The minimum useful record
For each trade, keep the date and time with an explicit time zone, instrument, account or environment (simulated, paper, or live), original thesis, size, requested price, fills, exit, fees, and reason for closing. Add screenshots or notes when they clarify something the numbers cannot show.
Separate plan from fact. “I meant to exit at X” and “the order filled at Y” are different pieces of information. Note whether a trade was manual or came from a rule or alert.
Classify the mistake before changing a strategy
A loss can come from a poor market thesis, a worse fill than expected, costs, a misconfigured rule, or a technical interruption. Consistent labels help prevent every loss from being assigned the same cause.
Do not start with dozens of labels. Try four: decision, execution, position management, and technical problem. Split a category later if it appears often enough to justify it.
A 20-minute weekly review
- Check that your history matches broker order states and fills.
- Separate simulated trades from live ones.
- Group trades by error type and total cost.
- Choose one change you can test in the coming week.
- Write down when you will review whether it helped.
The goal is not to tell a convincing story from a small sample. It is to preserve enough evidence to notice patterns when you have a reasonable sample.
Protect the data in your journal
A journal may contain account numbers, positions, and personal schedules. Before sharing a screenshot, hide identifiers, balances, and access data. If you compare results publicly, explain the sample size, environment, and whether costs are included.