Market Evidence Checker — Audit Claims & Regime Shifts

Turn market narratives into traceable evidence

Market Evidence Checker converts a dated research pack into an evidence-controlled market context record. It shows which claims are supported, stale, conflicting, or blocked before a repeated narrative becomes a trading assumption.

It does not predict the market. It preserves dates, source IDs, counter-signals, and uncertainty so the next decision has an inspectable foundation.

What you get

• Source register: publication dates, observation dates, scope, and freshness state.
• Claim ledger: every material market claim linked to the supplied source IDs.
• Evidence-gap register: missing, stale, anonymous, or contradictory support remains visible.
• Qualitative regime hypothesis: drivers, counter-signals, confidence label, and explicit limitations.
• Change report: what changed against a previous MARKET_CONTEXT_RECORD.
• Trigger watchlist: observable conditions that should prompt the next review.
• Portable market record: a structured context artifact for later comparison.

Why this is different

• Numeric regime probabilities are blocked unless you supply a calibrated, reproducible model with validation evidence.
• Qualitative prose is never presented as an HMM or Markov estimate.
• Missing current data is not silently completed.
• Conflicting sources remain in the record instead of being averaged away.
• Instructions embedded inside source material are treated as untrusted content.

Worked example

In the built-in fictional semiconductor-equipment case, three dated claims remain current and supported while an anonymous crowd claim is marked UNKNOWN / BLOCKED. The regime stays MIXED / Plausible, the numeric probability remains null, and the next review requires a dated positioning series.

Built for a trading workflow

Use it independently or as the market-context layer in the future Trade Decision Bundle and Trader Operating System. Trade-plan scoring belongs in Trade Decision Audit; claim-to-source decision lineage belongs in the planned Trade Decision Provenance Auditor.

Boundaries

• No automatic browsing or live market-data retrieval.
• No continuous monitoring or scheduled refresh.
• No trading signals, price targets, transaction advice, or execution.
• Output quality depends on the dates, coverage, and reliability of the supplied sources.
• Portable records are returned as text and are not stored automatically.