We propose Bell tests for collider decays using independently sampled analysis axes as settings and detector regions to assign binary or null outcomes. We show that a Bell violation requires postselection, opening a loophole. However, under measuremen
We propose Bell tests for collider decays using independently sampled analysis axes as settings and detector regions to assign binary or null outcomes. We show that a Bell violation requires postselection, opening a loophole. However, under measurement independence and operational equivalence of acceptance events across settings, any local model exploiting the loophole must be nonclassical in the sense of measurement contextuality. We derive a Bell inequality violated by singlet correlations when the effective spin-analyzing powers are sufficiently high. Its violation then witnesses Bell nonlocality or measurement contextuality.