How to improve the bounty success rate?

Question

I've been giving out a few bounties recently on some tough to answer questions, but they've mostly been dismal failures without even generating an answer.

The data on all bounties seems to back this up, with roughly 25%¹ of all bounties on questions expiring without even having an answer with 2 upvotes needed for an automatic reward of half the bounty. Some didn't even generate a single answer.

¹The Holiday Bounty Giveaway probably skewed the numbers in favor of successful bounties by a bit, so real bounty success is probably worse than the data.

Answer 1

For some questions, identify-this-game especially, the ability to answer the question falls on the OP being able to provide the AH-HA moment.

ITG questions about Flash/Online games from the late 90's are going to prove pretty difficult to answer too, since they were likely very very localized ... not generally available for download and the website may have gone into the ether many years ago.

It'd go so far as to suggest that ITGs about browser games should be automatically closed as too localized unless they are available as downloads or open source or similar.

Some of the questions I looked at recently are phrased "Is it possible"/"Can this happen" and seem to have an answer of "No." and it's likely impossible to prove a negative without a definitive statement from the developers.

For some of these questions, perhaps the answer isn't a bounty, but closing.

Answer 2

Other issues I'm seeing on the active bounty list include;

• Very difficult questions on obscure games or technologies. This means one or two other people may have played that game on the site, but it generally takes a larger community to tackle the tough questions. (example)
• Unanswerable questions because the ability to measure or determine the answer without being a developer on the game is exceedingly challenging.
• Stephen's answer to confusing identify-this-game (example)
• Bad questions: Too localized, not enough information, not reproducible. For bug-related questions, these would be closed in bug-tracking software. (example)
• The not-enough-award-for-a-very-challenging-question dilemma. Without precise character modelling for example, this (example) becomes a nightmare to figure out.

So what is the solution? Well this site has a very good answer to question ratio, people aren't incited by bounties so easily in my opinion as other sites simply because if they can't answer it without a bounty, they can't answer it with a bounty either.

Perhaps the bounty system doesn't work very well on this site for "questions that didn't receive enough attention". I find bounties more interesting for questions that are already answered but could;

• Use a more well thought out answer, with images, almost like a mini strategy guide or
• Questions with answers that are no longer correct because of game patches. A bountry draws attention to a question that is significantly different because of a patch in order to prevent a duplicate question from being asked.