What started as a friendly bet with a fellow trader changed the lives of many and started a whole school of thought in trading. Richard Dennis believed that anyone could become a trader if given a working system and mentoring, and William Eckhardt took him on that bet. After publishing an ad in the Wall Street Journal, he recruited 14 inexperienced traders to start his experiment. Those traders were the “turtles,” as Dennis called them, whom he taught his simple breakout trading system.
The Turtle trading system was quite simple and easy to follow. Nowadays, even a novice programmer can write an algorithm to automate the strategy.
Basics of the System
Dennis’ system was buying breakouts of 20 and 55-day highs, and selling breakdowns of 20 and 55-day lows. The system is clearly part of the trend following school of thought in trading, starkly contrasting to mean reversion, which they were firmly against. The original Wall Street Journal ad had a true/false statement to weed out traders with a mean reversion way of thinking, “The big money in trading is made when one can get long at lows after a big downtrend.”
The system exclusively traded futures contracts, all except grains and meats. They avoid meats because of rampant price manipulation of meats futures in the 1980s, which has presumably ended.
They managed risk by risking a fixed amount of capital per trade, and using average true range to quantify risk, and ensure they were placing stops correctly.
Entry Rules
There were two time frames that the Turtles traded, one based on a 20-day lookback window, the other on a longer term 55-day lookback. The way they traded these time frames slightly diverge.
20-day Breakouts
Longs:
Buy when price trades one tick above the 20-day high.
Shorts:
Short when price trades one tick below the 20-day low.
Additional rule for shorts and longs: If the previous 20-day breakout trade was successful, do not take this trade.
55-day Breakouts
Longs:
Buy when price trades one tick above the 55-day high
Shorts:
Short when price trades one tick below the 55-day low.
No additional rules based on previous breakout success.
Stop Losses
The Turtles used a quantified stop loss. To figure out what price their stops would be placed at, they used a volatility measure Dennis called N.
N is the 20-day exponential moving average of the average true range indicator. Looking at an ATR indicator, you can see it is quite choppy and overstates recent volatility, hence why Dennis smoothed it with a moving average.
For example, the S&P E-mini’s N level is at 44.43 points. Risking 44.43 points would represent 1% of account equity, and Turtles risked 2% of their account per trade, so they would have 2N stops.
Position Sizing
The Turtles used quite advanced position sizing techniques for their day, and even today, compared to many traders.
Each position was limited to a maximum loss of 2% of account equity.
In order to quantify the amount of contracts they should trade, they came up with a formula called Dollar Volatility, which took N, their risk parameter, and multiplied it by the dollars per point of a contract. This number is Dollar Volatility.
After arriving at Dollar Volatility, the would divide 1% of their account by Dollar Volatility.
For example:
S&P futures move in 25 cent increments, called ticks. Each tick is $12.50 per contract. So, a $0.25 * 4 = $50. The S&P’s dollars per point is $50. Now we multiply that by it’s N, which is 44.43. We get 2,221.50, which is our dollar volatility for the S&P E-minis.
Let’s assume we have a $1,000,000 account value. 1% of the account is $10,000. Divide this by the dollar volatility of 2,221.50, and we get 4.5. The Turtles would round down their contract size, so in this case, you would trade 4 contracts.
Exit Rules
In addition to their stop losses, the Turtles had other exit criteria to improve their trade expectancy. The exits differed slightly from the 20-day to 55-day system.
20-day Breakout
Long:
Exit at a 10-day low.
Short:
Exit at a 10-day high
Exit trade intraday, not at close.
Handling Correlation
Many of the markets the Turtles traded were closely correlated, because of this, trades would often occur together. So, if crude oil broke out, Natural Gas might follow shortly after. In order to ensure they weren’t too exposed to one sector of the market, they would set limits to how many units, or contracts they can be long/short.
Here is a table taken from Curtis Faith whitepaper on the Turtle strategy:
Level | Type | Max Units |
1 | Single market | 4 Units |
2 | Closely Correlated Markets | 6 Units |
3 | Loosely Correlated Markets | 10 Units |
4 | Single Direction – Long/Short | 12 Units |
Summary
Richard Dennis often said that you could publish trading rules in the newspaper and it wouldn’t matter because people couldn’t follow them. There are plenty of positive expectancy trading systems out there, but there are much fewer tradings successfully trading them. When large amounts of capital are on the line, the human reptile brain is engaged and their emotions force them to override their system, ruining their expectancy.
There were multiple failed traders who were fired from the program because they could not follow the simple rules. Even though an experienced trader worth over $100 million was mentoring them on following the trading rules, and acquainting them with the inconsistencies of each market, some still couldn’t bare to override the rules, thinking they know better. If the Turtles teach you anything, it’s to follow your system.