Beyond the problem of not counting votes as contributions (the knights of /new should be rewarded), I feel this scheme has a major downside in using the Wilson lower bound: namely, it rewards quick comments.
It's well known that the number of votes a submission or comment gets is very time-sensitive, and the Wilson bounds are substantially affected by that number. This leads to little-seen content gaining their author very little influence, which may be purposeful but probably isn't.
A further little-mentioned complication is the existence of thresholded voting: some people distribute votes freely, while others manage their vote histories by limiting themselves to voting on exceptional content, and yet others use contextual voting, touching only misrated content. Given the dependence on voting, I feel this aspect has not been sufficiently investigated.
It's well known that the number of votes a submission or comment gets is very time-sensitive, and the Wilson bounds are substantially affected by that number. This leads to little-seen content gaining their author very little influence, which may be purposeful but probably isn't.
A further little-mentioned complication is the existence of thresholded voting: some people distribute votes freely, while others manage their vote histories by limiting themselves to voting on exceptional content, and yet others use contextual voting, touching only misrated content. Given the dependence on voting, I feel this aspect has not been sufficiently investigated.