Sometimes that study of Economics pays dividends…
Some of the premises to establish a solid value (for hobbyDB, our Estimated Value) are the following –
As our plan to improve the quality of our Estimated Values some of the Price Squad members have added 1,000s of prices from other sources including StockX. This included this sale –
Looks good until you realized that this rare Funko Pop normally sells well above $1,000. Here the last 4 sales documented –
While StockX transactions get verified (the seller ships the item to one of StockX locations where it gets checked by an expert, repacked, and send to the buyer) we assume that it is highly likely that the seller in this case was NOT well-informed (and that one buyer got very lucky indeed). This makes this transaction not suitable to be used for the calculation of our Estimated Value. We therefore decided to remove this Price Point – a decision absolutely not taken lightly (we love to give as much data as possible) but in the interest of a fair representation of values.
Comments as always very welcome!
Would it possibly make sense to have the displayed value be the median rather than the average? I think that would mitigate the effect of outliers like this. Math has never been my strong point though so feel free to shoot down that thought, haha.
It is a much more complex formula than average and takes the time elapsed since a transaction and other factors. An average would be bad for an item that sold for $1 twenty years ago and now sells for $10,000.
We also need to take into account external events (for example the death of a celebrity such as Ruth Bader Ginsberg or Chadwick Boseman) where we want to ensure we react fast on new prices and where a new much higher Price Point is not an outlier but a new baseline.
Overall creating realistic Estimated Values is hard but gets better with experience and more prices (hence thanks to all of our Price Squad members!).
Gotcha, that all makes sense. There’s a lot more going into it than I realized. Yet another reason hobbyDB is amazing 🙂
I was also thinking median value would be useful if you treated all the data equally. With your complex algorithm, that is not the case. Being a data guy, I do not like to delete data either, even if it seems like an outlier. The example $168 is a real sales value – is it still shown, or is it deleted and only the “more realistic” higher values shown? It’s fine if you delete it from the calculation, but to delete it completely from the price history might not be wise. What happens if the next sales value is again in the $100s (even if it is not likely)?
We agree and keep all data and are working on showing the Price Points and the reasons why they have been excluded separately so that users can see for themselves.
Thanks – that’s a good choice, but much more work for y’all to do! 🙂
I have added it to the list… 😉 You can always check our roadmap here.
Thanks for the roadmap. It’s nice to see your idea list (and bug list). Thanks for making them public..
Enter your email address if you like to receive notifications of new posts by email.
hobbyDB is a participant in Amazon, eBay, Entertainment Earth, and other affiliate advertising programs designed to provide a means for sites to earn advertising fees by advertising and linking to other websites. While we also accept consideration from Brands or Retailers (which is then clearly marked as Sponsored Content) all editorial opinions are our own.