Four hypotheses are worth stating formally against the released record. None is a finding, and they are not equally supported: one of them, the Program, breaks on evidence the corpus itself supplies. This page states each one formally, lists what it predicts, checks every prediction against named documents, presents the evidence against, and says what would falsify it. The documents are one click away throughout; read them and disagree.
Any theory of this corpus has to account for three patterns that survive every filtering pass we have run on the corpus. The third was added after the August 2026 release, which supplied the first controlled evidence for it.
The 1946–59 files are overwhelmingly discs: the FBI headquarters saucer file alone spans ten sections of them, and the Air Force’s own 1955 statistical study categorized thousands of reports in shape classes that barely exist in the modern record. The 2020s record is dominated by spheres and orbs on infrared. Underneath the drift, the luminous-orb family appears in every era: the wartime night-phenomena file, the 1949–50 green fireballs over the nuclear labs, and the modern FLIR record. The full data is charted on Morphology.
Across independent witnesses, sensors, and decades, the same behaviors recur: parent objects releasing or absorbing smaller ones; objects that approach or orient on observers; objects that vanish from tracking at the moment of characterization; objects visible to one sensor band and invisible to others; and a clustering on nuclear and defense infrastructure that includes the Soviet side of the Cold War. The catalog is documented behavior by behavior on Behaviors.
This is not the drift in Pattern one, which unfolds across decades and which sensor-era arguments can absorb. This is divergence inside a single event. In 2026 two witnesses standing together watched the same lights: to the naked eye they were dull red dashes, and through night optical devices the same objects were circles wrapped in the halo an image intensifier puts around a point source. The FBI commissioned renderings of both views and published them side by side. The same interview records that one witness saw dashes where his companion saw triangles or diamonds, and attributes the difference to the first witness’s astigmatism. Thermal behaves no better: the Gulf of Oman orbs read cold in both infrared polarities, objects in the western United States read black-hot on one night and thermally hot on the next in the same sector, and the red lights of that same 2026 encounter could not be acquired on thermal at all while remaining bright to the eye and to night optics.
Stated on its own, this pattern is a discriminator rather than a curiosity: the Mirror predicts it, and the Survey is neutral on it.
The phenomenon co-varies with its observers: what is reported tracks the sensing technology, expectations, and conceptual categories of the era doing the reporting, in a way that outruns simple reporting bias.
Intellectual lineage: Jacques Vallée’s control-system conjecture, which predates every modern release. The corpus’s relationship to Vallée’s specific claims is examined on The Theorists; the books themselves are outside the release and are treated as context, not evidence.
Visual-report era: discs, in the thousands, formally categorized by the Air Force’s own statisticians. Naval radar-FLIR era: tic-tacs and cylinders. Drone-IR era: spheres and orbs. The drift is the cleanest large-scale pattern in the corpus, and it tracks the instrument as much as the decade.
August 2026 update. The mechanism has now been demonstrated inside a single event rather than inferred across decades: the same objects, the same night, the same witnesses, rendered as dashes to the naked eye and as haloed circles through night optics, with the Bureau commissioning both pictures. That is the strongest evidence this prediction has. It is also, read honestly, the strongest evidence for the cheaper rival stated below, because a demonstration that the instrument shapes the report raises the evidence for the mechanism while lowering the need for the hypothesis. The 1948 Air Materiel Command file cuts the other way: its four-class taxonomy of discs, cigars, spheres and “balls of light” was compiled from unaided human eyes, and three of those four classes are still in the modern sensor record.
The D25 diamond was detected on shortwave infrared only: invisible to the eye and to standard IR on the same aircraft. The cold-IR objects read cooler than ambient air while flying. The East China Sea object registered as a lens-flare-like signature across two different sensors. The record contains no object that was ever pinned down consistently across bands.
August 2026 update. Four distinct thermal behaviours now sit in one tranche, two of them in the same region within days: cold in both white-hot and black-hot polarity over the Gulf of Oman; black-hot in the western United States; thermally hot in that same western sector the following night; and, in the same 2026 encounter, objects bright to the eye and to night optics that could not be acquired on thermal at all. Either the phenomena are heterogeneous, or thermal signature is not a stable property of them. Both readings support this prediction.
AARO’s own analysis of its best-documented modern case leaves roughly forty percent of the reported phenomena unresolved. The IC’s Cheyenne Mountain assessment reached only “possible backscatter.” But “mixed” is the honest grade: the record also contains cases the system did resolve toward the mundane, which a strong Mirror reading would not expect the phenomenon to permit.
