Twelve videos of the tests, from both sides
A curated shelf of third-party footage taken from YouTube: flat-earth field tests and flat-earth readings of the sky, set next to the mainstream measurements and method critiques that answer them. Every title, channel and identifier on this page was checked against the video itself before it was listed.
Every clip on this page was made by somebody else. The archive did not film them, did not commission them, and does not vouch for their methods, their editing or their conclusions. A video is a record of what one camera pointed at, by one crew, on one day, and that is all it is — flattering to whichever side it happens to suit.
So we list them by argument rather than by verdict. Each card carries a tag saying which side it argues: a flat-earth test or flat-earth reading, a mainstream measurement, mainstream explainer or mainstream critique, or expedition footage shot by neither camp for an argument's sake. Nothing here is presented as proof of anything, because nothing here is a controlled measurement: the reader can check what was done, and the honest answer is that most of these runs are too short, too warm or too hazy to separate the two readings on their own. Where a test could not have come out against the person running it, that is worth knowing too.
Two further rules we apply to ourselves: we leave out anything that attacks, mocks or names a private individual, and we leave out anything that mixes the shape of the earth in with slurs or with hostility toward a group of people. Several well-known channels on both sides fail one or both of those tests, and are not here.
Playback is deliberate. Nothing is fetched from YouTube until you click a poster: no thumbnail, no player, no cookie. The posters below are drawn by this site and served from this site, not hotlinked from anywhere, and the player that appears on a click comes from youtube-nocookie.com. The privacy page sets that out in full.
The order is by subject, not by weight, and each group holds two videos that disagree with each other. We have not trimmed the claims: if a channel says its clip proves the earth is flat, that is what the card says it says. Reading the two sides of a group in sequence is the point of the page.
Laser and sightline tests over water or ice
A laser fired low across water, or a sightline held level across a lake, is the flat-earth test most often cited and the one most sensitive to the air just above the surface. Both runs below are shown as their channels present them.
Water level and lake surveying
Water is the flat-earth argument's favourite instrument, because a standing body of water is said to seek its own level. The pair below sets the water-level method beside a lake survey carried out against a published curvature figure.
The sun's apparent size and the sunset
If the sun is a small body circling above a plane, its apparent width must shrink as it moves away. The apparent width is easy to measure, and both videos below are about exactly that, from opposite directions.
Star trails and the two celestial poles
Long exposures show the sky turning about two centres in opposite senses, one in each hemisphere — the single hardest observation for a level plane to place. One video below treats it as the globe's evidence, the other explains it on a plane.
The 24-hour sun in Antarctica
The December 2024 expedition to the Antarctic interior was organised so that people on both sides of the argument could see the solstice sun for themselves. Both clips below are from Union Glacier that December; the first is from the expedition's own channel.
Mainstream critiques of flat-earth test methods
These two are not about the shape of the earth directly: they are about whether the tests above can bear the weight put on them, and they are the reason this page marks a test as a test rather than as a result.
The standard account does not rest on any of the clips above, and geodesy would not change if every one of them were deleted. The earth's figure is carried by measurements that never appear on YouTube at all: national levelling networks and triangulation, ellipsoidal datums every satellite fix is computed on, GNSS timing geometry, gravimetry, seismology, VLBI and laser ranging to reflectors on the Moon. Those lines of work agree with each other and put specific numbers on the curvature — a drop near 8 cm per kilometre squared, a dip of the horizon that grows with the square root of height, and a sun whose angular width stays at about half a degree all day.
On the flat-earth videos specifically, the mainstream reply is not that the observers are lying. It is that a laser grazed along a lake, or an object watched until its waterline disappears, is measured through the atmosphere, and the layer of air just above water bends light enough to cancel the effect being looked for. That objection is aired in full in the last group on this page, from the people who make it. Readers who want the site's own account of each test will find it on Try it yourself, and the objections at full strength on Objections answered.
How to read a test video
Most of the arguments on this page are really arguments about field method, so it pays to ask the same seven questions of every clip in the list, whichever side it is on.
- Is the camera at water level, and where exactly? A lens a metre above the surface and a lens at eye height are different experiments. If the height is not stated, the run cannot be compared with any prediction.
- Was the geometry written down in advance? A result that was predicted before the measurement is evidence; one that is fitted to the footage afterwards is not. Several of the better clips below show the arithmetic on screen first, and those are worth more than the more emphatic ones.
- Is there a witness, and is there a second camera? A single operator filming his own equipment is one unreliable element away from an unfalsifiable claim. Independent observers at both ends, or a second record of the same instant, cost almost nothing and settle most disputes about what happened.
- How rigidly was the laser or the level mounted? A tripod on a wooden dock, a wind-loaded mast, a tape measure under load, or a mount that is touched between readings will sweep the beam further than the curvature does over the same distance.
- What was the temperature gradient? Ask for air and water temperatures at both ends and for the time of day. A cool surface under warm air, or the reverse, bends a grazing sightline by more than the drop being tested, which is why the cautious runs happen early on still mornings and why a run that never records a gradient cannot settle anything.
- Is it a same-time comparison? Both ends of a sightline have to be recorded at the same moment, with the clocks shown, or a boat, a tide or a passing aircraft can account for what appears to move.
- Could the test have come out against the person running it, and do they say what would have counted? That is the question the whole page turns on. Credit where a channel states its own falsifying result in advance; treat a clip that cannot fail as a demonstration rather than a measurement.
Run those seven questions over the list above and it sorts itself. The measurement videos state their parameters; the argument videos, on both sides, mostly do not, and the two critiques in the last group are the reason to care about the difference. The archive's own tests, written so that a reader can fail them, are on Try it yourself.
This page presents the flat-earth interpretation as its subject matter. It is a belief-based reading of the observations, not a scientific finding — see About for how we frame claims and for the mainstream sources.