Based on past experiences I don't think that you would be happy with anyone's opinion but your own...

Cheers

dbartman wrote:  The Tsojconths prior to year 2001 have been removed from the valuations to keep the pricing more in line with current trends.  

I am left speechless with mirth……………

Note that the recent sale values are quite high compared to the earlier years.

By your own statement, it seems to me that by allowing the values from the years 2001-2003 (perhaps more) into your "evaluations' the resulting lower value from those remote years skews your results to near uselessness.......

Recent values?
What constitutes recent?
"Current Trends"?

To me it seems that the trend in recent years (The past 3) is that a condition 10 "Tsojconth" goes for between $1850-$2060.05............

There.................
you have your evaluation.............

Not being an Expert though what the hell do I know?  

Please explain the obvious to me if you would be so kind...........

Hi,

what does surprise is how many sales there have been in the past 7 years.

Shame theres no data on how many of these were the same copy changing hands more than once.

Cheers,
malc

dbartman wrote:Okay, someone asked for the raw data a while back in this thread.  This should be almost all of the Tsojconths that have been sold publicly since 2001.  

Thanks for the data.

There are significant "year", "quarter", and "condition" effects in the data.  The year effect isn't surprising (there's a trend), but the quarter effect might be spurious since the data aren't evenly distributed among quarters.  Still, makes you consider buying your next Tsojconth in the 4th quarter, doesn't it?  

They were both private IIRC.
I have wondered why private sales were not included. If anything those would be a better barometer than eBay, I would think.

A couple things to clear up:

>>>I am left speechless with mirth

If I remember correctly, the 2004 conversations on this matter centered on the fact that awareness of the rarity of D&D collectibles increased exponentially over the years as more and more veteran collectors compared notes and figured out just how rare things are.  The Acaeum was at the heart of that.  If valuations prior to 2001 are considered helpful to the data, however, they include a $78 Tsojconth (1999), a $375 near mint Tamoachan (1996), a $155 Ghost Tower (1999), a $64 Dwarven Glory (2000), etc.  The data could be included and lightly weighted but I'm not sure how worthwhile it would be when considering valuations in the current century.

>>> To me it seems that the trend in recent years (The past 3) is that a condition 10 "Tsojconth" goes for between $1850-$2060.05

Actually, the data is:
$2,060.05 2007 Q1 10
$1,120.75 2006 Q4 10
$965.00 2006 Q3 10
$1,870.00 2006 Q1 10
$1,895.00  2006 Q1 10

Which would give an average value of $1,582.16 for recent years, so you were off by about 25%.

On the matter of private sales, there are issues of confirmation and civility that tend to crop up continually.  eBay is the only venue where there's a public record of the product's visual and the final price.  If people who conduct private sales would be willing to provide billing documents (Paypal?) and condition assessments, the data would be valuable; otherwise it's hearsay and rather difficult to defend down the road.

Personally I think the real issue is that the market is changing and becoming more like it was in the 90s.  People are gravitating toward private sales and gamecon auctions, due to eBay's asshattery and a number of other things.  The hobby focus is fine, but it makes it a lot harder to find people willing to do valuations and perform the calculations.  It also lessens the value of posted numbers, since the numbers necessarily focus only on that part of the market that is reliably trackable.  It's probably better to go with a sale history and your gut, rather than the Acaeum's guideline, if you're going to submit the guideline to some form of rigorous testing methodology.

I would recommend that the people who are interested start recording a sales history thread for items that have widespread interest (R5?).  There's already threads tracking virtually every R5 that hits the market; private sellers could provide input on their own recognizance.  Then the price data would be available for everyone to haggle over with no one sticking their neck out and saying "It's worth about X" and getting chopped.  The times, they are a-changing …

I agree that the field is too small to provide reliable sales data, and that even calculation-based valuations are a shot in the dark.  We went round and round on this in '04 and it was decided that sparse data was better than zero data, and that we should devise a system that gets as much mileage out of the sparse data as possible.  

The issue of price confirmation on private sales is one that is a bit touchy; eBay provides firm proof-of-price on all items except for BINs, because the implication is that even if the buyer didn't pay (rare), there was a mutual consensus of value reached by two or more individuals.  After all, an item can't sell for $960 unless someone else bids $950.

There is also the side issue of price manipulation which is quite easy to perform, especially if three or more people report sales without proof that may or may not have taken place.  Then there's the issue of favoritism, "Oh, X would never do that, but Z might ... we need proof on sales done by Z but not by X ..."  Or even worse, "Christ!  You think I'd report a sale just to manipulate the price?  I'm going to shame you into accepting unproven data into the valuations!"  Etc., etc.  It gets way too personal way too quickly, so the entire ball of wax (snakes?) was sidestepped as a whole.  I was on both sides of the issue in 2004 and I think private sales have much more legitimacy in 2008 than in 2004, but it's no longer my concern.

But I do strongly advocate a grass-roots movement to record sales and make them public.  After all, the data would allow people to draw their own conclusions.  Some people only want 6-month data (!) to reflect on current valuations.  Some want 3 years, some 5, some 10.  Condition is basically nebulous when there is only a scan to serve as a discussion point, especially when the seller clams up suspecting outing or devaluation.  It would also be valuable to the VB as a double-check of their own research using volunteered time.

I'll close simply by saying that it's much easier to criticize than it is to support, and it's easier to destroy than it is to create.  It's also easier to undermine someone else's system than it is to structure your own against flaws.  And it's ... well.  I think the point is made.     At issue, everyone wants to either possess or criticize the data, but very few want to do the work, especially over a long period of time.

Of course I've been proven wrong before and will be happy to have it occur again.  All that would take is some positive-minded dedication from 5+ Acaeum members with time on their hands!

It isn't a price guide!  (The following is one man's rant, but it makes me sleep better at night about the numbers.)

One generally indisputable fact is that the only way to determine the precise value of a particular item at a moment in time is to sell it.  For example, the Tsojconth in the thread-starting auction was worth exactly $1169.57 USD at that moment in time -- assuming I recall the number correctly.

The problem is that people tend to extrapolate from that value and assume that identical items are worth that amount, even though they haven't actually sold.  This is incorrect for a variety of reasons, including the weather in Peru, the war in Iraq, and last but not least the previous sale of the aforementioned Tsojconth.  That's why individual sales aren't very useful for calculating valuations, and why the Acaeum uses historical weighted means.  

It's very important to note that the valuations are means.  Whatever the level of precision/complexity in their calculation, they will always represent an average value, not a maximum or a minimum.  So they will seldom if ever be an accurate representation of what an item will sell for, as that number is not an average.

So the main problem is that the values aren't understood for what they are.  "Acaeum lists this item at $x" is a fundamentally useless bit of information without additional information and/or experience watching the market.

One thing I think that would be tremendously valuable for those that want more accurate pricing (but much more time-consuming to calculate for the VB and perhaps not in their mandate) would be standard deviations.  You should never pay more/sell for less than a few standard deviations from the mean -- perhaps three.

serleran wrote:Yes, but wouldn't the problem with that lie in gathering enough samples of data to draw a valid deviation? I mean, if you have two points of entry, a sale at 1000 and another at 2500, the SD is going to be much more significant than if you have thirty sales where the mean will settle to a more probabilistic curve. But, its been a very long time since I did anything with statistics, so, if I'm off, just ignore me.

The more data you have the more precise a model you can build, but I'm not sure it's less "valid", it's simply less precise.  You have to work with what data you have, and apply common sense, naturally.