I think a better approach would be using the data that is being accumulated taking in consideration the users number of self-votes, up-votes, comments, and posts publish some detailed analysis on those clearly not acting in the best interest of the community that can be manually flagged.

For example, lets say these are running stats for the past 7 days of two users that both have 5000 SP:
@user1

posts: 5
self-upvoted posts: 5
comments: 50
self-upvoted comments: 2
upvoted posts: 65
upvoted comments: 30

@user2

posts: 56
self-upvoted posts: 56
comments: 14
self-upvoted comments: 14
upvoted posts: 0
upvoted comments: 0

Who is a drain on the reward pool? Who could be accused of block chain bloat?

Because @smackdown.kitty is flagging everyone their effectivness is diminished when flagging @user2 who is just here for the money. They'll leave when they find it is no longer profitable to be here.

The flag may have a negative affect on @user1's experience here in the community. Especially if they are a new user that bought their STEEM or an older user who spent months earning it.