I'm digging up this old thread because I'm interested in building a predictive statistical model, or econometrics type model, to estimate total print runs of vintage sets. It should be doable if I have a couple of known, or semi-known, sets that I can start with. I'll have to control for multiple factors, but I think it should be doable using pop report data as long as I can build a separate ensemble model that predicts the likelihood of a card being submitted in the first place (very doable, as the likelihood of a card being submitted for grading depends on demand and survival, as well as likelihoods of resubmission). From there, I should be able to estimate total print runs that at least pass the smell test.
Has any progress been made on this project in recent years?
__________________
If it's not perfectly centered, I probably don't want it.
|