40

I propose that high-priority flags, particularly spam flags, be refunded as soon as they are deemed helpful.

For instance, on sites like Server Fault, which see less frequent activity, obvious spam posts can remain visible for longer than they should.

By refunding helpful spam flags, we could keep fighting spam more effectively.

Example: the top 5 questions are spam, with the 5th one already being an hour old
enter image description here

Improve this question
5
  • 13
    It only took me about 6 hours to burn through 100 SF flags on spam today. Commented Mar 21, 2025 at 10:12
  • 3
    I would argue that the other flag types could also be refunded, assuming they are marked helpful withinin the 24h window. I see no real disadvantage for a helpful flagger to have more flags at their disposal. Commented Mar 21, 2025 at 12:40
  • 3
    @Velvet Yes, just wanted to highlight the fact, that a lot of flags do not get handled within the 24-hour window, which means they won't affect anything we are discussing here. Commented Mar 21, 2025 at 13:30
  • 2
    We get a rep refund when a downvoted answer gets deleted, so it would be consistent to get a flag refund when a spam or R/A post gets deleted. Commented Mar 22, 2025 at 3:20
  • 3
    @PM2Ring even more appropriate analogy is that we get a vote return if a post is deleted. And that only counts within the same day. So, you can still only vote 40 times total a day in the end on active posts. But if you vote today and the post is deleted tomorrow, you don't get the vote back, so you can't vote 41 times on active posts. Commented Mar 22, 2025 at 6:14

6 Answers 6

Reset to default
35

I like this idea. Let's be honest - our flagging system isn't keeping up with the sheer levels of spam, and flag limits let people ease into flagging. If you're doing valid flags in quantity and quality, putting roadblocks on virtuous behaviour does not make sense. Especially with folks chipping in on sites they're not regulars on, or donating flags to SmokeDetector/Charcoal - we can do a lot more good if we don't have people running out (and on a bad day, people can use up over 100 flags).

While we can get flag limits adjusted, this seems more efficient, especially since spammers seem to be able to fire up dozens of accounts, and tank ~10-20 flags each before getting question/answer banned, or mauled by a passing moderator.

Improve this answer
20

I have reservations. This should have more conditions than just the flags being marked helpful, to avoid abuse of the flags themselves, because the flags can be marked helpful by non-mods. I'm not sure what conditions would be best. Secret algorithm? Limit to situations characteristic of spam waves, such as the helpful flags need to be on posts that were posted recently by low rep users?

I would rather see more measures to limit the amount of spam that can be generated by a single individual or organization. Right now it's a whole lot.

It is good that the refund idea would let the manual defenses ramp up to match a flood of spam, and that it could do so even on small or slow sites.

Improve this answer
9
  • 13
    there's some merit there to having additional criteria Commented Mar 21, 2025 at 13:05
  • 3
    One criterion could be that none (or very few) of the spam flags are disputed or revoked. Commented Mar 21, 2025 at 13:21
  • 4
    How about limiting it to people who've cast a threshold number of helpful spam (or rude/abusive) flags, say 30 to 50, with a very low rate of disputed spam & R/A flags? Commented Mar 21, 2025 at 13:38
  • 2
    @Velvet Do you mean declined instead of revoked? I think dispute would be OK, but a decline should count as a criterion instead. Commented Mar 21, 2025 at 13:39
  • 2
    I might be missing something here, but how exactly could this be abused? It would require at least 4 people to be working together to cast spam flags on the same posts. Is this something that ever happens currently? Commented Mar 21, 2025 at 14:19
  • 12
    @cigien yes, it's possible to create a sockpuppet network and rapidly upvote content from socks to allow them to cast flags, then use those accounts to cause mischief including flag nuking content. The sockpuppet network can be caught eventually but it can still cause havoc before it's stopped. And the perpetrator(s) can just repeat this the next day, if they so wish. We've seen similar activity already. Commented Mar 21, 2025 at 14:25
  • @VLAZ Ah, interesting. I hadn't realized that. Makes sense to have guardrails then. Thanks! Commented Mar 21, 2025 at 14:26
  • 10
    Additionally, I think that comment flags should be excluded entirely. We don't see comments being abused like posts are, at least not in a way where having more comment flags is necessary. (For example, you can just ignore spam comments under spam posts; it's handled well enough by deleting the post.) Commented Mar 21, 2025 at 14:50
  • 3
    @Laurel you should definitely write an answer explaining how and why comment flags should be excluded from this FR (it's already status-review). Oh, and I totally agree, by the way. Commented Mar 21, 2025 at 23:25
15

This absolutely makes sense to me and looks like really a spot-on solution instead of generally increasing per-user flag limits.

