Mean Reversion Reference Workbench
Define the unchanged reference, eligible deviation and later distance before calling a price move recovery. Statistical behavior and trade economics remain separate.
Your evidence states are not a provider score, audit opinion or performance finding.

Primary reference documents
References support source definitions and mechanisms, not a completed review, measured return or endorsement. The questions are this publication's editorial record design.
Evidence states do not fit or verify the reference
The workbench asks for identified sources, reference scope, calibration, population and operational evidence. Manually marking an item documented records the user's claim and reference, not an independently authenticated finding. The three filled cards give complete question examples while retaining absent actual source and master. Keep unresolved fields unresolved and preserve import/export identity; a familiar URL is not a substitute for the relevant record. A workbench result is not a provider score, model admission or certification of a reversion edge.
Operational questions remain separately cloud-owned
Actual model delivery needs its served cloud artifact, subscriber route and certification distinct from a research LIVE label. Production data refreshes, publishers, monitors and verifiers must remain functional with personal computers off. Local educational forms and temporary QA servers are not those pipelines. This edition creates no scheduler, tunnel, PC fallback or extra paid capacity. It publishes static research questions and authenticated website evidence rather than claiming that the proprietary model chain has been implemented.