For readers of Your Broker Is Lying to You · 6 AI prompts · Free
Six prompts. Paste them into any AI. The AI does not get to freelance.
Your Broker Is Lying to You is a Mark Fields field manual about why 70–85% of self-directed retail traders lose money every year. The book is the field manual. The bundle is the operating system — six prompts that turn it into something you actually use.
These are not signals. These are not trade ideas. These are six ways to make the AI your structured thinking partner — before, during, and after a trade.
The prompts work with the free tier of ChatGPT, Claude, and Gemini. No special setup. No email gate.
How to use this pack
You do not need an AI subscription. A free ChatGPT account works. A free Claude account works. Gemini works. The prompts are written so the model knows what you want and in what format. Paste, read, act.
You will not paste these into the trading app. You will paste them into a separate AI window, away from the trade, after you have written down what you are about to do. The AI is a thinking partner, not a signal service.
1. Pre-Trade Reasoning Prompt
Use before clicking buy.
I am about to enter a trade. Here are the details: Ticker / instrument: [e.g. EUR/USD, AAPL, BTC] Direction: [long / short] Entry price: [X] Stop loss: [Y] Position size: [Z] Setup / reason for the trade: [describe in 1-2 sentences what made you see this] Before I commit real money, give me 5 reasons this might be a bad idea. Rank them by severity (1 = minor concern, 5 = fatal flaw). Be specific. Do not tell me whether to take the trade. Just make me aware of what I am missing. Output format: 1. [Concern] — Severity [1-5] — [specific evidence or reasoning] 2. [Concern] — Severity [1-5] — [specific evidence or reasoning] ... 5. [Concern] — Severity [1-5] — [specific evidence or reasoning]
Why this prompt exists: every trade you take will feel right at the moment you take it. The trade that loses money always had reasons not to take it. You did not look at them because you were already committed. This prompt forces the AI to surface the reasons you are refusing to see.
2. Position Sizing Sanity Check Prompt
Use after you have calculated your position size.
Sanity check my position size calculation. Account size: [€X] Max risk per trade: [Y%] Entry price: [A] Stop loss price: [B] I calculated position size = [Z units] = [W € total position value]. Check my math. Show me the calculation step by step. Flag any logical error. Tell me what position size I should have, and explain any difference from what I had. Output format: - Your input as I read it: [restate] - Position size I should take: [X units] - Calculation shown: [step by step] - Difference from your number: [Y units, Z%] - Verdict: [Correct / Wrong / You missed something — explain]
Why this prompt exists: a position size error is the difference between losing 1% of your account and losing 10%. A second pair of eyes, even an AI’s, catches the dumb arithmetic that the heat of the moment makes invisible.
3. Post-Trade Review Prompt
Use after you close a trade.
Review this trade I just closed. Trade details: - Date: [open and close] - Instrument: [X] - Direction: [long / short] - Entry: [price] - Exit: [price] - Position size: [units] - P/L: [€ amount / % of account] - Setup / thesis: [what made me take the trade] Review it for me in three sections: 1. WHAT WORKED — what did I get right about this trade? Be specific. If I lost money, do not skip this section. There is always something I got right. 2. WHAT DID NOT WORK — what did I get wrong? Be specific. If I made money, do not skip this section. There is always something I got wrong, or right for the wrong reason. 3. NEXT TIME — if I see the same setup again tomorrow, what would I do differently? Be specific. Avoid generic advice.
Why this prompt exists: a journal entry is not a review. A review is what you wish you had known before the trade. Without it, the next trade is a guess.
4. Drawdown Trigger Check Prompt
Use when you are losing and you should not be.
I am bleeding and I need a second opinion before I do something I will regret. Current state: - Account size: [€X] - Equity now: [€Y] - Drawdown: [Z% from peak] - My max drawdown limit: [W%] - Last 5 trades (P/L each): [list] - Time of day: [HH:MM] - Market mood: [briefly] Should I keep trading today, or stop? Output a clear yes/no answer first, then the reasoning, then the specific conditions under which I am allowed to resume. Output format: - Verdict: [STOP / CONTINUE] - Reasoning: [3-5 sentences, specific to my numbers above] - Conditions to resume: [specific triggers — e.g. "next session, fresh equity, position size cut in half, no revenge trades"] Do not be encouraging. Do not be diplomatic. Be specific.
Why this prompt exists: the worst trades in any retail trader’s history are the trades taken in the 48 hours after a blow-up. This prompt is a circuit breaker. It is the most important prompt in the pack.
5. AI Skepticism Prompt
Use when an AI has just given you a trade idea.
An AI just gave me this trade signal: - Instrument: [X] - Direction: [long / short] - Entry / target / stop: [numbers] - Reasoning given: [paraphrase what the AI said] Do not tell me whether to take the trade. Tell me 3 different ways this signal could be wrong. For each one, explain how I would have discovered it BEFORE the trade, not after. Output format: 1. Failure mode: [what goes wrong] — How I could have spotted it: [specific check] 2. Failure mode: [what goes wrong] — How I could have spotted it: [specific check] 3. Failure mode: [what goes wrong] — How I could have spotted it: [specific check] At the end, add: "None of these mean the trade is wrong. They mean you have not done your job yet."
Why this prompt exists: an AI that gives you a trade idea is not advising you. It is generating text. The trade idea is the input; your job is to interrogate the input until you find the failure modes. If you cannot find three, you have not looked hard enough.
6. Pattern Recognition Prompt (use sparingly)
Use monthly, with extreme caution.
Here are my last [N, e.g. 20] closed trades: [Paste your trade journal entries — one per line, with date, ticker, direction, P/L] Look at the data and describe patterns you see. Do not recommend what to do. Describe what you observe. Be specific. Do not invent patterns that the data does not support. Output format: - Pattern 1: [observation, with specific trades that support it] - Pattern 2: [observation, with specific trades that support it] - Pattern 3: [observation, with specific trades that support it] - Caveat: "Correlation is not causation. Even real patterns can disappear. Treat this as a hypothesis to test, not a strategy to execute."
Why this prompt exists: pattern recognition is useful when you have enough data. It is dangerous when you do not. With 5 trades, every pattern is noise. With 100 trades, some patterns are real. The caveat at the end is not optional — it is the most important sentence in the prompt.
What this pack is not
This is not a signal service. This is not a course. This is not AI trading. This is not going to make you money.
This is six prompts that force you to think before, during, and after you act, with an AI as a structured thinking partner. The book teaches you what to think about. The prompts teach you how to think about it out loud.
If you want a service that gives you signals, there are thousands. Most of them lose money. You are statistically the loser. Design everything around that.