageThe floor tells you where you are wrong, which means the box's own height tells you what the trade risks. Sizing from that height is the whole discipline; everything else on this page is consequences.6 min read
- Where boxes form: leading stocks, daily closes, and twenty minutes after dinnerDarvas ran this method on cabled closing prices from hotel rooms across the world. What that constraint proves about the method's data needs is still its best fit description: single stocks, daily bars, and not much of your day.5 min read
- When the boxes stop being boxes: chop, quiet breaks and the discretion problemEvery breakout system pays for ranges; this one adds two failure modes of its own — the judgement hiding in the box definition, and gap risk in exactly the stocks it selects for.6 min read
- The dancer who beat Wall Street by cable: the Darvas box explained from zeroThe box method explained without assuming you know what a breakout, a channel or a stop order is — starting with the story, because for once the story is the argument.7 min read
The ideas behind it
This system assumes you already know these. Each one is explained from scratch in Investing 101.
Compare with
- Donchian Channel BreakoutBuy when price makes a new four-week high, sell when it makes a new four-week low, and do nothing else.
- Minervini SEPA Trend TemplateFilter for stocks already in a textbook uptrend, wait for their volatility to contract into a tight pivot, and buy the breakout with a stop close enough to size against.
These are documented methods described for study. Nothing here is investment advice, a recommendation, or a claim about future returns — every system on this page has losing periods, and the pages say where.
Reading about a system is not having one.
Plutux is where you write your own rules down, test them against real data, and keep the record your memory would otherwise rewrite. Join the waitlist for early access.