A Trade Plan Adherence Score measures whether observable decisions matched rules written before the trade. It does not diagnose discipline, psychology, or strategy quality from P&L. A compliant loss remains compliant; a profitable rule-break remains a rule-break.
“I followed my plan” is too vague to audit. Which entry condition was required? What size did the plan allow? Could a stop move under any documented exception? Which exit method applied? Was the opportunity eligible at all? Mechanical scoring begins only after those questions have answerable fields.
The score is useful because it separates three problems that otherwise blur together: a strategy can be followed and lose, a viable strategy can be executed inconsistently, or the evidence can be too incomplete to tell. The method below reports adherence, review coverage, and critical exceptions separately so a clean percentage cannot hide missing data.
The Five-Component Adherence Framework
These five components cover many discretionary plans, but they are a template rather than a universal law. Keep a component only if the plan defines it before the measurement window. Replace or omit anything that does not belong to the strategy, and never add a rule after seeing the outcome.
Component 1: Entry Criteria Compliance
List the minimum observable conditions for an eligible entry: named setup, market or session filter, trigger, invalidation, and any required context. Score the rule from evidence captured before or at entry. “The chart looked good” is not a scorable predicate; “close above the premarked level while the session filter is active” can be checked.
Keep setup quality separate from compliance. A low-quality but authorized setup may follow the plan, while a beautiful unplanned trade may violate it. If the plan uses grades, define the grade rubric before the window and record the version.
Component 2: Position Sizing Compliance
Compare planned risk and allowed sizing logic with actual quantity and aggregate exposure. Include scale-ins, correlated positions, concurrent risk, account currency, and the exact account rule when relevant. A trade is not size-compliant merely because its final loss was small.
If sizing is variable, store the inputs and tier that authorized the size. The trading plan template shows how to turn a preference into a predeclared rule instead of reconstructing permission after the trade.
Component 3: Stop and Invalidation Compliance
Define whether a protective order is required, where invalidation belongs, which stop changes are allowed, and who or what authorizes an exception. Compare the planned stop, actual order history, modifications, and exit evidence. Do not infer a stop violation from a losing outcome or compliance from a profitable one.
Some plans use time, volatility, or structural invalidation rather than a fixed price. That can still be scored if the trigger and required action were written and timestamped. Unknown order history should remain unknown.
Component 4: Management and Exit Compliance
Identify the exit family that applied: fixed target, scale-out schedule, trailing logic, time stop, discretionary decision within documented boundaries, or another explicit method. Then compare the actual management sequence with that rule. An early exit is not automatically wrong if the plan authorized it, and hitting a target does not excuse an unauthorized size change.
Component 5: Opportunity-Selection Discipline
Executed trades can show which ineligible setups were taken. They cannot reveal valid setups that were seen and skipped unless you keep an opportunity log, alert record, or other frozen candidate set. Report these two directions separately:
- Unauthorized-entry rate: entered opportunities that failed the eligibility rule.
- Qualified-skip rate: eligible opportunities not taken, measured only when the full opportunity set exists.
Without a denominator of opportunities, label skip discipline Not measured. Do not treat absence from the trade ledger as evidence that no valid setup appeared.
Calculating the Adherence Score
Start with counts. Percentages are the display layer, not the evidence. Choose a frozen period and one plan version; when the plan changes, begin a new measurement window.
Step 1: Classify each eligible record
For every reviewed trade or opportunity, mark each applicable rule as pass, fail, or unknown/not applicable. Store the source used for the classification. If your journal uses the summary labels Followed, Partially, and Broke, retain all three instead of converting “partial” to whatever number makes the score look cleaner.
Step 2: Report coverage beside adherence
Review coverage = classified eligible records ÷ all eligible records × 100.
Strict followed rate = Followed ÷ (Followed + Partially + Broke) × 100.
Also report partial and broke rates using the same classified denominator. Unknown records reduce coverage; they do not become passes and do not silently disappear. A high followed rate with weak coverage is an incomplete measurement, not evidence of strong adherence.
Step 3: Calculate rule and component scores
For a binary rule, rule adherence = passes ÷ (passes + failures) × 100. A component with several rules can use a simple average or predefined weights. If weights are used, document them before the window, make them sum to one, and explain why they reflect the plan’s consequences. There is no evidence-backed universal weighting that fits every strategy.
A composite can be written as TPAS = Σ(component weight × component score). Publish the component scores, weights, counts, coverage, and critical failures beside the composite. Otherwise the single number is impossible to audit.
Illustrative calculation
Suppose a frozen window contains 24 eligible trades. Twenty-one have a completed plan-adherence review: 17 Followed, 3 Partially, and 1 Broke. Review coverage is 21 ÷ 24 = 87.5%. Among classified trades, the strict followed rate is 17 ÷ 21 ≈ 81.0%, the partial rate is 14.3%, and the broke rate is 4.8%.
Those numbers do not say whether the strategy is profitable, whether a trader is “disciplined,” or whether the next trade will comply. They say exactly how the reviewed records were classified. The three unreviewed trades stay visible as a coverage gap.
Step 4: Keep critical rules outside the average
A weighted average can hide one severe breach behind many easy passes. Define critical rules—such as maximum account risk, prohibited products, daily-loss stop, or mandatory protective orders—as gates. Report the number and exact evidence of critical failures even when the composite is high. For funded accounts, the firm’s current official rule remains authoritative; a journal score does not change the program consequence.
