On 17 Feb 2001, at 10:12, Gary Reese wrote:
> << I do NOT capture each and every e-Bay auction, so when you're
> looking at the data consider it only a representative sample. >>
>
> Subtle, but highly important point for anyone analyzing the
> information.
Which is why I mentioned it.
> But it is only "representative" if the sample was obtained in a random
> manner.
How 'bout more or less random? :-) So that everyone knows my
"methodology," I collect stuff that's of general interest to me, personally.
That means "just 'bout everything." However, because of the sheer
volume of Oly/Zuiko auctions on e-Bay, and the fact that I manually
enter each record into the database, there's no *way* I could get
*everything.*
So, while the database doesn't have "everything," it does have a pretty
large (and growing) sample of what's gone on. No, one *can't* do a
statistically valid analysis of the auctions. But, if one studies the
databases, one can get a "pretty good idea" of what stuff goes for.
> If there is a pattern to what is excluded, then the information is no
> longer "representative."
I suppose there's one pattern to the exclusions... E-Bay has a 20 item
limit to what one can "watch" so when my 20 item "quota" is filled I
can't add anything else. No, I don't generally bid on stuff just to keep
track of it. I only bid on stuff I'm actually trying to buy.
> Additionally, one may no longer be able to determine the commoness or
> rarity of any particular item - especially if one differentially records
> rare items at the expense of common items.
That isn't the intent of the database. It's merely intended to give me
(and now others) a general idea of what the price ranges are. But, as
we all know, e-Bay prices are known fluctuate wildly (e.g. that T20 I saw
go for US$75+ awhile back), so there's no way anyone can predict what
something will go for just because there's a "usual price range" for it.
> << As for rating systems, I DON'T believe in them or use them, as
> they're too subjective and open to interpretation. >>
>
> Condition is everything when it comes to appraising an items value
> (Garage Sales 101).
Absolutely!
> If one doesn't use some standard to judge condition, the price range
> becomes ambiguously broad and the mean becomes confounded by average
> condition (LN- for a 8mm Zuiko Fisheye, but BGN for an OM-1).
That's correct, too, but who ever said that the rating "systems" people
use on e-Bay are standardized in any way???? That, in fact, is the
whole problem... what some people call "mint" isn't even "excellent" to
me. I've seen *way* too many items described as "mint" and then when
you read the description you learn about numerous flaws (meters off,
dents on lens shades, etc.).
So, rather than try and rely on someone's "rating" of their item, I
carefully read the description and look at the pictures. Then, if I'm not
sure of the item's condition, I e-mail the seller and ask LOTS of
questions. And, of course, I look for those warning phrases such as "I
don't know anything about cameras" or "I think it works, but haven't
used it in ages."
Does that mean I don't bid on stuff that may be in questionable shape?
Definately not! I'll sometimes "gamble" on an auction, but I sure *don't*
"bet the farm" on questionable items.
> OK, this might seem like nit-picking from yet another person who just
> so happens to have a specialized professional knowledge.
Not at all! I'm glad you've raised your points. I'm more than happy to
share my "methodology" for collecting the data, and I sure *don't* want
anybody to give the data more weight than it deserves.
> Statistics is the only mathematics which ever captured my interested.
Well, then, you know that statistics can be used to prove just about
anything, if that's what the person crunching the numbers wants. So, to
sum up, let me make it absolutely clear that the data in my e-Bay
auction database is NOT a complete and statistically accurate
compilation of e-Bay prices for Olympus/Zuiko stuff.
Regards,
Chris
---
I'm *not* a Zuikoholic.... I'm a Zuikohobbiest!
Chris O'Neill (coneill@xxxxxxxxxxx)
http://www.nucleus.com/~coneill
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|