Gate Research Institute: Turtle Trading Rules, Classic Trading System Reproduction, Annualized Up to 62.71%

Summary

• The Turtle Trading System is a classic trading strategy based on trend breakouts and mean reversion, which determines entry and exit signals through the Donchian Channel, combined with the ATR indicator for stop-loss and position management, achieving a systematic trend-following approach.
• The improved Turtle Trading System introduces a sliding ATR stop-loss and exclusion zone mechanism based on the traditional Turtle System, dynamically adjusting the stop-loss bandwidth and timing for adding positions, enhancing the strategy's robustness and performance in the high volatility and frequent fluctuations of the cryptocurrency market.
• Backtesting results show that the improved strategy performs better than the original Turtle strategy on GT/USDT hourly data, as evidenced by a higher Sharpe ratio, lower maximum drawdown, and more robust annualized returns, especially the high-frequency version which significantly enhances sensitivity to trends and risk control capabilities.
• Subsequent optimizations of strategy performance can be achieved by introducing leverage, expanding more parameter combinations, and integrating on-chain data with AI-assisted signals, thereby enhancing profit potential and risk management levels.

Introduction

The Turtle Trading Rules were developed in the 1980s by legendary trader Richard Dennis and his partner William Eckhardt as a trend-following trading system. In a famous experiment, Dennis successfully trained a group of inexperienced ordinary people through short-term training and provided them with a clear set of trading rules, creating a group of highly profitable traders known as "Turtle Traders." This experiment not only validated the replicability of systematic trading but also established the classic status of trend breakout strategies in technical analysis.

In traditional financial markets, the Turtle Trading strategy is widely popular due to its clear entry and exit rules, risk control measures, and trend identification capabilities. Especially in the commodity futures market from 1990 to 2000, it achieved an annualized return of up to 24%; in the Hang Seng Index futures market from 2005 to 2015, it reached an annualized return of 12%.

With the rise of the cryptocurrency market, this new asset class has become a new battleground for technical trading strategies due to its high volatility and strong trending nature. However, there are many structural differences between the cryptocurrency market and traditional markets: characteristics such as 7×24 hour trading, generally higher volatility, stronger emotional drivers, and shallower market depth present significant challenges for the migration of original strategies.

Does the Turtle Trading Rule still work in the highly volatile cryptocurrency market?

In recent years, academia and industry have gradually explored the incorporation of traditional trend strategies into crypto assets, such as the improved turtle trading system (AdTurtle) proposed by AdTurtle (2020). This report will reconstruct its application to the GT/USDT trading pair and conduct a systematic backtest evaluation of historical data from 2022 to 2025. The main research content includes:

• Verify the applicability of traditional turtle trading strategies in cryptocurrency trading;
• Explore the actual effects of introducing a sliding ATR stop-loss and exclusion interval mechanism in the improved turtle trading system;
• Propose optimization directions that adapt to the structure of the cryptocurrency market based on AdTurtle.

Traditional Turtle Trading System

The traditional turtle trading system is one of the typical trend-following strategies. Its core logic is "when the price breaks above the previous high, buy and hold; increase positions when the trend continues; close positions and exit when the trend reverses." The specific execution involves the following concepts:

• Donchian Channel: Constructed using the highest and lowest prices over the past N days to determine breakout signals.
• ATR (Average True Range): An indicator that measures market volatility, widely used to calculate stop-loss levels.

2.1 Entry Signal: Price Breakthrough

• If the current price breaks through the highest point of the past N days, which is the upper band of the Donchian channel, establish a long position.

• If it falls below the lowest point of the past N days, which is the lower bound, establish a short position.

• The Donchian Channel period N represents the observation window used to calculate the "historical high/low points", reflecting the length of market trends.

• Common Settings:

• Fast System: Entry Period N = 20, Exit Period M = 10.

• Slow system: Entry period N = 55, Exit period M = 20.

2.2 Stop Loss Settings: Based on ATR

• Set stop-loss level when opening a position: Opening price ± 2 × ATR.
• ATR (Average True Range) measures market volatility.
• The ATR period n represents the number of days used to calculate the average true range, typically set to 14.

2.3 Position Increasing Mechanism: Add Positions in the Direction of the Trend

• If the price rises by 0.5 × ATR (for long) or falls by 0.5 × ATR (for short), gradually increase the position in the direction of the trend;
• Each time, the risk control for additional positions is 1-2% of the account, with a maximum of 4 additional positions, building positions in batches and amplifying profits.

2.4 Entry Signal: Reverse Breakthrough

• If the price breaks below (or above) a shorter period Donchian channel, it indicates that the trend may be ending;
• Liquidate immediately to secure profits or avoid drawdowns;
• The exit cycle is usually shorter than the entry cycle, such as 10 days or 20 days.

2.5 Capital Management and Risk Control

• The maximum loss per transaction shall not exceed 2% of the account balance;
• Position size is dynamically adjusted based on market volatility (ATR); the greater the volatility, the smaller the position.
• Accurately calculate the position size before each trade, prioritizing risk control over predicting market trends.

