Turtle Trading: the complete rules, and what happens when you run them today

Turtle Trading is a trend-following strategy with completely closed rules: you buy when the price breaks the high of the last 55 days, you exit when it breaks the low of the last 20 days or hits a stop of twice ATR, and you risk a fixed percentage of capital on every trade. There's no profit target. Whoever runs, runs. When we ran these rules on 184 American stocks over the last three years, they achieved +98% with a maximum drawdown of -15.9%compared to +95% and-18.8% for buying and holding the index. The same return, less pain, and a lot more work.

On this page: the story behind the name, the exact rules, what the numbers say and don't say, and why the Turtles' exit doesn't necessarily fit your entry. We tested this on our engine, and the answer was surprising.

The story: a bet on one question

In 1983, two Chicago commodity traders, Richard Dennis and William Eckhardt, argued over whether successful trading is an innate talent or a skill that can be taught. Dennis placed a newspaper ad, selected a group of people with no background, taught them for two weeks and provided them with capital. He called them "Turtles," after a turtle farm he had seen in Singapore: "We will grow traders just like they grow turtles." Reportedly, the group earned large sums in the years that followed. The rules leaked to the public in the 2000s, and since then they have been the benchmark of trend following.

The important lesson from the story isn't the return but what was required to achieve it: blind execution of rules that lose most of the time, for months, without changing them. Most of the original Turtles who failed didn't fail because of the rules but because they didn't stick to them.

The rules, as we ran them

• Entry: Daily close above the high of the previous 55 days. Buy at the open of the next day. No guessing, no "looks strong."
• Initial stop: Twice ATR below the entry price. This is the trade's R, and this distance determines how many units are bought.
• Exit: A close below the low of the previous 20 days, or hitting the stop. There is no profit target at all. A good trade can hold for months.
• Size: Risk of 0.5% of capital per trade (the original: 1% per unit, adding units along the way). A cap of 5% of capital per asset, up to 20 positions at once.
• Regime: Entries only when the S&P 500 is above 200 average. The original Turtles also traded short; we measured long only.

The numbers: the same return as the index, via a completely different route

The measurement: 184 American stocks and ETFs, daily candles, May 2023 through September 2026 (840 trading days), execution at the open following the signal, cost of 0.05% per side, long only and without leverage. The universe is today's and not 2023's — stocks that collapsed and disappeared are not included, and this biases every result upward. The market rose for most of the period.

What this means in practice

• Two-thirds of the trades are losers. This isn't a glitch, it's the mechanism: many small attempts, few big ones. The average profit per winning trade is several times the average loss, and that's what keeps the profit factor at 1.83.
• There was a red year. 2025 ended at-5.3% while the index rose 18%. Whoever gives up after a year like that doesn't get the 2026. This is the real test of the method, and it's psychological, not mathematical.
• The smaller drawdown is the real profit. The Turtles were in cash during parts of the declines, thanks to the exit on a broken low. A similar return with less pain is worth more than an identical return with more pain.
• Survivorship bias works in the method's favor. A universe containing only stocks that survived to this day gives breakouts a better chance of continuing. The real number is probably lower.

The Turtles' exit on a different entry: we tested it, and it didn't work

The temptation is clear: take the Turtles' smart exit (no target, 20-day low breakdown) and attach it to any Breakoutstrategy. We did exactly that on our engine's swing setups, over two years and 185 stocks, and measured a fixed target against a channel exit on the exact same setups. The result: the channel exit produced bigger runners (30 trades of four R or more, compared to 4 with a fixed target), but the win rate dropped from a third to less than that, holding time lengthened by 50%, and the average profit per trade didn't increase. In an account with a limited number of positions, the result was worse: the runners occupied slots for weeks, and what took their place measured negative.

The lesson: an exit belongs to its entry. A 55-day breakout filters long trends, and only those are worth letting run. A short-horizon breakout on the daily timeframe filters short moves, and those are worth selling at a target. Whoever copies half a method gets half the edge, sometimes less.