The case at its strongest

The Standout Cases

Every one of the 216 analyzed PURSUE documents was put through the same filter: witness credibility, sensor corroboration, specificity of the anomaly, and institutional routing. 174 did not survive it. These are the 28 that did, and this is what they agree on.

A note on yardsticks: these rate the strength of documents, not incidents. The Evidence Database rates each incident as a standalone case, so a document can stand out here for its specificity or provenance while the matching incident there still rates only circumstantial. The two pages weigh different things by design. And the filter cuts both ways: the 21 documents rated contested are examined, reason by reason, on The Contested Cases.

All four tranches · Seven agencies · 1949–2026

28
Strong-rated
174
Did not rate strong
6
Agencies represented
18
From 2020 or later
7
Involve parent-child behavior
5
Nuclear-adjacent sites
199
documents analyzed
174
did not rate strong
28
survived maximum skepticism

Noise to signal: roughly 7 to 1. About one document in eight rates as strong. This page is only that residue.

The anatomy of the 28

Where the strongest documents come from, in six clusters.

Modern military reports
9
Western U.S. Event
6
CIA / historical
4
Nuclear complex
4
FBI field reports
3
ODNI narratives
2

Six agencies, 1949–2026. Eighteen of the twenty-five are from 2020 or later; the rest anchor the same signatures back to the founding of the nuclear complex.

What the 28 agree on

Cross-cutting patterns that appear independently across agencies, eras, and sensor types.

7
Parent-child ejection
A larger object launches smaller ones, documented by AARO field agents (2023), a senior intelligence officer (2025), an Army FLIR platform (2026), and FBI agents (2026). Four independent institutional sources, three years.
5
Nuclear-adjacent geography
Sandia Base (1949), Pantex Plant (modern), western U.S. weapons test range (2025), Cheyenne Mountain (2022), and the Western U.S. restricted site (2023). The same geography recurs over 75 years.
3
90-degree turns
Kazakhstan (1994): "corkscrews and 90-degree turns under very high G's." Greece (2023): "multiple 90-degree turns at 80 mph." Iraq (2023): "maneuvering quickly" at FL600+. Same kinematic signature, three separate theatres.
3
Multi-sensor corroboration
The USPER Statement and ODNI-UAP-D001 both document the same event with FLIR, NVG, radar, and naked eye simultaneously. The Western U.S. Event adds rangefinder measurements to five first-person accounts.
2
Sphere with rigid appendage
UAE (2024): "glowing hot spherical object with a vertical, unwavering cylindrical pole/bar." Greece (2024): "diamond-shaped with a non-maneuvering probe at the bottom." Identical structural signature, separate theatres.
1
Official advanced-capability flag
Syria November 2023 (D74): the MISREP explicitly flags "Advanced Capabilities: YES, TRAVELED [redacted]." One of the few cases where the filing crew annotated the report with a formal capability assessment.

Same behavior, independent sources

The skeptic's strongest objection is reporting bias: one culture seeing what it expects to see. The rebuttal is the grid below. The same behaviors recur across separate agencies, decades, and sensor types, with no document referencing another. That independence is the engine of the case.

Recurring behaviorDept of WarODNIFBIStateEnergyEra span
Parent-child ejection● 2023, 2026● 2025● 20262023–2026
Nuclear-adjacent geography● 1949, 2023● 2025● 2022● modern1949–2025
90-degree turns● 2023● 19941994–2023
Multi-sensor corroboration● 2023● 20252023–2025
Sphere with rigid appendage● 20242024
Official advanced-capability flag● 20232023

Two behaviors, parent-child ejection and nuclear-adjacent geography, each surface across three or four independent agencies. The CIA's four strong documents sit outside this behavioral grid by design: they are the documentary foundation (the 1952 Scientific Advisory Panel, Blue Book Special Report No. 14), not field observations.

