The Turtle Trading Rules: What a Two-Week Course Actually Taught
Idea clave
- The experiment's claim was that rules beat talent. The rules were published afterwards, and they still work poorly for most people — which tells you where the difficulty actually lives.
- The entry is a two-line rule and the least important one. Position sizing and the exit are what produced the returns.
- The method loses on most trades and survives on a handful of very large winners. If you cannot sit through that, none of the rest applies.
Ruta de aprendizajeBuild a trend-following system: ride winners, cut everything elsePaso 5 de 9
Antes que esta:Trend Following: What a 22-Year Study Actually Shows
Basado en Way of the Turtle — Curtis Faith, 2007
The bet that started it
Richard Dennis believed trading could be taught. His partner William Eckhardt believed it could not. To settle it, Dennis advertised for trainees in 1983, took two groups of near-beginners, taught them a mechanical system in about two weeks, and gave them real money. He called them Turtles.
Dennis won the argument. The interesting part is what the rules turned out to be once they were published.
Curtis Faith was one of the trainees, and Way of the Turtle is his account of the system and of why most of the group underperformed anyway. That second half is the useful half, and it is the part that gets skipped.
The entry is two lines, and it is the least important rule
That is it. A price channel breakout, borrowed almost unchanged from Richard Donchian's work three decades earlier. There is no confirmation indicator, no volume filter, no candlestick pattern. It fires on a level that anyone can compute.
The reason to be blunt about the entry's unimportance is that entries are what beginners copy. The Turtle entry has been public for decades and has not stopped working in the sense that matters — it never had much edge on its own. What it had was everything that came after it.
N: the number that makes different markets comparable
Position size was never chosen. It was computed from how much the market moves in a day.
The Turtles measured a market's daily range — the 20-day average true range — and called it N. One unit of position was defined as the number of contracts for which a one-N move equalled 1% of the account.
This is the rule that does the most work, and it is the one worth stealing whatever you trade. It replaces "how many shares feels right" with a calculation whose answer changes when the market changes — automatically shrinking positions when things get wild, which is precisely when discretion fails.
| Problem | Fixed share count | Sized by N |
|---|---|---|
| Volatility doubles | Your risk silently doubles | Position halves; risk stays put |
| Comparing two markets | Not comparable at all | One unit means the same thing in both |
| A quiet market you like | Under-risked without noticing | Position grows to match the opportunity |
Adding to winners, and the stop that follows them up
Two things are happening here at once, and they pull in opposite directions. The additions increase exposure to a move that is working. The stop, marching up behind them, caps what the enlarged position can give back. Without the second half the first half is just a way to be maximally long at the top.
The exit was the hard part, and it was designed to feel wrong
A 10-day-low exit gives back a large piece of every winner, by design. That give-back is the price of never missing the trades that pay for the year.
Every trend-following exit faces the same trade-off: leave early and you clip the tail, leave late and you return open profit. The Turtles chose late, and Faith is candid that this is where most of the group failed — not by breaking the entry rule, but by taking profits before the exit signal because holding hurt.
Exiting when it feels safe
- Locks in the small and mid-sized winners
- Cuts the tail that pays for everything
- Feels good every single time
Exiting on the rule
- Gives back a chunk of every winner
- Keeps you in the one trade that matters
- Feels bad on the way out, every time
What the record actually looks like
The Turtle rules lose on the majority of trades. That is not a flaw being tolerated — it is structural. A breakout entry is wrong most of the time by construction, and the system is built so that being wrong costs 1R and being right occasionally costs nothing at all.
Which is why the honest way to read the experiment is not "rules beat talent" but something narrower: these rules, executed without deviation, beat the discretion of near-beginners. Faith's own account is that several trainees had the same rules and materially worse results, and that the difference was execution during drawdowns. That finding is stronger evidence than the headline one.
What transfers to an ordinary account, and what does not
Does not transfer
- Trading 20+ futures markets at once
- The capital to hold four units through a 2N stop
- Someone else's money and a salary
Transfers directly
- Size from volatility, not from conviction
- Define the exit before the entry
- Write the rule; do not renegotiate it live
The diversification is the piece most often ignored. A trend system on one instrument is a lottery ticket; the Turtles were running the same rules across dozens of uncorrelated markets, which is what turned a 35%-win-rate method into something with a survivable equity curve. Copying the rules onto a single ticker copies the losses and not the mechanism.
If you want to borrow one thing, borrow N. Sizing by volatility is the rule with the best ratio of benefit to complexity, and it works the same way in a stock account as it did in 1983 pork bellies. Then write the whole thing down — see your first trading plan for the shape that takes.
Prueba esta semana
- Compute the 20-day average true range for one instrument you follow. That is its N.
- Work out the position size for which a one-N move equals 1% of your account.
- Compare it to the size you would normally take. If the difference is large, that is the whole lesson.
- Write your exit rule down before your next entry, and record whether you followed it.
Preguntas frecuentes
What are the Turtle trading rules?
A mechanical trend-following system: enter on a 20-day price breakout, size positions so that one average daily range equals 1% of the account, add up to four units as the trade moves in your favour, stop out at 2N below the last entry, and exit on a 10-day low. Every part is computable with no judgement.
Do the Turtle rules still work today?
Trend following as a category still works and still has long flat periods. The specific parameters have been public since the 1990s and are widely traded, so the edge is thinner than it was. What has not changed is the sizing and exit discipline, which is the part that was doing the work.
What does N mean in Turtle trading?
N is the 20-day average true range — roughly, how far the market moves in a normal day. It is used as the unit for both position sizing and stop placement, so that one unit of risk means the same thing across markets with very different volatility.
What was the Turtles' win rate?
Low, as with any breakout system — the majority of trades were small losses. The method's profitability came from a small number of very large winners, which is why an exit rule that gives back open profit was preferred to one that takes it early.
Can I use Turtle rules for stocks instead of futures?
The sizing and exit logic transfers cleanly. The diversification does not: the Turtles ran the same rules across dozens of markets that did not move together, and a single stock or a basket of correlated ones gives you the losing months without the mechanism that paid for them.