in your scenario, every single person in the group posted.  the group is in effect now.  the only people i've seen  with lots of comments are the people who already had lots of comments (as in several pages).  so far as i've seen, people who get 3 to 6 comments now have 6 to 9 comments.  if the group grows huge... frankly, i don't know.  i think it depends on the culture.  i'm taking an average of 15 to 20 minutes per comment.  i'm not sure tons of people will do that.  if you're worried about the culture of the group shifting, i'd  say join in (even without doing the favorites) and start posting the kind of reviews you think should be happening.

and you're arguing on both sides.  you're saying the intent of the group needs to be clear because people will be scared to join but then saying tons of people will join to be popular.  it can't really be both.

it amazes me how many people are already deciding what will happen without actually checking what's going on now.  nothing like this initiative has been tried, but people keep talking about negative effects.  whereas just about every other solution suggested has been tried for other initiatives (critique thread, critique forum, critique gallery) and all pretty much crashed and burned.  multiple times.  this is the main forum; people won't go elsewhere unless there's a really pressing need.  hence the complete lack of traffic in the developers and w.i.p. forums.  it's not that nobody's doing that work and posting about it, it's that they post to modeling software threads at most and usually post here.  tons of people use more digital manipulation, digital painting and photographs than poser, but they fight  not to have to post to mixed medium because they get no views there.  i've seen people post to the poser gallery who are rendering in bryce and vue.  and threads  get messy, go off topic, then die.  right now the subscription thread has turned into a discussion.  that's what threads are for.  i've never seen any long thread actually stay completely on topic, let alone receive only the type of posts it was supposed to have.  so we could easily be stuck wading through pages and pages of posts (and tons of useless ebots when the inevitable disagreement came along, as it does in even the most friendly of group threads)  just to have to find the few image postings. 

and frankly, favorites last.  why do these buddy networks work?  because people make lots of favorites and get notices and keep posting to each other's pics.  their method works for generating feedback, because it works proactively.   in point of fact, it is the absolutely only way i've  seen a commenting / response group work.  including at artzone, which seems to be rapidly dying (based on the lack of updates i've seen in the many, many people i've bookmarked and the few i've added as friends, the stabilization of the number of users, and other signs of lowered activity).  even though it has actual community tools like groups you can join with membership controlled by group leaders.  i've watched everything else die pretty quickly.  there's a reason blogs have their own rss feeds instead of just posting to one main blog directory.  in fact, blogs became hugely popular precisely when rss feeds and rss readers became popular. 

no matter what people's intent, unless there is an image by image notification external to the wealth of unpopular and unfrequented forums or galleries,  and an ability to comment right then and there, i do not believe they will consistently comment.  i don't know it will work this way, but at least that's had some success.

it really doesn't matter how you slice it, either.  the important thing is always going to be the culture of the group.   there are zero things you can do to get a group of people to view your work and absolutely eliminate the possibility of it becoming a way to become popular.  community interaction helps make you popular.  whether this will actually be an influential  subcommunity, i don't know.  but i really doubt it.

right now the strategy is wait and see.  personally, i want this to succeed, and since every alternative solution people have raised, i've seen fail spectacularly, i think trying something that has been proven to work is better.