if a ring introduces noise (meaning they attempt to avoid the "not clever" trap you describe) how does this effect the measurement? I can assume that introducti

if a ring introduces noise (meaning they attempt to avoid the "not clever" trap you describe) how does this effect the measurement? I can assume that introduction of noise is somewhat counter to the greed and objective of the ring leader. But now the objective would change to be meerly introducing enough noise to putting the metric on this ring in a state of question. With that though, of course this is better than a model that requires no noise. Forcing noise from the edges reduces the set of entries requiring human review and, still, human review might reveal with sufficient weight the existence of a possible ring. But, this is an evolution to the cat+mouse stage. Hence, rings could still survive if clever? With that question (ring survival with measured noise) set aside, it seems if the cat+mouse were to come in earnest that reputation dependent models would be a driver for mapping user "types" (social, political, behavioral models) in a much more accurate manner than the current driver of advertising (google wanting to increase price of adds)