METHODOLOGY

A case enters the archive only when it can be tied to documentation that a reader can open and check independently. All 181 cases currently carry at least one linked document and at least one linked article - 243 source documents in total. Sources are used in this order of preference:

1. Primary government records. Project Blue Book, Project Sign and Project Grudge files held at the National Archives; CIA and FBI reading-room releases; AARO reports; Navy and Air Force incident reports; the PURSUE tranches published at war.gov. These carry the most weight because they are the record the government kept for itself, not a statement made for the public.
2. Sworn or on-the-record testimony. Congressional hearings, signed affidavits, and named first-person accounts from people whose service or employment record can be verified.
3. Contemporaneous reporting. Press coverage published within days of the incident, which fixes what was claimed at the time rather than what the story grew into later.
4. Investigator case files. Work by named researchers - Hynek, McDonald, Maccabee, Vallée and others - where the underlying interviews and measurements are documented.

Anonymous claims with no corroborating record do not get a case page. Where a case is famous but the evidence is contested, it is included with a low score and the contest described, rather than excluded - a reader looking up Aurora, Texas is better served by a sourced page explaining why it is doubted than by silence.

The 1-10 score measures how well evidenced the incident is - not how exotic it is, and not how likely a non-human explanation may be. A high score means multiple independent lines of evidence agree that something happened and that the record of it survives. It is an editorial judgment applied consistently, not the output of a formula.

The score rises with:

• Instrumented corroboration - radar returns, gun-camera or sensor footage, physical trace evidence
• Multiple independent witnesses, especially trained observers such as pilots, radar operators and military personnel
• Contemporaneous documentation, filed before the account could be shaped by publicity
• Official acknowledgement that the event occurred, even where the explanation is disputed

The score falls with:

• Single-witness accounts with no physical or instrumental record
• Long gaps between the event and the first account of it
• Documentation that traces back to a single retelling rather than independent sources
• A conventional explanation that fits the reported facts

Counts are computed from the live case file at build time, so this table cannot drift away from the data it describes.

Cases are tagged with the Hynek close-encounter scale, extended with the categories modern reporting actually uses:

• CE1 - object observed at close range, no interaction with the environment
• CE2 - physical effect left behind: ground traces, electromagnetic interference, injury
• CE3 - occupants or entities reported
• CE4 - reported abduction or perceptual displacement
• CE5 - reported deliberate two-way contact
• radar - the object was tracked on radar or another sensor
• trace-evidence - measurable physical residue was recovered or recorded
• mass-sighting - independently reported by many separated witnesses
• military / government - service personnel or official bodies were directly involved

A case marked declassified is one where a government document about the incident has been released and is linked on the case page. It means the paperwork is public - not that the incident was officially endorsed as unexplained.

Stating the gaps is part of the method. Measured against the current 181 cases:

• 55% of cases are United States incidents. That reflects where records were kept and released, not where sightings happened. Latin American, African, South Asian and Chinese cases are underrepresented.
• 25% of cases are from 2020 onward. The disclosure era generates dense documentation; the 1990s and 2000s are the thinnest decades here.
• English-language sources dominate. Non-English primary documents are underused, which compounds the geographic skew.
• The news pulse is machine-classified. Items on the /pulse feed are aggregated from public RSS feeds and sorted into signal categories by keyword rules. They are not vetted to the standard applied to case dossiers, and appearing in the feed is not an endorsement of the claim.

Case records are revised when a source is retracted, a document is superseded by a fuller release, or a reader shows that a detail is wrong. Corrections are made to the case record itself rather than appended as a note, so the page always reflects the current best understanding, and a materially changed case has its credibility score re-examined at the same time.

Links to primary documents are checked periodically; where a source has moved or been withdrawn, the link is repointed to an archived copy rather than quietly deleted.

UFO Explorer takes no position on the origin of the phenomena it catalogues. It does not sell access, does not host user submissions, and does not treat a compelling story as evidence. Cases that skeptics have explained convincingly are kept in the archive with the explanation attached, because an archive that removes its own inconvenient entries cannot be trusted on the rest.