Two stocks flash a fresh high on the same morning. One is a quiet utility that moves twenty cents on a busy day. The other is a chip name that swings five dollars before lunch. You want the same rule to tell you how many shares to buy and where the stop sits, without eyeballing each one and talking yourself into a different answer. That’s the problem the turtle trading rules were built to solve, and they solve it with a single number that every other rule leans on.
The turtle trading rules are a complete, documented trend-following system: a full set of instructions for what to buy, how much to buy, where to add, and where to exit. What makes them worth studying next to the other named systems on this site is that there’s nothing discretionary in them. The whole method is built around measuring volatility rather than guessing at a fixed dollar or percentage stop.
The volatility unit that runs the whole system
The number is called N. It’s a rolling average of the true range over the last twenty days, which is the same measurement most charting packages label ATR. True range for a single day is the largest of three distances: today’s high minus today’s low, today’s high minus yesterday’s close, and yesterday’s close minus today’s low. Average that across twenty sessions and you get a running read on how far the instrument travels in an ordinary day.
Here is the move that makes the whole thing coherent. Every other rule is expressed as a multiple of N, never as a fixed dollar amount or a flat percentage. The stop is written as 2N rather than a flat two dollars. Adds come every 0.5N of progress rather than at every fixed point of gain. Because N is measured from the instrument’s own recent range, one rule set stretches or tightens on its own. A quiet name gets tight spacing and a close stop in dollar terms. A violent name gets wide spacing and a distant stop. You write one rule, and the market fills in the distance.
Why not just use a flat percentage stop, say eight percent under the entry, the way a lot of swing systems do? Because eight percent means something completely different on a sleepy staple than on a biotech that routinely travels six percent in a session. A fixed percentage over-risks the quiet name and under-risks the wild one. N settles it by measuring each instrument against itself, so 2N is roughly the same amount of ordinary market noise whether the ticker is calm or frantic.
How the turtle trading rules size a position
Sizing starts from risk, not from a share count. One unit is built so that a 1N move against the position costs a fixed, small slice of account equity. The Turtles used one percent. On a $100,000 account that’s $1,000 of risk per unit for every 1N the price can travel against you.
Let’s work it through. If a stock’s N is $2.00, one unit is $1,000 divided by $2.00, or 500 shares. A 1N adverse move of two dollars on 500 shares costs exactly $1,000, one percent of the account. Now take a quiet stock with an N of $0.50: the same $1,000 of risk buys 2,000 shares. Take a volatile name with an N of $5.00 and the unit shrinks to 200 shares. Identical dollar risk, three very different share counts. When I run that math the spread is the entire point, because the volatile name and the quiet name land ten times apart in size for the same $1,000 at stake. It’s the same reasoning behind the volatility-based position sizing covered elsewhere on the site.
One misread to kill early. N says nothing about direction. A high N doesn’t mean a stock is weak or about to fall, and a low N doesn’t mean it’s safe to hold. N measures travel, not bias. It sizes the trade and spaces the stops, while the breakout rules decide which way you’re pointed.
The two breakout entries
Entries run on price channels, the same Donchian channel construction of a rolling highest-high and lowest-low. Two systems sit stacked together.
System One is the faster of the two. A long entry triggers on a new twenty-day high, and a short triggers on a new twenty-day low. System Two is slower and enters on a fifty-five-day high or low, acting as the failsafe. One filter trips people up. System One skips its signal when the previous twenty-day breakout would have been a winner. The reasoning is blunt and worth sitting with: the biggest trends often start right after a breakout that chopped a few earlier attempts to pieces, so the rules deliberately stand aside on the twenty-day signal that follows a win and let the fifty-five-day system catch the move instead. You’ll pass on breakouts that look clean, on purpose.
So the textbook line that a breakout is an entry runs far too broad for this system. Inside a strong trend, most twenty-day breakouts are skipped by design, and the trade you actually take is often the fifty-five-day one nobody was watching. Miss that detail and you’re running a different system that happens to borrow the same channels.
Adding units as the trend pays
A single unit is the start, not the finished position. As the trade moves in your favor the rules add units one at a time, every half-N of progress, up to a ceiling of four units in one market. Say a long fills at $50.00 with an N of $2.00. The first add comes at $51.00, one half-N higher. The next at $52.00. The fourth and final unit at $53.00. Four units, spaced a half-N apart, all pointed the same way.
