@GreyWolfNH - in theory you are right, but it only really works if everyone (or at least a large majority) use the same set of criteria to rate the cards. Unfortunately this is not the case and people have been rating cards as 0 simply because there was a ***** feature that they did not like - for example taking shoes off, or keeping them on - or because they (but nobody else) had a failure to download a clip. Also some people give 0 stars to cards that they got as part of a random TGIF pack.
All the more reason we need the largest possible sample size, it will eliminate the likelihood of statistical outliers ruining a card's overall rating.
These days I completely ignore the rating and I would suggest to others they do too and make you own judgement about the cards.
When I see a card that has a rating lower than 4, I usually investigate why before buying the card. Most often there are technical issues with the card itself, such as faulty sound or video clips causing the low rating.
Of course, I ignore the rating for any card that has been released for a short time, as it takes many hundreds of ratings to arrive at a good average.
As was pointed out earlier, after only five ratings, a card had a seemingly unfair rating of 1.5 applied, but after 82 votes, the same card is sitting at a rating of 4.67, which I'm sure is a more appropriate rating.
This is my point, as the sample size increases, the rating more effectively shows the overall opinion of the card itself.
This is further complicated by the fact, provided by @Rex, that over 90% (I think he said 95%) of all ratings were for either five stars or a zero stars, The result is that almost all cards have overall ratings between 4 and 5 (and would all cluster very close to 5 if you could filter out the "anomalous" awards of zero stars) so if you want your own rating scale to be compatible with the overall ratings you can pretty much only use 4, 4.5 or 5 star for most of your collection - this is too coarse grained if you are trying to put your entire collection into some sort of order.
It would be a very coarse grained system if everybody only ever used a 4.5 or 5 star rating, but if you search through the entire iStripper library, ratings run the entire gamut of numbers available, admittedly clustering far closer to the 4-5 star range. However, each rating is presented with two decimal places, effectively giving a zero-one hundred range between four and five.
Aww geez, someone stop me before I statistically analyze every card on the platform.
I'd better stop before I cringe.
😰 
