A trading journal is useful when it preserves what was known when a decision was made, what actually happened, and what you noticed later. Those are different records. If they are collapsed into one note after the outcome is known, a winning trade can make an unclear decision look disciplined and a losing trade can make a valid process look broken.
The aim is not to produce a prettier diary or prove that journaling improves returns. It is to leave enough honest evidence to review a defined process. Keep unknown evidence visible. Keep a changed rule separate from the rule that existed at the time. Keep a broker-confirmed fill separate from the price you hoped to receive.
Record the trade in four stages
Use the same fields for a completed trade, a skipped candidate, and a missed valid signal where they apply. A template is not a strategy, allocation rule, or recommendation to trade.
| Stage | Record | Why it belongs there |
|---|---|---|
| Before | Setup name and version; instrument; evaluation time and session; market context; trigger; checklist facts and unknowns; planned entry, invalidation, and risk assumptions | Preserves the decision-time record before the result can influence it |
| During | Order type; submitted time; actual fills; partial fills; fill prices; commissions, fees, and other recorded costs; changes and their reason | Separates the plan from the execution the broker actually gave you |
| After | Exit fills; exit reason; realised result; maximum favourable and adverse excursion where your data supports it; screenshots and source records | Captures the outcome without claiming why it happened |
| Review | Adherence; data-quality flags; review tags; rule-version changes; questions for a comparable sample | Turns individual records into a later audit instead of a story about one trade |
Maximum favourable excursion is the largest unrealised gain seen while the position was open. Maximum adverse excursion is the largest unrealised loss. They can be useful descriptive fields when the data is complete and their measurement is defined. They do not prove that a different exit would have been better.
Before: write what was available, not what later became obvious
The before-entry record should be completed at the point your written process calls for a decision. For a daily setup, that may be after a regular-session close. For an intraday process, it may be after a completed bar. State the time zone and session rather than assuming a chart timestamp means the same thing everywhere.
At minimum, capture:
- the setup's name and exact version;
- the instrument, market, timeframe, and data source;
- the trigger and the values used to assess it;
- relevant context required by the setup, including any facts that were unknown or stale;
- the planned entry condition and invalidation condition, if they were defined;
- any exposure, liquidity, event, or personal-process check your rules require;
- a screenshot or saved chart link that can be reopened later, where practical.
Do not replace an unknown with “clear” because no issue was noticed. “Earnings date not verified” and “no scheduled earnings event” are not the same record. What “unknown” should mean in a trading checklist explains why the missing-evidence state needs its own treatment.
The journal should also preserve skipped candidates. If a trigger occurred but an event check was unknown, a rule blocked the idea, or you chose not to act, that record helps later distinguish opportunity selection from execution. Without it, a journal can only show the trades that made it through the gate.
During: use the actual execution record
The planned entry is not the entry. Record the order details and then record what was filled: submitted time, order type, quantity, fill price, partial fills, cancelled quantity, and any change made after submission. Include the reason for a discretionary change without trying to make it sound better after the fact.
For US securities, a broker-dealer generally provides a confirmation after a transaction with information such as the security, quantity, and price. Investor.gov also recommends checking transaction records and confirmations. Your broker and instrument may present different fields, so retain the source record rather than relying only on a manually retyped number.
Costs deserve their own fields. A result that excludes commissions, fees, or other transaction costs is not comparable with one that includes them. Use the costs shown in the account record where available, and flag any estimate rather than presenting it as final.
After: capture the result without judging the setup from one outcome
After the position is closed, record the exit fills, exit reason, realised result, and the source used for each figure. If the plan changed, write when it changed and why. A later review needs to know whether the trade followed the original version, an explicit exception, or an improvised rule.
Screenshots can help preserve context, but label their time. A screenshot taken after the exit cannot show the chart as it appeared before entry. If a charting platform revises historical data, note that possibility instead of treating every image as immutable evidence.
Avoid turning the result into a diagnosis. A loss does not establish that the setup failed. A gain does not establish that the decision was good. One outcome may reflect the setup, the execution, position management, costs, a market move, or plain variance. The journal's job is to preserve enough detail to ask better questions later.
Worked example: one entry, then a weekly review
This is a hypothetical record, not a recommended trade.
| Field | Journal entry |
|---|---|
| Setup | Daily closing breakout, version 1.2 |
| Decision time | 16:15 New York time, regular session complete |
| Trigger facts | Close 51.20; defined level 50.00; completed volume 3.6m against a 1.8m 20-session baseline |
| Unknown or exception | Earnings-calendar source not confirmed at review time |
| Planned action | No order. The written rule blocks candidates with an unverified event check. |
| Outcome field | No position opened; later price movement is recorded separately, if at all |
| Review tags | skipped, event-unknown, rule-followed, daily-breakout-v1.2 |
The useful review is not “would this have made money?” That question is vulnerable to hindsight and can reward breaking the rule. At the end of a week, compare records with the same setup version and a similar decision state:
- Were all eligible candidates recorded, including skips and misses?
- Which facts were repeatedly unknown, and was the data source, timing, or rule definition the problem?
- Did discretionary changes cluster around particular conditions or emotions?
- Did actual fills and costs differ consistently from the assumptions used before entry?
- Are you comparing the same setup version, market, session, and treatment of exits?
- Is the sample large and varied enough to support any conclusion, or is the honest answer still unknown?
That is a review framework, not a promise that the answers will improve a result. A small or selectively recorded sample can still mislead. Trading expectancy: why win rate can mislead you covers why outcome averages need consistent rules, costs, and sample limitations.
Protect privacy and data quality
A journal can contain sensitive account and personal information. Keep only the fields needed for the review, restrict access, and avoid publishing account numbers, exact balances, addresses, or broker login information in screenshots. Crop or redact before sharing an example with anyone else.
For data quality, distinguish original source records from manual notes and later calculations. Record the source, time, timezone, and any transformation where it matters. If a fill, fee, event date, or chart value cannot be verified, mark it unknown or estimated. Do not silently repair the record once the outcome is known.
There is also a boundary between a journal and tax, regulatory, or broker records. A personal template may be useful for reviewing a process, but it does not replace official confirmations, statements, or professional advice on reporting obligations.
Where HeraldGoat fits
HeraldGoat is pre-launch. It is being built to monitor named playbooks, follow each Setup through its changing state, and make selected trigger and checklist facts inspectable. That alert history may help preserve what fired and what context was available at the time.
Journaling, performance calculation, execution assessment, and trade review are outside the current MVP promise. HeraldGoat does not decide whether to enter or exit, record a broker fill, calculate profitability, or determine whether a trader followed their plan.
Continue the review
- See why a historical outcome average has limits
- Separate profitability from one signal
- Keep a passed checklist separate from confirmation
- Browse all practical trading guides
If inspectable trigger history would help you review a defined process, you can join the launch waitlist. Joining does not imply immediate product access.
Sources
- Investor.gov: Broker-dealer record-keeping requirements: explains why investors should retain transaction records and check account statements and trade confirmations.
- Investor.gov: How to read confirmation statements: outlines transaction details and confirmation information for US broker-dealer securities transactions.
- Investor.gov: Better understanding your brokerage account statement: describes transaction details and costs that may appear in account records.
Method note: the worked entry, prices, volume, rule, and tags are hypothetical. They are a record-keeping example, not a backtest, a recommended strategy, or evidence of an expected return.
This guide is for educational information only. It is not investment advice, a recommendation, or a promise of trading results. Trading and investing involve risk, including the possible loss of capital.