I had an old question get closed recently and noticed in my reputation display that I got back +2 for a "undownvote" a couple of days later. It seems odd that a person would or could remove a down vote only days after the question was closed after such a length of time.
I didn't notice any other edits or votes recently on that question. The downvote apparently occurred on the same day it was closed.
Does closure count as an edit that would allow someone to change a vote cast long ago? If so, is that intended? Did the recent (and perhaps ongoing) changes to the reputation calculation code somehow forgive the down vote as a result of the recent closure? That seems to be an error.
Here's a link to the question: Getting back on track after disruptions
Note: the question hasn't been deleted, just closed, though I expect it soon will be now that I've posted a link to it on meta. I don't know if it would be helpful to the team to wait until they've had a chance to investigate the issue, but on the outside chance that it would be, could you please wait until this is resolved before casting your delete vote?
Edit: I should also add that the reason I noticed this is because my rep went from ...,908 to ...,875. I looked to see why and saw another question on which I had an accepted, 2 x upvoted answer had been deleted. This should have been a -35 drop, but it showed up as a -33 drop. That's what made me start looking for the discrepancy. According to the timeline the "undownvote" happened 11 hours ago so I have no idea why it just showed up, either.