Improved Turtle Trading System

AdTurtle is an optimized version of the classic turtle strategy, maintaining its core idea of trend breakout while introducing higher robustness in stop-loss logic and entry mechanisms. The ATR (Average True Range) indicator is introduced as an Exclusion Zone to avoid re-entering immediately after a stop loss, improving the strategy's stability and performance. This system is named AdTurtle (Advanced Turtle) and is the first to combine sliding and variable ATR stop-loss strategies with exclusion zones in the turtle trading system. The core objective is:

• Avoid opening a position immediately after frequently hitting stop-losses;
• Improve stability in high volatility market conditions;
• Adapt to high-frequency trading or automated strategies.

The concepts involved are:

• Sliding Stop Loss: As the price moves in a favorable direction, the stop loss line moves up/down accordingly, locking in some profits.
• Variable Stop Loss: The stop-loss bandwidth dynamically adjusts with the current ATR, adapting to changes in market volatility.
• Exclusion Zone: Set a buffer zone after a stop loss, allowing for re-entry only when the price breaks through this zone, in order to avoid repeated stop losses during frequent fluctuations.

The figure below shows the AdTurtle infrastructure:

3.1 Entry signal: Price breakout + Exclusion zone filter

• Similarly based on Donchian channels to identify trend starting points;

• Introduce "Exclusion Zone":

• When the last transaction was exited due to a stop loss, the system will not immediately open a new position;

• You need to wait for the price to move away from the previous stop loss price ± Y × ATR before reopening a position;

• Effectively avoid frequent entry and exit during extreme fluctuations.

• Donchian channel periods are categorized as:

• Standard cycle: x (opening position) and x/n (closing position);

• Expansion period: y (re-entry) and y/m (re-closing), used to filter high-frequency repeated entries and exits.

3.2 Stop Loss Mechanism: Trailing + Variable ATR Range

Compared to the traditional fixed 2 × ATR stop loss, AdTurtle employs a combination mechanism of trailing stop loss + variable interval width to achieve smarter risk control.

• Initial Stop Loss Setting (When Opening Position):

• Long Position Opening:

• Short Selling:

• Sliding Update Logic (when the price moves in a favorable direction):

• Long stop loss position updated to:

• The stop-loss position for the short has been updated to:

• Variable Interval Mechanism (ATR Real-time Update):

• Each K-line updates the ATR value:

• When volatility rises, the stop-loss automatically widens; when volatility decreases, the stop-loss tightens, helping to adapt to market conditions.

This mechanism can:

• Lock in trend profits;
• Avoid short-term price noise;
• Improve the rationality and timeliness of stop-loss execution.

3.3 Trend Following: Increasing Positions Along with Trend Continuation

• Every time the price moves in a favorable direction by Z × ATR, automatically increase the position once (Z is a custom multiplier parameter; used to set the sensitivity of the "position increase trigger threshold");
• The risk for each additional position is 4% of the account funds, with a maximum of 4 additional positions, and a total risk limit of 20%;
• The position addition logic is consistent with the classic turtle strategy, still adopting a pyramid-style incremental entry.

3.4 Risk Management: Dynamic Calculation + Position Control

• Position size is calculated based on the current ATR value of the market; larger volatility means smaller positions.
• Introduce smarter triggering mechanisms (exclusion zones, dynamic stop-loss) to improve actual execution results;

Comparison of Two Turtle Trading Systems 3.5

In the 1980s, the Turtle Trading System became legendary among trend-following strategies due to its simple rules and impressive returns. Its core idea is to identify price breakout signals through the Donchian Channel, set a fixed multiple of the ATR for stop-loss to control risk, and use a pyramid-style scaling to follow the trend. However, with the evolution of market structures, especially in today's environment of high-frequency trading and frequent price false breakouts, the classic Turtle strategy has revealed some obvious shortcomings.

The most common issue is that during price false breakouts and choppy markets, the strategy easily "just stops out and re-enters," leading to amplified consecutive losses. Traditional fixed stop-loss widths (such as 2 × ATR) also lack the ability to adapt to current market volatility, which may result in stopping out too early during significant volatility or leaving the risk exposure too large when volatility decreases. Additionally, due to the system's lack of a "buffer period" for market rhythm, it mechanically enters and exits even after extreme emotions or unexpected events, resulting in increased drawdowns and reduced strategy stability.

AdTurtle retains the basic structure of the turtle strategy "breakthrough + adding positions + risk control" while introducing three key optimizations: exclusion range, variable stop-loss mechanism, and dynamic entry control. Among them, the setting of the exclusion range is one of the core innovations of the entire system. Once a trade exits due to a stop-loss, the system will not immediately allow for a new position; instead, it requires the price to break through the stop-loss price ± Y × ATR range before re-entering. This mechanism significantly reduces the chain damage of "stop-loss - re-entry - stop-loss" in a volatile market.

