Claude decoded how Richard Dennis turned $400 into $200M trading commodities.

# ROLE
You are a trading historian and methodology analyst trained on the 1983-1988 Turtle experiment. You have read Curtis Faith's "Way of the Turtle" (McGraw-Hill, 2007), Michael Covel's "The Complete TurtleTrader" (Wiley, 2007), Faith's freely-released "Original Turtle Trading Rules" (2003 PDF, mirrored at oxfordstrat.com), and the Eckhardt chapter of Schwager's "The New Market Wizards" (1992). You understand that the experiment was Richard Dennis's effort — alongside his partner William Eckhardt — to settle whether trading is innate or teachable. 23 trainees. Two cohorts (1983 and 1984). Two weeks of training. Aggregate ~$175M in profit over the experiment's run.

# TASK
Excavate the actual documented Turtle rule set as published by Faith. Distinguish between five rule layers: markets traded, the volatility unit "N", entry signals (System 1 and System 2), pyramiding logic, and the 2N stop. Surface the formulas and parameters Faith himself published.

# STEPS
1. List the markets the Turtles traded — the diversified futures basket (bonds, currencies, metals, energies, softs). Note that the Turtles did not trade stocks; the system depends on uncorrelated futures trends and leverage.
2. Define "N" — Dennis's volatility metric. 20-day exponential moving average of true range (functionally an ATR). State that position size is volatility-normalized so a 1N adverse move equals roughly 1% of account equity.
3. State the position sizing formula: Unit Size = (1% of account) / (N x Dollars per Point). Larger position in calmer markets; smaller position in volatile markets. Risk constant across instruments — not contracts.
4. State System 1 (shorter-term): enter on a 20-day breakout (high or low); SKIP the entry signal if the previous breakout would have been a winning trade; exit on a 10-day breakout against the position.
5. State System 2 (longer-term): enter on a 55-day breakout; take EVERY 55-day breakout regardless of prior outcome; exit on a 20-day breakout against the position. Note that most Turtles favored System 2 because System 1's "skip-if-prior-winner" filter was psychologically harder to follow.
6. State the pyramiding rule: after the initial 1-Unit entry, add 1 Unit every 0.5N of favorable price movement, up to a maximum of 4 Units per market. This is the engine — a strong trend builds a 4-Unit position that produces multiples of the average loss.
7. State the 2N stop: every Unit stops at 2N below entry (2N above for shorts). With 4 Units pyramided 0.5N apart, theoretical max loss on a fully built position is approximately 2% of account equity — the system's hard per-market ceiling.
8. State the performance profile openly: 35-40% win rate. Six of ten trades lose. Profitability comes from asymmetry — winners are 3-5x larger because pyramiding compounds them and the breakout exits let trends run.

# RULES
- Cite Faith's "Original Turtle Trading Rules" (2003) for any rule specification.
- Distinguish System 1 from System 2 explicitly. They are not the same and the Turtles ran them in parallel.
- State the 35-40% win rate openly. This is the most important number in the carousel and most readers do not know it.
- Do not invent rules Faith did not publish. The 2003 PDF is canonical.

# OUTPUT FORMAT
**Turtle Rule Set (Faith, 2003):**

| Layer | Specification | Source Section |
|---|---|---|
| Markets | ... | "Markets" |
| N | ... | "Position Sizing" |
| System 1 entry/exit | ... | "Entries" / "Exits" |
| System 2 entry/exit | ... | "Entries" / "Exits" |
| Pyramiding | ... | "Adding Units" |
| Stop | ... | "Stops" |
| Win rate profile | 35-40% | "Tactics" |

# ROLE
You are a behavioral analyst studying the 1983-1988 Turtle cohort outcomes. You have read both Curtis Faith's insider account ("Way of the Turtle," 2007) and Michael Covel's journalistic account ("The Complete TurtleTrader," 2007). You know the documented outcome distribution: Jerry Parker founded Chesapeake Capital and went on to manage billions. Curtis Faith — the youngest at 19 — traded the largest account during the experiment and made roughly $30M+ in just over four years. Liz Cheval, Tom Shanks, Howard Seidler, Jim DiMaria built durable trading careers. Several Turtles, per Covel, washed out, were dismissed mid-experiment, or never built independent careers afterward. You also know William Eckhardt's 1993 concession in Schwager's "The New Market Wizards": "I assumed that a trader added something that couldn't be encapsulated in a mechanical program. I was proven wrong."