The part that protects you is the stop climbing with every add. When a new unit goes on, the stop for the entire position is lifted to sit 2N below the most recent entry. By the time the fourth unit fills at $53.00, the whole position rides a stop near $49.00 instead of the original $46.00. The early units, bought cheap, are no longer exposed to the full starting risk. You pyramid size into a trend while ratcheting the collective stop forward, so a reversal can’t walk the position all the way back to the first entry’s original line.
Where the stop goes
The Turtles ran two different exits, and it helps to keep them apart. The first is the money stop: 2N from the entry, or from the average entry once units are stacked. That’s the disaster brake, the line where the trade is simply wrong and the position comes off. The second is the trend exit, taken on an opposite channel breakout. System One long trades exit on a new ten-day low, and System Two long trades exit on a new twenty-day low. The volatility stop caps the loss, while the channel exit is what actually banks a long trend once it rolls over.
Don’t read the 2N stop as a guaranteed maximum loss. It’s the resting order, not a promise. A stock can gap straight through it overnight on earnings and fill far below, so the realized loss runs well past 2N. Futures gap less often than single stocks, which is one reason the number behaved better in the market the rules were written for. On an equity, 2N is where you intend to be out, not where you’re certain to get out.
Why the rules stay mechanical
Every input above is countable. N is a number. The entry is a channel print. The add is a half-N step. The stop is a 2N distance. Nothing asks you to judge whether a chart looks toppy or a breakout feels weak. That’s deliberate. The discretionary veto, the gut call that says skip this one, is exactly what makes traders miss the handful of trends that pay for everything else. Systematic trend traders built their records on taking every valid signal, good-looking or not, because the winners aren’t knowable in advance. Ed Seykota, one of the most quoted names in mechanical trend trading, is blunt on the point: the system works because you follow it, not because you overrule it on the trades that scare you.
I find this the hardest part to hold. The rules will have you buying a fifty-five-day high that already looks extended and skipping a tidy twenty-day breakout that any eye would take. Sit with the sizing and stop math long enough and the discipline gets easier. The point was never the single trade. It’s one rule set run across hundreds of them, where the edge is small and only shows up if you take them all.
Porting the turtle trading rules to equities
The rules were tuned for liquid futures: currencies, commodities, rates, and index contracts that trend for months and trade nearly around the clock. Point them at single stocks without adjustment and a few seams open up.
- Gaps. Equities gap on earnings and news far more violently than the futures the Turtles traded. A 2N stop assumes an orderly exit near your level, and an overnight gap can blow ten or twenty percent through it. Size built on N understates the real tail risk on any name carrying an earnings date inside the holding window.
- Corporate actions. Splits, spinoffs, and large dividends distort both N and the channel highs and lows. Feed the rules unadjusted prices and a split reads as a crash, firing a phantom breakout or a phantom stop. Split-adjusted, dividend-aware data isn’t optional here.
- Trend persistence. Commodities and currencies trend cleaner and longer than the average stock, which mean-reverts and chops more. The twenty-day and fifty-five-day windows that fit soybeans will whipsaw on a range-bound equity, so the breakout lengths and the N-multiples usually need testing and lengthening on the specific universe.
- Concentration. The original system spread four-unit pyramids across twenty-plus uncorrelated markets. Stack four units into one stock inside a book of correlated equities and the real risk sits far above what the per-trade 1N math suggests.
None of this breaks the framework. It means the volatility unit and the breakout lengths are starting points to adapt and validate, rather than constants to copy from a 1983 futures desk onto a 2026 equity screen.
Running one rule set across every ticker
Strip the story away and the turtle trading rules are a single idea worked out to its end: measure each instrument’s own volatility, call it N, and express sizing, entries, adds, and stops as multiples of that one number. Do that and a quiet stock and a violent one trade under identical rules, each sized and stopped in its own units. The parameters are dated and the futures roots show through, yet the spine, letting volatility set the distances instead of a fixed dollar guess, is as useful on a stock chart today as it was on a soybean ticket in 1983. Learn the pattern. Ride the trend. Keep the gains.
Educational content only. Not investment advice. Trading involves risk. You are responsible for your decisions.