19
$\begingroup$

To better understand the advantages and disadvantages of having MO questions featured in the list of hot network questions, as discussed here and here, it would be useful if Stack Exchange could release concrete data regarding how much traffic is generated by users clicking MO questions from this list compared to traffic coming from other likely sources, such as math.SE or reddit.

$\endgroup$
  • 7
    $\begingroup$ SE is going to overhaul the hot questions algorithm in the near future, it might make sense to see if those changes alleviate the concerns from the MO community. $\endgroup$ – Mad Scientist Jan 10 '14 at 18:47
  • 3
    $\begingroup$ @MadScientist: Is this algorithm overhaul going to be "if site=MO then ignore; elseif site=MSE then score=score/3; elseif as before" :-) $\endgroup$ – Asaf Karagila Jan 10 '14 at 23:07
  • 4
    $\begingroup$ @AsafKaragila I have some hope that CW questions might be excluded in the future, that would help with the soft question that get disproportionally popular. $\endgroup$ – Mad Scientist Jan 11 '14 at 8:04
  • 3
    $\begingroup$ @MadScientist is there some dedicated discussion of this going on somewhere to which one might contribute? CW is one thing, I agree. But another thing in my opinion is that the hot-formula is quite strange in that it gives in some sense most visibility to question that already got many answers (assuming score is not to bad overall). And when they reach the list they get still more...so then they get protected to block further answers...quite absurd, in a fun way :-) $\endgroup$ – user9072 Jan 11 '14 at 13:47
  • 2
    $\begingroup$ @quid As far as I understood SE is going to look at the whole algorithm, so they might also fix the rather strong weight it places on a large number of answers. There is no public discussion specifically for this, but there are quite a few older posts on MSO about the flaws of the algorithm. $\endgroup$ – Mad Scientist Jan 11 '14 at 14:16
  • $\begingroup$ @quid: The thing which seems strangest about the hot questions algorithm, to me, is that it is not personalised at all. So someone who is interested mainly in cooking and Judaism will get the same hot questions as someone interested mainly in theoretical computer science and aviation. So, everyone sees many questions they don’t care about, or that they care about but are not qualified to discuss at the site’s level. The obvious simple fix would be to restrict the hot-list a user sees to sites they are already members of, or established on. This would clearly be too drastic: [cont’d] $\endgroup$ – Peter LeFanu Lumsdaine Jan 16 '14 at 13:04
  • $\begingroup$ one of the main functions of the hot list is to advertise sites within the network, and hence attract new members to sites. And this function would need to be retained. But a more refined version of this could surely be helpful: weight a user’s hotlist, giving preference to sites related to their previous activity. So, for instance, a prolific MSE or TCS contributor might see a MO question on their hot list; but most other users would not. $\endgroup$ – Peter LeFanu Lumsdaine Jan 16 '14 at 13:07
  • $\begingroup$ @PeterLeFanuLumsdaine I am under the impression that what you suggest is possible if a user wants it, via them creating a filter see stackexchange.com/questions on the right. Likely one could try to guess such things also, but then this is not that easy especially for the more general sites, I think; and people do have diverse and virtually 'inpredictable' interest. In that sense, I do not find this that strange, it is sort of analog of (general) bestseller list, music and movie charts and so on. $\endgroup$ – user9072 Jan 16 '14 at 13:27
  • 2
    $\begingroup$ Related. BTW has there been any serious attempt of MathOverflow to opt out of this whole unfortunate hot list issue by now? I suspect posting on MSO would be completely futile, but maybe a direct communications between people on both (MathOverflow and SE) sites who negotiated MathOverflow's joining the SE network could help? $\endgroup$ – Dilaton Feb 8 '14 at 21:41
  • 4
    $\begingroup$ @Dilaton: Since the measurable effects are small to negligible, there has been little interest in doing that. $\endgroup$ – François G. Dorais Feb 8 '14 at 22:05
  • $\begingroup$ @FrançoisG.Dorais I see, then it is ok. $\endgroup$ – Dilaton Feb 8 '14 at 22:16
10
$\begingroup$

These are rough numbers, I'm afraid - I'm combining data from Google Analytics and our internal event tracking system, so there's a reasonable chance we're measuring slightly different things. Hopefully, this is sufficient to give you a rough comparison:

In the past 30 days, MathOverflow has seen:

  • 26,189 visits from Mathematics Stack Exchange
  • 10,386 visits from Reddit
  • 8,785 visits from Stack Overflow
  • 3,667 visits from the various hot lists on various Stack Exchange sites - note that these will have contributed to BOTH of the "visits from Mathematics" and "visits from Stack Overflow" totals as well.
$\endgroup$

You must log in to answer this question.

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