Is this actually a problem? If we find that Aurors seem to be abusing the procedure, then we'll modify the procedure to prevent its abuse.
This seems to be a reasonable heuristic for cases like this where the potential problem is as silly as someone would get undeserved LW karma. But I would be careful of generalizing it to cases where the potential problem is serious.
Also, in the HPMOR scene you are alluding to, Bones was saying that it was up to her to decide if the procedure was being abused, not up to the aurors whose job it was to implement it, and that the downsides of it being abused were insignificant compared to the downsides of not following it when it is needed.
Bones was saying that it was up to her to decide if the procedure was being abused, not up to the aurors whose job it was to implement it
The specific error the Auror was making (if we are to take a consequentialist look and ignore the settled rules) was to act on a hypothetical problem that turns out not to be real, so Bones's superior expertise would be used, in particular, to determine whether the hypothetical problem is real.
I don't think many people actually care that much about karma. To care that much about karma yet still be willing to deliberately make mistakes that will be corrected seems like an unlikely combination. The positive good that comes from encouraging people to retract errors is likely a better thing than the tiny chance that some people will deliberately make such mistakes to gain more karma.
Regardless of whether people care enough to do this to get karma, I definitely don't care if others do it. It doesn't impact me if someone has sneakily gained 5, 10, or 100 points of karma.
I tend to think that admitting one's mistakes will cost status/social capital/perceived reliability in any case, so giving the person karma hopefully cancels out that disincentive.
This seems like a disguised disagreement about whether karma's function is to signal good posts (to observers) or reward posters.
Under the first perspective, the dynamic you identify is dysfunctional. But I think most people here view thumbs up as "I want more of this type of post," which is closer to the second perspective. Under the analysis of the second perspective, one should net a reward if one does something difficult.
One consensus of this community is that it is harder to admit one was wrong than to avoid the error in the first place. If you disagree, then admissions of error after errors should not net karma.
I suspect that for this to work repeatedly, you would either have space things out, in terms of both time and intervening genuine contributions, or you would have to make increasingly interesting mistakes.
If someone were constantly making and then retracting the same kinds of mistakes, I would notice that the retractions don't actually represent learning not to make that sort of mistake, and would not be happy with it.
IME, retracting a mistake gets you less karma than you lost, or possibly a gain of 1. Even in the latter case, I wouldn't make deliberate mistakes to gain karma, because seeing a comment at -1 hurts significantly more than gaining 1 point feels good. (Yes I realize this is loss aversion bias. But since far more upvotes are given than downvotes, a downvote has more information than an upvote.)
To follow the policy I'd have to downvote this post but I upvoted it because I really like this policy.
What about requests for explanation of drive-by downvotes? I usually place them in a different mental category to general complaints/comments about the karma system
Personally, I'm OK with people humbly asking for reasons, as long as there is no implications that the downvotes are somehow "wrong".
I downvote most talk about karma (complaints, "I bet you'll downvote me", etc) as well, because I want people to do so less. But I support requests for explanations of downvotes, because I want people to explain their downvotes.
EDIT: Okay, I'll get aboard the irony train: Why was I downvoted?
Downvotes shouldn't be given for being factually incorrect. That would stifle disagreement. High-quality comments can be wrong, and low-quality comments can be correct.
I would like to see more people updating their beliefs based on new evidence. Thus, I upvote such comments for now.
An easy way to win cheep karma on LW: