How to Use ChatGPT to Analyze Your Trades (With Prompts)

Most traders who try using ChatGPT for trade analysis give up after a few sessions. Not because ChatGPT isn't useful — it is — but because generic prompts produce generic answers. "Was this a good trade?" gets you a hedged non-answer. The right prompts get you something actionable.

This guide covers exactly how to prompt ChatGPT for trade analysis, what it's genuinely good at, and what to do with its output.

Before you start: set the context

ChatGPT doesn't know anything about your trading style, risk tolerance, or the specific market you trade unless you tell it. Start every trade analysis session with a brief context block:

I'm a swing trader focused on crypto (mainly BTC and ETH). I use technical analysis — primarily price action, support/resistance levels, and volume. I trade on the 4H timeframe with occasional daily confirmation. My risk per trade is 1–2% of account. I keep a trade journal and I'm trying to identify patterns in my mistakes.

You don't need to write this every time — save it in a template and paste it at the start of each session. This context dramatically improves the quality of ChatGPT's feedback.

Prompt 1: Single trade debrief

Use this after a notable trade — win or loss — to get structured feedback.

Prompt:

Here's a trade I took. Please analyze it honestly and tell me what I did well, what I could have done differently, and whether the setup was valid given my criteria.

Trade:
- Symbol: BTCUSDT
- Direction: Long
- Timeframe: 4H
- Setup: Support bounce at $61,200 — price had tested this level three times and held. RSI was oversold on 4H.
- Entry: $61,350
- Stop: $60,800 (below support)
- Target: $63,500 (previous resistance)
- P&L: +$240 (hit target)
- Mood at entry: Confident, had been watching this level for two days
- What actually happened: Clean bounce, held stop easily, hit target within 18 hours

Question: Was the setup reasoning sound? And was there anything I should have done differently?

The key is specificity. The more detail you give about your reasoning, the more useful the feedback.

Prompt 2: Post-loss analysis

This is where ChatGPT is most valuable. After a losing trade, emotions cloud judgment. An outside perspective helps.

Prompt:

I took a loss today and I want to understand whether it was a process failure or just bad luck. Here's the trade:

- Symbol: EUR/USD
- Direction: Short
- Setup: I saw a bearish engulfing candle at a resistance level on the daily chart. Also, the 50 EMA was acting as resistance.
- Entry: 1.0845
- Stop: 1.0890 (above the wick)
- Target: 1.0780
- P&L: -$180 (stopped out, then price reversed and went to my target anyway)
- Mood: I was frustrated after a small loss earlier in the day. I entered this trade within 30 minutes of the previous one.

Was this a valid setup? Was the timing suspicious? Did my mood affect my execution?

Pay particular attention to ChatGPT's response about the timing — the 30-minute gap after a previous loss is a revenge trading flag, and a good model will call it out.

Prompt 3: Pattern analysis from multiple trades

This works best when you paste in a week or month of trades in structured form. Create a simple format like this:

Prompt:

Here are my last 20 trades. For each, I've listed: date, symbol, setup type, direction, and P&L. Please:
1. Calculate my win rate overall and by setup type
2. Identify any patterns in when I lose (setup type, day of week, direction, etc.)
3. Tell me which setup type is performing best and worst
4. Flag any potential behavioral patterns (revenge trading, overtrading, etc.)

Trades:
2026-05-01 | BTCUSDT | Breakout | Long | +$340
2026-05-01 | ETHUSDT | Trend continuation | Long | -$120
2026-05-02 | EUR/USD | Support bounce | Short | +$95
2026-05-03 | BTCUSDT | Reversal | Short | -$280
2026-05-03 | BTCUSDT | Reversal | Short | -$190
...

Important caveat: Always verify ChatGPT's arithmetic. It gets the calculations wrong often enough that you shouldn't trust them for decisions without checking. Use this for directional insight, not precise statistics.

Prompt 4: Identify your best setup

After several weeks of data, use this to find your edge:

Prompt:

Based on the trades I've shown you, which setup type am I most consistently profitable on? I want to understand:
1. Which setup shows the best win rate for me
2. Which setup shows the best average P&L per trade
3. What the common characteristics of my winning trades are (time of day, market condition, asset type, etc.)

Then suggest what it might look like if I focused exclusively on my best setup for the next month as an experiment.

Prompt 5: Rules from recurring mistakes

Once your journal has surfaced a pattern, use ChatGPT to turn it into an actionable rule:

Prompt:

I've noticed a pattern in my trading journal: I consistently lose money in the 30 minutes after a losing trade. I seem to take lower-quality setups, enter too early, and don't follow my stop discipline. This happens about 3–4 times per month and costs me roughly $400–600 each time it happens.

Please help me:
1. Name this pattern (so I can track it specifically in my journal)
2. Create a specific, concrete rule that addresses it
3. Suggest a quick mental check I can do before entering any trade that follows a loss

ChatGPT excels at this kind of structured rule-making. The answers aren't always novel — but having a specific rule, in writing, that came from your own data is far more motivating than generic advice.

What to do with ChatGPT's output

The goal isn't to follow everything ChatGPT says. It's to get a second opinion that helps you see your own blind spots.

After a ChatGPT session:

1. Copy any specific rules or insights into your trading journal
2. If ChatGPT identified a mistake, add a note to that specific trade entry
3. If it identified a pattern, create a tag for it and start tracking it explicitly

The feedback loop: journal → identify pattern → ChatGPT analysis → create rule → track in journal. Over time, you build a set of personalized rules based on your actual data.

Using ChartPilot + ChatGPT together

ChartPilot handles the structured data layer: every trade logged with setup type, P&L, mood, and notes. When you want to go deeper on a specific pattern, export or paste that data into ChatGPT for qualitative analysis.

More practically: use ChartPilot's AI analysis before the trade (structured chart read, bias, scenarios) and use ChatGPT after the trade (debrief, pattern analysis, rule-making). The two tools are complementary, not competing.

Log your next trade in ChartPilot →