Time-symmetry without retrocausality: how the quantum can withhold the solace

It has been suggested that some of the puzzles of QM are resolved if we allow that there is retrocausality in the quantum world. In particular, it has been claimed that this approach offers a path to a Lorentz-invariant explanation of Bell correlations, and other manifestations of quantum "nonlocality", without action-at-a-distance. Some writers have suggested that this proposal can be supported by an appeal to time-symmetry, claiming that if QM were made "more time-symmetric", retrocausality would be a natural consequence. Critics object that there is complete time-symmetry in classical physics, and yet no apparent retrocausality. Why should QM be any different? In this note I call attention to a respect in which QM is different, under some assumptions about quantum ontology. Under these assumptions, the option of time-symmetry without retrocausality is not available in QM, for reasons intimately connected with the fundamental differences between classical and quantum physics (especially the role of discreteness in the latter)