We've occasionally talked about the voting procedures of the Academy Awards for Best Picture. In short, they recently changed to using instant runoff voting, and I think that was a pretty poor choice; they should use range (AKA score) voting. So imagine my delight when it was pointed out to me that the nominees for this year's award for best visual effects will be determined by re-weighted range voting! (Rule 7b.)
I'm not entirely sure that a method of proportional representation in the best way to pick nominees for a single-winner election, but this is sill fantastic news. I wish I knew how the idea of RRV made its way into Hollywood, but regardless of how it happened, I'm ecstatic that it has.
Re-weighted Range Voting uses the same ballot as normal range voting, namely, each voter can give each candidate any score within a given range. The first winner is determined the same as well, it's simply the candidate with the highest score. The next step—and the way in which the method becomes proportional—is that each voter who voted for the winner has all of the scores on their ballot re-weighted downward, based on how highly they scored the winner. If you gave the winner the maximum score, the rest of your ballot's weight is cut in half, and the lower the score you gave them, the more strength your ballot retains. And then the second winner is whichever has the highest re-weighted score. Each additional winner further re-weights every ballot (if you gave the maximum score to the first two winners, the rest of your ballot is cut to a third, and so on) so that each time a voter gets their way, they have less say on the following rounds.
RRV is, of course, perfectly compatible with an approval-style ballot, since approval is equivalent to range, with a "range" of just 0 and 1. Re-weighted approval could therefore be a very easy and effective way to implement proportional representation.
