I have the SVA1032x (not hacked, not molested in any way) and I can tell you that what I saw during your calibration was an internittent contact. During the calibration, attaching the short and load will change the trace (even though it down not necessarily stay on the chart until cal is finished). You saw it yourself, an unstable reading while attaching the short (which became stable when you played with it), and an unstable reading while attaching the load (which you totally missed).
Furthermore, Telonic is not the originating company, they are just a certified dealer. There is no way to know if your "free" adapters were decent or if they are cheapest ones China can offer (leaning towards the latter).
The male's pin can be straight as a laser beam, in fact, that makes it worse, because when the female's hole stretches open, the male's pin barely touches. When you hit enter with that load attached, it wasn't touching, and that was clearly evident in the video. So, why does it jump around so much, easy.... You told it that an open circuit was 50Ω, and therefore, it put that open at the center. Now, as any VNA user can attest, these things rapidly lose accuracy as the impedance gets higher than about 1kΩ, so as it tries to extrapolate the impedances between the open in the middle and the open on the chart, that entire range is in the VNA's unstable territory. Just entertain me and try a different adapter, what do you have to lose? You can drop $3400 on a piece of test gear but want to avoid spending $20 on a decent adapter?
Good luck with that!