The audit checks whether the submitted evidence is internally consistent and whether the result survives basic cost, concentration, and stability pressure. It does not recommend an instrument, entry, exit, or allocation.
Data provenanceSource, timestamp, session, duplicate, and missing-field reviewAcceptable
Baseline reproductionSubmitted result compared with documented assumptionsReproduced with notes
Cost stressFees, spread, slippage, turnover, and sensitivity rangeFragile
Out-of-sample reviewTrain/sample split and later-period behaviourMixed
ConcentrationTop-trade, month, session, and regime dependencyReview required
Kill testsRule perturbation and adverse assumption checks where data supports itPartially failed
Assumption register
The report makes assumptions visible before judging the result.
Execution timing
Entry and exit timestamps treated as end-of-bar unless proven otherwise.
Costs
Commission, spread, and slippage are stressed above the submitted baseline.
Data handling
Missing sessions are documented before any score is interpreted.
Scope limit
No future return, instrument, allocation, or live-use conclusion is made.
Verdict frame
RejectParkNeeds more evidenceWorth further review
Sample findings
Data source is documented, but session handling needs stronger proof.
Cost stress materially reduces the submitted expectancy.
Out-of-sample stability is mixed across regimes.
Top-trade dependency is below the reject threshold in this sample.
No conclusion is made about future returns or what to trade.
What the report never claims
A report does not prove future returns, tell you what to trade, or clear a strategy for live use. It only records what the submitted evidence can and cannot support.
Next step
Submit your own evidence for a Reality Check.
Use the intake form to describe the strategy and the evidence you have. Public V1 does not accept files; follow-up happens after scope is agreed.
The artifact is a fuller static sample of the report format: cover, verdict frame, evidence matrix, assumption register, rejection triggers, and research-only boundaries.
