RESEARCH — WHAT WAE IS LEARNING

MEASURED, NOT CLAIMED.

WAE researches what it operates. This page separates what exists from what is planned — and what is planned from what is validated. Nothing here is marketing certainty.

R-01

ESTABLISHED — EVIDENCE METHODOLOGY

THE EVIDENCE KERNEL

A zero-dependency typed provenance system. Records carry entities, activities, and relations in a PROV-O-shaped model. Ingestion is checksum-bound: source bytes are matched to declared SHA-256 digests before any aggregation. Verification is an explicit invariant — a claim can be marked verified only when a complete, passed verification chain binds eligible real evidence to the claim. Invalid canonical states fail closed.

Adversarial referee tests attack the kernel with fake evidence, fixture masquerades, temporal inversions, and tampered caches. The current public consequence is the verified record WAE-CLM-2026-002 and the honestly asserted WAE-CLM-2026-001.

Inspect the records →
R-02

ESTABLISHED — FIELD MEASUREMENT

FIELD SURVEYS

Operating data is reduced to safe derivatives: identifiers and captions stripped, per-asset magnitudes kept, distributions computed. The field survey of the Wild Animal Encounters export — 1,269 assets ranked by view magnitude — is published as a measured instrument on the homepage, with a visible data register for verification.

The safe derivative commits to no raw rows and no account handles. Its reduction method is recorded in the artifact itself.

See the survey →
R-03

OPEN — PROSPECTIVE VALIDATION NOT ESTABLISHED

PREDICTIVE DIRECTION

WAE's predictive direction is real but early: historical evidence exists; prospective validation does not. No calibration claims, no causal-prediction claims, and no cross-brand transfer claims are made anywhere on this site, because current evidence does not support them.

When a prospective result is validated, it will appear here as a record with a chain — not as a headline.

THE DISTINCTION THAT MATTERS

Historical evidence says what happened. Frozen model evidence says what a model would have said. Prospective validation says what is actually predicted before outcomes — and it has not happened yet. These three are never collapsed into one another on this site.