Trader Process

Trading Journal: How to Tag Setups, Mistakes and Market Conditions for Analysable Data

A practical trading journal structure for tagging setups, mistakes and market conditions to compare decision quality and results.

Trading Journal: How to Tag Setups, Mistakes and Market Conditions for Analysable Data

An analysable trading journal is not a free-form diary. It is a structured dataset in which the plan, actual execution, one primary setup, market conditions, process mistakes and result are recorded consistently for every trade. The core principle is to keep facts, interpretation and outcomes separate.

If the tag “good trade” means a profit in one entry, correct execution of the plan in another and a pleasant feeling in a third, the comparison cannot tell you anything reliable. By contrast, a predefined vocabulary helps answer practical questions: under which conditions a particular setup is executed with discipline, which mistakes recur, and whether differences in results remain after commissions and other costs are included.

What a Trading Journal Should Actually Measure

The purpose of a journal is not to prove that a trader is good or bad. It should help separate four things:

  1. Opportunity: which setup and market context were visible before entry.
  2. Decision: what the plan specified—the entry, initial stop-loss, target and position size.
  3. Execution: what actually happened with the orders and position management.
  4. Result: profit or loss after attributable costs, preferably also expressed in standardised risk units.

This separation prevents two common interpretation errors. A profitable trade does not automatically become a high-quality decision if the plan was breached. Conversely, a losing trade is not a process mistake if it met the predefined conditions. The result is one observation. Process quality is what the trader can verify and repeat.

The Minimum Data Model for One Trade

You can start with a spreadsheet or database. The tool matters less than consistent field definitions. One row should represent one completed trade. If a position is added to or closed in stages, executions can be stored in a separate table, but they should retain a shared trade_id in the summary.

Field group

What to store

Why it is needed

Identity

trade_id, instrument, direction

Prevents records from being confused and allows orders to be linked

Time

Plan, entry and exit times, time zone

Makes it possible to analyse session and event context

Plan

Planned entry, initial stop-loss, target, size

Records the decision before the result can influence it

Execution

Actual execution prices, quantity, order type

Shows the difference between the plan and execution

Costs

Commissions and other known transaction costs

Prevents gross results from being confused with net results

Classification

Setup, market regime, volatility, session

Creates comparable segments

Process

Plan adherence, mistake tags

Allows controllable actions to be analysed

Result

Net P/L and realised R

Standardises trades of different sizes

Evidence

Screenshot before and after the trade, brief note

Preserves context for auditing purposes

For a linear instrument, the initial planned cash risk can be calculated in simplified form as |plānotā ieeja − sākotnējais stop| × daudzums. Realised R is the net result divided by the initial planned cash risk. Derivatives, currency pairs and contracts may require calculations for point value, contract multiplier and currency conversion. If the initial risk is not fixed or the data is unreliable, it is better to mark it as “not calculable” than to create a seemingly precise R.

How to Create Setup Tags

A setup is a predefined combination of conditions under which a strategy allows a trade to be considered. It is not a description of the trade’s result. In practice, assign one primary setup tag to each trade, such as breakout_retest or range_rejection. Store additional characteristics in separate fields rather than creating ever-longer tags such as breakout_retest_high_vol_morning_good.

Each setup needs a short definition card:

  •          a name and fixed code;
  •          mandatory conditions;
  •          conditions that invalidate the setup;
  •          the permitted entry principle;
  •          at least one visual example of a valid and invalid case;
  •          the definition version and effective date.

The version matters because trades before and after a rule change are no longer fully equivalent observations. Historical entries should not be silently renamed to match the new interpretation. For example, you can retain setup_version = 1 while using version 2 from then on.

If you cannot determine before entry which setup the opportunity matches, use other or unknown and assess later whether a new category is genuinely needed. Creating a new tag for one unusual trade usually fragments the data rather than improving the analysis.

Separate Market Conditions into Independent Dimensions

A single slikts tirgus tag cannot be analysed because it combines direction, volatility, liquidity and personal judgement. It is more useful to create several controlled dimensions.

Dimension

Possible values

Definition principle

Market regime

trend, range, transition, unknown

Determine using predefined price-structure criteria

Volatility class

low, normal, high

Link to one consistent indicator and comparison period

Session

pre-market, opening phase, mid-session, close

Define boundaries in the time zone of the specific market

Liquidity state

normal, reduced, unknown

Use a predefined spread or volume criterion where data is available

Event risk

scheduled event nearby, none, unknown

Record using a predefined calendar and time window

It is important to classify conditions using information that was available when the decision was made. If a trend is assigned only after the price has already moved decisively in one direction, hindsight bias is introduced. Use unknown for an unclear case. It is better-quality data than a forced guess.

Absolute labels such as “high” and “low” can mean different things for different instruments. The vocabulary should therefore specify the calculation rule, data source and time period. Otherwise, the meaning of the same tag changes over time.

