How does many-worlds theory explain Bell's inequality? How are the correlations in the experiment proposed by John Bell explained by many-worlds theory?
I understand that the assumption of a measurement having a single outcome, is not supported by many-worlds. And hence it may violate Bell's inequality.
But it still has to explain why correlations occur when the experiment is performed.
It would be helpful if the answer is in the context of this experiment setup, which has 1 particle source, 2 detectors, 3 switches and 2 possible outcomes for every measurement. How does many worlds explain (in the mentioned experiment) that when switches are in same position, the measured outcome is equal, but when switches are different then the outcome is random.
 A: Bell's inequality occurs in the statistical spread of multiple possibilities. Its implication for the many-worlds theory is therefore in the statistical spread of the (events which spawned their) new universes.
Thus, each world retains local realism and the nonlocal aspect of quantum theory manifests only in the spawning pattern of the new universes.
Bell's theorem shows that the statistical distribution of quantum measurements will differ, depending on whether reality is local or nonlocal. For example if you have pairs of entangled photons and measure their polarisations under certain conditions, the statistical distribution of polarizations will differ depending on the nature of reality. We all know that reality has been measured in this way and has shown itself to be nonlocal.
According to the many worlds interpretation, on each measurement the world divides. By the end of the experiment you have followed a long chain of world-divisions. The statistical nature of that chain is just the nonlocal distribution which we all know and love. But in each individual world, every photon and every polarization is locally real. Nonlocality influences only the statistics of world-division, not anything else.
