Estimating rarity or number of units in circulation by frequency on eBay

[orb85750]

I always wonder how rare a piece of Amiga hardware (or software) is.  Some stuff you see up on eBay all the time, while other items you're lucky to find a few times per year, or even less.  Has anyone devised a formula to estimate the number of units still in circulation based on frequency of appearance on eBay?  Seems like a rough formulaic estimate could be gotten with just a few reliable known data points.  Any leads on this idea?  Someone somewhere must have already done it.  -Dave

[orb85750]

Rudei wrote:
Perhaps you should give it a go? :-)

Membership of sites may give an indication of active users but as for pieces of hardware, I guess that would very much depend on the actual hardware itself.

eg. A1200s - very common
    Blizzard 030s - quite common
    A1000s - very rare
    VideoToasters - rocking horse pooh
    Walker prototype - urban myth :-)

Rude!

OK, so if nobody has done this so far, maybe I should give it a go once I have a little time.  It would be nice to know roughly how many of each model/unit was made in the first place, to get some decent data points.  For example, number of Commodore A1200, A2000, A2500, SX-64, etc.   Does anyone know of any good sources of the total production numbers for these and other systems.  Of course, they don't have to be just Commodore or Amiga stuff either, for the purposes of formulating an equation -- but I don't want to spend all my time looking up Atari ST crap (for example).  -Dave

[orb85750]

murple wrote:
I sort of hope you don't do this, orb. It may be an interesting mental challenge but for the reasons I already gave (as well as periodic fluctuations... I think you'd need to collect statistics over a couple years at least to take into account the way things sometimes seem to hit ebay in waves), I don't think you're going to get really meaningful accurate data.

Problem is the minute you release the numbers, every douchey seller on ebay is going to start using your rarity stats as a justification to jack up their prices.

Well, I only wish I had such an influence on the free market!  I'm essentially trying to do a simple inversion.  If I have a few reliable data points regarding the number of items in circulation of, say, products X, Y, and Z and they come up on eBay with frequency f(X), f(Y), and f(Z), it would be nice to use eBay frequency of appearance for other products in order to estimate their total circulation.  Seems harmless enough to me.  But to be done well, I agree that it will require many data points and plenty of time watching eBay!