Thanks for the analysis!
There are probably a variety of potential (and unpopular) ways to try to reduce attendance bias and afterglow in the ratings. I wonder what effect a "cooling off" period would have - not allowing a show to be rated until a few days or a week has passed after the show.

As far as biases over time periods of a few years or longer, such as those of newer phans coming to .net, I think those could only be controlled for as part of a statistical analysis. I want to delve into this as soon as I get some free time.