August 2026 update: this prediction got weaker. The tranche added three clean resolutions. A 1963 Brazilian crash-and-body claim ran from a radio broadcast to a White House space-policy office to a diplomatic cable to a consul’s finding that the story was “apparently fabricated in Rio,” in twelve days. The 1947 War Department review of the Scandinavian ghost rockets concluded that most reports were meteors or fireworks, with at most a handful of real V-1 tests. And the 1948 Air Materiel Command study resolved much of its own 180-incident sample to balloons and to Venus. The weight of cases the system did close has roughly doubled.
Sensor rendering explains much of the drift for free. Any distant point-source on infrared renders as a round blob. An era whose primary sensor is FLIR will produce “orbs” from a heterogeneous population of objects without any co-variance at all. This rival explains the 2020s orb dominance more cheaply than the Mirror does.
But note what the rival cannot explain: the same decade’s civilian eyewitnesses, filming on phones in the visual band, also report orbs, with FBI-assessed credibility. Two independent channels converging on the same form weakens the pure-artifact reading; it does not rescue the Mirror specifically.
The unfalsifiability trap. A theory in which the phenomenon adapts to observation can absorb any observation. Stated carelessly, the Mirror is unkillable, which makes it useless. Hence the falsifier below.
A stable, repeatable, multi-band characterization of a single object (visual + IR + SWIR + radar agreeing on form, size, and emission), or a future tranche showing shape distributions that fail to track newly fielded sensor modalities. The D25 case is already uncomfortable for the Mirror: an object visible only to the newest band looks less like a reflection of the observer and more like a physical emission profile.
The core phenomenon behaves like distributed, autonomous, observation-oriented reconnaissance of strategic infrastructure: patient, repetitive, indifferent to being seen, and resistant to being measured.
The 1949–50 green fireballs over the Sandia–Los Alamos complex. A modern unidentified-object incident report from the Pantex nuclear weapons plant. The 2022 Cheyenne Mountain object. A September 2023 sighting at a restricted test facility. The Western U.S. event beside a sensitive national-security site. And the check that matters: CIA reporting places the same phenomenon over the Soviet Union’s own weapons-testing range, which no theory of purely American secret technology survives. Tranche 4 hardens both ends of this line: the verbatim 1949 Los Alamos conference transcript and the full 2015 Pantex incident report are now primary sources rather than summaries.
August 2026 update. Two independent witnesses reported triangles over Colorado Springs within the same month of 2023, which places them over NORAD, U.S. Space Command and Cheyenne Mountain, one year after the Cheyenne Mountain object already in this corpus. A 2002 account extends the pattern to Bagram Air Base, a deployed main operating base rather than a domestic nuclear site. And the Gulf of Oman encounter happened during a live-fire exercise, with the orbs first found hovering directly over the flare the gunship had dropped as its own mock target.
Federal agents watched orange orbs release smaller red orbs on two separate days. A gunship crew recorded a possible sub-object detachment over the East China Sea. Civilian witnesses filmed two orbs moving “as though tethered” that appeared to merge. Unrelated observers, unrelated sensors, same grammar.
August 2026 update, with a distinction worth keeping. An AC-130J crew logged pairs and threes manoeuvring in coordinated undulating formation, and the source’s own comment on the pair reads “they knew they were flying together.” In 2026 a witness watched six to ten lights appear overhead, apparently “sync up” with a light already being tracked, and depart east-southeast together. Both are new channels. But this is coordination and collective departure, not strictly parent-and-child dispersal, and the two should not be blurred: the coordination limb of this prediction strengthened more than the dispersal limb.
A senior intelligence officer’s helicopter was approached within an estimated ten feet; the object then split in two. Objects hold co-altitude with aircraft for minutes, then leave. The same record shows disappearance at the moment of tracking, single-band visibility, and not one recovered artifact. Things that hide do not approach helicopters. Things that pose do not stay unmeasurable for eighty years.
August 2026 update, and a refinement. Two cold orbs held station over an aircraft’s flare for about fifteen seconds, then left during the mechanical delay between the trigger pull and the cannon’s recoil, before the round, the muzzle flash or the noise. Taken at face value that is a response to the act of firing rather than to being observed, which sharpens this prediction into something more discriminating: indifferent to being seen, responsive to being acted upon. It also cannot carry much weight alone, being one crew’s account of sub-second timing, and human timing judgment at that scale is poor. The complementary half arrived the same day: 2026 objects that approached to within 400 metres, one apparently six feet off the ground, while remaining unacquirable on thermal.