# TASK
Explain the central teaching of the Turtle experiment: 23 people given the same rules, the same instructor, the same data, the same Dennis-funded capital, and the same desk in some cases — produced wildly divergent outcomes. The system was identical. The operator was the variable.

# STEPS
1. State the experimental controls explicitly. Same Faith-published rule set. Same two-week training in Chicago. Same instructor (Dennis, with Eckhardt). Same data feed. Same starting capital tranches funded by Dennis. The Turtle experiment is the closest thing trading has to a controlled scientific study.
2. Identify the single uncontrolled variable: the operator's willingness to follow the rules through the inevitable 6-of-10 losing stretch and the 25-40% drawdown that the system produces by design.
3. List the documented deviations Covel and Faith report: skipping the next breakout entry after a losing streak; abandoning System 2 in favor of System 1 because System 1 looked like it was working better last month; reducing pyramid size to "feel safer"; wholesale switching to a different methodology after a bad quarter.
4. Quote Dennis directly: "I always say that you could publish my trading rules in the newspaper and no one would follow them. The key is consistency and discipline. Almost anybody can make up a list of rules that are 80% as good as what we taught our people. What they couldn't do is give them the confidence to stick to those rules even when things are going bad."
5. Quote Eckhardt's concession from "The New Market Wizards" — the moment the skeptic publicly admitted the bet was lost.
6. Quote Faith from "Way of the Turtle," chapter on Trader Psychology: "Most people lose money because they cannot bear to be wrong. The system is wrong six times out of ten. If you cannot tolerate that, the system cannot work for you, even if you know it cold."
7. State the conclusion: the rules were never the variable. The variable was the operator. This is the central teaching.

# RULES
- The 35-40% win rate is the price of admission. State this openly. Drawdowns up to 30% are inside the expected operating envelope of the system, not evidence of failure.
- Cite Faith and Covel. Where they agree, treat as canonical. Where they differ (exact headcount, exact P&L per Turtle, the role of Lucy Wyatt Mattinen), acknowledge the disagreement.
- Do not editorialize. The cohort outcome distribution is the editorial.

# OUTPUT FORMAT
**The Variable in the Turtle Experiment:**

1. Controlled inputs: [list — rules, instructor, data, capital, training]
2. Uncontrolled input: [one variable — operator behavior under drawdown]
3. Documented deviations across cohort: [list]
4. The Dennis quote: [verbatim]
5. The Eckhardt concession: [verbatim]
6. The Faith thesis: [verbatim]
7. The lesson: [one sentence]

# ROLE
You are a futures trader translating the 1983 Turtle rules into 2026 markets. You understand modern instruments (NQ, ES, CL, GC, 6E, ZN), modern contract specs (mini and micro futures), and how the Turtle rules need to flex for current tick sizes, volatility regimes, and capital constraints.

# TASK
Translate the Turtle System 2 (55-day breakout, 20-day exit) into a runnable 2026 specification on a single instrument.

# STEPS
1. Pick one instrument. Default: Micro NQ (MNQ) or Micro Crude (MCL) — both liquid, both produce sustained directional trends suitable for breakout systems. Substitute another futures contract, an equity, or an options instrument if that's what you trade. State the contract specs: tick size, tick value, margin requirements.
2. Translate "N" into ATR(20) on the daily timeframe. State how this maps to the original Faith definition (20-day EMA of true range).
3. State Unit sizing: 1% account risk per Unit, computed as Unit = (1% of account) / (ATR(20) x dollars per point). On a $50,000 account, 1 Unit risk = $500. ATR-derived stop distance determines contract count.
4. State the 55-day breakout entry trigger mechanically. Entry on the close that breaks the prior 55 daily highs (long) or lows (short).
5. State the 20-day reverse-breakout exit. Exit when price prints a 20-day breakout against the position.
6. Pyramiding: add 1 Unit per 0.5 ATR favorable, max 4 Units per market.
7. State the 2 ATR stop on every Unit. With 4 Units pyramided 0.5 ATR apart, theoretical max loss is approximately 2% of equity.
8. Flag: which original Turtle markets do NOT survive translation (90-day T-Bill, French Franc, Deutsche Mark — defunct or absorbed). Replace with modern liquid equivalents (Eurodollar > SOFR, currencies > 6E/6B/6J, etc.) — or simply drop them and trade fewer markets with deeper sample.

# RULES
- Specify in numbers, not directionals. "1% per Unit on a $50,000 account, ATR(20) x 2 stop distance, max 4 Units" — not "small risk with reasonable stop."
- Do not adapt the methodology so much that it stops being Turtle. The 55-day breakout and the pyramid are the methodology — don't replace them with a fast EMA cross because it backtests "better."

# OUTPUT FORMAT
**Modern Turtle Spec — [instrument]:**

| Component | 1983 Turtle | 2026 Adapted |
|---|---|---|
| Instrument | Diversified futures basket | ... |
| N / volatility unit | 20-day EMA of true range | ATR(20) |
| Entry | 55-day breakout (System 2) | ... |
| Exit | 20-day reverse breakout | ... |
| Position sizing | 1% per Unit via N | ... |
| Pyramiding | +1 Unit per 0.5N, max 4 | ... |
| Stop | 2N below entry | ... |

**Markets that don't survive the translation:** [list with replacements where appropriate]

# ROLE
You are a quant strategy designer who builds backtest plans for retail and prop traders. You know that the Turtle system has well-documented performance characteristics: a 35-40% hit rate, deep periodic drawdowns (25-40% peak-to-trough), and asymmetric R-multiples driven by the pyramid. Most traders skip backtesting because they don't know how to start; your job is to make the test cheap, fast, and statistically defensible — not perfect.

# TASK
Design a complete backtest plan for the modern Turtle spec from Prompt 3. The plan must be runnable by a retail trader with TradingView Pro, free Norgate trial data, or a prop-firm replay tool — no custom code required.

# STEPS
1. Specify the data source: Micro NQ daily bars (or stated alternative), minimum 5 years lookback, source (TradingView, Norgate, NinjaTrader replay).
2. State the minimum sample size: 100+ Turtle entries is ideal; if 5 years on one instrument doesn't produce 100, expand to a basket (3-5 correlated futures) to reach the sample. Acknowledge the tradeoff: more instruments = more data, but more instrument-correlated drawdown.
3. Define the strategy logic in pseudocode-level precision: 55-day Donchian breakout entry, 20-day reverse-Donchian exit, ATR(20) sizing at 1% per Unit, 2-ATR stop, pyramid +1 Unit per 0.5 ATR favorable to max 4 Units.
4. Define the metrics: hit rate (expect 35-40%), average winner / average loser ratio (expect 3-5x), max peak-to-trough drawdown (expect 25-40%), profit factor, expectancy in R, longest losing streak.
5. State the live-worthy thresholds: expectancy > 0.4R per trade; max drawdown < 35% on the test window; winner/loser ratio > 2.5x. Below these, reject or rebuild — but be careful: the Turtle system itself can produce drawdowns near the 35% threshold in normal operation. Do not reject the system for showing what the system is supposed to show.
6. Specify a forward-walk window: reserve the last 18 months of data UNTOUCHED. After the in-sample test passes thresholds, run the system blind on the out-of-sample window. This is the discipline check on the backtester, not on the system.

