The Moon landing, and what the mirrors cannot
prove.

The boundary

Two questions wear the same name, and the run picked one.

“The Moon Landing Hoax” can mean an inquiry into whether the claims hold up, or an inquiry into who believes them and why. Those need different evidence, and the run recorded the ambiguity before choosing.

It went with the evidentiary reading: the claims and their support, tested against the physical, photographic and third-party record. It also recorded a second ambiguity — whether the question covers Apollo 11 alone or all six crewed landings — and assumed all six, because hoax arguments range across the programme.

That first choice matters more than it looks, and chapter six is where the bill arrives. Having scoped itself to evidence rather than to the belief, the run then failed to retrieve anything on most of the evidence that hoax arguments actually use. It went to laser ranging and orbital imagery, and to almost nothing else.

Out of scope before any source was read: uncrewed lunar missions unrelated to Apollo, conspiracy-theory psychology beyond this case, other space-programme controversies, and spacecraft engineering detail beyond what the specific claims require.

The run scoped itself to the evidence, then retrieved almost none of the evidence the argument is actually about.

Competing explanations

Three explanations, and one of them collected nothing at all.

Before any source was read the run wrote down three rival accounts and, for each, the observation that would destroy it. Fourteen passages went on to declare a bearing on all three.

H1 is the documented history. H2 is fabrication in whole or significant part, with the independent verification either misread or not actually independent. H3 is a claim about the claim: that the controversy is a traceable cultural narrative rather than a response to real gaps in the record.

H2's row has no positive bearing of any strength. Not one of the fourteen passages, read by a run that was actively searching for material against the official account, supports fabrication. Ten bear weakly against it and two strongly. That is about as one-sided as this archive gets.

H3 is the column to be careful with. Four weak positives, nine non-discriminating, one negative — and the run's own reading is that H3 is only half-testable here. Its evidentiary half is weakly supported, because the domain most often invoked in hoax argument turns out to be well-modelled and non-anomalous. Its causal half — that the phenomenon traces to a specific origin and evolved rhetorically from there — is untested, because nothing in this ledger is about the hoax literature at all.

Zero passages bear for fabrication. That is a fact about this ledger before it is a fact about the Moon.

The measurement still running

The strongest evidence here is an experiment you could rerun tonight.

Four observatories on Earth fire short laser pulses at five retroreflector sites on the Moon, and the pulses come back. Not an archive to be trusted — a measurement, still operating, whose success depends on hardware actually sitting at fixed lunar coordinates.

The precision is the part that does the work. Early ranges carried metre-level uncertainty; within a few months of 1969 the measurement went from hundreds of metres to metres to decimetres, reached 2 cm in 1995, and is under a centimetre today. Current analyses fit modern ranges with a 9 mm weighted rms residual — 2.3 parts in a hundred billion of the distance. Half a century of lunar geophysics, geodesy and gravitational physics has been built on top of it.

The arrays are also mutually distinguishable, which closes an obvious alternative. Apollo 11 and 14 each carry 100 circular corner cubes 3.8 cm across; Apollo 15 carries 300; Lunokhod 1 and 2 carry 14 larger triangular cubes whose exiting beam is hexagonal rather than circular. The Apollo returns cannot be Soviet hardware relabelled, and fabricated data would have to stay mutually consistent across four independent observatories for fifty years.

This is the argument that usually ends the conversation, and on its own terms it holds. The next chapter is about the terms.

Nine millimetres, from four observatories, for fifty years.

The step to the side

Laser ranging constrains where the hardware is, not who carried it.

This is the run's second finding, and it is the reason this dossier exists. Every laser-ranging passage that carries precision information is silent about how the arrays got there.

The crew attribution in the record is real, but look at where it lives. It is descriptive prose inside the same technical reviews — “placed on the Moon in 1969 by the Apollo 11 astronauts”, “Apollo astronauts and Russian rovers deployed” — not a property of the returns. Nothing about a 9 mm residual distinguishes an array a person set down from an array something else set down. So H2's robotic-delivery variant passes through the entire laser-ranging ledger untouched.

And the proof that robotic emplacement was possible at the time is sitting in the same dataset. Lunokhod 1 and 2 placed working retroreflectors without a crew, and their returns are being measured alongside the Apollo ones. The strongest evidence for the landings contains, as a component, a demonstration of the cheapest alternative to them.

What the ranging record does do is narrow the claim. It moves H2 from “the returns are faked or misattributed” to “the arrays were delivered without crew” — a much more specific and much more expensive thing to assert. The run also filed an attack on its own reasoning here: corner-cube arrays might be deliverable by hard-impact penetrators or tethered deployment from orbit, and that possibility fails only if optical modelling shows the arrays need soft-landing orientation and levelling no unguided system of that era could achieve.

The best evidence for the landings includes, as a component, a working demonstration of the cheapest alternative to them.

What discriminates

A Japanese orbiter photographed ground the engine had scoured.

