Post-session routine & journaling

16:00 ET (04:00 ICT, winter). The NQ cash session is over. You made $180, or lost $120. Either way, the next 15 minutes matter more than the last three hours. Trading produces data; the review turns data into adjustments; adjustments are the only way your execution improves. Skip the review and you pay the tuition without attending the class. If you trade from Vietnam, do the review when you wake, before the next session and before you read anyone else’s market commentary.

Key terms

Post-session review (đánh giá sau phiên): a fixed, short routine run after every session that compares what you planned with what you did House rule.

Trade journal (nhật ký giao dịch): one row per trade with the same fields every time, so that a month of rows can be counted Definition. It is a dataset, not a diary.

Placebo level (mức giá giả dược): a random price at the same distance from the open as one of your levels, scored with the same written rule Definition. It tells you how often any price at that distance gets touched or holds.

The 15-minute review

Same order every night House rule:

1. Confirm every trade is logged with a screenshot (journal below).
2. Morning sheet vs reality: score the sheet’s levels against their placebos (below), then write where your read broke.
3. Plan-adherence score: how many trades fully followed the plan?
4. One lesson, one sentence. Not five. “I moved breakeven too early twice; rule: not before +1R.”

Score adherence, not P&L House rule. A +3R day with three plan violations is a bad day that got lucky; a −1R day with perfect adherence is a good day that paid tuition. Adherence is what you control. P&L includes luck, and grading yourself on luck teaches the wrong lessons (M7.07).

The journal: 10 fields, every trade

Ten fields, under two minutes a trade, logged within a few minutes of the exit House rule:

date/time (ET) · instrument · setup name · entry model · planned entry/stop/target · actual fills · size · R planned vs R realised · screenshot · one-line note (why taken, deviation if any).

Log it before you see how the rest of the session goes. Once people know an outcome, they remember having expected it: hindsight bias is a replicated finding in general psychology MeasuredOther market (Fischhoff 1975). That it distorts traders’ own journals is plausible, not measured on traders PlausibleOther market. A row written in the minute after the exit is the cheapest guard.

Two rows from one illustrative MNQ session (synthetic prices; MNQ is $0.50 a tick, $2 a point; the trader’s usual 1R is 8 points on 1 MNQ = $16):

The journal is honest or it is fiction. The second row is the one that teaches you something, so it gets logged with the same care as the first.

Planned vs realised R is the column that matters most Definition. The gap between them is your execution leak: early exits, slipped stops, moved targets and unplanned trades all live there. A method with +0.4R planned and −0.1R realised has an execution problem before it has a strategy problem. Realised R is measured against the planned risk: in row 1, 19.5 points of gain on an 8-point plan risk is +2.4R.

Tag deviations, don’t hide them House rule. Every broken rule gets a tag: EARLY_EXIT, MOVED_STOP, NO_SETUP, OVERSIZE, SKIPPED_COOLDOWN. Once a month, sort by tag: your most frequent tag is the first thing to fix.

Score the morning sheet against a placebo

“Expectation vs reality” needs a yardstick, or every reaction at one of your lines feels like proof. Add one block to the nightly review. For each level on the morning sheet, put a placebo at the same distance from the 09:30 ET open on the other side, then mark both after the close with one written rule: touched or not; if touched, held or broke (M4.07 gives the exact rule and the drill). Over weeks you get two numbers: how often your levels held, and how often random prices at the same distance held.

Expect them to be close. That is what our lab found for the usual pre-market level families.

So what goes in the “where did the read break” line? Not “the level failed”: levels by themselves hold at the random rate. Journal what you saw at the level (the order flow, the context) and what you did about it. That is the part that might carry an edge, and your journal is where you collect it.

Monthly review: numbers, with their sample size

Once a month, one hour House rule: per-setup expectancy (M6.08), win rate, average R planned vs realised, adherence %, the top three deviation tags. One page of numbers beats a year of feelings about “how I’m doing”.

