Dr.AddMedPHL said:

The dose-response is real but shallow at the top.

I will push back on the "any working dose is fine" framing. The maintenance evidence sits overwhelmingly at the top studied dose, and the extension data shows regain tracking dose reduction rather than tracking stopping. Holding low is reasonable; pretending it is evidentially equivalent is not.

The figures, for anyone assembling their own picture. Worth knowing that the injection site changes very little. Abdomen, thigh and upper arm are bioequivalent for semaglutide, so a site change is not a plausible explanation for a bad week.

LipidDoc_ATL said:

I will push back on the "any working dose is fine" framing.

Adding the part of the answer the thread has not reached. The useful move here is to separate what is established from what is widely repeated. Those two sets overlap less than the confident tone of most write-ups suggests, and the second set is where nearly all the disagreement on this board comes from.

A narrower follow-up, since the general answer is now clear:

How would you tell the difference between that and the alternative explanation?

Closing the loop on my own question.

Update. I did go to 2.4mg in the end, and the honest report is that it bought me less than the step before it and cost me two bad weeks. Worth knowing rather than worth repeating.