After experimenting a bit more, I think I see why this issue exists.
1) It's too easy to register a note-off velocity of 1 on the VPC1 (hey, maybe that's what the "1" stands for...). I can't even get it to register 2 through 4 at all, and 5, 6 are tough to hit.
2) Pianoteq seems to be trying to compensate for a MIDI keyboard limitation: a simple note-off velocity is crude compared with how a piano works. On a piano, if you very slowly release a key, there's a point where the damper just barely starts to touch the strings, but doesn't fully dampen them. You can keep the dampers in this position if you hold your finger there (or your foot on the damper pedal), or you can even back the dampers off the strings again without fully damping them. None of that can be transmitted by a MIDI keyboard (any I know of). Of course you wouldn't do this in normal playing, so it's not a big issue.
3) Pianoteq seems to do a pretty good job working around this limitation. But I think, with that limitation in mind, Pianoteq is a little too extreme with these long decays for low note-off values, and 1 shouldn't act like the damper never went down (although maybe it's still damping the note a tiny bit?).
4) No matter how #3 is tuned, the sustain pedal should still reset those dampers. I'm guessing that was just overlooked.
5) For now there's an easy fix on my end. If I set the note-off curve starting point to 0,8 it all seems to respond quite naturally (for the VPC1). Anything lower than that (like 0,4) feels exaggerated.
As an aside about the VPC1, the response of the action is definitely a step up from the MP8 overall. It's action is a bit lighter, less loud and clunky feeling; the key return is a little sluggish, but not as much as the MP8; and the response at playing at the back of the keys (at the "fallboard") is better, but not great. I would still prefer a much longer key/pivot point. One area where the MP8 was superior was its wider range of key force, particularly on the upper end. On the VPC1 it's too easy to register a note-on velocity of 127, and so no matter what touch curve you use, you're always hitting that ceiling sooner than you should.
