The Turtle Strategy: Volatility-Based Position Sizing for Trend Following
In the mid-1980s, legendary commodity traders Richard Dennis and William Eckhardt engaged in a bet that would change financial history. Dennis believed that great traders could be "grown" like turtles on a farm in Singapore, while Eckhardt argued that trading ability was an innate talent. They recruited a group of novices, taught them a specific set of rules, and handed them millions of dollars to trade. The "Turtles" went on to earn hundreds of millions of dollars, proving that a systematic approach to market behavior could outperform individual intuition.
While many focus on the Turtle’s entry signals—breakouts of the 20-day or 55-day highs—the true engine of their success was their proprietary approach to position sizing. This system did not treat all assets as equal. Instead, it used a sophisticated mathematical model to adjust exposure based on market volatility, ensuring that a trade in a stable currency and a trade in a volatile commodity carried the same "economic weight" in the portfolio.
The Bedrock of Risk: Understanding "N"
Before a Turtle could execute a single trade, they had to calculate a value known as "N". In modern terms, this is effectively a 20-day Exponential Moving Average of the Average True Range (ATR). It represents the typical price fluctuation of an asset on any given day. By defining volatility first, the system could normalize risk across disparate markets like gold, oil, and the S&P 500.
The calculation of N followed a strict protocol to ensure that recent price action was weighted more heavily than older data. This allowed the system to adapt quickly when a quiet market suddenly turned turbulent. If the volatility increased, the N value rose, and the required position size for new trades automatically shrank. This built-in defensive mechanism is what allowed the Turtles to survive massive market swings without blowing out their accounts.
How to Calculate N
1. Calculate the True Range (TR) for the day, which is the greatest of: (Current High - Current Low), (Current High - Previous Close), or (Previous Close - Current Low).
2. Calculate the 20-day EMA of these TR values. This result is N.
Note: Because N represents the daily dollar-volatility per contract, it serves as the denominator for every sizing decision in the system.
The Unit Calculation: Scaling by Volatility
The core innovation of the Turtle system was the "Unit." Rather than buying a random number of shares or contracts, the Turtles bought units of risk. One unit was defined as 1% of the total account equity. The goal was for each unit to represent the same dollar-amount of risk based on the asset's current volatility.
To determine how many contracts or shares made up a single unit, the Turtles used a specific formula. This formula ensured that if the price moved by one "N" (one typical daily range), the trader would lose exactly 1% of their account. This creates a level playing field across all markets.
The Unit Sizing Formula
Unit Size = (1% of Account Equity) / (N x Dollars Per Point)
Let us look at a practical calculation for a 1,000,000 USD account trading Gold:
- 1% of Account Equity = 10,000 USD
- Current N (Volatility) = 25.00
- Dollars Per Point (Gold Contract) = 100 USD
- Calculation: 10,000 / (25 x 100) = 4 Contracts
If the same trader moved to a less volatile market where N was only 10.00, the unit size would automatically increase to 10 contracts. This ensures the dollar-risk per trade remains identical regardless of the instrument.
The 1% Threshold: Protecting the Principal
The use of a 1% risk threshold per unit seems conservative to many retail traders, but it is the cornerstone of capital preservation. In trend-following strategies, win rates are typically low (often between 35% and 45%). To survive a long string of losses—what traders call a "drawdown"—the risk per trade must be kept small enough that the account is not decimated before the next big trend arrives.
By limiting initial risk to 1%, a Turtle could survive a 10-trade losing streak and still have approximately 90% of their capital intact. This psychological and financial cushion is vital. If a trader risks 5% or 10% per trade, a standard losing streak leads to a "mathematical death spiral" where the gains required to break even become exponentially larger than the losses incurred.
Small Account Perspective
For accounts under 50,000 USD, 1% risk may result in share counts so small they are difficult to execute with high commission costs. Modern traders often adapt by using 0.5% risk or switching to micro-contracts.
Institutional Perspective
Large funds often use even smaller risk per unit (0.25% or 0.1%) because their total capital is so large that even 1% slippage in a single asset could move the entire market.
Pyramiding: The Art of Adding to Winning Trades
The Turtles did not just place a single trade and walk away. If the market moved in their favor, they "pyramided" into the position, adding more units at specific intervals. This allowed them to maximize profits during the massive, multi-month trends that define the commodity and currency markets.
The rule for pyramiding was strictly tied to volatility. A new unit was added every time the price moved 1/2 N (half of the daily volatility) above the previous entry price. A maximum of four units could be held in any single market. This scaling-in technique ensures that the largest positions are only held in the strongest trends.
| Entry Step | Price Level Trigger | Position Size | Total Risk (N units) |
|---|---|---|---|
| Initial Entry | Breakout Level | 1 Unit | 1.0 N |
| First Add | Entry + 0.5 N | 2 Units | 1.5 N |
| Second Add | Entry + 1.0 N | 3 Units | 2.0 N |
| Third Add | Entry + 1.5 N | 4 Units (Maximum) | 2.5 N |
It is important to note that as units were added, the stop loss for all existing units was moved up. This ensured that even with a four-unit position, the total risk remained controlled. If a market reversed sharply after the fourth unit was added, the trader would exit with a small profit or a controlled loss, rather than a catastrophic blow to the account.
Portfolio Limits: The Diversification Hard-Cap
The greatest risk to any trader is not a single bad trade, but a correlated cluster of trades. For example, if a trader is long on Gold, Silver, and Platinum, they aren't diversified; they have one giant bet on precious metals. The Turtles used strict exposure limits to prevent this "systemic" risk from destroying the portfolio.
No more than 4 units could be held in any single market (e.g., Gold). This prevents a single outlier event in one asset from ruining the month.
No more than 6 units in closely correlated markets (e.g., heating oil and crude oil). This acknowledges that these markets often move as a single block.
No more than 10 units in loosely correlated markets (e.g., Gold and Copper). While different, they both respond to general commodity demand.
No more than 12 units total could be held in one direction (Long or Short). This prevents a total market "melt-up" or "melt-down" from catching the trader with too much exposure in one direction.
Volatility-Based Stops: The Logical Exit
In the Turtle system, the stop loss was always a function of N. The standard stop was placed 2N below the entry price. If the asset was volatile, the stop was wide; if the asset was quiet, the stop was tight. This ensured that the trader would only be "stopped out" if the market behavior actually changed, rather than just hitting a random price point.
As mentioned in the pyramiding section, when a new unit was added, the stops for all previous units were moved to 2N below the newest entry. This "trail" effectively locks in some protection as the trade matures. This logical connection between position size, entry price, and stop distance created a mathematically sound environment where every variable was accounted for.
The Modern Turtle: Adaptations for Today’s Markets
Markets have changed since the 1980s. Volatility is higher, mean-reversion is more common, and electronic trading has reduced the duration of many trends. However, the principles of Turtle sizing remain as relevant as ever. Professional hedge funds and proprietary desks still use volatility-adjusted sizing and "unit" logic to manage billions of dollars.
Today’s traders often adapt the system by using a 0.5% risk per unit to account for the increased "noise" in modern markets. Some also use "Total Risk" caps that reduce the size of all units if the total portfolio drawdown exceeds a certain percentage (e.g., if the account is down 10%, all new unit sizes are cut by half). This "dynamic sizing" is the evolution of the original Turtle rules, focusing on the preservation of the equity curve over any single trade's outcome.