The Flagship Case

The Western U.S. Event, 8 documents

The single most-documented case in the corpus. A restricted national-security site in the western United States, October 2023. Six trained observers across multiple vantage points, an AARO field investigation, five first-person written narratives, and a formal AARO unresolved-case analysis concluding that roughly 40 percent of the phenomena has no plausible explanation. The parent-child ejection signature is documented by five independent witnesses independently describing the same behavior.

Modern Tactical Record

DoD Mission Reports, 8 documents

Active-duty ISR crews and AFSOC platforms across six theatres: Greece, Syria, Iraq, UAE, East China Sea, Persian Gulf, and Kazakhstan. Trained military observers with calibrated sensors. Each report was formally filed through AARO and declassified under the PURSUE program. The kinematic signatures, 90-degree turns, co-altitude high-speed approaches, and disappearance from targeting pods, are documented independently in separate theatres with no shared observation chain. A Navy Range Fouler report (Tranche 4) adds the U.S. East Coast: a roughly 28-year veteran aviator's high-speed rectangular contact that outran the sensor field of view, captured on video.

First-Person Accounts

Senior Intelligence Officer Encounter, 2 documents

The same event, documented twice: the USPER Statement (Tranche 1, SECRET//NOFORN, the handwritten account) and ODNI-UAP-D001 (Tranche 2, the prose narrative filed through ODNI channels). A senior US intelligence community official on a dedicated UAP investigation mission, over a western U.S. weapons test range, late 2025. FLIR, NVG, radar, and naked eye across multiple observer positions. An orb that split in two, approached the helicopter to within ten feet, and sped away.

FBI Modern Files

FBI Field Reports, 3 documents

Three modern FBI field reports rated Strong: the Cheyenne Mountain sighting (February 2022, multiple Army soldiers, daylight, motionless bean-shaped object) and two corroborating FD-302s from the Northeastern U.S. (February 2026, paired red orbs). Both the Cheyenne Mountain location, adjacent to the primary nuclear command center, and the paired-orb morphology in the FBI 2026 reports echo patterns found in the DoD and ODNI record.

Nuclear Complex Record

The Nuclear Complex, 4 documents

A 116-page SECRET-classified Sandia Base correspondence file from 1949, documenting a sustained multi-agency investigation of unexplained aerial phenomena over the New Mexico nuclear complex. Tranche 4 supplies its founding primary source: the verbatim transcript of the 16 February 1949 Los Alamos conference, with Teller, Bradbury, Reines, and LaPaz in the room, concluding the green fireballs were most likely not conventional meteorites. And two Pantex Plant records from Texas, the highest-value nuclear weapons assembly facility in the country: the existing radar-tower image and, now, the full incident report behind it, a silent object with no identifiable propulsion that evaded the Protective Force. The nuclear-proximity association spans 77 years and two separate agencies.

CIA Institutional Record

The CIA Foundation, 4 documents

Four CIA documents that define the institutional context rather than the phenomena themselves. The 1953 Robertson Panel recommended public debunking and surveillance of civilian UFO groups, and its recommendations are still traceable in the institutional behavior documented across the rest of the corpus. The U-2/OXCART history explains a documented share of Cold War-era sightings. Project Blue Book Special Report No. 14 found statistically that the better the report quality, the more likely it remained an "unknown." The Australian Defence assessment argues that the problem has genuine intelligence dimensions no allied government can assess alone.

What 28 documents can and cannot prove

The 28 Strong-rated documents establish, across six agencies and 77 years, that something is happening that trained observers cannot explain, that the institutional response has been consistent and documented, and that the kinematic signatures recur across independent observation chains with no shared personnel or equipment.

They do not prove what the phenomena are. They cannot, because no classified sensor data has been released, no physical sample has been publicly confirmed, and the redaction wall around the most specific reports remains intact. The Standout Cases are the strongest cards in a hand that does not yet include the trump card.

What they do prove is that the question is real, the witnesses are credible, the sensors are calibrated, and the institutional seriousness is documented. That is a different and more defensible claim than the one this field usually makes.