Cryptoin, how are you factoring in avg. chips for the unaccounted? is that avg counted as orders that are not seen in numerical order" like 1 2 3-6 7 (5 and 6 get tagged with an avg chip count?) i may just be having a blonde moment here,,,lol
Basically I have an average chip count from the orders that I have data from. From there I apply a reduction in the average chip count based on the % of orders that are cancelled, unpaid and just missing due to order number jumps. For example.
Avg. Chip Count = 200 chips
% of Orders Cancelled (based on existing data) = 25%
% of Orders Unpaid (based on existing data) = 5%
% of Order #s unaccounted for that do not exist period due to number jumping (based on the statement that Terrahash made that all the orders placed through 6/19 will be covered by 40,000 chips or less) = 80%
Now let's assume the order #s started at 1 and went to 100 and we only have data for 50 of them. So there are potentially 50 other orders but we know this not to be true due to the order number jumping. So we would assume 80% of those 50 orders are non existent. Leaving us with 10 actual missing order numbers. Of that 10 we can assume about 25% we're cancelled and 5% are unpaid. That leaves us with 7 orders at an average order amount of 200 chips or 1400 chips for 50 missing order #s which gives us an average of 28 chips per missing order #.
The way my formulas work is they apply this average chip count per missing order number #.
Wow thats a great breakdown of your formulations. One of the reasons I ask, is in my case, I had a total of 9 order invoice numbers created durring my attempts to get paypal to go thru, begining with order # 502 and then randomly from there more were created.. #'s. In the end only 2 were actually processed. The remaining 7 invoice numbers starting with 502 and so on, showed cancelled, and eventually were purged from the system it seems, never to be used again. So those invoice numbers became not existant in the end. So are you applying the 28 chip count to those or are those not counted at all in your tabulations? and Im sure others had the same experience when placing orders with paypal to some extent. Simply put, is my first attempt order # 502 anywhere in the accounting? ok, thanks and again, keep up the great job you are doing with this, since we have nothing else to lean on right now
Well those should be covered by the cancellation %. The 28 chip count is an overall average for a missing order. This average takes into account non-existant, unpaid, and cancelled orders.
If I only have 50 of 100 numbers that remaining 50 I am missing is assumed to have 28 chips per order. In reality, there will actually only exists around 7 orders with an average 200 chips for each of those 7.
In regards to the accuracy of the cancelled order %, the assumption here is that people are reporting their cancelled orders to me. It is likely that there are a higher percentage of cancelled orders than I account for since it is less likely that people will report them.
If that is the case my numbers will be slightly high but the affect would still be rather negligible.
If you would like to know how accurate the prediction is for your particular order I suggest referencing the % of total orders accounted for. The higher that percentage is the higher the accuracy of the prediction.
All in all you can pretty much guarantee at this point that the first 10,000 chips will be used around the 500 order mark. 20,000 chip order mark will likely fall in the low 600s.