A Mistake Tag Describes a Verifiable Deviation from the Plan

A mistake is not “the trade lost money”. A mistake tag should identify an observable action that can be compared with a previously recorded rule. Examples include:

  •          entry_without_trigger — entry before the mandatory signal;
  •          late_entry — entry outside the zone permitted by the strategy;
  •          oversized_position — size exceeds the planned amount or the amount allowed by the rules;
  •          stop_widened — the initial stop was moved in the direction of greater risk without a rule permitting it;
  •          unplanned_exit — the position was closed for a reason not included in the management plan;
  •          missing_plan — mandatory plan fields had not been completed before entry.

One trade may have several mistake tags, but each needs a precise definition. Emotions such as fear or FOMO can be recorded in a separate self-assessment field. They should not automatically be declared the proven cause of the trade. The observable mistake is “entry without a signal”; FOMO may be the trader’s hypothesis about the reason for that action.

Alongside mistake tags, include a plan_adherence field with the values yes, partial, no or unknown. This makes it possible to select cleanly executed trades even while the mistake categories are still being refined.

When to Complete Each Part of the Journal

The correct recording sequence is more important to data quality than lengthy notes.

Before the trade, record the setup, context, planned entry, initial stop, target, size and cancellation condition. These fields should not be overwritten after the result is known; add a separate, auditable note for any correction.

During execution, retain the actual order data. Where possible, import it from the broker rather than entering it manually. Imported data still needs to be checked: time zones, partial fills, commissions and instrument symbols may differ between systems.

After the trade, assign mistake tags, rate plan adherence and add a brief note. A screenshot is useful if it shows the structure visible at the time of entry and marks the planned levels, rather than showing only the final P/L.

The quality of the journal itself should also be reviewed periodically: missing fields, new free-text variants, duplicates and tags interpreted differently across entries.

How to Turn Tags into Analysis

For an initial report, one row for each setups × tirgus režīms combination is sufficient. The columns can show:

  •          number of trades;
  •          plan-adherence rate;
  •          mean and median realised R;
  •          proportion of profitable trades;
  •          mean positive and negative R;
  •          proportion of trades with mistakes;
  •          net costs, where available.

The proportion of profitable trades is not sufficient on its own: frequent small gains can coexist with less frequent but much larger losses. Mean net R directly summarises standardised outcomes, but it should still be considered alongside the number of observations, dispersion and data quality.

The next comparison is trades without process mistakes versus trades with mistakes within the same setup and similar conditions. This may reveal an association, but not necessarily a cause. For example, difficult market conditions may increase both the frequency of mistakes and losses. A journal is an observational dataset, not a controlled experiment.

Do not draw a firm conclusion from a few trades or search for an ever-narrower filter until only successful historical examples remain. The more combinations you test, the easier it is to mistake a random coincidence for an edge. First formulate a hypothesis from the journal, then test it with new, previously unseen trades or an appropriate strategy-testing process.

A Short Hypothetical Example

Suppose the journal contains a trade with the primary setup breakout_retest, the regime trend, volatility high and session open. The plan had been prepared, but the actual entry occurred outside the permitted zone. Assign late_entry to the trade while storing the result separately—for example, as realised R.

If the trade is profitable, late_entry is not removed. If it loses money, the mistake is not supplemented with the tag bad_trade, because that would rename an outcome as a process diagnosis. The entry can later be compared with trades using the same setup in a similar context where the entry followed the plan.

This structure answers a precise question: “How did outcomes and execution quality differ when the entry rule was followed?” It does not answer the broad and unclear question: “Why am I unsuccessful?”

Limitations and Risks

A trading journal does not make a strategy profitable or guarantee future results. FINRA emphasises that day trading can be extremely risky, while ESMA and CFTC investor materials encourage investors to understand product, leverage, cost and fraud risks. A journal helps review the process, but it does not replace risk limits, broker statements or an appropriate understanding of the specific instrument.

Data limitations must also be considered:

  •          missed opportunities are not recorded, so the journal describes selected trades rather than the full set of opportunities;
  •          hindsight bias can affect tags assigned after the result is known;
  •          strategies and market conditions change over time;
  •          incorrect commissions, currency conversion or contract multipliers distort P/L and R;
  •          broker exports, screenshots and notes may contain sensitive financial or account data.

Access to journal files should be restricted, backups should be created, and account numbers, personal data and other unnecessary sensitive information should be removed before sharing. Tax or legal records should use documents that meet the applicable requirements; a personal analysis journal is not automatically considered an official record.

Practical Implementation Checklist

  •          Create one fixed trade_id for each trade.
  •          Separate the plan, execution, context, mistakes and result.
  •          Select one primary setup for each trade.
  •          Store market conditions in separate dimensions.
  •          Write a definition and an example of an invalid case for every tag.
  •          Include unknown so that guessing is unnecessary.
  •          Retain versions of the tag vocabulary.
  •          Always show the number of observations and missing data in the analysis.
  •          If a hypothesis emerged from historical data, test it with new observations.

A good trading journal reduces room for interpretation. When what was known before the trade, what the trader did and what the result was are recorded consistently, the journal becomes an auditable record of the decision-making system. A list of wins and losses is only one part of it.