Stefan asked:
"Therefore my question: is your theory falsifiable..."
Of course it is. Just demonstrate that the code in my paper, either (1) does not in fact do what I claim it will do, when executed on your own computer, or (2), that it does in fact do what I claim, but not for any of the reasons that I have claimed. And consider this: I choose to attack this particular "quantum correlation" issue and provided my code, precisely to make it very simple (just download the code and run it) for everyone to "hit me with your best shot." So here we are, four years later, the code has now been viewed by thousands, rewritten and reproduced by a few, and yet no one has publicly claimed to have falsified it.
In this regard, note that I am claiming to have falsified Bell's theorem and related Bell tests, via (2) above and not by (1); "Quantum Correlations" really do happen; but they are being caused by an entirely different mechanism than that supposed by Bell and others. I am merely exploiting a "loophole", that they failed to close. And they failed to close it, precisely because they failed to realize that it even existed, which is exactly what Einstein et al predicted would happen with the EPR paradox. So there is no great surprise that this has eventually happened.
The issue has nothing to do with "undetectable physical mechanisms". It is simple, World War II era radar signal detection theory, combined with Shannon's World War II era Information Theory. Here is a simple, Korean "exclusion zone", detection problem, that may help to put this all in perspective:
Your mission (and it is not always impossible), should you decide to accept it, is to determine that, if an intruder ever appears in the exclusion zone separating North and South Korea, then correctly decide if they entered from North Korea (up) or from South Korea (down). You are allowed to adjust these two parameters as required; (1) the "bandwidth" of the exclusion zone - how wide it is, and (2) the time interval between successive observations looking for any possible intruder. Thus, for example, if you think the intruder will be on foot and running no faster than 30 km/h, then you might wish to make the width of the exclusion zone wide enough to ensure that such a runner could not possibly run from either edge of the zone and across a "decision point" in the middle of the zone, between the time-intervals between observations, that you have also chosen. On the other hand, if the intruder is in a jet fighter, screaming along at 2000 km/h, you might wish to construct a wider exclusion zone than would be necessary in the case of the runner, and employ shorter time-intervals between observations, to ensure that even such a fast intrusion would never prevent you from correctly deciding if the intruder is either an "up" intruder or a "down" intruder; they cannot cross the decision "threshold" before being observed.
The point is, your ability to correctly decide the "up" or "down" intrusion event, is highly dependent on the time-bandwidth product that has been built into the detection mechanism. That is what Shannon's Information Theory is all about. So what is going to happen, in the case in which you reduce both the time-interval and the bandwidth of the detector, to values so very small, that even a tiny amount of "noise" anywhere in the situation, makes it impossible, even in principle, to correctly decide the issue? You had better pray, very hard, that such non-zero amounts of noise can ever possibly exist, in the real world! Unfortunately, that prayer, the prayer of all physicists, does not appear to have been answered; "identical" particles (the "intruders") do not seem to exhibit identical "noise" - hence they appear to behave as if they are a bit "fuzzy" around the edges - just enough to make it impossible to ever correctly decide the detection problem, when the "exclusion zone" is effectively smaller than the "fuzz" around the edges. So even though you may frequently succeed at detecting the existence of any intruders within the "exclusion zone", you may nevertheless fail to correctly decide the "up" versus "down" problem in most intrusion events.
That is the situation being systematically constructed within my code; it is constructing "locally real" objects, that have just the right amount of "noise" to prevent correct "up" versus "down" bit-decisions, whenever the polarized objects being measured have their polarization axis misaligned with the axis of the detector. And that brings us back to the original subject matter of this discussion - is there a preferred reference frame for making such measurements? Yes there is - the polarization axis of the entity to be measured. Because the entire detection process has been so delicately balanced - at the very limit (Shannon's limit) of what can been done, that ever Bell test will collapse (produce too many bit-errors) when "lope-sided", off-axis measurements are made.
Rob McEachern