How to use AI to review your trading journal (a 15-minute weekly routine)
Your journal already knows when you break your rules. It's just hard to see in a long list. AI can sort it in seconds — if you log the right things and ask the right way.
Why AI fits this job
Reviewing trades is boring work. You scroll through rows of numbers, looking for something that repeats.
That's exactly what AI is good at: finding patterns in data you give it. It doesn't need to know the future. It only needs to read your past — honestly logged.
It won't tell you what to trade next. It will show you how you trade, which is the part you can change.
Step 1: Log the right columns
AI can only find what you write down. A useful journal has these columns:
| Column | Example | Why AI needs it |
|---|---|---|
| Date and time | Tue 10:05 ET | Finds your best and worst times |
| Market | EUR/USD | Spots where you break rules most |
| Setup (one sentence) | Pullback to a turning point, with the bigger trend | Groups trades by idea |
| Followed my rules? | Yes / No | The most important column |
| Feeling before | Calm, FOMO, revenge, bored, tired | Links feelings to choices |
| Result in units of risk | +2, −1, −0.5 | Compares trades fairly |
| Error type | Moved stop, early exit, none | Names the mistake |
| Note | "Entered before the candle closed" | Gives context |
Logging results in units of risk (often written R) instead of dollars matters. A −1 means "lost the amount I planned to risk." That way a small account and a big one read the same.
No journal yet? Grab our free journal template, or read how to keep a trading journal.
Step 2: Clean it before you paste
Before you paste anything into an AI chat:
- Delete account numbers, broker names tied to your login, and anything personal.
- Remove dollar balances. Units of risk are enough.
- Keep at least 20 trades. Fewer than that and "patterns" are mostly luck.
Small samples lie. Three losses on Mondays is not a Monday problem. Ask AI to say how many trades each pattern is based on, and treat anything under 10 as a hint, not a finding.
Step 3: Paste this review prompt
Copy this into any AI chat assistant (like ChatGPT, Claude or Gemini), then paste your table under it:
You are reviewing a beginner's practice trading journal. Only describe what is in the data. Don't suggest new strategies, markets or indicators, and don't predict anything.
1. How many trades, and what share followed the rules?
2. Compare trades that followed the rules with trades that didn't: count, average result in units of risk.
3. Which feelings came before rule breaks most often?
4. Which times of day or markets have the most rule breaks?
5. Which error type shows up most?
6. Name the ONE habit that, if fixed, would affect the most trades. Say how many trades it touched.
For each point, say how many trades it's based on. If a pattern is based on fewer than 10 trades, label it "too few to trust".
[paste journal table here]
Notice the limits. "Don't suggest new strategies" stops it from sending you chasing the next shiny idea. Your job this week is to follow your plan better, not swap it.
Step 4: Read it like a coach, not a judge
AI might say something like: "9 of your 12 rule breaks came after 2 losses in a row, and 7 had the feeling 'revenge'."
That's not a verdict on you. It's useful news. It points to a single fix: a daily loss limit — stop for the day after 2 losses.
Check the claim yourself. Filter your sheet and count. AI sometimes miscounts, and the habit is only worth changing if the pattern is real.
Step 5: Turn it into one rule
Pick one change per week. Not five. Write it as a rule you can check with yes or no:
- "After 2 losses, I stop for the day."
- "I only enter after the candle body closes past the level."
- "If my feeling is FOMO, I wait one full candle before deciding."
Add it to your pre-trade checklist. Next week, ask AI one question: "Did I follow the new rule? How often?"
A worked example
Here's an illustration of a review, with made-up practice data:
- 30 practice trades. 21 followed the rules, 9 didn't.
- Rule-following trades averaged +0.4 units of risk. Rule breaks averaged −0.9.
- 6 of the 9 rule breaks happened after 11:30 am ET.
- Most common error: entering before the candle closed (5 times).
What the trader takes from it: "Stop practicing after 11:30, and wait for the close." That's a math example from made-up numbers, not a promise — real results vary, and any trade can lose.
What AI can't see in your journal
- What you didn't log. Skipped trades and "I almost took it" moments matter. Log missed and skipped trades too.
- The chart. It reads your words, not the setup. Your one-sentence setup note is all it has.
- Whether your method works long term. 30 trades can't tell you that. Keep practicing and keep logging.
- How you feel right now. That check is yours, before every trade. See trading psychology for beginners.
Make it a weekly routine
- Sunday, 15 minutes.
- Export or copy the week's trades.
- Clean out personal data.
- Run the review prompt.
- Check the top finding yourself.
- Write one rule for next week.
For deeper self-reflection, pair it with our journaling prompts for traders. AI finds the pattern; the prompts help you understand why it happens.
Ask about behaviour, not money. "When do I break my rules?" teaches you more than "how do I make more?" — and it's a question AI can actually answer from your data.
Questions
Can AI analyse my trading journal?
Yes. Paste in a clean table of your trades and ask it to describe patterns — rule breaks, feelings, times, error types. Check its counts yourself, since AI can miscount.
How many trades do I need before reviewing with AI?
At least 20 to start, and more is better. Patterns based on fewer than 10 trades are hints, not findings.
Is it safe to upload my trades to ChatGPT?
Remove account numbers, balances and personal details first. Trade details in units of risk are all the AI needs.
What should I do with what AI finds?
Pick one habit, turn it into a yes-or-no rule, add it to your checklist and check next week whether you followed it.
Want the next step? We'll walk you through it.
Four free lessons by email, one every few days — each with a 10-minute exercise on a real chart.
Send Me the Free Lessons 4 short emails · No spam · Unsubscribe anytimeOr see the whole method, step by step:
See What’s Inside the Course Try everything free first. Decide when you're ready. · Open the free chart tool (works in the TradingView phone app)One chart idea a day, in plain English. Follow along: