Comment on "Strong Evidence for the Normal Neutrino Hierarchy"

In the preprint arxiv:1703.03425 "strong evidence" for the normal neutrino mass ordering is claimed. The authors obtain Bayesian odds of 42:1 in favour of the normal ordering. Their conclusion is based on adopting a flat logarithmic prior for the three neutrino masses. Such an assumption favours a hierarchical spectrum for the masses, which is much easier to accommodate for the normal mass ordering, and hence their prior assumption makes the inverted ordering much less likely a priori. We argue that the claimed "evidence" for normal ordering is almost entirely driven by the adopted prior and not due to the data itself.