A rigorous framework for investigating Unidentified Anomalous Phenomena using multi-domain analysis, Bayesian hypothesis testing, and radical epistemic humility.
We do not know what is happening. The goal is not to "solve" UAP but to reduce uncertainty through rigorous analysis.
Treat witnesses with respect while treating claims as unproven. Test extraordinary claims with extraordinary evidence standards.
Physics, psychology, folklore, intelligence analysis, and atmospheric science. Single-discipline analysis misses crucial patterns.
Maintain 10 competing hypotheses with equal initial weight. Update priors systematically as new evidence emerges.
Primary sources with clear provenance. Document chain-of-custody gaps explicitly. Re-verify claims when possible.
Evidence is classified by quality, not by what it supports. Extraordinary claims require Tier 1 evidence.
Multi-sensor data with chain-of-custody, corroboration, metadata, and expert analysis from trained observers.
Multiple independent credible witnesses with consistent details and partial sensor or documentation support.
Single-witness testimony or low-quality media without metadata. Most reported sightings fall here.
Rumor, anonymous claims, "a friend in intel," unverifiable documents. Cannot be used for hypothesis testing.
We maintain multiple explanations with equal initial weight. None are privileged—evidence updates all priors.
Cases clustered by observable patterns—kinematics, sensors, location—not by conclusions.
November 14, 2004 • Southern California Coast
Multi-day encounter between USS Nimitz Carrier Strike Group and a white, Tic Tac-shaped object. Cmdr. David Fravor and Lt. Cmdr. Alex Dietrich observed the object exhibit extreme acceleration (100g+), with no visible propulsion, exhaust, or control surfaces.
Apply tier classification. Document chain-of-custody. Capture all metadata. Log alternative explanations.
Cluster by observable patterns: kinematics, luminosity, radar, encounters, geography, time.
Maintain 10 hypotheses. Track confidence. Update with Bayesian methods. Design discriminant tests.
Produce ranked case lists, hypothesis matrices, next-action plans, and narrative risk assessments.
"The phenomenon adapts itself to the cultural expectations of each epoch."
— Jacques Vallée