Quote - "Of course, this does not take into account the usefulness of morphs, only their number."

I don't doubt your theoretical numbers, but that's not how Poser works.

Hm, i think that is exactly how poser works.

Quote - A cr2 only stores morphs at full strenght which are actually useful. The theoretical "trash-morphs" are all ignored.

But if a morph is useful or not, is dependent of the particular application. If i do a portrait, all the toe-wiggle morphs are not useful. How should poser know that the toes are not useful?

Quote - A high res mesh can have more "useful" morphs than a low res mesh because it allows greater detail and you are not limited to predetermined edgeloops.

Yes, but "not being limited" means more flexibility. And more flexibility always comes at a price. With software this price is often paid in memory, CPU power or usability.

Quote - And once you go into custom morph creation with magnets or morphbrush, you have a lot more possible useful combinations with almost zero overhead, because only the finished morph is actually stored in Poser, not all possible combinations.

I agree that single morph has almost zero overhead (like a single vertex more or less has zero overhead). The overhead comes into existence when you put many many of them into a figure. For example, on my machine i notice a slowdown when i am loading all of the V4++ morphs into V4 at once. Your laptop might be a bit better, so you might notice it a bit later, depending on the circumstances, but it surely would not go unnoticed. I even tried to subdivide V4 up to a million polygons once, because i thought it would be a good idea so i would not need displacement maps any more, but it was unusable (even without morphs).

Quote - And of course, yes, a high res mesh uses more space for each morph.

But that's what injection morphs are there for.

:-)