I think it’s a very creative approach and a useful addition to the data set, but I’m hesitant to assume it’s “best” esp. just based on a one-time survey and without some time for people to poke at what underlies it.
For example, it’s pretty unlikely the horde population of OB is 10% the size of Mankrik’s. I can’t say for certain either, but (anecdotally) based on the last time I logged in I’d wager it’s much less than that, whether or not it’s as bad as the 2% Ironforge suggests. But I could be wrong.
If the question you’re trying to get at is population, then this is a proxy with some potential utility, and it’s worth someone pursuing it a bit more and flushing it out for sure.
In general it’s pretty good. I’ve noticed PvE servers will typically have more auctions than PvP servers per capita, likely because PvP servers have more people focused on PvPing and not farming and selling things.
Do you assume that say Sulf or OB has more single listings vs Faerlina or Grobb or Bene?
Or are you assuming that all the AH’s are inflated?
Either you make the assumption that all servers have the same % of single listings (best option) or you are making the assumption that certain servers have more (which would need a reason why they have)
Inherently, people can post auctions in many different stack sizes, which will result in either more or less auctions for the same amount of resources being sold.
So, sure, I guess you could say that every server’s number is more than likely inflated, but I’m not going to draw any conclusions from that.
I was just poking holes in this methodology for gauging server health/activity.
Stack sizes is one of those things that is more or less a toss out because every single server will have roughly the same amount of it going on.
It’s basically one of those possible vs probable type things. Is it possible that one server is more skewed because of stack size listings? Yes. Is it probable that it only happens on one server more than another to really make any difference? Doubtful.