You are operating from the assumption first of all that these ratings stop scams from happening. That is arguable at best.
I never made that assumption. I simply pointed out the number of negative feedbacks being left which, under your system, would each require discussion.
If people aren't even going to read the person's ratings having a red mark in a sea of red marks is not going to signal much and actually ends up providing cover for them. Furthermore that level of complete lack of due diligence guarantees a user will eventually be robbed regardless of what anyone else does.
I see your point, but we have no evidence that too many red ratings provide cover. I agree there are a minority of users who will robbed/scammed regardless, but that's not an argument to remove the ratings from the majority of users who find them helpful.
Second you are assuming that every one of those ratings was valid, beneficial, and needed to be made.
Again, I never stated that. Those ratings could all be nonsense, but they would still all require discussion prior to reaching that conclusion under your system.
Also the whole point is there are less negative ratings left.
I would like that too. I think there are too many negative ratings left (i.e. all of them) for differing opinions or points of view. I just don't think the system or methods you have outlined are going to get us there.