The turtle trading strategy is one of the most famous experiments in systematic trading history. In 1983, traders Richard Dennis and William Eckhardt ran an experiment: could great traders be trained from scratch using a defined set of mechanical rules, or was trading ability innate? They recruited 23 people with no trading experience — the “turtles” — taught them a complete systematic strategy, and gave them capital to trade. The experiment produced extraordinary results and the rules became the foundation of modern trend-following algorithmic trading.
What Is the Turtle Trading Strategy?
The turtle trading strategy is a rules-based trend-following system built on Donchian Channel breakouts, ATR-based position sizing, and systematic exit rules. It contains no discretion. Every entry, position size, and exit is determined by the rules alone. Dennis taught the system in two weeks. The turtles then went on to collectively generate over $100 million in profits.
The system was designed around the core belief that markets trend, that breakouts from defined ranges signal the beginning of trends, and that disciplined position sizing controls the risk of being wrong. Those three principles remain the foundation of most modern systematic trend-following strategies.
The Entry Rules: Donchian Channel Breakouts
The turtles used two systems with different breakout periods.
System 1 (shorter-term): Enter long when price breaks above the highest high of the previous 20 bars. Enter short when price breaks below the lowest low of the previous 20 bars. Exit when price reaches the 10-bar opposite band — the 10-bar low for long positions, the 10-bar high for short positions. System 1 has a filter: skip the entry if the previous System 1 trade in the same direction was a winning trade. This prevents chasing trends that are already extended.
System 2 (longer-term): Enter long on a 55-bar high breakout. Enter short on a 55-bar low breakout. Exit on a 20-bar opposite band. System 2 has no filter — every breakout is taken. It is slower to enter but captures the major trends that System 1’s filter occasionally misses.
These are Donchian Channel entries. The channel plots the highest high and lowest low over the lookback period. A breakout occurs when price closes outside the channel boundary.
Position Sizing: The ATR Unit
The turtles’ position sizing system was built around the concept of the N unit — where N is the 20-period ATR of the traded asset. The system sized each position to risk 1% of account capital per N move against the trade.
In practice: calculate 1% of account capital. Divide by the current ATR. That gives the number of units to trade. A high-ATR asset (volatile) produces a smaller position size. A low-ATR asset (quiet) produces a larger position size. The result is consistent dollar risk per trade regardless of asset volatility — a volatility-normalised position sizing approach.
The turtles were also allowed to pyramid into winning trades: add one additional unit at each 0.5N move in the direction of the trade, up to four units total. The initial stop was set at 2N below entry for long positions (2N above for shorts). Each new unit added maintained the same 2N stop from the most recent entry — meaning the stop on earlier units was moved up to reduce risk as the position grew.
Why the Strategy Works — and When It Fails
The turtle system works because it cuts losses short and lets profits run. Most breakout attempts fail — the system takes many small losses. But the winning trades that follow genuine trend breakouts produce returns large enough to cover all the small losses and generate overall profit. The edge is not in the win rate (which is typically below 50%) but in the asymmetry of wins versus losses.
The system fails during extended choppy or ranging markets. When price repeatedly breaks above the 20-bar high and reverses immediately, the system takes a series of small losses in quick succession. These periods of drawdown are expected and were documented by the turtles themselves. The system requires patience to hold through these periods without abandoning the rules.
The other failure mode is correlation: if all positions are taken in correlated assets during the same trend, a reversal in that trend hits all positions simultaneously. The turtles managed this through portfolio construction — trading uncorrelated assets to avoid concentration. See our post on strategy correlation for how this applies to modern systematic portfolios.
How to Build the Turtle Trading Strategy in Arrow Algo
Arrow Algo’s no-code builder includes all the blocks needed to replicate the core turtle system.
Add a Donchian Channel block and set the period to 20. It outputs the upper band (20-bar high), lower band (20-bar low), and midpoint. Add a condition block: enter long when the current close is above the upper band, enter short when it is below the lower band.
Add an ATR block set to period 20. This is your N value. Use a divide block to calculate 1% of account capital divided by the ATR output — this produces your unit size. Connect it to your order size input.
For the exit, add a second Donchian Channel block set to period 10. Exit long positions when price breaks below the 10-bar lower band. Exit short positions when price breaks above the 10-bar upper band. Connect this condition to your close position block.
For the stop loss, use a subtract block: entry price minus 2× ATR for long stops, entry price plus 2× ATR for short stops. Connect this to your stop-loss field in the order block. Backtest the strategy across different assets and timeframes in Arrow Algo — trend-following systems benefit from diversification, so test on multiple instruments and review how the drawdown periods correlate between them.
What Are the Key Takeaways?
- The turtle trading strategy is a mechanical breakout trend-following system built on Donchian Channel highs/lows, ATR-based position sizing, and systematic exits
- System 1 uses a 20-bar breakout entry with a filter that skips entries after a winning trade; System 2 uses a 55-bar breakout with no filter
- Position sizing uses 1% of account capital per ATR unit — producing consistent dollar risk regardless of asset volatility
- The strategy has a below-50% win rate but profits from the asymmetry between many small losses and occasional large winning trades
- Extended ranging markets produce losing streaks — the strategy requires patience and consistent rule-following to survive them
- Arrow Algo’s Donchian Channel, ATR, and condition blocks let you build the complete turtle system without code
Disclaimer: This content is for educational purposes only and does not constitute financial advice. Trading involves significant risk and you should only trade with capital you can afford to lose. Past performance is not indicative of future results. Always conduct your own research before making any trading decisions.
Ready to build your own automated trading strategies without writing a single line of code? Start for free at Arrow Algo and join thousands of traders who’ve made the switch to systematic trading.