The only evidence in this ledger that bears on a landing event rather than on hardware presence came from a counter-search aimed at the fabrication hypothesis — and it came from JAXA, not NASA.

Two things, from SELENE/Kaguya. A reflectivity halo brighter than the surrounding regolith was confirmed at the Apollo 15 site and attributed to the lunar module's descent engine exhaust plume. And a 3D terrain-camera reconstruction of the same site reproduces the mountain and hill profiles in a photograph taken from the surface by the crew, AS15-82-11122HR. A plume-scoured patch is a mechanical consequence of a powered descent at that exact spot; a surface-perspective photograph matching independently derived topography is evidence that a camera was standing there.

Together those refute “nothing landed there” and “the photographs were shot on Earth” — the two claims the laser ranging could not touch. They are also the only two passages in the entire run that earned the strongest available bearing against fabrication.

Then the limit, which the run states as plainly as the finding. The terrain camera resolves 10 metres. Rocks are explicitly not resolvable, and a lander, equipment or a person even less so. The halo constrains that an engine fired at that location; the topographic match constrains that a camera was at that location. Neither distinguishes a crewed landing from an uncrewed powered landing carrying a camera.

The chain below is H1's own required conditions — the winning hypothesis, audited. One is supported. One is the weak link. Three are simply not addressed by anything this run retrieved, including the lunar samples, which the ledger never touches at all.

The camera resolves ten metres. A person is not ten metres.

What was never reached

Nine of twelve subquestions were answered by no passage at all.

The run asked itself twelve questions before searching. Three got answers. The other nine are the ones the argument is usually about.

Unanswered by any passage in this ledger: where the hoax claims originated and who the earliest proponents were; how the claims evolved; the photographic and film anomalies — shadows, lighting, flag movement, missing stars, cross-hair alignment — and their explanations; the Van Allen belt and radiation dosimetry claims; the technical feasibility arguments about Apollo-era rocketry and life support; the documentary and testimonial evidence about keeping thousands of people quiet; where the two sides disagree most sharply; which hoax claims lack a rebuttal; and what proponents still cite as unresolved.

Part of that is retrieval and part is the gate. Seven adversarial searches ran and two returned. Both Soviet tracking searches came back empty, and so did the search for peer-reviewed literature identifying genuine unresolved Apollo anomalies, and the search for technical objections predating the 1976 book. Then, of the 23 readable passages the run did collect, nine were refused by the verbatim gate and appear nowhere — a 39% refusal rate, the highest in this archive.

The consequence is specific rather than rhetorical. An article on the origins and persistence of the hoax narrative is sitting in this run's source list, snapshotted and assessed, and no passage was ever read out of it. That is why the run reports H3's causal half as wholly untested while the source that would test it is printed in the endmatter.

The source that would answer where the claim came from is in the ledger. Nothing was read out of it.

The best-supported reading

The strongest finding, printed with the attack made on it.

The run's leading finding is the laser-ranging record itself, typed observed. Then the run named the way it could be wrong, and the objection goes at the word “independent” rather than at the physics.

The attack is filed as measurement error. Ranging fits run on standard geophysical and orbital models that carry assumed coordinates and target properties. If the processing software across those four observatories shares common ephemeris models or systematic calibration baselines, then a 9 mm residual might be measuring internal model consistency rather than four independent confirmations of where the reflectors are.

The failure condition was written down in advance and is a specific piece of work someone could do: independent re-analysis of the raw timing data using unconstrained orbital parameter estimators, failing to recover distinct point-source reflections at the published coordinates within the stated error bounds. Nothing in this ledger runs that re-analysis.

Seven attacks are on the record against eight asserted findings, including one arguing the Apollo 15 halo could be a natural albedo anomaly from micro-meteoroid impact or dust redistribution, and one arguing that reading H3 through laser ranging is selection bias, since precision geodesy is the domain least likely to show the macroscopic anomalies skeptics point at. This run does not record whether any person reviewed any stage of it, and the page says so.

The score

Eight findings asserted, every one unknown, and 14 out of 100.

Eight findings were asserted, three of them typed observed rather than inferred. Every one computed to a confidence of unknown, and the evidence score came to the same 14 out of 100 as every publication before it — for the same reason.

Zero of twelve snapshots reached an upstream origin, so corroboration and provenance both scored zero. One measurement did improve on the archive: one source of twelve was assessed as primary data — the NASA ranging dataset itself, the first primary-data source in this series. It does not move the score, because provenance and primary-data classification are separate measurements and neither repairs the other.

Falsification scored 16 of 20 for two of seven counter-searches returning, and both of those returns did real work: one produced the only evidence here that bears on a landing rather than on hardware. Saturation scored zero — round one found three source families, round two found eight and all eight were new, and the counts on all three explanations moved when it was added.

The measurement that most limits this page is the verbatim gate. Of 23 readable passages, nine were refused; the entire dossier rests on the fourteen that survived, and those came from just four of the twelve stored sources. Eight snapshots produced nothing at all.