Original editorial revision: 9 October 2026. Expanded edition: 14 October 2026.
Educational diagnostics, not a signal service. A standardized price displacement is not a calibrated probability; invented examples and user-supplied observations are not verified market performance.
Observation-count, wall-clock and completed-session horizons are different policies. Freeze the chosen clock and maximum hold before evaluation. A half-life estimate is not permission to extend a losing trade or redefine the record after seeing its exit.
A rolling displacement does not supply a normal-tail probability. Mean-reversion evidence needs a specified process and conditional relationship, not just a price below a mean. Stationarity diagnostics, distance recovery and executed profit must be examined separately.
A process-to-reference map
Choose the process before choosing the destination: price level, return and fitted spread residual describe different hypotheses. Then identify the causal source and reference clock. A prior average is available only after its permitted observations exist; a fitted equilibrium has training scope and parameters. The cards distinguish these designs from a later changed destination. None supplies authentic data or a fitted hedge ratio. A fixed positive-price exercise is intentionally a narrower lane than actual multi-leg research, and its clear limitations should accompany any derived label.
An endpoint-to-horizon map
Strict closeness, inclusive band arrival and touch/cross can occur at different steps. The new workbook shows all three on complete authored paths. The horizon fixes which observations may answer the question; later declarations remain visible but unused. A null event within a complete horizon is not the same as missing data or proof of never recovering. Observation count requires an actual calendar before becoming elapsed time. Choosing an endpoint after inspecting the path changes the hypothesis, even if the arithmetic for the new endpoint is correct.
A diagnostic-to-financial map
Descriptive distance and assumed expectation decay do not establish tradable profit. Economic evaluation needs raw causal entry ownership, a frozen candidate, its single predeclared native exit, actual fill convention and complete accounting. Event overlap and repeated selection can also undermine a naive interpretation of counts. Keep those evidence objects separate instead of presenting a neat chart as a complete experiment. The original lab and worksheet remain available as diagnostics, not substitutes for the authoritative full instrument factory or independent master.
A research-to-release map
A completed independent process may accept no model, which is a different state from an unimplemented or blocked master. Genuine production delivery separately requires code, registry, deployed cloud identity, subscriber route and certification. A file labelled LIVE cannot answer those stages. The education site deploys static content only and must remain served when workstations are off. It adds no production data refresh or PC fallback. This method map makes each missing artifact explicit before any marketed result or operational badge is allowed to claim it.
Keep the outcome operator with its source
Store strict closeness and inclusive band equality explicitly. A table of hit indices without the reference and operators is incomplete, even when every index is computed correctly. Retain those definitions before connecting the table to any future genuine model experiment.