I've actually been emailing other academics personally behind the scenes, those who study commons systems or networks, to what they think about different rewards rules, participation rules, etc. Linear versus exponential for consensus, reputation distributions, stake distributions, etc.

Most amazingly enough haven't heard about steem, despite steem having the most transactions of any blockchain and being the only large consensus economy.

I've been collecting comments like this, analyzing. It's not obvious what is the best rule, because structured systems are not intuitive. Caffeine blocks receptors, inhibits parts of the brain . . . which inhibit other parts of the brain . . . so it ends up activating, stimulating.

Voting on someone over 70 at 100% only gives that person a 10% vote in STU while a 100% vote on a person with a rep under 40 might count for 100% of the worth. Obviously the scale could be adjusted for lower voting percentages.

The scaling for vote percentages would be the key. F(x1, x2, ..., xN ; k) = x is dominated by k if the x vector is monotonic improvement related to F, and k is not.