[ancients] Statistics : Auction estimates vs prices realized

Excellent! I knew I was going to be throwing in my lot with the right people.
There's no way we can fail... we'll be minting so much money we best start talking to Steve about leasing space in his vault.
We can take further discussions offline, but we should think of a motto for our joint services. I like "Bid With Coinfidence!"

 

Log in or Sign up to hide this ad.

I've done this for several auctions as of late (didn't have the data on-hand, however). Here's a more computationally-drive analysis of NAC sales 70-72 from last May, with me attempting to determine some conclusions from a purely mathematical perspective. Let me know if any of it is unclear:







Perceived overpriced lots:


Highest prices:


Highest over estimate:


Highest over estimate (if > 10K starting price):


Lowest versus estimate (if > 4K starting price):

 

Thoroughly laughed out loud at that!

 

Joe, you know there's medication for your condition, right?

Seriously... nice work. I hope you keep posting these.

 

Everything just looks fancier with graphs and spreadsheets!
Seriously though, thanks for sharing those, AJ.

 

Questions...

First graph. ">1K", ">4K" etc. Are the columns limited? In other words, does the "1K" column contain data for coins with opening bids from 1K to infinity or just from 1K to 3,999K?

Second picture, first two graphs: what is the x axis? Percentage over estimate?

Looking at this data is fun. For me, I think it would be most useful to compare similar data from all major auction houses. It might help me predict when an "Overachiever" bid might be required. It might also lead me to not chase coins at from certain houses, or at least to not get my hopes up.

 

I'm glad you like it! I don't go this indepth for every sale but I enjoy this type of analysis a bit more than I should...

The X axis on that one is the lot number - I was looking to determine the distribution of lots throughout the sale, in an attempt to see if there is any commonality or logic to the layout of expensive vs inexpensive lots. I've seen that some auction houses tend to intermix very expensive coins with very inexpensive coins, and the inexpensive ones have their prices elevated somewhat between the more expensive examples.

 

Whoops, I shouldn't have edited the above post because you were answering one question while I added another. In case you didn't see that question:

 

The range is from 1K to infinity rather than 1K to 4K. However, I can pretty easily get different ranges if you'd like it split in any varying ways

 

Perhaps incorporating moon charts would be helpful. (People do crazy things during full moons, ask any ER doc )

 

....now my head hurts.

has anyone does the statistics that shows that estimated price is under valued (or not) for some place or everywhere in general?

 

For the graphs "Price Distribution" and "Realized vs. Estimate" it would be interesting to set the y-axis to be logarithmic. This would provide more separation for the lower-priced lots and might show some interesting trends.

One thing is clear, though: NAC is expert at setting "estimated" prices along the same curve that the higher hammer prices will bring. That's one reason why they consistently have the rarest and most sought-after coins.

 

Now THIS is humor. Brilliant, clever satire without resorting to goofy pictures or adolescent male hyuck-hyuck jokes.

Well done!

 

Yes this was funny; however, being a few years older than 60 (humor), I sometimes like to act like an adolescent.

 

I think there's room for both here.

 

Z-bro => your stats rocked, brother!!

 

Rumor has it that CACA will be adding PhotoCertification services in the near future...

 

Hmm...I think you're right.

 

My brother was laughing so hard over this, as was I.

Nice work TIF.

 

Somehow this post showed up as a recent one, didn't realise how old it is lol. But

This was brilliant!

I wonder if anyone has continued to do analysis of current auctions to see how they stack up against these from 2014..