We have been going down a giant rabbit hole here trying to understand all of this… and it goes so deep.
Only way to count flags for trust level is for spam_count or inappropriate_count to be larger than 0. So this whole topic started from a bit of a misunderstanding…
- when a post is flagged as spam/inappropriate count goes up.
- if mod says … oops this was a bad flag… count goes down.
You always could do whatever you wanted with “something else…” flags, and it would not count towards TL or anything. Only impact was in the UX on the user page. Where you actually may want to find this information…
Which also opens the hornets nest that was being opened in the OP.
something else... this post is terrible and abusive
vs
something else... this post is fantastic person needs a badge
In both cases the flag is helpful
in the first case the flag is an indicator of a bad post and arguably should be a TL penalty.
in the second it is not.
I get this is all massively complicated, overly complicated, but I am not sure about the OP here. It is proposing adding even more complexity on top of this complexity:
Arguably our TL algorithm is broken at the moment… Admin said
Hey, something is wrong with this post per big red box
But we just did nothing with that info.
The incoherence here is around:
- Is the review flag and actions on flags about the post being bad vs good
OR
- is the review flag and actions on flag about the flag and the flag being useful or not
Breaking it down a bit more:
- What should count as penalty for TL?
- What hyperlinks should we show on a user page? (every time a user was flagged regardless of outcome? every time a user flagged regardless of outcome? outcome based links flag is good vs bad?)
The more you peel this onion the more layers there are and the more complex all of this feels.