The original Turtle Trading rules, explained

A quick note on provenance. The rules below were taught to the Turtle class of 1983 and 1984 by Richard Dennis and William Eckhardt. They were kept confidential for years, leaked in pieces during the 1990s, and eventually published in full, most completely by original Turtle Curtis Faith in Way of the Turtle. What follows is an educational summary of that public record.

The five questions every complete system answers

Eckhardt taught the Turtles that a complete trading system removes every decision from the heat of the moment. It has to answer five questions:

1. Markets: what do you buy or sell?
2. Position sizing: how much do you buy or sell?
3. Entries: when do you buy or sell?
4. Stops: when do you get out of a losing position?
5. Exits: when do you get out of a winning position?

Most traders obsess over entries. The Turtles were taught that sizing and exits matter far more, and the structure of the rules reflects that.

1. Markets: liquid futures, broadly diversified

The Turtles traded large, liquid futures markets where their size would not move prices. The list reportedly included U.S. Treasury bonds and bills, major currencies (the Swiss franc, British pound, Japanese yen, and Canadian dollar), the S&P 500, gold, silver, copper, crude oil, heating oil, unleaded gas, and softs such as coffee, cocoa, sugar, and cotton.

Diversification was the point. Trend followers do not know which market will trend next, so they watch many markets at once and let the breakouts decide.

2. Position sizing: the N formula

This is the heart of the system. The Turtles measured every market's volatility with N, a 20-day exponential moving average of the true range. Today the same concept is known as the Average True Range, or ATR. N answers a simple question: how much does this market typically move in a day?

Positions were bought in units, sized so that one N of price movement equaled roughly 1% of the account. The formula:

The effect is that a volatile market gets a small position and a quiet market gets a larger one, so every unit carries roughly the same risk. A trader who sizes every market identically is, without realizing it, betting far more on the volatile ones.

3. Entries: Donchian breakouts

The Turtles ran two parallel systems and could allocate between them. Both entered on Donchian channel breakouts, price exceeding its highest high or lowest low over a lookback window.

System 1 (shorter term)

• Buy when price exceeds the highest high of the previous 20 days; sell short on a break below the 20-day low.
• The filter: skip the signal if the previous breakout in that market would have been a winner, even if it was not taken. After a winning breakout, the odds of the next one succeeding were judged lower.
• A failsafe 55-day breakout was always taken, so no major trend could be missed entirely.

System 2 (longer term)

• Buy when price exceeds the highest high of the previous 55 days; short the mirror-image low.
• Every signal is taken, no filter.

There is no prediction in either system. A breakout does not mean a trend will follow; most breakouts fail. It simply means a trend might be starting, and the Turtles could not afford to miss the one that does not stop.

4. Stops: 2N and no exceptions

Every unit had a stop placed 2N from the entry price. Combined with the sizing formula, a 2N loss equals roughly 2% of the account per unit. The stop was a hard rule. Eckhardt taught that the trades you most want to hold past the stop are precisely the ones that go on to do the most damage.

5. Pyramiding: pressing the winners

When a position moved ½N in their favor, the Turtles added another unit, and continued adding every ½N up to a maximum of 4 units in a single market. Stops on earlier units were raised as new units were added, so a full position that immediately reversed lost far less than four separate 2% bets.

Risk was also capped across the portfolio with unit limits, reportedly:

• 4 units in any single market
• 6 units in closely correlated markets (for example, heating oil and crude)
• 10 units in loosely correlated markets
• 12 units total in any one direction, long or short

6. Exits: give the trend room

Winning positions were not closed at targets. There were no targets. System 1 exited longs when price broke below the 10-day low (10-day high for shorts); System 2 used the 20-day opposite extreme. These exits deliberately give back a chunk of open profit at the end of every trend. In exchange, they keep the trader aboard for the rare move that runs for months, which is where all of the system's profit comes from.

Why the system made money

The win rate was poor. Most breakouts failed, and losing months outnumbered winning ones for some Turtles. The system profited because its losses were structurally small (2N stops, volatility-normalized units) while its wins were structurally unlimited (no targets, wide trailing exits, pyramided positions). One captured trend in silver or crude could pay for dozens of failed breakouts.

That asymmetry is also why the system was psychologically brutal. Following it meant taking signal after signal through a losing streak, then watching a huge open profit shrink before the exit finally triggered. The rules were public knowledge within a decade, yet few people could follow them. That gap between knowing and doing is the real lesson of the experiment, and it is covered in depth in the books we recommend.

Frequently asked questions