The position (FMP) and orientation (FMO) of the foramen magnum have been used as proxies for locomotion and posture in extant and extinct primates. Several indices have been designed to quantify FMP and FMO but their application has led to conflicting results. Here, we test six widely used indices and two approaches (univariate and multivariate) for their capability to discriminate between postural and locomotor types in extant primates and fossil hominins. We then look at the locomotion of australopithecines and Homo on the base of these new findings. The following measurements are used: the opisthocranion-prosthion (OP-PR) and the opisthocranion-glabella (OP-GL) indices, the basion-biporion (BA-BP) and basion-bicarotid chords, the foramen magnum angle (FMA), and the basion-sphenoccipital ratio. After exploring the indices variability using principal component analysis, pairwise comparisons are performed to test for the association between each index and the locomotor and postural habits. Cranial size and phylogeny are taken into account. Our analysis indicates that none of the indices or approaches provides complete discrimination across locomotor and postural categories, although some differences are highlighted. FMA and BA-BP distinguish respectively obligate and facultative bipeds from all other groups. For what concerns posture, orthogrades and pronogrades differ with respects to OP-PR, OP-GL, and FMA. Although the multivariate approach seems to have some discrimination power, the results are most likely driven by facial and neurocranial variability embedded in some of the indices. These results demonstrate that indices relying on the anteroposterior positioning of the foramen may not be appropriate proxies for locomotion among primates. The assumptions about locomotor and postural habits in fossil hominins based on foramen magnum indices should be revised in light of these new findings.