1. What to record
Per trade, minimum: date, market, direction, setup name, entry, stop, target, size, risk %, and - the part everyone skips - why this trade in one sentence, and what the plan said at that moment. Screenshots of the chart at entry cost nothing and reveal much later.
2. Grading adherence
One field: followed plan / partial / broken. This single column, filled honestly, is worth more than any indicator. After 30+ trades, compare adherence buckets: if your broken-rule trades lose systematically, the journal just made you money without a single prediction.
Remember: "You followed your stop in 18 of 20 trades" is a better performance report than "you made $300". It measures the thing you control.
3. The patterns worth hunting
Three recurring finds: overtrading after losses ( revenge trades within an hour of a stop), the FOMO entry (no setup, rising market, note "felt left out"), and stop-moves (wider stops after the fact). All three are invisible in the P&L and obvious in the journal.
4. Reviews beat records
A journal that is only written is a diary. The monthly review - one hour, adherence stats, biggest rule-break, whether the plan still fits the market - is where the compounding happens.
5. When this fails
Self-reporting corrupts: "kind of followed" slides toward "followed". Write the plan's rule verbatim next to the trade; compare letters, not feelings. Honest journals are uncomfortable by design.
Risk: Journals that only record wins die quietly. If breaking rules feels too bad to write down, that discomfort is the signal that the system works.