Then read every number with its sample size Definition. A month of 20–30 trades is a small sample. At 22 trades and a 50% win rate, the 95% interval on the win rate is about ±21 percentage points; at 50 trades it is still about ±14 (M7.10). Expectancy is just as loose: with a typical spread of about 1.2R a trade, 22 trades give a standard error of about 0.26R, so a measured +0.31R a trade is compatible with anything from about −0.2R to +0.8R. One month can show you where you leak (the tags are counts of your own behaviour, and they are hard to argue with); it cannot tell you whether the method works.

Worked example: a Friday review

Week: 14 trades, +1.8R net, adherence 9 of 14.

• Adherent trades (9): +2.6R, planned and realised the same.
• EARLY_EXIT (3): banked +0.9R where the written ladder (M6.11) would have made +2.4R. Leak: −1.5R.
• SKIPPED_COOLDOWN (2): two trades taken straight after stops, against the written cooldown, −1.7R. Under the plan they would not exist. Leak: −1.7R.

Plan as written: 2.6 + 2.4 + 0 = +5.0R. Realised: 2.6 + 0.9 − 1.7 = +1.8R. That is the waterfall above.

Morning-sheet scorecard for the week: 24 levels, 11 touched, 4 held; their 24 placebos, 10 touched, 4 held. Four of 11 against four of 10 is no difference, and with so few touches it could not show one anyway. Nothing about the levels needs changing; the read that broke twice was entering at a level before the rejection bar closed, and that goes in the notes.

One lesson: “Profit-taking follows the ladder, not the P&L flash. Ladder levels go on the chart before entry.” One plan change for Sunday (M7.04): add that line to the routine. The next week reads adherence 12 of 15. That is one week; keep the EARLY_EXIT count for a month before you call the leak fixed.

Prop-firm note. If you trade an evaluation, log fills and slippage from the platform’s own fill report, and mark whether the trade was on a simulated account. Simulated fills can be kinder than a live queue, and firms prohibit order handling designed to exploit them; the PROP course covers the rules.

Common mistakes

Common mistake: Journaling only the losers, or writing essays instead of rows. Selective journals produce selective data, and the 10-field row exists because friction kills journals: if it takes 10 minutes a trade, you will stop by Wednesday. Two minutes, every trade, winners included.

Common mistake: “My levels worked this week.” A level that held three times is three touches, and random prices at the same distance hold about as often. Only a running count against a placebo tells you anything, and it takes many weeks of touches.

Common mistake: Judging the method on one month. Twenty-two trades at +0.31R a trade is not proof of an edge, and twenty-two trades at −0.2R is not proof there is none. Fix the top tag; let the sample grow.

Drill: find this week’s leak

Take your last 10 or more trades (or a sim week).

1. Fill the R planned and R realised columns for every trade. For a trade with no plan, R planned is 0.
2. Tag each deviation.
3. Sum the gap (planned minus realised) by tag, and draw it as a waterfall like the figure above.
4. Write one sentence naming the top tag and one rule that removes it.

Recap

• Nightly, 15 minutes: confirm logs, score the sheet against placebos, adherence score, one written lesson. Score adherence, not P&L.
• The 10-field journal, logged within minutes of the exit; the planned-vs-realised R gap, sorted by deviation tag, shows where execution leaks.
• Our lab measured no edge for pre-market levels over random prices at the same distance, so journal what you saw at the level, not whether the line held.
• A month of trades shows your leaks; it is too small a sample to prove or disprove the method.

Further reading

• Fischhoff, B. (1975). Hindsight ≠ foresight: the effect of outcome knowledge on judgment under uncertainty. Journal of Experimental Psychology: Human Perception and Performance, 1(3), 288–299. Why the row is written before you know how the session ends.
• Dr. Brett Steenbarger on CBT tools for traders: being your own coach, and what the journal is for.
• Our lab: the usual pre-market level families held after a touch at the same rate as random prices at the same distance, over 721 sessions on the Nasdaq-100 and S&P 500 ETFs, with a 46-session check on NQ and ES futures pointing the same way.

Connects to: M7.02 (the session being reviewed) → M7.04 (the plan the review improves); M4.07 (the level-scoring drill); M7.11 (the weekly review).

Educational content only — not financial advice. Futures trading involves substantial risk of loss.