6

Lately, I am getting the impression that on relatively low quality questions comments requesting clarification and "what has your prior research shown you so far" are getting fewer in numbers. It seems they were replaced with immediate upvotes. Securing those questions a spot on the Hot Network Question (HNQ) list.

HNQs can have quite pleasant effects: reputation for the posters, exposure for the site on the network etc.

But this has also effects that seem concerning:

This site gets outside exposure on the network primarily with low quality questions, which attract often quite a bunch of votes, newbie answers of comparable quality lacking sources and so on.
While HNQs are primarily about popularity of the topic, not about the quality of the posts, and the results predictable, we place those posts on that list. Before there is time to improve the post (that is: if the original poster improves the question along the lines we should desire) a huge bandwagon of voters has already stampeded over the whole thread – distorting the StackExchange voting model and this site's standards.

Why? Why are there so widely differing standards on posts sometimes requesting prior research and requesting sources in questions or answers? Shouldn't we aim at a little more consistency on desirable quality standards?


As a rule of thumb I would say that if a question is low quality, like "does not show prior research effort" we should seriously restrain us from upvoting too quickly. Assuming that as the main reason for HNQ placement.
As a second rule of thumb I would say that if an answer has "no sources" a "sources would improve this answer" banner should always be added.
As a third rule of thumb I suppose that an answer that is judged as "good enough as to not needing sources" then the question falls automatically into the category of "too basic" and should be closed.


It seems that the hotness algorithm is not fully published and different for different sites (depending on its size on the network) and different for the sidebar and the hot notwork page. (Two metaSE discussions on this).

From what I understand about this it seems clear that the huge disproportionate voting effects from HNQ only start to pile in once we upvoted a poor question quickly, after we answer such a question quickly and then upvote that answer equally quickly. The situation gets worse if there are multiple answers.

The conclusions to draw from that seem difficult. I guess we do not want to stifle HNQ effects completely and especially not for relatively new users. But how to organise the community better into knowing this, into restraint for quick voting and quick answering.
And how to work with that algorithm?:
A question might be "bad" but with a good core. Holding that quickly will prevent it from going hot forever, despite the core and perhaps the eventually fixed version deserving a wide audience.

  • 1
    Agree fully with ROTs 1 & 2, not so sure about 3 as, sometimes, it's a matter of putting two not obviously (to many people) related bits of common knowledge together. That said, there are far too many answers getting upvoted which have no sources or links. – Lars Bosteen Jul 16 '18 at 0:14
  • 4
    Food for thought: Perhaps we might ponder the implications of the SE network at large routinely loving questions that regulars here consider subpar. I'm not saying one side or the other is wrong. Its just something that's been niggling at me. – T.E.D. Jul 16 '18 at 13:53
  • 1
    @T.E.D. It might just be that since they can't downvote, all we see from them is pure love. Does History HNQ get more upvotes than questions from other communities? – Semaphore Jul 16 '18 at 13:59
  • 2
    @Semaphore - Like getting dogpiled by a horde of puppies? – T.E.D. Jul 16 '18 at 14:17
  • Just to add to my pwnded confusion: I give a fair but necessarily vague answer, since I do not have the definite answer -> HNQ // I give a sourced and reffed exact answer 2 upvotes. –– Still baffled. – LangLangC Sep 13 '18 at 0:01
12

I agree with your central concern, and I agree that we shouldn't upvote poor questions too easily.

However, I think you have the voting situation reversed. What in fact happens, I believe, is that a question become HNQ, and then receive a bunch of upvotes, probably because most HNQ viewers can only upvote. It seems to only take a couple of votes from History to start the process.

Instead, notice that questions do not become HNQ without being answered. So the key here is actually that we're answering poor questions too quickly. Poor questions by their nature invites quick answers: we've all seen ones that do nothing but cite Wikipedia articles, or one that simply correct basic misunderstandings, or worse (but thankfully rarely), "answers" that simply wonder what an unclear question actually meant.

The proper way to prevent this is to put the question on hold. This is something the community has tended to be reluctant to do, even when the same users leave comments requesting changes or clarifications. Hence, the present issue.

Therefore, the community needs to be proactive about putting subpar questions on hold. Fix the question, nominate for reopening, and then answer it.


Alternatively, it's also possible for the moderators to adopt a more pro-active stance and preemptively put questions on hold pending edits. A substantial change like this would have to be supported by community consensus, though.

  • I do not know the exact algo for hotness, but I guess you are right that I neglected the answer issue. What's the algo? 1 A and 2 upvotes each, quickly? It's a real bummer that older, that is matured or properly improved Qs never get hot. – LangLangC Jul 16 '18 at 10:02
  • 4
    @LangLangC The formula is published somewhere on the Meta, but basically it's all about having the most answers within six hours of posting the question. The question just needs to have > 0 score, and answer scores aren't very important. And yes, its a shame that the formula is completely biased towards speed, which works for some sites, but not for those looking for a considered question/answers like History. – Semaphore Jul 16 '18 at 10:16
  • 5
    I think this answer might be important enough for public discussion of alternatives/courses of action. I would love to see us make two changes (a) repair and reopen a greater percentage of closed questions (b) close more questions more quickly (and with more courtesy). – Mark C. Wallace Jul 16 '18 at 13:12
  • 1
    Seems to me that there are 2 different things to make the community aware of: Do comment quickly, do VtC quickly, try to fix a post quickly. – But 2nd: to avoid HNQ-badness: do not vote as quickly as commonly prescribed by the network ("early and often") but wait at least 6 hrs from posting-date to upvote on apparently poor but salvageable Qs and also do not answer Qs as long as they are in a bad state? – LangLangC Jul 25 '18 at 0:21
  • 2
    @LangLangC I think do not answer, and flag poor low quality answers for deletion is the main recourse we have. It's impossible to prevent stray upvotes, but answers take marginally more effort and without them a question wouldn't make it on to HNQ. My observation is that poor questions tend to have simple or obvious answers, and leads to lots of people trying to whirl up a quick answer, leading to HNQ and a feedback loop of upvoting. – Semaphore Jul 25 '18 at 4:12
  • Hm. Granted that I may have embarked on a lonely crusade against a single windmillI: tried and failed on all accounts. Commented (deleted, yet someone else replaced that with his own, almost identical in intent and highly upvoted), downvoted, VtC'd (others disagreed), voiced concerns in chat. No response towards fixing the Q from anyone. –– This needs either a boost or I have to take a step back from bothering altogether. – LangLangC Sep 22 '18 at 21:51

You must log in to answer this question.

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