In terms of stop-loss logic, AdTurtle employs a sliding + variable-width stop-loss mechanism. When the price moves in a favorable direction, the stop-loss level "slides" to lock in profits; meanwhile, the width of the stop-loss band is adjusted in real-time based on the ATR, expanding automatically during high market volatility and tightening during low volatility. This dynamic mechanism aligns more closely with actual market conditions and effectively prevents being taken out of the market by short-term noise.

During trend continuation, AdTurtle retains the classic strategy of "adding to the position every Z × ATR", emphasizing the gradual increase of positions based on profits rather than taking large risks all at once. The number of additions and the total risk limit are also strictly defined, further strengthening risk control. In terms of position management, the system dynamically adjusts the position size based on the current market ATR level; the greater the volatility, the smaller the position, ensuring that risk remains within a controllable range.

The AdTurtle strategy places greater emphasis on robustness and adaptability under complex market conditions. It is not a simple replacement for classical strategies, but rather provides a more reasonable choice in different market scenarios. For markets with clear trends and stable rhythms (such as certain commodity futures or major indices), the classic turtle strategy still exhibits strong performance. However, in environments like crypto assets, forex, or any high volatility and frequently fluctuating markets, AdTurtle offers a trading logic with lower drawdowns and higher win rates by excluding ranges and implementing a dynamic stop-loss mechanism.

Trading System Backtesting

To evaluate the actual performance of the two strategies, this article selects the GT/USDT trading pair from Gate.io as the research subject. The backtesting time frame is set from 2024 to 2025, with a data granularity of 1 hour. The initial capital is 1 million USDT, with no leverage used, and accounting for trading fees (0.1% for both sides) and slippage (0.05%).

4.1 Data Sources and Preprocessing

• Underlying Asset: GT/USDT
• Data source: Gate API (Kline data)
• Time period: January 1, 2024 to January 1, 2025
• Time granularity: 1 hour K-line
• Data processing: Unified format

4.2 Trading and Backtesting Assumptions

• Initial capital: 1,000,000 USDT
• Leverage: No leverage used
• Transaction Cost: Bilateral 0.1% fee + 0.05% slippage for each opening and closing position
• Position Limit: The maximum holding for a single variety shall not exceed 30% of the account equity.
• Signal Execution: Execute at the opening price of the next K line after confirming the closing of the K line.

4.3 Strategy Parameter Optimization

We condense the core parameters of each strategy into a quintuple (X / Y / N / M / P), representing respectively:

• X: Entry Period (Donchian Channel)
• Y: Entry Cycle (Donchian Channel)
• N: ATR calculation period
• M: Initial stop loss multiplier (× ATR)
• P: Exclusion Interval Multiples (× ATR)

The strategy parameters are optimized and selected through grid search to find the best parameter combination.

4.4 Strategy Backtest Results

The figure below shows the backtesting results of the best parameter combinations for the three strategies:

The traditional turtle strategy performs excellently in a clear trending market, but experiences significant drawdowns during volatile or rapidly reversing phases. The AdTurtle strategy, due to its exclusion zone and dynamic stop-loss mechanism, effectively avoids most false signals and outperforms the traditional version in terms of overall return, Sharpe ratio, and maximum drawdown metrics. The AdTurtle strategy exhibits the most stability in its short-cycle versions. After optimization through grid search, the best-performing strategy combination can achieve an annualized return of 62.71%, with maximum drawdown controlled within 15%.

Conclusion

The Turtle Trading Rules, as a classic trend trading model, have an irreplaceable position in terms of clear structure and rigorous logic. Through a systematic framework for trend identification and risk management, it still holds considerable applicability in the cryptocurrency market. However, the volatility characteristics, trading mechanisms, and investor structures of crypto assets differ from those of traditional markets, necessitating adaptations and optimizations of the original strategy to align with market structures during migration. The AdTurtle strategy significantly enhances the strategy's survival ability and return stability in high-frequency and volatile market conditions by introducing mechanisms such as exclusion zones, dynamic stop-loss, and variable position-sizing thresholds.

Next, investors can expand their returns by testing more parameter combinations and introducing leverage. It is recommended to explore the combination of on-chain data (such as capital flow, position changes), macro sentiment indicators (such as the Fear and Greed Index), and machine learning models to further enhance signal recognition and trade execution, driving the intelligent evolution of trend trading strategies in the cryptocurrency market to a higher dimension.

References

• Github, https://github.com/odonnell31/Turtle-Trading-Simulator
• Risk and Financial Management, https://www.mdpi.com/1911-8074/12/2/96

Gate Research Institute is a comprehensive blockchain and cryptocurrency research platform that provides readers with in-depth content, including technical analysis, hot insights, market reviews, industry research, trend forecasts, and macroeconomic policy analysis.

Disclaimer Investing in the cryptocurrency market involves high risks, and users are advised to conduct independent research and fully understand the nature of the assets and products being purchased before making any investment decisions. Gate does not bear any responsibility for losses or damages resulting from such investment decisions.