Just can't rely on results. Going to have to simply use spreadsheets (which is terrible... but reliable). I've had various problems over the years so yesterday I sat down and rely analyzed things.
Have an assembly set to "Geometry and Parameters", It has 3 subassemblies, each set to "Fully assigned". All models are set to IPS. When I run mass properties for the top assembly, the results DO add up properly when the last two models are suppressed. The results are also correct when just the last model is suppressed. However, when I resume the last model (ergo, all three models), the "total mass" is less than the sum of the three.
Even if this is some kind of "corruption problem" (it shouldn't be because, like the others, the last model is "fully assigned") I don't see any way to rely on Creo for mass properties when that kind of thing can happen. Perhaps for small models where errors may be evident... but not for large models like rockets, cars, boats, large machinery, etc.
If it can't even add masses correctly, I have no confidence at all in the inertial information.