Bell's theorem - for or against Hidden Variables?

[FrediFizzx]

The point is not if the experiment is "unphysical", but that a theory which fulfills some properties (like EPR realism and Einstein causality) has to predict that the Ak(a)Bk(b) do not depend on the question if a measurement is actually done or not.  So we can evaluation them using an experiment where Ak(a)Bk(b) is measured, and extend the result to cases where it is not measured.

It is preposterous to propose a hypothesis that is impossible to test.  [...]  All the "Bell" experiments do is confirm that the predictions of QM are correct and nothing more.  They do not test locality or realism.