6
$\begingroup$

Multiple of our high-activity questions have appeared over the last week on Hot Network Questions (HNQ) on the Stack Exchange front page, leading to an increase in traffic. Is this desirable at the current time? Should we get our site excluded until we have a better handle of things?

(I thought I remembered reading that an SE has an option to decline to be included from HNQ, but am failing to find where it was mentioned. A past discussion on making changes to the formula didn't lead to any answers, or a definite response on whether it could be done.)

$\endgroup$
5
  • 5
    $\begingroup$ Personally I don't see why this would be useful for us. Many current good contributors found the site via recent hot posts. Let's focus on dragging down the kinds of material we don't want flaunted on the HNQ list $\endgroup$
    – d'alar'cop
    Commented Nov 20, 2014 at 1:15
  • $\begingroup$ @d'alar'cop I think it's just a question of volume. The amount of stuff coming is overwhelming our ability to downvote, close, edit, or suggest to posters how to improve their content. I've heard "15x increase" thrown about. I don't think we can act fast enough to quality-control the firehose of activity on an HNQ question as it pours in. $\endgroup$
    – xnor
    Commented Nov 20, 2014 at 1:32
  • $\begingroup$ good retort. I guess you're right... I just want highlight that some featured material is pure puzzling gold $\endgroup$
    – d'alar'cop
    Commented Nov 20, 2014 at 1:35
  • $\begingroup$ @d'alar'cop Yeah, it's too bad. I only found out about puzzling from HNQ. But I think it's needed now as an emergency measure. $\endgroup$
    – xnor
    Commented Nov 20, 2014 at 1:37
  • 2
    $\begingroup$ it may also be useful to take into consideration that just now (due to all the commotion) people have started banding together to enforce quality in a much more effective manner than before (or simply taking personal responsibility for it e.g. me)... or indeed the influx of people are learning $\endgroup$
    – d'alar'cop
    Commented Nov 20, 2014 at 1:39

2 Answers 2

Reset to default
7
$\begingroup$

Word from the Stacks Above says this is not an option.

While I'd like to post a longer answer, that's pretty much all there is to say. I checked with Stack Exchange staff, and they said it isn't something they can do.

Even if we wanted to, modifying the way we appear on Hot Network Questions is just... not a request that could be fulfilled, unfortunately.

$\endgroup$
5
  • 1
    $\begingroup$ What about modifying the formula? $\endgroup$
    – d'alar'cop
    Commented Nov 20, 2014 at 1:53
  • $\begingroup$ @d'alar'cop Same answer, unfortunately. It's probably just too intrinsic to change. $\endgroup$
    – user20
    Commented Nov 20, 2014 at 1:54
  • $\begingroup$ @Emrakul You should also post on this meta question to say it's not an option. $\endgroup$
    – xnor
    Commented Nov 20, 2014 at 2:50
  • $\begingroup$ @xnor Thanks! Closed as a duplicate to here. $\endgroup$
    – user20
    Commented Nov 20, 2014 at 3:16
  • $\begingroup$ Maybe one could use this condition to build an argument that the overall formula/method in SE for calculating hotness is itself flawed? $\endgroup$
    – d'alar'cop
    Commented Nov 20, 2014 at 4:15
5
$\begingroup$

We should temporarily remove ourselves

HNQ has been showing our worst side to the rest of Stack Exchange. The "hotness" formula heavily weights popularity and activity over a short period of time, which systematically rewards bad puzzles: ones that spawn many different short answers because the question is ambiguous and anything goes. Meanwhile, puzzles that are well-specified, hard puzzles, ones that take a while to write up answers, or expert-level questions about puzzling aren't showcased, giving a skewed image of our site.

I think the influx of traffic from HNQ is in part to blame for the recent dip in quality. The mass of new, lower-quality answers and questions has overwhelmed the existing user base and made it hard to enforce existing policies due to their sheer number.

HNQ creates a bad image for our site. Showing off poor questions drives traffic to these poor questions, and invites users who will post poor answers and further poor questions. This creates a self-reinforcing cycle of low quality. Sure, quality questions are sometimes exposed and quality users learn of the site, but right now the odds are stacked the other way.

We may lose exposure, but that's not what we need now. We have plenty of activity but low quality. We can afford to lose some exposure until we get a better handle on quality and are able to enforce the existing policies from Meta.

Once this influx dies down and we have policy and quality better under control, we should flip the switch back on and show off our new, better site to the world.

Edit: Adding my comment on why I think just pruning lower-quality content doesn't suffice now:

I think it's just a question of volume. The amount of stuff coming is overwhelming our ability to downvote, close, edit, or suggest to posters how to improve their content. I've heard "15x increase" thrown about. I don't think we can act fast enough to quality-control the firehose of activity on an HNQ question as it pours in.

$\endgroup$
3
  • 1
    $\begingroup$ Just for context -- as you may know, but some readers may not -- other sites have had similar problems with red tides of attention from the HQ list. I think the master complaint question is In network hot questions formula, discard answers when voting evidence indicates that these are not good data points; gnat, the author there, takes this as his pet issue, and I believe he's linked most of the other issues to that post at this point. $\endgroup$
    – jscs
    Commented Nov 20, 2014 at 3:46
  • $\begingroup$ @JoshCaswell Thanks for the pointer. I was aware of the issue on other sites, but not the general discussion on SE Meta. Indeed, it seems like a systematic problem with the hotness formula. $\endgroup$
    – xnor
    Commented Nov 20, 2014 at 5:57
  • $\begingroup$ @JoshCaswell, thanks for the useful info. It's interesting (and a bit of a relief) to hear that other sites have had the same problem in the past. $\endgroup$
    – A E
    Commented Nov 20, 2014 at 16:54

You must log in to answer this question.

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