How to Interpret TPAS Without Invented Tiers
| Adherence display | Coverage | Critical failures | Valid conclusion |
|---|---|---|---|
| High | High | None observed | The reviewed window closely matched the frozen plan; strategy quality is still a separate question. |
| High | Low | Unknown | The result is coverage-limited; complete missing reviews before interpreting it. |
| Mixed | High | None observed | Inspect the lowest component and repeated rule-level failure. |
| Any | Any | One or more | Address the exact gated breach; the average must not neutralize it. |
| Improving | Changing | Changing | Check whether the plan, opportunity mix, or review process changed before crediting an intervention. |
There is no context-free “good score.” Set the operational threshold from the strategy and account constraints before the window, then evaluate it with counts and uncertainty. A short or highly repetitive sample can move sharply from one record; a larger heterogeneous sample can conceal a component-specific problem. Report the denominator rather than relying on a magic trade count.
Why Subjective Adherence Rating Fails as an Audit
A retrospective impression is not reproducible. Another reviewer cannot see which plan version you remembered, which trades you excluded, or how profitable exceptions changed the label. This does not prove a particular psychological mechanism; it establishes an evidence problem.
Outcome contamination
When the result is already known, it is easy to relabel a profitable violation as judgment and a compliant loss as a mistake. Freeze the rule and classify the action before reviewing P&L. Then analyze adherence and outcome in separate columns.
Denominator drift
A trader may remember only executed trades, only reviewed trades, or only the memorable part of a period. Mechanical scoring exposes all three counts: eligible, classified, and unknown. Without them, two identical percentages can describe very different evidence.
Plan-version drift
If rules change during the period, a trade can appear compliant under today’s plan but not the plan active at entry. Attach a plan version and effective timestamp. Do not retroactively grade old decisions under a new rule.
The Journal-First Discipline
The minimum record is not a mood score. It is the plan version, eligible setup, planned entry or trigger, invalidation or stop rule, allowed size, applicable management rule, actual execution, and review classification. Add a mistake tag or note as trader-supplied context, not as an automatically discovered cause.
Capture pre-trade fields before execution where possible. Capture modifications and fills from the authoritative source. Complete the review while the evidence is available, but never fill a missing field from confidence alone. The trade quality score guide explains how to keep process quality distinct from outcome quality.
Adherence Improvement Protocol
The purpose of measurement is to test a controllable process change, not to punish a label.
- Freeze the baseline. Choose one plan version, account scope, period, and eligibility definition.
- Find one repeated, evidenced failure. Inspect the rule-level records, not only the lowest headline percentage.
- Write a reversible control. Examples include a pre-entry size check, an order-state check, or a required invalidation field. Choose the control from the observed failure rather than a universal prescription.
- Define success and counter-evidence. State which rate should change, what must stay constant, and what result would make you remove or revise the control.
- Run a new frozen window. Keep plan version, classification, coverage, and critical gates visible. Do not combine windows when the plan or evidence pipeline changed.
- Compare like with like. Check setup, session, instrument, market regime, and account context before attributing a difference to the control.
No universal number of days or trades guarantees improvement. Continue until the exact opportunity occurs often enough to evaluate the control with the uncertainty you are willing to accept. If the pattern disappears because the market or strategy changed, report that limitation.
Who Should Prioritize Adherence Scoring
- Traders changing strategy after a drawdown: first separate compliant losses from execution deviations; the setup failure analysis workflow handles the next strategy question.
- Traders with inconsistent risk: compare documented sizing and concurrent exposure with actual execution before interpreting performance.
- Funded traders: treat critical firm rules as gates and verify them against the exact current program terms.
- Mentors and teams: agree on rule definitions and evidence so two reviewers can classify the same record consistently.
- Systematic traders: compare the execution assumptions in the research process with live order and fill evidence.
An adherence score should not replace the broader trading discipline framework. It is one audit: whether observed execution matched a frozen plan.
Measure Plan Adherence in TSB
TSB is our product. A completed trade review can record the setup, execution grade, and plan-adherence state as Followed, Partially, or Broke; partial or broken reviews can require a trader-supplied mistake tag. The review contract treats plan adherence as covering the planned entry, stop, target, and invalidation.
The evidence layer can report review coverage plus the share of reviewed trades marked Followed or Broke. It must not infer a missing plan, skipped opportunity, emotion, or causal explanation. If the selected evidence set lacks plan-adherence reviews, that gap stays visible instead of being filled by the model.
TSB recognizes 331 import profiles and has processed 600K+ imported trades. Imported execution scale does not automatically create plan evidence: the plan version, opportunity set, and review labels still need to exist. Prop-rule tracking and a lifetime-access route can support the workflow, while the exact current product scope and price render from canonical server truth below.
Start with one rule: define it before the next eligible trade, import or record the execution, classify the review, and keep unknowns visible. Open your trade review
Methodology Note
- Framework status: the five components are an editorial template, not a validated universal psychometric scale.
- Weights: no standard weighting is asserted. Any composite weights must be strategy-specific, predeclared, and published beside the result.
- Partial classification: Followed, Partially, and Broke remain separate in the core summary; an arbitrary numeric value is not silently assigned to Partial.
- Missing evidence: unknown records reduce coverage and never count as compliant.
- Outcome boundary: adherence describes plan match, not profitability, personality, future behavior, or the quality of the plan.
See our editorial evidence methodology for the distinction between verified facts, first-party observations, inferences, hypotheses, and opinions.
Final Verdict: Measure Adherence, Don't Rate It
A useful TPAS is reproducible. It names the plan version, eligible records, rule definitions, classified counts, unknowns, coverage, component scores, weights, and critical failures. It keeps outcome separate and starts a new window when the plan changes.
That makes the score less theatrical and far more useful. You can see whether the evidence supports a repeated process deviation, test one control, and reject the explanation when a matched comparison does not improve. The score does not need an invented “elite” tier to do its job.
