Limits, sizing, stops, and bias guards compound only if you remember what you did. Journaling is an operations log connecting calibration tools to the signal-reading module ahead—not therapy, not a screenshot dump.
The module closes where professional trading always closes: records. Without rows you cannot score calibration, cannot prove venue choice worked, and cannot learn from Module 08’s signal reading with a straight face.
What to record
Every trade deserves a row before you submit the order: date and timezone, venue and contract, side, market price at decision, your pre-trade probability, edge after fees, size in dollars and percent of bankroll, order type and fill average, thesis in three sentences, invalidation trigger, emotion score one to five, and resolution outcome when settled. Optional tags for correlation cluster, strategy sleeve (passive, momentum, contrarian, arb, hedge), and a depth snapshot when size was large.
Changing probability after outcome is cheating; lock the pre-trade column. Narrative without numbers is not a journal row—it is marketing to your future self.
Metrics that matter
Realized profit and loss net of fees, return on capital at risk, Brier score on resolved forecasts, calibration bins (do your sixty-percent claims happen sixty percent of the time?), edge capture versus entry edge, and count of rule violations (market chase on a cooling-room day). Hit rate alone misleads: seventy percent wins with small average win and rare huge losses is still ruin.
Expectancy per dollar risked beats win rate: average win times win rate minus average loss times loss rate, divided by average risk.
A narrative row example
Kalshi YES on a March rate cut—entered mid-January afternoon. Decision price 57¢, your probability 64%, edge about six cents after fees, size $420 (2.1% of twenty-thousand bankroll), limit bid filled 58¢, thesis tied to futures path and dot plot, invalidation if officials hold with no cut language by February first, emotion 2, planned hold unless invalidation. Months later you learn whether 64% was calibrated, not whether you felt smart.
Monthly review example
Ten resolved trades; bankroll up five percent. Brier 0.19 is decent. Average stated probability on winners was 71% with outcomes one; on losers 62% with outcomes zero—suggests overconfidence on losing tickets. Three rule violations were market chases on CPI day. Kalshi sleeve positive; capped venue negative after fees. Action: shave three points off pre-trade politics probabilities for thirty days; ban market orders on macro release minutes.
Templates by strategy
Passive rows add planned hold duration and roll decision dates. Momentum rows add catalyst timestamp and trend-break exit. Contrarian rows log crowd price and mispricing mechanism. Arb rows list both legs, basis, and capital lock time. Hedge rows note offset contract and net exposure. Portfolio rows record cluster percent before and after the trade.
Rhythm
Daily: two minutes per trade with minimum fields. Weekly: fifteen minutes on open risk by theme versus portfolio caps, fix missing theses. Monthly: forty-five minutes on Brier, venue comparison, best and worst process decisions, at most one new rule.
Closing the loop with earlier modules
Did price at entry match later consensus? Did slippage match the depth you saved? Did arb trades clear basis in time? Did positive expected-value tickets add P&L net of fees? Any resolution surprise versus thesis? Did venue caps block good ideas? Answers live in the journal, not in vibe memory.
Bridge to reading the signals
The next module interprets spikes, volume, open interest, spreads, and time decay. Your journal is the control group: when you traded, what you believed, whether the move should have changed your probability. Compare spike timestamps to your entries; test whether high volume days matched informed flow for you; record spreads at entry and exit; map days-to-resolution against mark moves.
Without logs, advanced analysis becomes pattern matching—and bias wins again.
Common journal failures
Logging only wins, changing probability after the fact, omitting fees, never reviewing, tracking too many metrics on trade eleven, and skipping emotion scores when they would embarrass you.
Module closure checklist
Aim for twenty logged trades or a deliberate paper set, at least one Brier computation, a written sizing policy referenced in entries, one thesis stop tested, bias tags in weekly review, and venue filled on every row before diving into signal reading.
Tools (keep simple)
A spreadsheet is enough: one row per trade, columns for the minimum fields, a monthly tab for Brier. Fancy dashboards without discipline are worse than a plain table you actually fill.
Process over outcome
Rank decisions by whether you followed sizing, limits, and thesis rules—not by whether YES won. A good process loss is tuition; a bad process win is luck that teaches the wrong lesson.
Module 07 integration
Passive, momentum, contrarian, arb, hedge, portfolio, execution, sizing, stops, and bias only compound when the journal connects them. This row is the scoreboard for the whole module.
Example questions for monthly review
Did I violate cooling-room rules on news days? Did capped venues show up as positive EV after fees? Did contrarian tags correlate with profit or just with activity? Did open risk ever exceed cluster caps the week before a catalyst? Honest answers change one rule; dishonest answers waste another month.
Key ideas
Log before submit; lock probability; review monthly; change one rule at a time. Calibration bridges forecasting math to reading the market’s signals next.
What comes next
Measurable discipline here makes market tape readable there.
Fields you can add later
After fifty trades, optional columns might include screenshot links, implied versus realized volatility of mark, or open interest at entry. Add complexity only when a monthly review question cannot be answered without it.
Misconceptions
“Journal after I win” preserves bias. “I remember” fails on twenty trades. “Brier is for academics” is wrong—it is for anyone who quotes probabilities. “More columns equals better” creates empty dashboards.
Start minimal, review monthly, grow fields only when you actually use them in decisions.
Next: Interpreting Price Spikes and Crashes