Detection probability "enhancement" and "unfair sampling" in Bell inequalities

TL;DR

This paper examines how hidden variable models can violate inhomogeneous Bell inequalities through no-enhancement assumptions and discusses the limitations of current experimental tests in conclusively ruling out local realism.

Contribution

It provides explicit examples of hidden variable models violating inhomogeneous Bell inequalities and critically analyzes the limitations of recent experimental tests.

Findings

Hidden variable models can violate inhomogeneous Bell inequalities without no-enhancement.

Experimental violations are extremely low, raising concerns about systematic errors.

Current data and models lack sufficient validation against quantum predictions.

Abstract

On one side, so far a great part of the evidence accepted as proof of the alleged quantum non-locality relied on inhomogeneous Bell inequalities involving an additional assumption (no-enhancement) whose role had not been sufficiently examined (in contrast to homogeneous inequalities, where the role of low detection rates is well acknowledged). A first contribution of this paper is to provide explicit examples of how a model of hidden local variables (LHV) defying no-enhancement is able to produce a violation of an inhomogeneous inequality, a possibility that so far was pointed out only in qualitative terms. On the other hand, recent tests have attempted to overcome this reliance on supplementary assumptions, but still give rise to doubt, due to: (i) violations are extremely low, which points to some systematical error (the fact that alleged random errors of the detectors seem to conspire to interfere exactly at such narrow strip around the frontier with local-realism and not a different part of the spectrum that QM would theoretically produce can hardly be considered just coincidence); (ii) the correspondence between the observations and one or another quantum state is not clear either, as recent models cannot be really validated against the available data (they have equal or more parameters than restrictions). Following (ii), it would be convenient to verify a more significant set of quantum predictions, instead of just the few counting rates involved in a particular Bell inequality. Finally, another type of "no-go" tests (an instance of which we examine here) also present subtleties which call for serious examination.