Hi Rory,

I was hoping to bring your attention to a comment thread starting with: space.meta.stackexchange.com/questions/3140/…

I think these comments reveal that the reasoning that Starship is applying when he casts delete votes. He says "[the] question still fails to answer the question, even if it gets closer" and then later states that in his opinion the answer does not even attempt to answer the question. I wanted to bring this thread in particular to your attention - in case you're not following along.

I have read through the question and answers and comments (and I will review those ten you linked) but Starship seems to be using the voting tools correctly in this case.

Your statement "I have to assume that delete votes on posts that have been improved are either bandwagon delete votes, or users who are voting-to-delete when they should be down-voting instead." is an incorrect assumption

It doesn't matter if a post has been improved if the voter thinks that in its current state it should be deleted. It doesn't matter that others may have voted to delete previously - that is irrelevant to how any member of the community may vote

regarding those posts you want undeleted:

48931 - I would delete as not an answer

53731 - definitely not an answer - it is Link Only, so not valuable here

58736 - link only post, so not an answer, which is why I deleted it

25541 - not an answer, possibly a comment

65688 - I would have voted down, but with the un-evidenced assumptions you have made, I can see why folks voted to delete. And remember, anybody with VTD privs is allowed to.

65907 - I would not have vtd'ed as I think you have improved the post sufficiently

66743 - this one is back in the undelete queue so wait and see, but I would probably just have downvoted, as I think your answer has assumptions that are wrong. But I get the comment around it not answering the question

19418 - not an answer

65398 - not an answer

I've summarised that in an answer on your meta post