My answer for the vanilla preconditioners question gets like 2 spam edits a day? Just curious as it is a majority of my notifications per day.


Here's the guidance I've gotten from other mods in Stack Exchange so far:

  • Edits of 400+ characters that are anonymous and substantially change the content of the post are auto-rejected.
  • The typical vandalism edits don't meet that threshold; they're too short.
  • One way to flag those edits for autorejection is to lengthen them.
  • However, it still appears that manual intervention is needed.

If it keeps happening, I will escalate the issue up the chain at Stack Exchange. Thank you for bringing it to our attention!

  • $\begingroup$ Thanks. They keep rolling in :) The above comment from @Godric suggests that there is a dashboard somewhere for edit tracking? Is that a mods only panel somewhere? Much thanks. $\endgroup$ – meawoppl Jan 14 '14 at 22:50
  • $\begingroup$ Might be worthing checking again. In the last two days (or there abouts) I have rejected this edit twice because it appeared in my review queue. And it just came up again, it's visible now as meawoppl says. $\endgroup$ – boyfarrell Jan 15 '14 at 3:07
  • $\begingroup$ Screenshot of the review panel i.imgur.com/9FgqrsC.png $\endgroup$ – boyfarrell Jan 15 '14 at 3:15
  • $\begingroup$ @meawoppl To clarify something Geoff wrote: anonymous edits that change too much of a post (like those spam posts do) are automatically and silently rejected, but only on posts that are more than 400 characters long in the first place (because otherwise “change too much” is too imprecise). So you can protect a particular post by fleshening it out to be >400 characters long. $\endgroup$ – Gilles 'SO- stop being evil' Jan 15 '14 at 12:29

We're seeing this everywhere on the network right now, and have systems in place to quickly block this as it surges. The problem is, it's coming from some very large distributed networks.

The good news is, by the time a few have been rejected as spam, the system has been trained on them and they'll quickly be blocked. This data is shared across the network, so if someone goes and gets themselves blocked for spam edits on Ubuntu, they'll also be blocked here.

Much like your inbox, however, once evil has been identified and filtered, there's still some evidence of what got them identified in your inbox that needs to be cleaned out - depending on the scale of the attack. You can rest assured that rejecting these edits as spam is having the desired effect, but even after being blocked some cleanup is still needed.

We're looking at ways to reduce this a bit, and I'm going to be examining the weights that scicomp carries in the global space and perhaps increase them a bit, but you're handling it quite well.

Just continue to reject these edits as spam, as you have been, and the system will quickly learn from you. If they go to the trouble of making accounts, even registered ones to do the same thing, it's important to spam flag the posts, and for moderators to ultimately destroy these accounts as spammers.

For the meantime, I've gotten into the system and found the hosts that were on the verge of being blocked for activity here and helped them along firmly into the blocked state. I'll continue to keep an eye on it here as I work to address this at the network level.

When we rolled this system out, it had profound and immediate effects. I envisioned two outcomes, the first being that spammers would simply opt to move onto easier targets, or perhaps a second outcome where spammers saw success here to be worthy of premium rates, which unfortunately seems to be the direction they're heading with increasingly sophisticated methods and networks.

Still, we're handling it quite well, and learning quite a bit as we do so. I'd personally like to thank them for their unwitting cooperation when it comes to making this system better, we couldn't have done it without their help.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .