25 ChatGPT Trading Prompts for Setups and Risk

These are 25 trading prompts for the execution side of the job: reading a setup on a chart, defining an entry, a stop and an invalidation price, sizing the position, and reviewing the trade afterwards. Each one names a ticker, a timeframe and exactly what to return, so you can paste it and swap in your own. Research prompts about filings and valuation live on a separate page.

Why prompt quality matters

ChatGPT does not know what you want unless you tell it precisely. A prompt like "is AAPL a good buy?" forces the model to guess your timeframe, risk tolerance, and what data it should weigh, so it returns a generic, hedge-everything answer. A prompt that assigns a role, supplies the chart context or the numbers, and asks for a specific format gets you something you can actually use.

The pattern that runs through every prompt below is role plus data plus task plus format. You tell the model who to be, give it the inputs, state exactly what to do, and define how the answer should look. That structure is what turns a chatbot into a focused analyst, and it is the single biggest lever on output quality, far more than clever phrasing.

10 prompts for reading a setup

These are execution prompts: what is the chart doing, where does the trade break, what would change the read. For research prompts about filings, margins and valuation, use our ChatGPT stock analysis prompts instead. Swap the tickers and timeframes for your own, and verify every level against the actual chart.

1. "Attached is a daily chart of AAPL. State the trend from the swing structure. List only the horizontal levels you can actually read from the price axis, then separately list any level you are estimating. Give one bullish and one bearish scenario, each with the price that would invalidate it. Do not tell me what to do."
2. "Attached is a 4-hour chart of EURUSD with a 20 and 50 EMA. Describe where price sits relative to both averages, whether the averages are converging or diverging, and what a re-test of the 50 EMA would need to look like for the uptrend read to survive."
3. "On this 15-minute NQ chart, identify the most recent higher low and the swing high that would confirm continuation if broken. Give the exact price of each as far as you can read it, and say which of the two you are less confident about."
4. "I am watching TSLA on the daily. Here are the last five daily candles: open, high, low, close for each [paste]. Describe the sequence in structural terms, name any recognisable candlestick formation, and say what the following session would have to do to negate it."
5. "Attached is a weekly chart of BTCUSD. Compare the current consolidation to the two previous consolidations visible on the chart in terms of duration and range width. Do not predict the breakout direction."
6. "Here is my read of the SPY daily chart [paste your own analysis]. Act as a skeptic. List the three strongest arguments against my read and the single piece of chart evidence that would settle it."
7. "Attached is a 1-hour chart of GOLD with RSI. State the RSI condition you can see, whether there is a visible divergence with price, and what specifically would confirm or kill that divergence. Say clearly if the RSI panel is not legible enough to answer."
8. "Given these levels on MSFT daily: resistance 512, support 486, current price 498 [substitute your own]. Frame an entry, a stop and a first target for both a long and a short scenario, show the risk to reward on each, and state which scenario the current structure favours and why."
9. "Attached are two charts of the same asset, a daily and a 1-hour. Say where the two timeframes agree and where they conflict, and which conflict would matter most for a swing entry."
10. "Build me a five-question pre-trade checklist specific to breakout trades on liquid US large caps, where each question has a clear yes or no answer from the chart alone."

8 prompts for risk and trade planning

Risk prompts work well because the math is deterministic, but always double-check the numbers.

11. "Act as a risk analyst. Account size [X], risk per trade 1%, entry [Y], stop [Z]. Calculate dollar risk, share count, and the position value."
12. "Given entry [X], stop [Y], and target [Z], calculate the risk-reward ratio and the win rate I'd need to break even."
13. "Explain how ATR-based stops work and how I'd set one given an ATR of [value] and a 2x multiple."
14. "Stress-test this trade plan [paste]. What happens to my account after three losses in a row at this risk level?"
15. "Explain the difference between the 1% and 2% risk-per-trade rules and the drawdown math behind each."
16. "I have [N] open positions [paste]. Point out where I might be over-concentrated in one sector or correlated names."
17. "Walk me through calculating breakeven price including fees for this trade [paste the details]."
18. "Build a pre-trade checklist that forces me to define entry, stop, target, and position size before I click buy."

7 prompts for journaling and review

Reviewing your own trades is where AI shines, because you supply the data and ask it to find patterns.

19. "Here are my last 20 trades [paste setup, result, and notes]. Group them by setup and tell me which performed best and worst."
20. "Review these losing trades [paste]. What mistakes repeat, and what single rule would have prevented the most damage?"
21. "From this trade log [paste], calculate my win rate, average win, average loss, and expectancy."
22. "I tend to [describe a habit, e.g. exit winners early]. Suggest a journaling prompt I can answer after each trade to catch it."
23. "Summarize my trading week from these notes [paste] in five bullet points, focused on process not outcome."
24. "Help me design a one-page trading journal template that captures the data needed to spot my edge or my leaks."
25. "Given this emotional note I wrote after a bad trade [paste], help me reframe what happened in process terms and what I'd change."

For the workflow around this, see our guide on how to keep a trading journal and our ChatGPT trading journal walkthrough.

The chart prompts above have one structural weakness no wording fixes: consistency. ChatGPT has no fixed definition of support and no memory of the previous session, so the same screenshot can return different levels on different days, which makes tracking a setup across a week impractical. Bullynx's chart analysis runs the same technical routine on every upload so levels, structure and invalidation come back in a comparable shape each time. It still reads only the image you give it, has no order flow or news feed, and does not issue buy or sell calls.

Prompt-engineering tips that improve every answer

A few habits sharpen any trading prompt.

• Assign a role. "Act as a risk analyst" or "act as a skeptic" focuses the model's tone and priorities.
• Supply the data. ChatGPT cannot see your chart or live prices. Paste the levels, the numbers, or the transcript.
• Define the output. Ask for a table, a checklist, or three bullet points. Vague requests get vague answers.
• State constraints. Tell it not to give buy or sell recommendations, and to flag uncertainty rather than guess.
• Iterate. Treat the first answer as a draft and refine with follow-ups like "now make the bear case stronger."

Follow-up prompts that sharpen the first answer

The first answer is a draft, not the final product. Prompting well is conversational, and a short follow-up usually beats trying to craft one perfect prompt.

• Strengthen one side: "Now make the bear case stronger and assume the breakout fails."
• Add depth: "Explain the second risk in more detail and what on the chart would confirm it."
• Reformat: "Put that into a comparison table."
• Pressure-test: "What is the weakest part of this analysis?"

Iterating this way also surfaces the model's reasoning, which is what makes the output checkable. Asking it to "show your reasoning" does double duty: it improves the answer and it gives you a map of the steps to verify first.

The real limits of ChatGPT for the markets

ChatGPT is a powerful aid with hard limits you must respect.

1. No live data by default. The base model can use stale information and does not know today's prices unless you give them.
2. It can hallucinate. Confident, wrong figures are common. Verify every number against a primary source.
3. It is not an advisor. It has no accountability, no view of your full situation, and no fiduciary duty.
4. It can sound authoritative while reasoning poorly. Polished prose is not the same as a sound analysis.

Putting these prompts to work

The 25 prompts above turn ChatGPT from a vague chatbot into a structured analysis, risk, and review aid, as long as you supply context and verify the output. The skill is not memorizing prompts but internalizing the pattern behind them: role, data, task, format. Master that, and you can write a sharp prompt for any situation the market throws at you.