Hmm. I haven't gotten into the numbers or anything, but I think there are other factors involved as well. From my personal experience, frequency of comments do matter. I'm not good at the whole data analysis thing, so I'll just put what I notice . . . My posts either do very well, or bomb horribly. There's not a lot of middle ground, though I do see that subject matter does influence their success rate. I have a decent number of followers (HI GUYS!) for such a young account. I've seen accounts much older than me don't have as many followers. I'm not saying this to brag though. I think the reason may be is that I comment more than some people. I've noticed that on days where I spend half my time in front of the computer, commenting away on Hubski, I'll gain 2 or 3 new followers. Sometimes I might lose one though (was it something I said)? The number of followers I have has done jack all for my post shares. It's still hit or miss. I don't post for shares though. I post what I love and what I would love talking about. Same thing for sharing. If I see something worth talking about, I'll share it. If I see something I'm dying to talk about, I'll share and comment on it. Lastly, ever since I've been doing social sites, so about 15 years now, I've always scrubbed my accounts and started a new one about every 5-8 months, sometimes ditching a site altogether for newer pastures (I don't miss you one bit FARK!). I'm torn on Hubski though. For once, I think my account is worth something. I don't know if I'll be able to bring myself to scrub it once it's time.
I agree a lot about the commenting. You make good intelligent comments, and just so you know, you were in the 20. And you bring up a good point, I should have discriminated between visibility (number of feeds it's in) and popularity (number of shares). And yes, you should try out alt accounts. For this one though, I unfollowed everyone, left a quick bio to say to ignore it, and logged out forever.
I could go through each post, look at the tags and who shared it, then sum all the followers for both. So technically, yes, but it would take so long, no significant number of data could ve taken in a time frame small enough that it wouldn't change pretty significantly while I did the sum. Nor would I know how to account for overlap. I'm sure there is some sort of code that could do it, basically a web crawler, which are used on lots of social networks, but I couldn't make one.