Learn / Risk & Position Sizing
What to record in a trading journal
A journal of outcomes tells you what happened. A journal that also records what you intended, what you rejected, and how far each trade went before reversing tells you why — and only the second kind can be acted on.
Most journals are the first kind.
The fields that make a review possible
Per trade, at entry
- Symbol, side, entry, stop, target, size
- Planned R — computed from entry, stop and target
- Risk as a percentage of the account
- Setup type — the tag everything useful gets grouped by later
- Reasoning, in your own words, at the time
- Invalidation — what would make this wrong, stated separately from the stop price
Per trade, at exit
- Actual fill prices for entry and exit, not the levels you asked for
- Realised R
- Exit reason — target, stop, discretionary, or time
- Fees and funding, separately from gross result
- Maximum favourable excursion — how far it went your way before reversing
The ones usually missing
Rejected trades. The highest-value rows in the file. Without them you can only evaluate the trades you took, which is half of what you do. “Were the trades I rejected worse than the ones I approved” is the question that measures your judgement, and it is unanswerable without this.
Maximum favourable excursion. Needed to evaluate trailing stops and partial exits without guessing. You cannot simulate a different exit rule without knowing where price went.
Rule breaks. Which rule, and what it cost. Recorded neutrally, not as confession — the point is the pattern, and the pattern is usually that they cluster after losses.
Which limits fired. Including the ones that passed. “The daily-loss breaker was at 0.4% of 2%” is information; silence is not.
What each field buys you
| Field | Question it answers |
|---|---|
| Planned vs realised R | How much slippage, fees and early exits cost |
| Setup tag | Which setups carry the edge and which are subsidised |
| Rejected trades | Whether your filtering adds value |
| Max favourable excursion | Whether a different exit rule would be better |
| Fees separately | Whether costs are eating a thin edge |
| Rule breaks | Where discipline fails, and when |
| Reasoning verbatim | Whether particular reasoning correlates with outcomes |
Every one of these is unavailable from a P&L.
The realised-versus-planned gap
If you record one thing beyond outcomes, record both R figures.
The gap between them is fees, slippage and early exits, and it is usually larger than expected and always in the same direction. Better still, it converges faster than expectancy because every trade contributes — so it is actionable on a small sample, unlike profitability. See how many trades before you can judge a strategy.
A planned 2R realising as 1.3R is a 35% tax on your edge, and it is fixable through execution rather than through better analysis.
Practical format
Structured, not prose. You will aggregate this. Line-delimited JSON or a spreadsheet works; a chat transcript does not.
One row per trade, with rejections as rows too, flagged.
Written at the time, not reconstructed. Reasoning recalled after the outcome is not the reasoning you had.
Local. A complete trading record is sensitive — see local vs hosted models if you plan to review it with a model.
The R integrity rules
Three conventions that keep the numbers comparable:
1R is always the original entry-to-stop distance. If you move a stop, R stops meaning anything unless you hold the denominator fixed.
Record scaled exits as components, not as a collapsed single figure — a half at 1R and a half at 3R is not a 2R trade in the sense your other rows use.
Flag trades where rules were broken so they can be excluded from strategy evaluation. Results from trades you did not follow your rules on do not evaluate your rules; they evaluate something else.
Reviewing it
Weekly over everything, including the quiet weeks. Reviews triggered only by bad outcomes produce a biased sample and a demoralising ritual.
A model is genuinely good at this — it reads every row with attention that does not vary, and it asks about the trades you skipped over. See reviewing your trade journal with a model.
FAQ
What should I record in a trading journal?
Entry-side: symbol, side, entry, stop, target, size, planned R, setup tag and your reasoning. Exit-side: actual fills, realised R, exit reason, fees and funding separately, and how far the trade went in your favour before reversing. Plus the trades you rejected, which most journals omit entirely.
Why record trades I did not take?
Because they are what measures your judgement. If the trades you reject are not systematically worse than the ones you approve, your filtering is costing time and adding nothing — and that is worth knowing. Without those rows you can only evaluate the trades you took.
What is maximum favourable excursion and why record it?
How far a trade moved in your favour before reversing. It lets you test alternative exit rules — trailing stops, partial exits, different targets — against your own history instead of guessing, which is otherwise impossible because the data is gone once the trade closes.
How long before a journal is useful?
The gap between planned and realised R is informative within a few dozen trades, because every trade contributes to it. Expectancy by setup needs substantially more, and more still for high-R approaches whose returns concentrate in few trades. Start recording before you know what you will want to ask.