@Magisch We've done the implementation details in the past. I have links to previous questions in the post. A process description is fine, but I'd much rather link it in one of the bullets in this post than rehash it on meta. "How to access the dashboards" is just providing a link, isn't it?
@Mithrandir (why isn't @tripleee pingable?) How successful have you been working with other sites and getting stuff taken down? I'd be happy to provide some stats on that too.
@Magisch This is sure to be asked (and I don't know myself) - if six flags were cast on a site were moderators weren't active for a few days how would it be fixed quickly? Only thing that comes to mind would be for someone to ping an employee, but that might not exactly be fast.
@Andy depends on the site - WordPress takes nudging (multiple reports for each site), Weebly takes them down, Wix absolutely refuses, GoDaddy and Namecheap are hopeless. Google Sites is hit and miss. Medium is good. TripAdvisor takes stuff down. I haven't had any luck with YouTube.
@Magisch No idea on mod rules - but I've had spam flags expire and currently have months old NAA flags on one site. Some small sites not having a mod present for a few days to a week seems pretty common to me - sometimes I check all their last activity dates.
> Fixed a race condition that resulted in [one post on Stack Overflow][6] receiving 6 flags. On [this post][7], the error was noticed in less than a minute and automatic flagging was [stopped][8]. Flagging remained off line for several hours while the issue was investigated and resolved. During that time, SmokeDetector remained online and reported potential spam via chatrooms as usual. The post being removed was spam and remained deleted, despite the error on our part in issuing to many flags.
@Magisch How's that? It's a brief summary of the problem, resolution and ultimate outcome.
The post has been updated.
And I just noticed @FrenzyLi 's list of improvements...I'll do those soon
@Makyen That pattern looks like it's already caught by Potentially bad ns for domain in answer and Potentially bad ns for domain in body; append -force if you really want to do that.
For 5 autoflags, have we ever discussed something like 4 flags, then another a minute later if the post is still alive and has no fp feedback? That'd achieve the goal of killing long lasting spam just as effectively, and with a lower chance for 1 human flag deletion (even though the odds of a bad post getting 5 flags is > 0.01%). And, if people are OK with that, it'd be easy to extend to 6 flags.
I don't know if we had the minute delay discussion. There was a discussion about doing the initial wave and then waiting for a confirmation TP to finish it off
There are certainly outliers, which we definitely want to fix. Having delayed flags would help alleviate that, but I don't really see how it's philosophically different from just waiting to get tp feedback from someone?
@thesecretmaster I think you misunderstand what people have been claiming is sockey. The issue was that casting autoflags was predicated on anyone casting a flag/feedback. Doing so effectively augments their flag. That's the sockey part.
But if nobody feeds back, the flags are cast anyways. The change here is from a human flag triggering an autoflag to a human being able to stop an autoflag. It's sockey in the same way !!/stopflagging is socky.
Personally, I think it's easier to sell and better for us to do it the way I suggested, but any human confirmation is good. The 1 user nuke will be hard to sell without emphasizing that it will be reviewed.
@quartata Yes, there has been conflicting feedback on autoflagged posts. The search on MS will significantly under-report how often it occurs, due to people going back and changing their feedback and/or conflicting feedback being invalidated. The most recent one that I'm aware of was 2 days ago.
@quartata Unfortunately, that essay writing service has pages with some of their essays. They are used as references in some posts, including by moderators.
@Magisch Assuming each flagger is independent, and taking the SO human accuracy of around 89%, there's an 0.0002% chance of that happening - which is less likely than autoflagging going wrong, even at 100.00%.
@quartata Oh, I agree that it was most likely spam. At best, it was a very poor attempt to answer with what is, at best, a dodgy source/reference. Personally, I custom mod-flagged to let the site moderators make the choice. It was deleted, and the custom mod-flag marked as helpful.
@Undo Good info. Thanks. How many do we not have delete times for? Also, it'd be interesting to know how many of that ~20% have feedback come in very soon after being deleted (e.g. a moderator seeing the SD post and spam deleting the post (e.g. using FDSC)).
@EriktheOutgolfer That pattern looks like it's already caught by Potentially bad keyword in answer and Potentially bad keyword in body; append -force if you really want to do that.
@EriktheOutgolfer That pattern looks like it's already caught by Potentially bad keyword in answer and Potentially bad keyword in body; append -force if you really want to do that.
I flagged this answer as spam, and I'm pretty confident it is - I've seen three such posts from that account in the Late Answers queue over the last hour or so (and the first two were deleted after I flagged them) with no other non-deleted answers on their account, it's a brand new library so the...
