We also show the impact that the number of fossils included in the unresolved FBD analysis has on node ages. Turns out that you need a substantial number of fossils (in our croc example, >20 in total) to have effective priors that are more informative than those induced by flat node calibrations.
