It seems it's now possible to place bounties on other people's questions. I noticed a certain user placing a lot of bounties recently, including one on a question of mine which already had an accepted answer with over 30 upvotes. Even stranger, the person who wrote that answer and the person placing the bounty are the same. This happened on another question as well. Here are the questions I'm talking about: 1, 2. I don't really know what's going on with this particular user, because she had some strange bounty behavior earlier in the week related to questions of Porton. Perhaps she's playing around to learn the system better. But it raised two questions for me.
First, what happens when a bounty of the sort above expires? I assume her answer will win, because it's already the optimal answer. So she'll get back the reputation she wagered and also whatever happens to you when you win a bounty (badges I guess). So in the end the bounty won't produce any new math and won't lead to a transfer of reputation from one user to another. It will however lead to more publicity for the question and answer, and that leads to lots of upvotes. Hence my second question: is this the sort of thing we want happening on MathOverflow? The same thing happens on a lesser scale when someone edits an old question, but the front-page turnover rate is fast so those only stay for an hour or two. Also, there's a disincentive because if you bump your old question/answer too many times it'll become CW. But a question with a bounty stays advertised for a whole week. I feel like using bounties to promote your best answers from the past is a misuse of the system, and I wonder if others agree and if there's anything we can do to dissuade that method of using bounties.