@Undo Are you suggesting that we could tap into what's not reported, but still deleted as spam? - That's right. I don't know how feasible that is. But basically there are two lists (deleted spam, and Smokey-detected spam). A comparison would show what was deleted but not detected. You could also catch false positives that way as well (Smokey-detected but not deleted).
@NickGammon We can already do the last thing, we have data on what's been deleted and what hasn't (within a certain window, but the vast, vast majority of spam is deleted within 20 minutes)
The first thing, if it were done network-wide, would require a level of integration with SE that we don't have yet
[ SmokeDetector ] URL in title, bad keyword in body, bad keyword in title, link at end of body, link following arrow in body, pattern-matching website in body, pattern-matching website in title: http://slimbodyketone.co.uk/dermagen-iq/ by user21609 on arduino.stackexchange.com (@NickGammon) …
1. [:30035937] <processed without return value> 2. [:30035922] That message is not a report. 3. [:30035880] <processed without return value> 4. [:30035840] <processed without return value> 5. [:30035836] <processed without return value> 6. [:30035772] <processed without return value>
[ SmokeDetector ] URL in title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, link at end of body, link following arrow in body: http://klereumcol.com/zymax-male-enhancement/ by user21616 on arduino.stackexchange.com (@NickGammon) …
@sandwich I really like your Spam tRacker extension, its really useful, even in the "on" mode, It gets more false positives that way, but sometimes, it also provides interesting posts to read
> Bad keyword in body, bad keyword in title, bad keyword in username ---------- Title - Position 15-32: Live Streaming</a Username - Position 15-32: Live Streaming</a Body - Position 15-32: Live Streaming</a
Someone is asking about URL manipulation and the guy recommends a stock investment advice firm. The first recommendation before the "...or you can use <$myproduct>" is from the OP
@sandwich did you mind doing me a favour and check if the Roomba ran today on Math? It didn't on EE so wondering if it was just there or more widespread.
How would y'all feel about throwing every "Offensive {} detected" post with tp feedback back into review? I have a sneaking suspicion some of them are wrongly tp'd.
> Even if a post is a bad post for some reason or another, it is probably not offensive. The Offensive flag is meant to be used only in extreme cases, like hate speech, abuse against people, or abuse of the system.
For example, if a user posts obscene images to the site, that should be flagged as offensive. But if someone says something bad about your favorite technology, that probably doesn't apply.
As a rule of thumb, if you can't justify something as being hate speech, or abuse, you shouldn't mark the post as offensive. Instead, you should downvote the post.
@ArtOfCode I think its worth noting down What to do with specific categorys of posts, like "Blatant spam? k" "Inappropiate self promotion but not quite there? Maybe not k"
Stuff like that
Maybe bring the new feedback categories while you're at it
Suppose a man is sitting in a train. Man is moving with train till the last journey as man is in contact with train. BUT WHAT ABOUT FLY which is also moving along with train with high speed. HOW IT IS POSSIBLE? (if fly is not sited anywhere till last journey).
