In 1983, a former lawyer and real estate investor named Richard Dennis made a bet with his business partner William Eckhardt. The wager was deceptively simple: could they take a group of complete novices and turn them into successful traders in just two weeks? Dennis believed that trading skills could be taught like a set of rules. Eckhardt believed it required innate talent. To settle the argument, they ran an advertisement in the Wall Street Journal, recruited a diverse group of applicants from a professional gambler to a security guard, and brought them to Chicago for a bootcamp. The experiment became legendary, and the group of trainees became known as the Turtles.
What the Turtles proved—and what Dennis had suspected all along—was that a purely mechanical, rule-based approach to speculation could generate extraordinary returns. The Turtle Trading System, as it came to be known, produced annualized returns of over 80% for some of its participants during the four-year experiment. The irony, however, is that the "secret" of their success had nothing to do with predicting where the market was going. It was about abandoning prediction altogether and embracing a simple, often uncomfortable, truth: you don't need to know what happens next; you only need to know what you will do when something happens.
I first encountered the Turtle story not through Dennis's original materials, but through a conversation with an old institutional trader in Singapore. Over a plate of Hainanese chicken rice, he told me, "The Turtles weren't geniuses. Their logic was genius. They had a 56% win rate, maybe less. But they were wealthy because their winning trades were five times larger than their losing trades." That ratio—the idea that you could be wrong almost half the time and still crush the market—struck me as both liberating and deeply counterintuitive.
The Core Doctrine: "Prices Make the Trend, Not the Other Way Around"
The fundamental principle behind the Turtle system is rooted in a concept called trend following. Most retail traders approach the market as a forecasting exercise. They listen to central bankers, analyze earnings reports, study geopolitical tensions, and then form a hypothesis: "I think the dollar will weaken." "I think gold will rally." "I think the Fed will cut rates." The Turtle approach inverts this hierarchy. It begins with a premise of radical ignorance. "We don't know what the market will do," Dennis often said. "It's arrogant to pretend we do."
Instead of forecasting, the Turtles designed a set of rules to react. They used two primary breakout systems:
That was it. There was no fundamental analysis. No judgment calls about "overbought" or "oversold" conditions. The system didn't care about news headlines or macroeconomic data releases. It only cared about one question: has the price moved to a new extreme? If yes, we act. If no, we wait.
I remember testing a version of this system myself on the GBP/JPY pair back in 2018. It was a brutal market, prone to sudden spikes and erratic reversals. The system generated a long signal at 148.50. Within three days, the position was up 200 pips. I felt like a genius. Then, over the next week, it gave back 150 pips before hitting my trailing stop. I ended with a 50-pip gain. The next signal was a short at 146.20. That one ran for 400 pips. My overall win rate that month was around 40%. But I was profitable because the short trade had paid for all the small losses and then some.
The Turtle Risk Management Rulebook
The secret to the Turtle success wasn't the entry signals. Those were public knowledge. The real edge lay in their uncompromising approach to risk. The Turtles operated with a set of rules that would make most modern retail traders squirm.
The 2% Rule (Per-Trade Risk)
The Turtles were not allowed to risk more than 2% of their account equity on any single trade. This is a standard rule today, but in the early 1980s, it was radical. Unlike position traders who set a fixed dollar stop-loss, the Turtles used a volatility-based stop. They calculated the Average True Range (ATR) over the last 20 days to determine how much a given instrument was currently "moving." The position size formula was:
Position Size = (Account Equity × 2%) / (ATR × Contract Multiplier)
For example, if an account had $1,000,000 and the ATR for crude oil was 1.5 points, with each point worth $1,000, the maximum risk per contract was ATR × $1,000 = $1,500. The 2% rule meant the trader could risk $20,000 on that trade, leading to roughly 13 contracts (20,000 / 1,500 ≈ 13.3). This dynamic sizing meant that in volatile markets, the Turtles traded smaller. In calm markets, they traded larger. It was a brilliant self-regulating mechanism that balanced risk across vastly different instruments.
The Trailing Stop (The "No-Touch" Rule)
The Turtles didn't use fixed stop-loss levels. Instead, they used a trailing stop based on a 10-day low (for long positions) or a 10-day high (for short positions). This meant the stop-loss moved up as the price rose, locking in profits, but it never moved down to "give the trade room." Once a position was in profit, the stop was always above breakeven.
However, there was a strict add-on rule: "Never add to a losing position." For the Turtles, adding to a loser was a cardinal sin. It was called "doubling down," and it was explicitly forbidden. If the price moved against them, they took the loss and moved on. If the price moved in their favor, they would pyramid into the trade, adding to the position at fixed intervals (e.g., every half-ATR move in their favor). This practice allowed them to maximize the winners while keeping the losers small. It is a psychological nightmare for most people, because it requires you to be scared when you're losing and brave when you're winning—the exact opposite of the natural human instinct.
The Humility of the 56% Win Rate
One of the most eye-opening statistics from the Turtle experiment is that their average win rate across all systems was only around 56%. This means they were wrong on 44 out of 100 trades. A random coin flip would give you 50%. A 56% win rate is barely above chance.
Yet, the Turtles generated compound annual returns that dwarfed most hedge funds. The mathematics of their success is undeniable. By maintaining a risk/reward ratio that consistently exceeded 3:1 on their winning trades, a 56% win rate becomes a license to print money. Let's run the math:
With that mathematics, you can withstand dozens of consecutive losses and still come out ahead.
I've seen this dynamic play out in a friend's trading account—a mechanical CTA (Commodity Trading Advisor) who manages about $50 million. His win rate hovers around 45% to 50% annually. But his average winner is roughly 4.2 times his average loser. When I asked him how he sleeps at night during a six-trade losing streak, he shrugged and said, "I don't look at the P&L. I look at the system. The system is valid until proven otherwise by drawdown statistics." This is the mindset of a professional: they trust their process more than their predictions.
The "Adaptation Problem" – An Original Perspective
The Turtle system, however, is not a "set it and forget it" machine. In the decades since 1983, markets have undergone profound structural changes. The rise of high-frequency trading (HFT), the explosion of passive index investing, and the unprecedented levels of central bank intervention have distorted the classic breakout patterns that the Turtles exploited.
My Exclusive Perspective: The biggest risk to trend-following systems today isn't the entry or exit rules. It's adaptation failure. The Turtles operated in a time of pure price discovery, where breakouts often signaled genuine shifts in supply and demand. Today, breakouts can be entirely artificial—the result of a single large algorithmic order, a news spike, or a "stop hunting" raid by institutional players. If you apply the 55-day breakout rule blindly today, you will get whipped around more frequently.
To address this, I've added a "friction filter" to my own version of the Turtle system. Before entering a trade, I check the average daily volume and compare it to the previous 20-day average volume. If the breakout occurs on volume that is below the 20-day average, I consider it a weak signal and often reduce my position size by half. This is not a rule from Dennis's original system; it is a modern adaptation.
This view is supported by research from the Bank for International Settlements (BIS). In their 2019 working paper on "Volume and Volatility in FX Markets," BIS analysts observed that "algorithmic and high-frequency trading have increased the incidence of false breakouts, as liquidity is quickly withdrawn during periods of elevated uncertainty." If a breakout doesn't have the volume to support it, you are essentially betting on a pump-and-dump scenario. Adding a volume filter helps weed out the noise. It also aligns with the broader principle of "knowing what you are up against." The market is no longer just speculators and hedgers; it's algorithms that exist to take your stop-loss. You have to be smarter.
The Exit Conundrum
One of the most difficult aspects of trend following is the exit. The Turtles had a simple rule: exit when the price closes below the 10-day low (for longs) or above the 10-day high (for shorts). But in practice, this often meant giving back a significant portion of your profit. I vividly recall a trade in the EUR/JPY cross in 2022. I entered a long at 137.50 on a 55-day breakout. The price surged to 145.80—a massive 830-pip move. My trailing stop was at the 10-day low, which was at 143.20. The price retraced, triggered my stop, and I walked away with 570 pips. It felt terrible to give back 260 pips. But the system was correct. If I had exited early, I would have missed the massive breakout in the next leg.
This is where the psychology of journaling, as we discussed in our previous exploration of Peter Brandt, intersects with the Turtle philosophy. Both systems are designed to eliminate the emotional bias of "I should have taken profits earlier." The Turtle system is built on the assumption that you will leave money on the table. The goal isn't to maximize every trade; the goal is to maximize the portfolio over a series of trades. The exit rule is designed to capture the bulk of a trend. The Turtles understood that a trend is composed of many small pullbacks. If you try to sell the top, you will inevitably sell too early.
Modern Implementation: A Practical Checklist
To translate the Turtle system into modern action, here is a practical checklist I use:
---
Reference
Covel, M. (2007). Trend Following: How Great Traders Make Millions in Up or Down Markets. FT Press.
Curtis, F. (2012). The Original Turtle Trading System.
Bank for International Settlements (BIS). (2019). "Volume and Volatility in FX Markets." BIS Working Papers.
Dennis, R. (1986). Interviews and lecture transcripts from the original Turtle experiment.
本文首发于FXEAR.com，原创内容，未经授权禁止转载。