According to Shog9's post, there are two and only two reasons why votes are removed when an account is deleted:

  1. Some thresholds relating to vote count and number of people affected are not breached.
  2. The aforementioned thresholds are breached, but the account has exhibited fraudulent activity at some point in its history.

It's easy to say differentiating these 2 reasons is unimportant. But some users experience arbitrary black-box shocks to their reputation, while others just see "more comfortable" 50-200 reputation losses.

Often it's observable in a popular tag a whole bunch of users who frequent that tag see a correlated loss, presumably because the deleted user was active in that tag.

So my question is: while it's important not to publicize the thresholds, wouldn't it make sense to differentiate between the 2 reasons above in reputation history? In other words, improve the messages to:

  • User has been removed: thresholds not breached
  • User has been removed: fraudulent activity

The benefits are:

  • Some broad user-based oversight that mods aren't fat-fingering their deletions. If we see dozens of users in a tag losing 100-200 rep on the same days on a regular basis due to "thresholds not breached", something's gone wrong.
  • Greater confidence in the process. Right now, it's a black box mechanism, usually a negative shock to users' reputation.
  • 3
    Effectively, this would boil down to two states: an employee was or was not involved in the deletion of a user. The thresholds only prevent moderator or self-initiated deletion from kicking in immediately, and send things to employees to review. I don't know if this would clear things up in the way that you'd like, because large point reductions like the one linked can still result from simultaneous deletion of a large sock puppet ring by moderators, without employee approval. – Brad Larson Dec 10 '18 at 16:50
  • @BradLarson, Thank you for the clarification. My understanding of Shog9's post is that it applies to all user deletions. That was not mentioned or implicit in the post, but the argument that non-employee approved deletions are not covered changes the problem completely. – jpp Dec 10 '18 at 16:52

I'm not sure it would be so great an improvement to get the extra information (I would be much more interested in which of my posts are affected, not that I'm requesting it). What could I do with it?

But it would have some drawbacks:

  1. Resources for this change could be put to better use on other features.
  2. With enough data, people may find out more about how high the thresholds are and then possibly abuse it to hide fraudulent activity.
  3. Sometimes you may find out, whose account was deleted and then you could conclude that they were engaged in voting fraud, although that's usually none of your business (it's between them and the moderators, especially once it has been handled).
  • These are good reasons, thank you! Probably (1) is the greatest sticking point. (3) was never my intention, for the record. For "what could I do with it", for very low probability events, e.g. top 10 users in a tag all losing ~50 reputation on the same day, you can query the reason. – jpp Dec 10 '18 at 13:23

What you're suggesting won't be helpful. Although not mentioned in that post, the documented process does not cover users deleted without employee approval. This would occur, for instance, in the case of a large sock puppet ring.

So, ultimately, the extra information will only be partially useful. Since there's no plan to reveal precisely the reputation lost to sock pocket rings being found and closed, there's limited added value in adding a feature only for cases when the user deletion is moderator or self-initiated.

  • 5
    FWIW, all deletions are subject to employee approval if they breach the thresholds alluded to in that post. The tricky bit with a large ring of sockpuppets is that there's a good chance no one sockpuppet will breach those thresholds (the puppetmaster does this intentionally to avoid detection), but in aggregate the full set of puppets represents a large number of deleted votes. On the flip side, a user engaged in voting fraud whose account is deleted may have voted for numerous innocent users as well; these votes aren't fraudulent, but they're not preserved because I can't separate them. – Shog9 Dec 10 '18 at 20:32

You must log in to answer this question.

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