At the moment it seems that UTC timezone is used for all calculations that involve "days" (Mortarboard, Epic, Legendary and other reputation/comment/question organization). I totally understand this from and ease-of-development perspective, but it seems pretty unfair to those poor users in the far-flung reaches of the US west coast (etc) because by the time you wake up the day is already almost half gone, and those of us with jobs can hardly be expected to spend all our time in the office on SO. I'd be curious to see a breakdown of these badges/user/timezone.

This issue has been raised before (where it was well received but declined) and Jon Skeet wrote an answer to a similar question essentially saying it would be a pain and would be subject to abuse (easy for him to say, as he's in the UK AND gets 200 points a second). However, recent changes to settings that limit the frequency with which you can change your timezone gets around the abuse issues, and I just can't imagine that the bright folks at Stack Exchange couldn't figure this one out. Hell, just create a setting what says "What UTC time would you like your 24 hour cycle to begin?"... any changes to this setting could be set to only take effect the following day, allowing no points-per-time-based badges in the intervening interval, or if that's too hard, just force uses to pay a point penalty to change this setting.

Whatever the mechanism, it seems like an obvious issue that should be addressed.

  • 2
    Maybe it's the Stockholm Syndrome speaking, but ... you'll get used to it. Even out here on the left coast.
    – Charles
    Feb 1, 2012 at 2:27
  • 1
    As for preventing abuse, just reset the user's day whenever they change their timezone, so all the counts for any badges start back at 0. Even if there's only a couple hours left for their "day". It's probably a lot harder than it sounds.
    – animuson StaffMod
    Feb 1, 2012 at 2:27
  • It would make an interesting study, but a user's timezone is not in the SEDE data. Nor is IP (which would be a nominal proxy). Most users do not fill in their location, either. Feb 1, 2012 at 2:31
  • @BrockAdams probably true. Though even with just those users who do have location data filled in, I imagine trends would emerge.
    – Ben D
    Feb 1, 2012 at 2:34
  • 1
    Hey there downvoters, any explanations? If you don't like the idea, fine, but I think it warrants a discussion.
    – Ben D
    Feb 1, 2012 at 2:36
  • 2
    The down voters are part of the discussion, just without words. Feb 1, 2012 at 2:45
  • @AndrewBarber: I don't disagree, and it's not that I particularly care about my Meta score... it just seems that with comments have been left are thus supportive or ambivalent, I'm curious to get a better idea of why the down votes.
    – Ben D
    Feb 1, 2012 at 2:49
  • Because they disagree, and are using the voting system to say so. Just like those who agree, and are using the voting system to say so. Feb 1, 2012 at 2:52
  • @Ben, voting works differently on Meta. Feb 1, 2012 at 2:54
  • @AndrewBarber I already said that I agree, but I do think that in the interests of constructive conversation no votes should come with an explanation if no existing explanation exists (I'm not saying force it, I'm just saying it's nice protocol). If I say "I think this is the best way of doing X is Y", and you say "yes!", it doesn't really warrant additional explanation. However, if you say "no!" then it naturally leads to a follow up question: "Why not?"
    – Ben D
    Feb 1, 2012 at 2:57
  • @MichaelPetrott Thanks! I was vaguely aware of this, but I appreciate the link. Even so, I maintain that without anyone suggesting why this shouldn't be so someone ought to throw some explanatory criticism into the discussion, especially now that earlier (2009) issues that were brought up have been addressed.
    – Ben D
    Feb 1, 2012 at 3:00
  • Well, the reasoning goes, why flood a question with discussion when you could simply indicate agreement or disagreement? It's clearly inspired by SO's question and answer (as opposed to discussion) ethos. Three users gave some good reasons, above - should I really need to chime in with "I vehemently agree/disagree"? Feb 1, 2012 at 3:12

3 Answers 3


Well, the awarding of badges is definitely correlated with a user's timezone, but it's not clear if this is due to clock bias or to the location of the most-industrialized countries.

