On the Poser spec, IMHO, the base right now that should be developed for is Poser 9 (2012 is included obviously).  Poser 9 is the minimum for newer script support, addon support, weight-mapping, sub-surface scattering and other things that really shine through for current generation products (not to be confused with Victoria's generation).  It's what you need if you wish to utilize Reality, Octane, DSON for Genesis in Poser (without baking it as a DS export) and that sort of thing.  So in my mind it's the base and as we go to further Poser versions, that will be more cemented.  For too long, Poser 4 has been the base and it's time to step it up.

But with saying that, all those toys and features are still options and that's the way it should be.  SM has a good track record for keeping things backwards compatible, so you can still use non-weighted content and you can mix and match it.  Poser's grouped object system may not be standard (when you look at games and other things like that) but it does build on what was there before instead of just replacing it and sorry, but I think that's a smart move.  Not only does it give youi the option of hybridization but it allows you to conform non-weighted content to weighted content and vice-versa.  And frankly, the all-in-one object is true in games but they are still broken down into weighted body parts, it's just that the divisions are cut on the rigging level (and can overlap).

Sorry but the OP's initial post does come off as something of a taunt.  Sure it's legitimate to wonder and discuss the pole position of where the various figures are at but the initial wording invites fighting.

You can dissect every figure and find flaws.  Certainly the queen bea Victoria 4 flaws too, even when she was new.  And by the way, I disagree that V3 was ousted immediately.  It took some time to gain momentum and there weren't replacements for Aiko 3, Stephanie Petite or The Girl ready out of the gate.  And speaking only for Aiko, the figure had a significant chunk of fans.  Plus, at that time, DAZ was the only game in town and we had a two-way road between Poser and DS (with already existing blocks).  Now we are dealing with a fragmented overview.  Out of that fragmentation, "most" users use Victoria 4, but I think that "most" is certainly smaller than it used to be.  But again, we only know what we can touch and see, so that means we only know by the response in our community, which is significantly smaller than the pool of larger buyers.

Scott is right in that any new figure is going to be ripped to shreds and even if you scan a real human and build an exact copy as a figure, someone will find flaws with it.  For every text-book dissection of the human anatomy, there are multiple exceptions.  There is no example of "every real woman" - there is just way too many differences from one person to the next.  You're not going to please everyone and the best you can do is hopefully please a large portion.

The best you could do for Dawn 2, if you cared to, is note what you don't like and give constructive reasons why.  This is what I would cite but before I put it out, I want to say that none of it is such a barrier that it can't be worked around or worked into.

• The extra body cuts don't seem to really add much benefit, particularly when it comes to being weight mapped.  I would go through and use the kiss principle to eliminate extra cuts if they aren't significantly adding anything positive.  Two necks was tried on V1/V2.  We even had extra buttock parts on V3.  I don't think it's really necessary and just adds more cuts to deal with for content development.  Dropping a Pelvis under hip is interesting but is it adding something so non-standard that causes extra work in conversion or rigging without any real benefit? - types of internal questions to ask.

• The jaw split from the head makes it automatically tricky to create head morphs or to produce work where weight maps are projected and need the "head" but not the jaw.  There's no real good reason to split the jaw when open mouth and jaw shifting morphs can already be designed.

• The big faerie secondary lashes make it additionally tricky to create head morphs.  I think it would have been better to just include faerie whatzits as morph targets.

I think those are the biggest hang-ups.  Yes, I could also say that DS and Poser in the WM department are too different and there are not enough cross-tools to get the job done and heal the divide.  But I personally don't mind if content divulges and some is available here and some over there, in terms of apps.  Unless there are two-way street tools to facilitate 2 different workflows of content creation and rigging, that's what you are going to have.

As far as which pizza reigns supreme, even if Dawn has a niche, as long as it is significant enough, that's fine.  Unless the tech gets closed off, which has been already noted, any non-V4 figure is going to be niche as long as content is still created for V4 to the degree we've had for 8 years.  That includes Roxie, Genesis and any other figure you want to write down.  Large niches, small niches, doesn't matter, they are still below the queen bea.

.