Covel actually publishes Richard Dennis's Turtle trading rules in full: breakout entries, volatility-based position sizing, pyramiding, drawdown cuts. But the rules take up one chapter out of fourteen. The rest is a Turtle-by-Turtle biography of who got rich and who blew up.
The one idea
The Turtles were taught that price is the only input that matters. No news, no fundamentals, no forecasting. They entered when a market broke out to a new high or low over a set lookback window, on the theory that a breakout is the earliest observable evidence a trend is starting. They had no bias toward being long or short: a market making new lows was as tradeable as one making new highs.
The second half of the method is what actually made the rules work: risk was sized off a volatility measure the Turtles called N (a twenty-day average of the true range, what most traders now call ATR), never off a dollar amount or a gut feeling. Every position risked a fixed 2 percent of account equity, so a corn contract and a Japanese yen contract carried the same dollar risk despite trading completely differently. Winners got more capital added as they moved, in fixed N-based increments, up to a five-unit cap per market.
The discipline was in the exit, not the entry. Every trade had a hard stop at two times N, and the Turtles took whichever exit hit first: the hard stop or the opposite breakout signal. William Eckhardt, who built the exit logic, told the Turtles liquidation criteria mattered more than entry timing: a bad entry with a good exit rule still made money over time.
What it actually teaches
The rules, as taught in the two-week 1983 training and reconstructed in Chapter 5 of the 2007 edition, ran in this order.
MEASURE VOLATILITY WITH N
N is the twenty-day moving average of the true range (the largest of: today's high minus today's low, yesterday's close to today's high, or yesterday's close to today's low). This is the book's name for what most traders now call the Average True Range. Every other rule below is expressed as a multiple of N, not as a dollar figure or a percentage price move.
ENTER ON A PRICE BREAKOUT
System One (S1) entered on a twenty-day (four-week) high or low, but skipped that signal if the prior S1 trade was a winner, forcing a wait for System Two. System Two (S2) entered on a fifty-five-day (eleven-week) high or low with no filter, and existed partly as a fail-safe so a filtered-out S1 trend could still be caught. Turtles split capital roughly half and half between the two systems.
SIZE THE POSITION TO 2 PERCENT RISK
Each 'unit' risked a fixed 2 percent of account equity. Contracts traded = (2% of equity) divided by (2N in dollar terms), rounded down. Worked example from the book: a $100,000 account trading corn with N of 7 cents (dollar value $350) risks $700 at 2N, so $2,000 / $700 = 2.67, rounded down to 2 contracts.
PYRAMID WINNERS EVERY 1N
Add one more unit each time price moves 1N in the trade's favor, up to a cap of four to five units per market. Each addition moves the stop on every unit in the position up to the new unit's 2N stop, so profit is protected as the position grows without capping how far a big trend can run.
EXIT ON WHICHEVER STOP HITS FIRST
Every position carries two live exits: a 2N hard stop, and the opposite-direction breakout exit (a two-week low for S1, a four-week low for S2). Whichever triggers first closes the trade. The book's worked live cattle example shows a five-unit pyramid built from 74.00 to 77.00 and exited at 84.50 for a 35.8 percent account gain, illustrating why the Turtles tolerated many small losing breakouts while waiting for one of these.
CUT UNIT SIZE IN A DRAWDOWN
For every 10 percent drawdown in account equity, cut the trading unit risk by 20 percent. A 2 percent unit becomes 1.6 percent after an 11 percent drawdown, then 1.28 percent after a 22 percent drawdown. Size is restored only as equity recovers, which the book argues protects against ruin during the long stretches when no market is trending.
CAP RISK ACROSS CORRELATED MARKETS
Total portfolio unit risk is calculated by taking the smaller of total long units or total short units, dividing it by two, and subtracting that from the larger side. A book with 7 long units and 3 short units carries 7 minus 1.5, or 5.5 units of total risk, which is how the Turtles held positions in twenty-plus markets without treating each one as full independent risk.
What it looks like Monday morning
Pull daily open, high, low, close for the markets you actually want to trade and compute a twenty-day true-range average by hand or in a spreadsheet: that single column is what every other Turtle rule is built from. Pick one breakout lookback (the book uses twenty and fifty-five days) and one risk percentage (the book uses 2 percent) and paper trade them mechanically for a stretch before risking money, because the rules only work if you take every signal, not the ones that feel right.
Stop reading market news for trade decisions. The book's whole case for breakouts rests on price already reflecting what a news story would tell you, days later and with an opinion attached. If that reflex is the one thing you keep after finishing the book, you've kept the part that matters most.
| Market | N in dollars | 2N stop | Account risk (2%, $100k) | Contracts traded |
|---|---|---|---|---|
| Corn | $350 | $700 | $2,000 | 2.0 |
| Lean Hogs | $420 | $840 | $3,000 | 3.0 |
| Japanese Yen | $725 | $1,500 | $1,875 | 1.0 |
| Ten-Year Notes | $525 | $1,050 | $2,000 | 1.0 |
Contracts = account risk divided by the 2N dollar stop, rounded down, which is how the book equalizes risk across unrelated markets.
Pure price systems are close enough to the North Pole that any departure tends to bring you farther south.Michael W. Covel, The Complete TurtleTrader
Where it fails
- The rules are one chapter out of fourteen. Chapter 5, 'The Rules,' runs pages 83 to 115 in this edition, about thirty pages. The other roughly two hundred pages are Turtle biography, courtroom disputes over who counts as a real Turtle, and Covel's broader trend-following advocacy. A reader who wants only the method is paying for a lot of narrative to get there.
- The System One filter has a documented hole, and the book admits it. S1 skips a breakout if the prior S1 trade won, which means the Turtles could sit out exactly the trade that turns into the big trend. The book's own fix, catching the missed move later on System Two's fifty-five-day signal, means a filtered-out trend is caught later and smaller. Covel calls this 'a built-in problem' in the text rather than resolving it.
- The post-program track record is survivorship-biased. The book profiles the Turtles who compounded into decades of managed-futures success (Jerry Parker, Salem Abraham, Liz Cheval) in detail, while Curtis Faith's failed Acceleration Capital hedge fund (shut down and fined by the CFTC in 2007 after an employee embezzled from client accounts) gets a few pages near the end. Readers never learn how many of the hundreds of Turtle-style copycats named nowhere in the book failed quietly.
- The performance data stops in August 2006. Tables 5.1 and 5.2, comparing Turtle-style trader expectations to stock indexes, run 'from inception to August 2006.' Trend-following CTAs went through a rough stretch from roughly 2011 to 2019 with flat-to-negative returns against a rising stock market, a period this edition cannot address because it predates the book.
- The correlation math assumes a diversified futures book most readers can't run. Table 5.16 lists the original twenty-one markets the Turtles traded, from T-bonds to cocoa to Swiss francs. The unit-limit and long/short correlation rules in this chapter assume margin and access across that many uncorrelated futures markets, which is a different starting point than a retail account trading a handful of ETFs or stocks.
Who it is for
Buy it if
You want the actual mechanical trend-following rule set, in enough numeric detail to code or backtest it yourself, and you don't mind reading two hundred pages of Turtle personalities and lawsuits to get to Chapter 5. It's also worth it if you want to understand volatility-based position sizing (the N/ATR concept) from the source that popularized it for retail traders.
Skip it if
You already know the Turtle rules from another source and want updated data, or you're looking for a how-to workbook rather than a business biography. Anyone trading a small, undiversified account should also expect the correlation and unit-limit sections to assume more market access than they have.
As an Amazon Associate I earn from qualifying purchases. It costs you nothing extra, and it does not change what gets recommended or what the verdict says.