Evidence, not hype

Validation

Current Confluence evidence pack for 13FLOW. This page is intentionally conservative: it separates mechanical dataset readiness from any alpha or investment-performance claim, so a reviewer can see exactly what is ready and what still needs review.

125feature rows
25tickers
100.0%60d return coverage
50.0%60d hit rate
01 · Scope25 liquid tickers with mature 13F plus Form 4 evidence.
02 · GateSchema and row checks pass before any public claim.
03 · MetricsDescriptive results are weak or neutral, not promotional.
04 · NextFull-universe adjusted-price review remains the required artifact.

Mechanical Evidence

Readable takeaway: the evidence pack is structurally ready for review, but it is not yet a published validation study.

FieldValue
Artifact statusvalid_minimum_schema
Evidence reviewmechanical_evidence_ready_for_review
Metrics statusminimum_schema_valid_metrics_unreviewed
Feature scope13f_form4_joined
Rows125
Tickers25
Row errors0
Rows with Form 4 accessions101
Rows with open-market buyers14
Tickers with open-market buyers9
20d forward-return coverage100.0%
60d forward-return coverage100.0%
120d forward-return coverage100.0%

Descriptive Metrics

These 60-day metrics are weak or neutral. They are useful as a review checkpoint, not as a performance promise.

MetricValue
Splittest
Modelfull_score
Horizon60 trading days
n113
Rank IC-0.003655
Permutation p-value0.964072
Top-bottom spread0.026515
Top-bottom spread CI95-0.078722 -> 0.110038
Hit rate50.0%
Mean forward return0.5%

weak_or_neutral_descriptive_metrics; this is not a validated alpha, forecast, probability or expected-return claim

What this proves

  • The local feature table is mechanically schema-valid.
  • The 25-ticker artifact joins 13F and reviewed Form 4 accessions.
  • Forward returns are complete for the 20d, 60d and 120d horizons on the mature window.
  • The artifact is ready for human methodology review.

What this does not prove

  • It does not prove validated alpha.
  • It does not provide a probability, price target or expected-return model.
  • It does not cover the full historical universe.
  • It does not complete price-source, delisting, liquidity, costs or no-lookahead review.

How to use this page

For analysts: use the validation boundary to decide how much weight to give a Confluence signal before reading the filings.
For reviewers: use the status, artifact hashes and source links as due-diligence evidence before relying on the data.
For operators: do not market output as validated alpha until the next artifact covers broader history, adjusted prices and no-lookahead controls.

Verification Links

SurfaceUse
/api/product-statusMachine-readable validation boundary
/api/methodology/appApplication methodology contract
/api/methodology/confluence-v1Frozen Confluence v1 contract
/statusDeployment and runtime proof
/proCommercial access boundary