@Mendel your posts in this thread repeat the same pattern:
You dismiss the demonstrated facts as "circular" or "not established," and insist the distant-glare explanation remains intact despite direct control-test failures and geometry mismatches.
Here are the key demonstrated facts you keep sidestepping (from the refinements thread + Go-Fast control test + independent stitching).
Strongly NOTE. This post in no way shape or form is to be construed as Mick agreeing with ANYTHING I am saying. I am NOT attempting to paraphrase him, nor put words in his mouth. Mick did a number of tests, reproducing at least what can be called similar results, NOT conclusions.
- LBFs horizon formula: It is explicitly based on an assumption of "what a pilot would likely see" according to LBF. It is an imagined, idealized pilot-view model, not an empirical measurement of the actual sensor.
- What demonstrates the additional pod roll: When directly asked what in the actual Gimbal footage demonstrates that LBFs derotation mechanism is physically present and operating, "What in the footage demonstrates this". Mick told me to stop asking the question. No answer can be given because the entire correction is built on LogicBear's imagined "what a pilot would likely see" formula - not on any observable, measurable behaviour in the video, pod telemetry.
- Circular use of the formula: LBFs formula is used to "correct"/derotate the clouds and horizon to make the distant-glare model fit, and then the resulting "good fit" is cited as proof that the formula (and therefore the distant model) is correct. This is using the formula to validate the formula - classic circular reasoning.
- Go-Fast control test (same flight, same system): Using the frustum + natural bank method reproduces the observed curved water path and downward pan exactly. "The formula" produces an impossible straight-line ground track, Micks result found this also. This falsifies the assumption that the camera behaves the same way for a distant target.
- Metallic-sphere example (Bellingcat stitch): Shows both the downward pan and the non-linear ground path expected, curved ground path, as demonstrated in the Go-Fast example.
- Mick's own results show the same downward motion: Mick reproduced the prominent downward motion - the exact behavior that appears in both the Go-Fast control test and the Bellingcat metallic-sphere stitch (nearby target), my 3D recreation of the reaper being over taken. This downward pan is a key signature of closing range on a nearby object, not a distant one. Its presence in Mick's run further supports the close-trajectory geometry and directly contradicts the idea that the camera motion is fully explained by a distant glare under pure pod-roll.
- Cloud triangulation + lines of bearing: Do not resolve the clouds at the distant ranges required by Sitrec. Mick's Sitrec demonstrating a distant plane is possible, truncates the line-of-sight spread by ~20 % and fails to triangulate. When Mick switched to the real values (bank/ speed) he obtained results consistent with mine and there is only an impossible distant plane path.
- Rapid range closure I demonstrated from, Line of Sight intersect points ~12 NM down to ~8 NM, then steep decrease after the reversal point. This matches the pilots' reports of a close, manoeuvring object.
- Reversal-point pivot analysis: Using that point as the pivot shows the exact range closure the pilots described. The close trajectory is in the data.
- Decoupled rotation (first ~20–22 s): Mick himself called this a "puzzle." It is not explained by pod roll or the refined derotation model; the object rotates independently of the horizon/clouds/camera motion.
These are not "theories." They are reproducible tests run on the actual video, the actual flight parameters, and the same sensor as Go-Fast.
Repeating "refine the model" while ignoring the control-test failure and the geometry mismatch is not engagement - it's avoidance.
The data shows range closure, independent rotation, and camera behaviour consistent with a real (not glare) nearby target.
[edited for caveat, format.]