Fantastic idea! looking forward to it.
My thoughts though - the terminology sounds counter intuitive - a regression rate of 100% should be full regression (bad), and a regression rate of 25% should be much better (good) - but this is just semantics, lol.
Just a comment - I like what you are doing here, but I feel that everyone should get a chance to be a guy that is able to hang on to his skill longer - not just over 80's - and I suggest an algorithm for that as well (perhaps far less likey for the under 80's, but not impossible), just my thoughts. (also that being the case, maybe even include the flash in the pan types too - allow for it: guys that seem decent early then just crash to further flesh out reality). -- and I would suggest to always error on the side of too few rather than too many regarding statistical outliers.
Anyways, keep doing what you are doing - this utility is absolutely MONUMENTAL to the game, imho.