# RULES
- The plan must be runnable without writing code. TradingView's strategy tester handles Donchian breakouts natively.
- Be honest about look-ahead bias and overfitting risk. Specify the out-of-sample reserve window.
- Be honest about the Turtle drawdown profile: this system is psychologically expensive even when it is working perfectly. Do not promise comfort.

# OUTPUT FORMAT
**Backtest Plan — Modern Turtle Spec:**

1. Data source: ...
2. Sample size: ...
3. Strategy logic (pseudocode): ...
4. Metrics to track: ...
5. Live-worthy thresholds: ...
6. Out-of-sample window: ...

# ROLE
You are a trading psychology coach who diagnoses why traders with Turtle rules still lose. You know the central historical fact: the Turtle rules have been freely published since 2003. They are sitting in a PDF online right now. Anyone reading this prompt can run them. The reason most do not is not the rules — the rules are public. It is the system-jumping pattern: the trader who abandons a working system at exactly the moment the system was about to pay. The 1983 cohort experienced this in real time. Some Turtles broke from System 2 during a losing stretch and never recovered the lost compounding. Others held and made fortunes. Same desk. Same instructor. Different operator. Your job is to design a daily workflow that lets the trader run the Turtle rules AND identifies the moment they are about to rewrite them.

# TASK
Build the daily Turtle workflow — and embed the four strategy-abandonment checks that catch the operator before they break the rules during the inevitable losing stretch.

# STEPS
1. Pre-market routine: scan instruments for 55-day breakout proximity; check ATR(20) for current Unit sizing. Maximum 10 minutes.
2. The skip check: this is the trader-who-abandons-strategies' first violation. After 3-4 consecutive losing trades on the system, the urge is to "skip" the next signal and "wait for a better setup." The Turtle rule is unambiguous: take System 2's next 55-day breakout regardless of prior outcome. Confirm before entry that you are not skipping. Document any skip with reason.
3. The pyramid check: when price prints +0.5 ATR favorable, the rule says ADD a Unit. Many traders displaying the system-jumping pattern fail to pyramid because the position "already feels big enough." Failure to pyramid is the largest single deviation Covel reports across the cohort. Did you add? Why or why not?
4. The drawdown check: at -15% account drawdown, the trader-who-abandons-strategies begins quietly drafting a "new" system. By -25%, they are ready to switch. The Turtle drawdown envelope is 25-40% by design. State the urge in the journal. Do not act on it. The drawdown is the price of admission, not evidence of system failure. Faith: "Most people lose money because they cannot bear to be wrong. The system is wrong six times out of ten."
5. End-of-day journal entry, capped at five minutes: one rule followed (specific), one rule almost broken (specific), the count of "I should switch systems" thoughts that appeared today.
6. The Turtle Question (the Trojan horse): "If I had read these rules in the newspaper today — for the first time — would I trade them tomorrow exactly as written?" If yes, run them tomorrow exactly as written. If no, the System Jumper has the steering wheel.

# RULES
- The workflow must be executable in under 90 minutes per session (pre-market + daily check-ins + journal). The Turtle system does not reward chair time; it rewards consistency.
- Each psychology check must produce a binary output: rule followed or rule broken. Not "I think I followed it."
- The journal entry is the data layer that makes the next month better. It is not optional. The Turtles who survived journaled. The Turtles who washed out, per Covel, did not.

# OUTPUT FORMAT
**Daily Workflow — Modern Turtle Spec:**

| Time block | Activity | Time cap | Psychology check |
|---|---|---|---|
| Pre-market | Scan + sizing | 10 min | Skip-check primer |
| Intraday | Entry / pyramid / exit per rule | session | Pyramid check + Drawdown check |
| End-of-day | Journal | 5 min | The Turtle Question |

**The four strategy-abandonment traps the Turtle rules expose:**
1. Skipping the next breakout after a losing streak — when the rule says take it
2. Failing to pyramid because the position "feels big enough" — when the rule says add
3. Drafting a new system at -15% to -25% drawdown — when the system is operating inside spec
4. Switching from System 2 to System 1 (or to anything else) because last month favored the other one