The New Mexico nuclear complex appears in 1949–50 and again in the modern DOE record. The Northeastern orb cases recur across 2021–2026 within a 25-mile radius. But the corpus is a curated sample of what agencies chose to release; revisit-rates computed from it inherit that curation. Graded mixed until a fuller release allows real base rates.
August 2026 update. The instance count rose without the objection moving: Cheyenne Mountain in 2022 followed by two Colorado Springs sightings in 2023, and the western U.S. site recurring from 2023 into 2026 with the same phenomenology and new witnesses. The grade stays mixed, because none of that is a base rate.
The selection effect is real and large. Sensitive sites are exactly where trained observers, high-end sensors, and mandatory reporting live. A phenomenon distributed uniformly over the Earth would still appear to cluster on the nuclear complex in a corpus assembled from military and FBI channels. The corpus cannot compute a base rate, and honesty requires saying the strongest Survey evidence sits on top of the strongest sampling bias.
The system sometimes resolves things. D7’s weapons-quality track ended in “looks like a balloon.” The ICA tried “backscatter” for Cheyenne Mountain. A survey hypothesis built only on the residue must acknowledge how much of the record is not residue.
No artifact, ever. Eighty years of alleged reconnaissance and the release contains no material, no isotope ratio, no wreckage claim beyond testimony. Absence of evidence is weak evidence of absence, but it is not nothing.
A future tranche with honest geographic base rates showing no site affinity once reporting density is controlled; or demonstration that the parent-child cases reduce to sensor artifacts and misidentified munitions; or a resolved mundane explanation for the Western U.S. event, which is currently the hypothesis’s load-bearing case.
A meaningful share of the residue is a natural, poorly characterised class of luminous atmospheric or electrical phenomenon — transient, self-organising, and without a solid body to recover, resolve, or crash.
Intellectual lineage: the long, unglamorous history of phenomena that science declined to accept until it could produce them. Ball lightning was dismissed for more than a century and is still not reliably modelled. Sprites and blue jets were folklore among aircrews until 1989. Nothing in this hypothesis requires anything new; it requires only that the atmosphere still holds behaviours nobody has funded the instruments to catch.
This is the prediction the August 2026 release turned from plausible to strong. The Air Materiel Command’s 1948 taxonomy sorted roughly 180 investigated incidents into four classes, and the fourth is not a shape at all: “balls of light with no apparent form attached.” That category was drawn from unaided human eyes, before radar tracking, before infrared, before every sensing technology later blamed for manufacturing orbs. The same form runs through the wartime night-phenomena file, the 1949 green fireballs over the New Mexico complex, and the modern infrared record. The commonest thing in eighty years of reporting is a light with no structure, and the government said so before it had the instruments that supposedly cause it.
A solid object of a given temperature behaves predictably across infrared. These do not. In 2026, red lights bright to the naked eye and bright through night optical devices could not be acquired on thermal at all, while an object seen in the same sector the previous night read thermally hot. Over the Gulf of Oman in 2021 the orbs read cold in both white-hot and black-hot polarity, which rules out a display artefact and describes something luminous with almost no bulk thermal mass. The Greek diamond was detected on shortwave infrared only, invisible to the eye and to standard IR on the same aircraft. For a hull this is baffling. For an ionised gas with a structured emission profile it is close to expected.
The founding case fits almost too well: the 1948–50 green fireballs concentrated over the New Mexico nuclear complex, at that moment the most radiologically and electromagnetically exotic ground on Earth. Pantex, Sandia, Los Alamos and the ICBM fields are all high-energy environments before they are strategic targets. But the confound is fatal to a clean reading, because those sites are both at once, and this hypothesis cannot separate them any better than the Survey can. Graded mixed, and it will stay mixed until someone reports the same phenomena over a high-voltage installation of no military value whatsoever.
This is the constraint every other hypothesis has to apologise for. Eighty years, thousands of reports, some at extremely close range, and the release contains no material, no isotope ratio, no fragment. The Swedes searched for ghost-rocket debris in 1946 and recovered nothing but material described as “nonmetallic slag.” AMC recorded in 1948 that “so far, no physical evidence of the existence of the unidentified sightings has been obtained.” For a craft hypothesis that record is a growing problem. For a discharge hypothesis it is simply what the phenomenon is: there was never anything solid to find.
It cannot explain structure. This is the hole, and it is large. Plasmas do not have recessed corner lights. Two unconnected witnesses in Colorado Springs described a solid triangle with a defined edge, one of them reporting a dark seam running around it. A Navy radar held a delta-shaped contact in excess of 3,800 knots for twenty-one minutes while an F-8 pilot watched it out-climb him. No reading of atmospheric electricity accommodates that, and pretending otherwise would be exactly the kind of stretching this page exists to refuse.
It struggles with apparent coordination. Six to ten lights appearing overhead and departing together on a shared heading; pairs manoeuvring around and under one another in what an experienced aircrew described as a deliberate formation. Self-organising plasmas can do surprising things, but “they knew they were flying together” is not a description of weather.
It is unfalsifiable if stated loosely. “Some unknown atmospheric process” can absorb any observation, which makes it useless. It earns its place here only because it makes the specific spectral predictions above, and those can fail.
One recovered artifact of any kind. A stable, structured object characterised consistently across visual, infrared, shortwave and radar in a single event. Or, most cheaply of all, a spectrum: emission lines would strongly support this reading, and a reflected or blackbody continuum would gut it. The Naval Photographic Interpretation Center asked for spectrophotometric analysis of the Utah film in 1953 and noted a preliminary test plan was underway. The released record does not show it was ever completed. Seventy-three years later that measurement still has not been published, and it is the single cheapest experiment in this entire field.
A meaningful share of the residue is human technology — ours or an adversary’s — observed by people not cleared to know it exists.
Intellectual lineage: the government’s own repeated instinct. This is the first explanation the apparatus reaches for every single time, and the corpus documents it reaching, from the Horten brothers in 1948 to a Black Hawk filmed as a scale reference in 2026.
The corpus documents the “is it ours?” test being run across seventy-eight years. In April 1948, mid-investigation, the Air Force arranged through the British Joint Services Mission to interview Dr Reimar Horten, designer of the German flying wing. In February 1965 the CIA office that ran the A-12 OXCART obtained the Navy’s radar film for the Puerto Rico contact, because the commanding officer at Ramey had reported it as a YF-12A. And the Agency’s own overhead-reconnaissance history concedes that U-2 and OXCART flights account for a documented share of Cold War sightings — a confirmed instance of this hypothesis being correct. In 2026 an FBI agent filmed a passing Army helicopter specifically to build a comparison reference. The reflex is continuous and it is written down.
The 2026 western material reads, in places, like a swarm profile: six to ten lights appearing overhead, appearing to “sync up” with a light already being tracked, departing together on a shared heading, one descending an estimated thousand feet and vanishing, another apparently six feet off the ground. And the detail that makes this hypothesis hard to dismiss is the surrounding apparatus: an agent was cued to look by a radio-frequency detection at 10,000 feet and directed to a specific bearing. Somebody at that site had already installed equipment to listen for something in the air. That is what you build when you are worried about drone incursion.
This is where the hypothesis breaks, and it breaks hard. The corpus now begins in 1946, with nearly a thousand Scandinavian reports of objects performing manoeuvres no 1946 airframe could execute, and a War Department assessment that found the Soviet-missile explanation unsupported by any evidence. By 1948 the Air Materiel Command had studied 180 incidents and concluded that the discs, cigars and balls of light were “not of domestic origin,” while its engineers judged that such configurations would have poor climb, altitude and range with the power plants then available. The most authoritative attribution test in the record was run by the people who would have known, and it came back negative.
The other side of the Cold War reported it too. A CIA information report places comparable phenomena over the Sary Shagan weapons testing range inside the USSR. No theory of American secret aircraft survives that, and a theory of Soviet aircraft has to explain why they were flying them over their own closed range for the benefit of nobody.
Eighty years is too long. Programs get cancelled, declassified, or leaked. The U-2 became public. OXCART became public. A continuous eight-decade black program spanning multiple nations, producing no defectors and no procurement trail in a released corpus that includes the Horten request and the OXCART memo, is a heavier lift than the thing it was invented to avoid.
It explains the modern residue better than the historical one. Which may simply mean it is true of part of the pile and false of the rest — a possibility this page takes seriously below.
Declassification of a program that accounts for a named case would confirm it, case by case, and the corpus already contains one such confirmation in the U-2 history. What would kill it as a general explanation is what the record has arguably supplied already: well-documented pre-1950 cases with performance no contemporary airframe could deliver, and comparable reporting from inside a closed adversary range.
It would be evasive to run four hypotheses and leave this one unnamed. So: stated plainly, weighed against the same record, and ranked where the evidence puts it, which is last.
Some part of the residue is the activity of an intelligence that did not originate on Earth.
What keeps it alive. One thing, and it is not the one usually cited. A CIA information report places comparable phenomena over the Sary Shagan weapons testing range inside the Soviet Union. If the same thing appears over Los Alamos and over a closed Soviet range, then every theory of American secret technology fails at once, and you are left with either a natural phenomenon that favours high-energy sites or something with a genuinely global and indifferent distribution. That is a real constraint and it does real work.
What holds it down. The absence of a single artifact in eighty years. Thousands of reports, hundreds at close range, one light apparently six feet off the ground beside two federal witnesses in 2026, and the release contains no material, no isotope ratio, no fragment. The best statement of the problem is the government’s own, written in 1948 by people actively entertaining the possibility: the nature of the objects “cannot be established until physical evidence, such as that which would result from a crash, has been obtained.” Seventy-eight years later that sentence is still the state of play.
And the same document did the work honestly. The Air Materiel Command considered the extraterrestrial hypothesis by name, plotted incident frequency against the approach curves of Mercury, Venus and Mars looking for a signal, found a periodic variation it flagged as possibly coincidental, and concluded that tangible evidence was “completely lacking.” That was the assessment of an organisation with every incentive to find something. Nothing released since has changed it.
Where it sits. Last of the five, and not close. Not because it is absurd, and not because of any prior about what is possible in the universe, but because it is the only reading on this page that predicts a physical trace and has had eighty years to produce one. It stays on the list because the Sary Shagan constraint is real, and because a page that quietly omitted it would be managing its readers in exactly the way this site criticises the apparatus for.
This section is speculation, and it is flagged as such. It is not a hypothesis with predictions; it is a doubt about the shape of all four above.
The Mirror, the Survey, the Discharge and the Program disagree about almost everything. They agree, without ever saying so, that the referent is a thing: something with a boundary, a position, and a velocity, which is either being observed, or observing, or discharging, or flying. Every investigation in the corpus makes the same assumption. Project Sign looked for aircraft. The Robertson Panel looked for aircraft. AARO looks for objects. The 1948 taxonomy sorts by “shape or physical configuration,” and the one category that resists it — balls of light with no apparent form attached — is precisely the category that never leaves the record.
Now consider the observational signature that eighty years of investigation actually produced. Formless luminosity as the invariant. Emission that varies inconsistently across sensor bands. No physical trace, ever. Resistance to stable characterisation that survives every improvement in instrumentation. Occasional behaviour that reads as responsive, always at the exact edge of measurement. Clustering where energy and observers both concentrate.
That is a poor description of objects. It is a better description of events: processes, transients, things that happen rather than things that are. You cannot recover wreckage from an event. You cannot get a consistent multi-band signature from something whose emission depends on its state. You cannot photograph the edges of something that has none, and you will keep producing kinematic figures that are hostage to an assumed distance, because there is no rigid body whose size would let you fix the range.
If some real fraction of the residue is a class of transient event rather than a fleet of craft, then the entire eighty-year history of failure stops being suspicious and becomes predictable. The investigations did not fail because they were sabotaged or underfunded, though both were sometimes true. They failed because they were built to catch the wrong kind of thing, and they have been rebuilt to catch the wrong kind of thing roughly once a decade since 1948.
This does not resolve anything. It is offered as a possibility that the corpus permits and no framework in it has tested, because no framework in it was designed to.
None of the four hypotheses needs a conspiracy, and the documents do not show one. What they show is an apparatus repeatedly choosing perception-management over resolution. The earliest instance is now the plainest. In November 1948, having studied roughly 180 incidents and found a residue it could not explain, the Air Materiel Command wrote that it was “not considered advisable to present to the press information on those objects which we cannot yet identify,” and that if the press insisted, it should be told that many objects had been identified as weather balloons or astral bodies. A routing note the same month recommended that no statement be released until press demands made it imperative. That is the policy, in the government’s own hand, four years before the earliest item previously cited here. And it is not only an American reflex: released in the same tranche, a January 1947 War Department review records the Swedish Defence Staff privately telling the U.S. Military Attaché it believed Soviet missiles were overflying Sweden while its public communiqués shifted to “celestial origin,” a reversal the review attributes to a fear of repercussions on Swedish–Soviet relations. Two governments, twenty-two months apart, converging on the same posture. The rest of the chain follows: the 1953 Robertson Panel recommending debunking and monitoring of civilian groups; the FBI filing UFO organizations under internal security for fifteen years; a 1963 NASA-to-State memo titled “Thoughts on the Space Alien Race Question” confirming the policy question was live at the highest levels; a 1998 senator’s constituent inquiry bouncing off the White House; and a committee of French defense officials concluding in 1999 that the U.S. was running systematic disinformation on the subject. This chain is consistent with all four hypotheses and proves none of them. It is documented in full on The System.