As mentioned by Journeyman Geek there're dedicated users who flag stuff on sites where they otherwise don't have a lot of activity. This is possible thanks to association bonus they can get on any SE community by participating on one of them, which enables them to flag (among other privileges), but not to have high flagging limits.

Though the limit of how many posts you can flag varies and can be as high as 100 flags, it still takes some work from user to reach high thresholds (+1 extra available flag per 2,000 rep gained or 10 flags deemed helpful is a long road). Although it feels logical to have this kind of limit mechanism for all other types of flag, spam is kinda different and may arrive in big amounts overwhelming flagging abilities of active users.

Improve this answer
12

The whole point of the flag limit is to stop abuse by, for example, new users, who don't fully know what should and shouldn't be flagged appropriately (or by the occasional troll).

But, if a flag is deemed as helpful, then there should be no reason as to why that user can't 'bypass' the limit, which isn't really even being bypassed because that flag is simply being refunded.

If someone — somehow — manages to use all of their flags unhelpfully, then that's on them. The rest of us should still have the opportunity to help.

Improve this answer
4
  • 10
    Potentially this could even allow flag limits to be lowered, increasing reactivity to abuse, without meaningfully impacting those who do flag appropriately. Commented Mar 21, 2025 at 12:28
  • 4
    Thanks for sharing this POV. To understand the concern a bit about "new users, who don't fully know what should and shouldn't be flagged appropriately" was that a problem before that the flag limits were designed to address? Or was the limits set up to preemptively stop potential issues from new users? (I'm not totally familiar with the history, so apologies if this is a well known issue with new users) Commented Mar 21, 2025 at 15:31
  • 1
    @EmmaBee Honestly, I'm a new user too, so I'm not too sure whether it used to be a problem, but I'm pretty sure the flag limit has indeed fixed it, if there was anything to fix. I highly doubt that this feature request will change that, because, as I said, if a flag is helpful, why not refund it, effectively removing the limit for helpful-flaggers. Commented Mar 21, 2025 at 21:49
  • 2
    @EmmaBee I've never come across a post explaining why 100 flags per day was chosen as the limit. The closest thing I've found is this blog post announcing the change: stackoverflow.blog/2011/01/14/improved-flagging Commented Mar 21, 2025 at 23:04
4

I think this solution is to the wrong problem. When there isn’t an ongoing spamwave (or a troll on the loose) even 10 flags per day is usually plenty.

If users are seeing more that 100 spam posts a day (on any site) the problem is not that they don’t have enough flags, it’s that there is way too much spam getting posted in the first place.

The current spamwave has been going on for several months. Something needs to be done to stop the spammer from creating accounts and posting spam at such a fast rate and in such high volume.

Improve this answer
4
  • 4
    Closer to a year or two, though the ramp up was recent, and the impact is variable depending on total available flags. Commented Mar 23, 2025 at 7:13
  • I can't get on board with this. It would be great if less spam made it onto the site, but that's not an argument against this proposal. It's possible to tackle spam from multiple angles. Commented Apr 15, 2025 at 9:07
  • @Anerdw My concern is that a small number of coordinated malicious sockpuppets can validate their own flags (even on perfectly good posts). Refunding would effectively mean unlimited flags, which could cause large-scale damage very quickly. Commented Apr 15, 2025 at 9:31
  • @ElementsInSpace That one I can see. It's a different argument, though. Commented Apr 15, 2025 at 10:12
3

It's always the case that prevention is better than cure - especially with spam, and especially when cleanup is manual vs automated spamming. That seems to be a separate question?

An alternative for flag refund that addresses @DanGetz

I have reservations. This should have more conditions than just the flags being marked helpful, to avoid abuse of the flags themselves, because [...]

If those extra conditions are not yet easily discerned, then here is an alternative.

  1. Spam flags are stored in the database fractionally, and rounded down to present to the user. (You have 19.325 flags, you see and can use 19)
  2. Spam flags are refunded fractionally as $SFRF, by some metric which can start off global and be tweaked as needed.

This answers the question by being a superset of yes and no answers,

  • When SFRF=0.0, the answer is no. This can be for and user, site or period of time.
  • When SFRF=1.0 the answer was yes.
  • Values in between can be used to introduce the feature gradually and see the impact; or if spammers find a way to exploit it for trolling when SRFR=1 then to reduce it a little and see if that incentive goes away.
  • Values SFRF>1.0 are possible if needed.

When dichotomies are troublesome, try using real numbers to break them.

The obvious disadvantage is complexity. (Hiding complexity with a layer of simplicity is not the same as avoiding the complexity!)

Improve this answer

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.