As of July 2013, Very Low Quality flags automatically add the flagged post to the Low Quality review queue. As of April 2014, Not An Answer flags do the same.
Each review task requires a minimum number of "Looks OK" reviews (2 on SO, 1 everywhere else) to be dequeued; this number increases by on...
Given the point that I'm judging before adding extra flags, it's doing no harm at all.
One can even argue that an extra flag on a truly NAA post can prevent an extra robo-reviewer before the post's wrongly dequeued, but that argument is a bit irrational.
@Undo How would you feel about a PR with a migration which adds a post_type field to every Post record based on link? And the fixes everything else to use that?
This way, we know that (e.g.) "6k posts were flagged by at least one C member", or "100 posts were flagged by 6 C members"
since "posts flagged by at least 3 Charcoals" is a subset of "posts flagged by at least 1 Charcoal", it makes sense to put the 6/5/4-flag bar under (below/beneath) 3/2/1
@ArtOfCode It might be useful to say "~n% of charcoalers flags aren't on this graph", a stat that you could get by checking number of charcoalers w/ > 200 feedbacks & api flags on divided by number of charcoalers w/ 200 feedbacks.
@Mithrandir Yeah, he had a jam sandwich for lunch and just can't get that last bit off his lip. He only made jam because he'd run out of ham and he spent the rest of the day being irritated about it and trying to buy some ham on the internet.
> That's mostly down to users individually choosing to use a certainty threshold lower than the recommended settings, which are 100% accurate to the nearest 0.01% - that's up to individual users, but, again, all certainty thresholds must be above 99.5%.
I know what that means...but it's worded uninutively
> We allow users to set their specific flagging conditions, provided they don't go below our 99.5% certainty. We, however, recommend a higher value of X.XX%. Posts that fall between those two ranges will see more false positives.
> We allow users to set their specific flagging conditions, provided they don't go below our 99.5% certainty. We, however, recommend a higher value of X.XX%. People who set their conditions under X.XX% may see more false positives autoflagged using their accounts than those with a condition of X.XX% and over.
> We allow users to set their own flagging conditions, provided they don't go below our baseline 99.50% certainty. We recommend, however, a higher value that has a certainty of 100.00% - those who set their conditions below that are likely to see more false positives flagged using their account.
@ArtOfCode I believe you were involved in the last MSE post too. It should probably be someone that was involved in that, just for a bit more continuity.
