I've commented before about how I felt it was unfair that the votes of users with higher SP had a greater weight when it comes to witness votes, but @lukestokes introduced me to the concept of sybil attacks & game theory which helped me understand why it might be necessary. The stats you have produced regarding the numbers of users who aren't casting witness votes are interesting.
I would be really interested to know, (were you to be given proxy voting privileges for all accounts you identified) how much difference it would make to the top 30 witness list. What number would you be? Would the combined power of all those minnows, trump the power of the few whale witness voters?
I'm about to set you as a proxy. Perhaps a few hundred thousand more will too, and we can see the results of the experiment!!!
Jimbo thank you very much for setting me as a proxy! According to http://www.steemreports.com/witness-voters/ the power of thousands of smaller voters generally ads up to one massive upvote from a few of the top voters. For example, the top witness @gtg has thousands of smaller votes which equal his one vote from @freedom and are slightly more than his one vote from @dan.