The four are not mutually exclusive, and the most likely answer is that the pile is heterogeneous and more than one of them is true of different parts of it. But they make different bets, and future releases can score them.
| Observation | Mirror | Survey | Discharge | Program |
|---|---|---|---|---|
| A stable multi-band characterization finally lands | Strongly against | Compatible | Strongly against | Predicted |
| One object characterized differently by two simultaneous sensors | Predicted | Neutral | Predicted | Against |
| A physical artifact is recovered, of any kind | Against | Compatible | Fatal | Predicted |
| Spectroscopy shows emission lines rather than reflected light | Neutral | Neutral | Strongly predicted | Fatal |
| Honest base rates confirm site affinity after controlling for sensors | Unexplained | Predicted | Predicted (energy, not strategy) | Predicted (test ranges) |
| A black program is declassified that accounts for named cases | Neutral | Weakened | Weakened | Confirmed, case by case |
| Well-documented pre-1950 cases with impossible performance hold up | Neutral | Predicted | Compatible | Fatal |
| The residue shrinks to zero under better data | Dies | Dies | Dies | Dies |
The last row is still the point: every hypothesis here is hostage to the residue, and all four should be. Two rows are worth more than the rest. An artifact would kill the Discharge outright and hand the field to the Program or to something stranger. A spectrum would do the opposite, and it is the cheapest measurement in this entire subject — the Navy asked for one in 1953 and the released record does not show it was ever completed. Seventy-three years of argument have run on data nobody paid a few thousand dollars to disambiguate.
Where the evidence actually points, in order. No numbers, because the corpus cannot support them and a decimal point would only invite an argument about the decimal.
First, the Discharge. It is the only reading whose central prediction — formless luminosity as the invariant form — was written down by the Air Force in 1948, from unaided human eyes, before every sensor later blamed for producing it. It is the only one for which eighty years without a single artifact is a prediction rather than an embarrassment. And it is the only one that makes sense of objects bright to the eye and to night optics while absent on thermal. Its hole is real and large: it cannot explain a solid triangle with recessed corner lights, or a radar track at 3,800 knots.
Second, the Program, for a substantial slice of the modern record and almost none of the historical one. The 2026 western material reads like a swarm profile watched by people who were not read in, at a site that had already installed equipment to listen for incursions. It dies on 1946, and it dies on the Soviet range.
Third, the Survey. It gained on every prediction in the August 2026 release and remains the best account of the behaviour that reads as purposeful. It still sits on top of the largest sampling bias in the corpus, and it still has no artifact.
Fourth, the Mirror. The strangest position of the four: its mechanism has never been better evidenced, and the same evidence has never done less for it. The 2026 renderings demonstrate sensor-dependent form inside a single event, which is exactly what it predicts — and exactly what the cheap rival predicts too, at a fraction of the theoretical cost. The tranche also closed three more cases outright, which a phenomenon that keeps resolution out of reach should not permit.
Last, non-human intelligence, for the reasons set out above, and not by much of a margin over nothing at all.
But the honest reading is not a ranking. It is that the pile is heterogeneous, and that this alone explains the history. Any single-explanation programme run against a mixed pile will resolve most cases, leave a stubborn remainder, and be accused of credulity or cover-up depending on who is looking. That is precisely what happened in 1948, 1953, 1969 and 2026. The failure is predicted by the mixture, without needing conspiracy or exotica — and the corpus records competent people noticing loose threads and never pulling them, over and over, for eighty years. The Navy asked for a spectrum in 1953 and did not get it. The Air Force planned a second interview with Langmuir about the balls of light and the record does not show it happening. AARO calls forty percent unresolved after first-stage analysis, which is a statement about how far the work got, not about what the phenomena are.
The best-evidenced fact in this entire collection is not about the sky. It is that a government spent eighty years unable to explain a small number of things, unwilling to say so plainly, and never funding the work that would settle it.
For the documents behind every claim above, open the cited files in place, or work through the full Corpus. For the strongest cases these hypotheses lean on, see The Standout Cases; for the ones we set aside, The Contested Cases; for how established UAP thinkers fare against the same record, The Theorists.