Here is the percent of users who have the Mortarboard badge versus their average comment-posting hour. (Note that comments were used instead of Q & A's because there are 19 percent more comments than posts. (SEDE doesn't make the user's timezone available.) )

User's Avg     Num Mortarboard   Num Users     Percent of
Comment hour   Badges           (who comment)  users with badge
------------   ---------------   ----------    ----------------
      0                1            2646             0.04
      1                0            3634             0.00
      2                1            3863             0.03
      3                3            4586             0.07
      4               11            5509             0.20
      5               21            6362             0.33
      6               32            8163             0.39
      7               69           10156             0.68
      8              144           13405             1.07
      9              284           16992             1.67
     10              396           19712             2.01
     11              589           22718             2.59
     12              991           27004             3.67
     13             1402           29632             4.73
     14             1448           29936             4.84
     15             1225           28194             4.34
     16              734           25060             2.93
     17              437           21580             2.03
     18              186           17411             1.07
     19               54           13776             0.39
     20               16           10789             0.15
     21                5            8568             0.06
     22                6            6107             0.10
     23                1            4419             0.02
     24                0             506             0.00
------------   ---------------   ----------    ----------------
     13.0           8056          340728             2.36

Reference: "Mortarboard badges as a function of a user's average commenting hour".

Note that the SEDE doesn't make the user's timezone available, hence trying to measure when a user is most active.

If SEDE is reporting these results in UTC, then it would seem that the bias is towards the UK workday. One would expect the bias to be towards the USA workday, due to a much greater number of Stack Overflow users. Could clock-rollover effects be the difference? There's too many factors and not enough data.

  • Your percentages seem to be very off. For example the first one should be 0.038%, not 0.2%.
    – user154510
    Feb 1, 2012 at 16:58
  • @MatthewRead: good catch, thanks. I'd accidentally pasted-in the values from the wrong column of my spreadsheet. If you'd refer to the linked query, you would see that it had the right figures (as does the answer now). ... The result and the shape of the data are the same, irregardless. Reputation-based badges are time biased. The UTC clock might have a small effect. Feb 1, 2012 at 20:34

by the time you wake up the day is already almost half gone

After that remaining half day, you get a head start on the next day. Is that any less unfair to the people elsewhere who are going to sleep at that point?

You could argue that it's more difficult to have a single good SE UTC day because you need to participate on two "physical" days. I would counter:

  • Some days I'm less busy in the afternoon and also the next morning. The UTC day gives me an advantage!
  • SE doesn't want to encourage people to post like crazy on a single day in order to hit the rep cap or whatever you're trying for. There are incentives for consistent valuable participation. Your participation obviously doesn't need to be 24/7, but sleeping between some of your posts shouldn't be a big deal.

I don't think the specific rollover point makes any difference.

  • 1
    I think this is a good point. It still will create a division between those who will have the bulk of their "SO Hours" in a single day vs split (and others in-between) but point well taken.
    – Ben D
    Feb 1, 2012 at 4:02
  • "I don't think the specific rollover point makes any difference." -- And, yet, the data suggests that it might. Certainly, when you post, greatly determines your likelihood of getting the badges the OP mentioned. It seems that upvotes (MB badges) favor those who work UK hours (0900 to 1700 GMT). Feb 1, 2012 at 6:59

Honestly, I've thought about making the same suggestion but for a different reason: I'm on the West coast of the United States, and the day change happens at 16:00. I can get two day visits just by showing up at 1555 and voting until 1605. For me to earn one of the dedicated badges I can actually get away with logging on every other day:

+--------------------+--------------------+--------------------+ ...
|   24 hours      <--|-->   24 hours      |                 <--| ...
+--------------------+--------------------+--------------------+ ...

                  ^^^^ ------ visits every 48 hours ------ ^^^^

I've got an extremely unfair advantage for these badges.

But I can't bring myself to care enough about the badges to .. uh .. care.

Voting is done by a 24-hour rolling window so it's not like moving the zero-time around could get me more than the pitiful 40 daily votes...

The obvious stability and sanity benefits of using UTC far outweigh switching to any other system, especially local-user-timezones.

You must log in to answer this question.

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