I think you're pretty spot on with your logic.
As for FMC predictions, I reason they must take the Mach/IAS transition into account otherwise we would always end up 'off profile' not long after the TOD. Another clue is that they ask for Mach no. & IAS as input...