A useful trading journal preserves what you believed before an outcome was visible. If the only note is written afterward—“I knew it was risky”—the record cannot distinguish a careful plan from a story fitted to the result.
Log the decision, including a decision to stay flat, separately from the simulated or observed result. The goal is not to prove that one trading style is better or diagnose a trader’s emotions. It is to make a hypothesis, its limits and the next question easier to inspect later.
Keep a before record and an after record
Before the decision, record a timestamp and timezone, market and instrument, timeframe and source, the setup you see, why it matters, what would invalidate it and how much planned risk the example uses. Record the planned entry and the assumptions behind your sizing. For a perpetual, include the venue, collateral mode and funding interval if known. When something is unknown, write “unknown.” A written stay-flat reason is valuable evidence too.
After the observation, record whether an order was actually simulated, observed or placed elsewhere. Keep planned and realized prices separate. Add fees, funding, exit reason, missing data, execution assumptions and any gap from the plan. An empty field must not imply a zero cost or a completed fill. For a software backtest, include its exact strategy and data version; for a real trade, follow the venue’s statement as the authoritative fill record.
At review time, note what changed in your explanation, which evidence supports or contradicts the first idea, and the next test. Do not overwrite the old rationale. If you correct a timestamp or interpretation, preserve the correction date and reason so the timeline remains readable.
A fictional journal entry
| Field | Example only |
|---|---|
| Prepared | 15 January 2026, 12:00 UTC |
| Market | BTC perpetual, fictional worksheet |
| Decision | Stay flat; no order or fill |
| Observation | Hypothetical next four-hour close remains above a stated level |
| Planned loss | $25 paper scenario assumption—not a suitable amount for anyone |
| Invalidation | Close below the written level; execution and funding not verified |
| Afterward | No realized fill or P&L; revisit after a later observation |
These values and the observation are fictional examples. They do not represent a market event, signal, historical test or suggested trade. A realistic blank sheet leaves realized entry, exit and costs empty until there is an actual recorded event.
Count process records with a stated denominator
Suppose a fictional monthly review contains 12 decision records, including deliberate decisions to stay flat. The reviewer had written eligibility rules in advance: a record needs a timestamped plan version and complete enough source data to classify. Four records fail those checks, so they remain visible as unclassifiable; they are not silently deleted or counted as compliant. Of the remaining eight, six match the recorded rule and two depart from it. Under this example’s definition, recorded adherence is 6 ÷ 8 = 75%.
That percentage describes only how those eight fictional records were classified. It does not measure whether the rule was sound, whether the sample represents later decisions, or whether the choices made money. A different eligibility rule can produce a different denominator. Write the rule before reviewing outcomes, keep excluded and missing records visible, and treat disputed classifications as uncertain rather than forcing them into a favorable category.
Keep this process count separate from financial results. A rule-following loss can be consistent with the written process; an out-of-plan gain does not validate the deviation. If you summarize results, state whether the sample is trades, signals or all eligible decisions, and show missing data and costs rather than letting a neat percentage hide them.
Review patterns without fooling yourself
Group records by rule version, market regime, setup or decision reason only when the sample supports that comparison. A few trades cannot establish a stable win rate. Outcomes can depend on each other: trades from one volatile week are not the same as many independent tests across regimes. Note when you changed a filter after seeing earlier results, because those records are now part of the design process.
Useful review questions include: Did I follow the rule I wrote? Were costs present in the plan? Which exit was anticipated? Did the evidence use the same dataset? Were my losing decisions consistent with the invalidation rule? What could contradict my current explanation? Count both trades and deliberate non-trades, and keep financial outcomes separate from process observations.
Recordkeeping can help reveal activity and cost patterns, but the journal itself is not evidence that a strategy works. Barber and Odean studied 66,465 U.S. households holding common stocks through one large discount broker from 1991–1996; the most active group earned 11.4% annualized while the market benchmark returned 17.9%. That historical equity sample does not describe crypto, perpetual futures, current markets or the effect of keeping a journal. It is a reason to preserve turnover and cost records when studying activity—not a forecast or a claim that journaling improves returns.
Choose a format you will actually preserve
The downloadable CSV journal template uses blank planned and observed fields and provides one explicitly fictional stay-flat example. Keep a private backup. Avoid putting email addresses, wallet secrets or account credentials in a trading journal. SuperQuant’s local Diary and paper records are separate from exchange statements and do not establish real fills or account balances.
A journal will not make every decision profitable or eliminate bias. It makes the evidence trail more complete so a future review can challenge an earlier belief. The CFTC’s virtual-currency trading risk advisory provides regulator-authored context on speculative and leveraged-market risks; it does not endorse journaling or establish that recordkeeping improves trading results. This is educational process material, not psychological treatment or individualized investment advice.