Skip to reversion research
Mean Reversion StrategyReference / Excursion / Evidence

Frozen-reference excursion workbook

Three complete event contracts, seven authored paths and every declared observation. No instrument research or profitable strategy is claimed.

One reference, several different endpoints

The packet distinguishes strict distance improvement, inclusive band arrival and touching or crossing from the original side. These events need not occur together. Its overshoot case crosses the reference immediately while moving farther away; its equality case crosses repeatedly without getting strictly closer. The default reaches the band and reference, then ends farther away again. None of these endpoints is an executable trade or a net account outcome. A scalar frozen destination and positive closes make the descriptive question explicit; actual spread, valuation or native financial targets require their own identity and research evidence.

Complete declarations expose excluded later observations

Each compact case includes every supplied close, the requested horizon and its independently reconstructed metrics. The ledger includes step zero and all future declarations, including five later values deliberately unused in the bounded case. Its inspected marker makes exclusion visible. The observation indices are authored, not actual dates or elapsed minutes. A null hit means not observed within the inspected horizon, not proven never to recover. The packet supplies no probability, standard error, statistical test or independent sample. Formula checks demonstrate software behavior rather than the financial validity of a proposed mean-reversion model.

Prior-window price reference

ObjectRecordBounded finding
ObjectAnonymous EDU positive closes; not a real instrument or adjusted price source.Written teaching contract; real evidence absent
AvailabilityA reference uses earlier completed observations only; decision input requires original availability.Written teaching contract; real evidence absent
ReferencePrior mean frozen at the event; no future or moving destination inserted.Written teaching contract; real evidence absent
ScalePrior population scale explicitly stated; zero scale means undefined z-score, not zero displacement.Written teaching contract; real evidence absent
EventA predeclared absolute-z threshold; actual thresholds/candidates have not been trained here.Written teaching contract; real evidence absent
OutcomeStrictly closer absolute price distance to the same reference. Crossing alone is a different endpoint.Written teaching contract; real evidence absent
HorizonFive authored future closes, one-based steps; observation time is not calendar minutes.Written teaching contract; real evidence absent
BandInclusive absolute price band of one EDU unit. This is not volatility probability or a stop.Written teaching contract; real evidence absent
PopulationAll seven authored cases, not a selected historical market ledger or independent sample.Written teaching contract; real evidence absent
Entry/exitNo financial entry study; native causal entry and one predeclared actual exit remain separate.Written teaching contract; real evidence absent
CompletionNo instrument research factory/master, fitted model, cloud model deployment or subscriber route.Written teaching contract; real evidence absent
Missing next sourceReal instrument identity, causal vintages, complete event ledger and full independently verified research before financial claims.Written teaching contract; real evidence absent

Spread reference with missing leg identity

ObjectRecordBounded finding
ObjectA proposed two-leg spread concept, not values calculated from actual leg history.Written teaching contract; real evidence absent
AvailabilityBoth legs require synchronized permitted observations and lawful source access.Written teaching contract; real evidence absent
ReferenceA proposed fixed training-only equilibrium; actual hedge-ratio fit absent.Written teaching contract; real evidence absent
ScaleResidual scale would have its own fit scope; single-price z-score is not automatically spread calibration.Written teaching contract; real evidence absent
EventSpread boundary remains proposed; no actual pair or selected opportunity is supplied.Written teaching contract; real evidence absent
OutcomeDistance recovery would require the same frozen residual/reference version.Written teaching contract; real evidence absent
HorizonActual joint leg calendar and label endpoint absent; no invented elapsed duration.Written teaching contract; real evidence absent
BandNo actual spread band calibrated. Single-price fixture inputs do not fill that gap.Written teaching contract; real evidence absent
PopulationNo full paired-data population, join failures or genuine validation set supplied.Written teaching contract; real evidence absent
Entry/exitBoth actual leg executions, borrow/costs and native exit absent; spread crossing is not net profit.Written teaching contract; real evidence absent
CompletionWritten teaching contract only, no independent master or production promotion.Written teaching contract; real evidence absent
Missing next sourceInstrument/leg route, frozen hedge ratio, causal joins, complete accounting and authorized full research.Written teaching contract; real evidence absent

Changed equilibrium, prior finding not inherited

ObjectRecordBounded finding
ObjectAn authored proposal to replace event reference 100 with a later 108.Written teaching contract; real evidence absent
AvailabilityThe later value cannot be used as though known at the original event.Written teaching contract; real evidence absent
ReferenceNew reference identity and a new question; original frozen path remains unchanged.Written teaching contract; real evidence absent
ScaleAny changed scaling needs its own causal source and fit record.Written teaching contract; real evidence absent
EventA revised filter cannot erase original failed or unresolved event rows.Written teaching contract; real evidence absent
OutcomeCloseness to the revised destination is not the original recovery label.Written teaching contract; real evidence absent
HorizonNo extension after inspecting later prices. Unhit endpoints stay not observed within the declared horizon.Written teaching contract; real evidence absent
BandChanging band after seeing the path consumes selection evidence.Written teaching contract; real evidence absent
PopulationOriginal and revised populations must stay attributable; neither is genuine market history here.Written teaching contract; real evidence absent
Entry/exitA reference revision grants no new financial entry ranking or exit tournament.Written teaching contract; real evidence absent
CompletionNo inherited independent master, served model identity or certification.Written teaching contract; real evidence absent
Missing next sourceAuthorized successor experiment with actual causality, frozen native policies and full independent completion.Written teaching contract; real evidence absent

Seven complete authored path cases

CaseFirst closer stepFirst band stepFirst touch/cross stepTerminal distance recovery %
Gradual recovery and later rebound13480.0
Crossing first, recovery later34175.0
Equal distance is not closerNoneNone10.0
Distance widens without a hitNoneNoneNone-75.0
Below-reference start and repeated band visits1220.0
Starts at reference, percentage undefinedNone00None
Later declared values excluded by horizon13None90.0
Two authored paths around reference 100, including an overshoot from 104 to 95
Authored observation paths, not genuine market prices or strategy returns.

The cards deliberately retain real-evidence gaps

The prior-price card describes a causal reference but has no real source. The spread card requires both actual legs, a frozen hedge ratio, synchronized source versions and accounting; it is not simulated by the positive-price lens. The changed-equilibrium card records why a later destination cannot replace the original label. All have null actual source, fitted parameters, independent master and fills. Their written completeness does not complete a genuine instrument task. Real research remains subject to its authoritative route, raw causal entry objective, frozen entry, one native exit and full independently verified master.

Downloads are educational artifacts, not production promotion

The site adds no quote fetch, calibration procedure, trading connection, daily pipeline or production model. Inputs remain visitor declarations and exports carry their limitations. Local software tests and temporary preview servers are not live refresh dependencies. Any real production release needs its separate cloud artifact, subscriber route and certification while personal computers are off. Marketing a documented model result would need the actual ledger and source scope; these fictional event cards cannot fill that gap. Keep them as question templates and known-answer examples rather than relabelling them as authentic historical performance.