Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 158 Stack Exchange communities.

What is meta?
Here's how it works:
  1. Any Stack Exchange user can ask a question
  2. The community provides support, votes on ideas, and reports bugs
  3. Your voice helps shape the way Stack Exchange operates

When bounty grace periods were introduced, Jeff commented:

I decided it can't hurt to extend the auto-award period 24 hours past the true end of the bounty.

There has now been time to observe how the grace period affects user behaviour:

  • without it, an offeror is implicitly encouraged to award her bounty to the first answer that satisfies its criteria (although she might hold off doing so until later in the bounty period, she must still proactively choose a deserving question knowing that others might yet follow); but

  • with it, I've more than once seen offerors say that whilst an answer is perfect they will hold-off awarding until the bounty has expired in case any better answers appear.

Whilst an undoubtedly minor point, such behaviour somewhat discourages (or at very least, does not promote) rapid answering of bounty questions—which would no doubt be better for both the questioner and the community (by "rapid" I do not mean hastily put-together sub-standard answers, as they are unlikely to have ever have been awarded the bounty); albeit that this effect is somewhat offset by the greater attention early answers will receive by virtue of being featured for longer and thereby may receive more upvotes.

It would be interesting to see some quantitative analysis of how the grace period has in fact changed behaviour (of both offerors and answerers); then consider whether grace periods are, in fact, more harmful than first thought.

share|improve this question
People often hold off until the last possible second, if only to get the most attention from the bounty status; a status which is gone by the time the grace period kicks in. And the chance they get an answer after the bounty period is much lower since it's no longer featured. I seriously doubt this is a common or significant "problem" – Ben Brocka Jun 7 '12 at 10:55
Frankly, I rarely saw any "rapid answering" for bounty questions - both before and after the change. The reason is mostly that the bounty questions start off at the bottom of the "featured" list and only get to the top of it when they are about to expire. Consequently, they also get the most views in the last two days or so. – Wladimir Palant Jun 7 '12 at 12:45
We're talking about the difference between 7 days and 8 days here -- why would it make such a big difference? – balpha Jun 7 '12 at 13:51
@balpha: I believe that it's psychologically quite different (hence curious to see any quantitative analysis to support/disprove that theory): the difference being that where a grace period exists, one invariably waits for the bounty to expire before awarding it; whereas without such a period one must award it at some point during the bounty's life, therefore much more likely to happen as soon as a satisfactory answer is received. It may be that users still wait toward the end of the period before making an award, but was this seen in reality? – eggyal Jun 7 '12 at 13:54
Are you aware that you cannot award a bounty in the first day? – balpha Jun 7 '12 at 13:56
@balpha: If that was directed to me, then yes I am... not sure how it affects this concern though? – eggyal Jun 7 '12 at 14:18
Point being that you can't award a bounty immediately. – balpha Jun 7 '12 at 14:20
@balpha: You're getting dangerously close to providing an answer. ;) – eggyal Jun 7 '12 at 14:36
oh, I'm working on one :) – balpha Jun 7 '12 at 14:40
up vote 11 down vote accepted

Here's the (Stack Overflow) data, starting from July 2010 (shortly after the new bounty system was introduced) until May of this year. It excludes auto-awarded bounties (in other words, only counts bounties that were actually awarded by the user). The month counted is the month in which the bounty was started.

Month    Number of bounties Average hours before awarding 
-------  ------------------ ----------------------------- 
 7/2010  359                110                           
 8/2010  439                113                           
 9/2010  410                114                           
10/2010  416                116                           
11/2010  417                112                           
12/2010  368                111                           
 1/2011  516                105                           
 2/2011  496                113                           
 3/2011  556                110                           
 4/2011  515                110                           
 5/2011  621                106                           
 6/2011  624                111                           
 7/2011  596                117                           
 8/2011  612                114                           
 9/2011  706                121                           
10/2011  722                120                           
11/2011  687                119                           
12/2011  648                118                           
 1/2012  768                120                           
 2/2012  740                118                           
 3/2012  876                117                           
 4/2012  744                116                           
 5/2012  770                114                           

Your quote from Jeff is from the end of June '11. So I would say yes, it did indeed increase the time between starting and awarding the bounty, but only by something like 5 hours on average.

That said, even if the increase was larger, I wouldn't see an issue with this.

I've more than once seen offerors say that whilst an answer is perfect they will hold-off awarding until the bounty has expired in case any better answers appear.

They're spending their own rep – why should they not wait and and see if something else comes along?

such behaviour somewhat discourages (or at very least, does not promote) rapid answering of bounty questions

For bounty-earning answers that were posted after the bounty was started, the average time between bounty start and answer posting fluctuates between 50 and 60 hours, with no noticable change around July '11.

share|improve this answer
Very thorough and detailed - thank you! – eggyal Jun 7 '12 at 15:05
Has the % of bounties that were not awarded changed? – agf Jun 7 '12 at 15:12
@agf No, that percentage stayed pretty constant as well (I first thought that a decrease of auto-awarded bounties might have been responsible for the increase in time-till-awarding, i.e. fewer people "forgetting" to award the bounty in time, but the numbers don't support this). – balpha Jun 7 '12 at 15:20

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .