Trading Journal Basics: What to Record, What to Ignore and Why Screenshots Beat Numbers
A broker statement records fills. A journal records decisions, and decisions are the only thing you can change. The eleven fields worth logging on every trade and every no-trade day, the numbers that mislead when they become the headline, and the case for a screenshot at entry over a row of prices.
A trading journal template has one job: to make the same information available about every trade, so that after thirty or fifty of them a pattern can show up. The broker already records the fills, the prices and the P&L. What it does not record is what you saw, what you planned and what you did instead, and those three things are where the money leaks. If you can already read a chart, the journal is where you find out whether you trade what you read.
The template below is the one I would hand to someone starting from zero. Eleven fields, in the order they happen, and a shorter list of things that look like data and are not.
The eleven fields
Before the trade
- Date, session and contract. The contract as the platform shows it (MNQ 12-26), and which window you traded: European hours, the cash open, the afternoon. Sessions behave differently and a journal that cannot split by session will average them into mush.
- Context. Two or three words: trend day or range day so far, which levels are marked and where price sits relative to them, whether a scheduled release falls inside your window. This is the field that later tells you which conditions your setup likes.
- Setup name and grade. The setup from your playbook by name, and a grade (A, B or C) given before entry, not after. Grading after the fact turns every winner into an A.
- The plan. Entry price, stop price, size, and one line on why the stop is where it is. The sizing post explains why the stop is a fact about the trade and size a fact about the account; the journal is where you check that you kept them apart.
During the trade
- Execution against the plan. Actual entry versus planned, whether the stop was moved and why, and how each exit happened: target, trail, manual, stop. One column per event is enough; the point is to see where plan and behavior diverge.
- A screenshot at entry. The chart as it looked when you clicked, with the levels and the bar type you were trading from. A second one at exit if the exit was discretionary.
- MAE and MFE. Maximum adverse excursion (how far the trade went against you before it finished) and maximum favorable excursion (how far it went your way). On NinjaTrader 8 the exported values are wrong on live accounts, so these come off the chart or a recording; they are the two numbers that tell you whether stops and targets are in the right place.
After the trade
- Result in R. Profit or loss divided by the planned risk, so a $150 win on a $100 risk is +1.5R. Points are fine as a second column. Dollars go last, because they change with size and size is not what you are reviewing.
- Error tag, from a fixed list. Chased, entered early, stop too tight, moved the stop, sized up, traded outside the window, traded through news, no error. One tag per trade, chosen from the same list every time. A free-text field here produces prose nobody reads.
- State, one line. How you felt going in and what you noticed in yourself during the trade: calm, rushed, still angry from the last one. Jared Tendler's C-B-A framework from The Mental Game of Trading is a good format: which level of your own A-game, B-game or C-game showed up.
- No-trade days. A day with no entry gets a row too: which setups were watched, why none qualified, whether you were tempted. A journal without no-trade days cannot tell you whether patience is working.
A layout that fits a spreadsheet
The fields above are columns. If you want a table to copy, this is it, with one hypothetical row so the types are clear.
| Date | Session | Contract | Context | Setup / grade | Plan (entry / stop / size) | Actual entry | Exit | MAE / MFE | Result R | Tag | State | Screenshot |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Sep 14 | EU | MNQ 12-26 | Range day, above VAH, no release in the window | Level reversal / B | 29,380 / 29,410 / 2 | 29,384 | Trail | 18 / 44 pts | +1.1 | none | calm, slightly early | entry-0914-1.png |
Two rules keep the sheet useful. The tag column and the grade column only accept values from a list you wrote down in advance, and every row has a screenshot, even the losers you would rather forget. Rows without a screenshot become numbers without a story within a week.
What to leave out
Some things feel like data and are noise.
Dollar P&L as the headline. It scales with contracts, so a two-lot loss looks twice as bad as the same mistake on one lot and a good week on size hides a bad week of decisions. Keep dollars, but review in R.
Win rate on its own. Fifty percent means nothing without the average win and the average loss beside it; the expectancy post has the arithmetic.
Indicator readings at entry. RSI was 31, the fast average crossed the slow one: unless the indicator is the setup, it is decoration, and it will look predictive in hindsight on every winner.
Other people's trades, market commentary, and adjectives. "Felt strong" is not a state, "chop" is not a context. If a word cannot be counted later, it does not go in a column; it can go in a notes field that you agree never to sort by.
Tags without a definition. "Bad entry" means four different things on four different days. The tag list is short because each tag has to be recognizable in a screenshot.
Why screenshots beat numbers
A row of prices tells you that you bought at 29,384 with a stop at 29,410. Three weeks later it cannot tell you whether the level was obvious, whether the bar that triggered you was the third test or the first, or whether the chart was already trending against you. The screenshot can, and review is the act of looking at what you saw with the outcome known.
Three things make the picture worth more than the numbers. Errors are visual: chasing looks like an entry far from the level, a tight stop looks like a stop inside the noise of the bar, and no tag describes either as well as the image does. MAE and MFE have to be read off the chart anyway if the platform export cannot be trusted. And a screenshot is the only record of the decision that does not depend on memory, which is the least reliable instrument on the desk.
The step past screenshots is a recording of the whole session. It costs disk space and nothing else, and it lets you scrub to the thirty seconds before the click, where most of the interesting behavior lives. The Market Replay post covers replaying the market; recording your own screen is the replay of you.
Reviewing it
A journal that is only written is a diary. Once a week, sort by tag and count. Once a month, sort by setup and grade and compare the R per trade of A setups with B and C. Do not change a rule before thirty trades of the same setup are in the sheet; a pattern in ten trades is usually a pattern in your mood that week.
My own journal started as a spreadsheet with roughly these columns and grew into an app because the tag list, the playbook and the state column kept wanting to be linked to each other. The columns did not change. What changed was that every session now begins with a plan and ends with a review, and the review is a series of screenshots with tags, not a P&L figure.