I see two main problems with a trajectory like the one you point:
1) you use the first 20s, but the last 14s have closer line of sights.
The first 20 seconds are only used to pick the true airspeed that matches XXL's stitched image. The lines I plotted in the above are taken at 1, 11, 21, and 31 seconds. The picture looks comparable if you use 4, 14, 24, and 34.
For the plane to maintain a steady trajectory (i.e. not strongly decelerate), it has to make a sharp turn. Because for a constant speed, the segments between each line of sight have to be of same length.
This is not quite right because the F-18 is turning, and in fact its turn rate changes quite a bit as it turns. In fact, I noticed something interesting: here's what the trajectories look like if you plot only the last 4 seconds, that is, right after the object crosses over in front of the F-18's nose:
Here's the same thing zoomed in near the F-18:
... These lines never touch. There's no intersection point. More properly, the intersection point is behind the F-18, which is inly possible if the camera is moving with the object. This means that as long as the modeled trajectory here is close to right, we can conclude the gimbal object is moving right to left from the F-18's perspective. With a straight-line trajectory, it has to be beyond the mess of intersections near the convergence point. This rules out anything closer than 30 Nm or so, minimum. In my analysis the lines only look plausible for a straight line trajectory if the object is further still. Here's one of the closest ones that could fit, at 60 Nm:
Same caveats as before apply: this is the best reconstruction I was able to make but there's a lot we don't know as far as wind shear and the real TAS, which I'm fixing to match XXL's reconstruction (which has its own set of assumptions), I had to use smoothed frame-by-frame azimuths in lieu of the unavailable raw data, so there's a large but difficult to quantify* set of possible trajectories that can be consistent here. Some new ones: the times are now 1, 12, 13, 34 (I wanted them to cover more of the video), and I made some accuracy improvements (that don't change much in the final product, but are worth noting). That said, the 10 Nm figure seems now very difficult to make consistent with any scenario.
* Difficult to display, too: using little fans instead of straight lines as ersatz error bars it might give the impression that the errors are uncorrelated.
But why would the cloud movement stop then? If the plane maintains its speed, we should still see cloud motion as the camera tracks it.
It never really stops, it merely slows because towards the end of the turn the F-18 and the object are moving roughly in the same direction. The fact that the lines after crossing the nose (when the sight lines are tangent to the path) diverge is consistent with the little remaining cloud movement here.
2) To me the glare hypothesis is valid if we see the jet exhaust from behind (ex : Chilean case glare). In your trajectory we see it from the side all along. Why we don't see any jet signature, and why would the size of the glare increase ?
That trajectory was just an example of how it didn't have to be that fast to be at some extreme distance, even if we keep the intersection point where it was. I agree that a trajectory as you described seems more likely and seems more consistent with the total of the evidence. But a jet engine is not the only thing that could create a glare like that: reflected sunlight is also a possibility, in which case it could be reflecting off the side of a fuselage, or something, it's just the jet engine seems a better match based on prior experience.
Like I said you can find a midrange point where the plane at normal speed has time to turn in a reasonable way. But midrange distance means it's less likely to not see a jet signature, especially when seen from the side (because it has to turn).
Again, there's no reason to believe it has to turn. "Midrange" here is some good 50 Nm or so. A Super Hornet, seem from the distance I indicated above (60 Nm), would have a wingspan spanning 0.007 degrees, or 1/50th of the ATFLIR 0.35° x 0.35° field of view.
About this big.
The trajectory of this reconstruction is at odd with the cloud movements, in the DCS reconstruction background clouds would go from left to right (i.e., they would open to the right, if they were in the simulation), but they go from right to left (open to the left).
Fair enough, that trajectory also disagrees with this new analysis because the sight lines would always cross even towards the end. Still you can easily mirror his trajectory across the intersection point and obtain a similar match, but with the right cloud movement and correct sight lines post-nose.
If they say it was in the 10NM range, stationary (at least at some point, or going against the wind), and it turns out it's a distant plane, that's not an accurate report to say the least.
bogus adjective
: not genuine : COUNTERFEIT, SHAM
a bogus claim
The evidence was completely bogus.
I'm assuming here that the operators described what they saw to the best of their ability. I don't think making a perfectly honest and understandable observational mistake merits the word "bogus", especially with something so weird that we're still talking about after several years and doing detailed quantitative analyses over a period of several months. They had seconds to decide what to do, and hours to think about it before writing a report.
The glare on that one is a different animal, with the clear signature of two exhausts.
Well, four exhausts. It's a four-engine plane, though it doesn't necessarily look like it because the glares sort of blur together. It's possible this is what's happening here, especially if we're dealing with a plane with engines mounted nearer the centerline, like a business jet or a fighter.
If doubt they brainstormed all day for two years on this one, it's probably just the time it took to get the report out.
But that's the thing, people are busy. I wouldn't be surprised if we're doing much more analysis work on this (with far less data) than they ever did in the UAPTF.