The violation of Bell's inequality can be understood as a conflict between the two principles we need to prove Bell's inequality — on one hand, realism in the sense of EPR and Bell, on the other hand, Einstein causality and relativistic symmetry in general.

As an introduction into the problem posed by the violation of Bell's inequality, we suggest to consider a simple game. As well, you can consider the formal proof.

The position of the mainstream of modern physics is the rejection of realism in favour of relativity. The aim of these pages is to argue against this choice. There is no good reason to give up realism. We consider refutations for some popular defenses of relativity – one quite general idea of refutation of various arguments which I have named the FTL phone argument, and a rejection of the argument that violations of Bell's inequality cannot be used to transfer information.

But, given the seriousness of the problem, it seems reasonable to consider it from a more general point of view. Thus, we consider some metaprinciples or criteria for choice between physical principles and apply them to this particular conflict:

- The criterion of minimal loss of restrictive power seems the best equivalent to Popper's criterion of empirical content for principles. It's application refutes realistic Einstein causality and a realistic version of relativistic symmetry. The only principle which survives is manifest relativistic symmetry.
- The criterion of generality obviously prefers realism.
- The criterion of incompatibility with other principles gives a long list of principles incompatible with relativistic symmetry. The only popular candidate for a conflict with realism – quantum theory – is compatible with realism, as proven by the pilot wave interpretation.
- Even the criterion of existence of compatible theories, which clearly prefers the mainstream principle, does not give any advantage for manifest relativistic symmetry.

The consequence of preserving realism is very nontrivial: If we do not give up realism, the violation of Bell's inequality for arbitrary pairs of events allows to derive the existence of some hidden absolute contemporaneity.

But does it help? Are there realistic theories which allow to replace current relativistic and quantum theories? There are.

- There is the pilot wave theory of de Broglie and Bohm. It is a realistic, deterministic theory, which makes the same predictions as quantum mechanics. Because the existence of this theory is an important point in favour of realism – it destroys the claims that quantum theory requires a rejection of realism – I have spend some time to defend it against some counterarguments. In particular, I refute arguments proposed by Streater, Mohrhoff and Motl.
- Nelsonian stochastics is also a realistic theory, which makes the same predictions as quantum mechanics. It is not deterministic. I have some sympathy for it, because it gives the "quantum potential" of pilot wave theory in a very natural way.
- A replacement for general relativity with preferred frame is my own theory of gravity — General Lorentz Ether Theory (GLET).
- Introducing a preferred frame allows a nice geometric interpretation of standard model fermions as Aff(3) ⊗ C ⊗ Λ(R
^{3}) and a corresponding ether model as a lattice of elementary cells, with configuration space Aff(3)(Z^{3}).

Here I agree with the mainstream:

- Starting from realism and Einstein locality, we can prove Bell's inequality.
- Bell's inequality is violated.
- Thus, Einstein-local realism is falsified.

Thus, realism needs to be non-Einstein-local.

That realism, therefore, needs a preferred frame, is not questioned. It is a quite natural consequence.

The disagreement is about the following: The mainstream rejects realism. I propose to accept realism, and, as a consequence, to accept the existence of a (hidden) preferred frame.

Let's note that I support explicitly the mainstream position against defenders of "local realism" (more accurate would be Einstein-local realism) and claims about the existence of "loopholes" in Bell's theorem. So I have already published three papers rejecting various published "refutations" of Bell's theorem:

- Schmelzer, I.: About a "nonlocal" local model considered by L. Vervoort, and the necessity to distinguish locality from Einstein locality, acc. by Foundations of Physics, arXiv:1610.03057. This paper is nice because I succeeded to publish there, as a side issue, the convention to name "Einstein locality" simply "locality":
**Moreover, a naming convention which forces us to name theories which are local in any physically important sense "non-local" is not only absurd, but can be even considered as Orwellian.***To classify the actual convention as "Orwellian" is justified not only because it requires to name a local theory non-local. It also shares another important aspect with newspeak -- it leaves some incorrect thoughts without words to talk about then: Indeed, the word "local" is the natural word to describe the class of models considered in this paper, with some much higher speed of information transfer in a hidden preferred frame, and to distinguish it from theories with really pathological locality and causality violations. And this is, indeed, a class of theories which is the closest thing to anathema in modern physics.* - Schmelzer, I.: About a "contextuality loophole" in Bell's theorem claimed to exist by Nieuwenhuizen, acc. by Foundations of Physics, DOI 10.1007/s10701-016-0047-2.
- Schmelzer, I.: Comments on a paper by B.Schulz about Bell's inequalities, Ann. Phys. (Berlin), 523, 576-579 (2011), arxiv:0910.4740.