19
$\begingroup$

I am just a relatively new user, so please correct me if I'm wrong.

If you read about the questions with most votes in MSE, you can find that until the $35$-th of question, the questions are all asked at least 4 years ago. Some of them are even about 8 years ago.

Does this mean new questions can't receive enough attention?

One of the reasons of this phenomenon, is, I think, is about the 'hot topics':
Classic results like Different methods to compute $\sum\limits_{k=1}^\infty \frac{1}{k^2}$ (Basel problem)
Some explanation to some important concept in mathematics, such as Is $\frac{\textrm{d}y}{\textrm{d}x}$ not a ratio?; What are imaginary numbers?
Or some 'interesting' results like Is this Batman equation for real?.

And interestingly, some of the problems are really off-topic, such as My sister absolutely refuses to learn math.

Does this mean oftenly, receiving many votes and attentions really doesn't mean it is a good question?

Are our users really focus on wrong position?

We should not just pay attention to familiar topic, or 'hot topic'.


Does it means we should revise our MSE system?

Maybe it is because we have lots more questions these days and so the average attention of problem is decreasing. But I think the number of good quality posts won't decrease.

Maybe we can let moderators or expert users to select good question (for month, years, or etc.?) Somethings like votes shouldn't be the most important.

Thank you for paying attentions to my question.

$\endgroup$
13
  • 3
    $\begingroup$ As usually I will point out that some of the points you raise have been discussed before. Here is a five years old question about decrease in upvotes: Statistics on upvoting. But there are also more recent posts: Is voting on questions getting harsher? or Has the upvoting really gone down? (And probably you can find further posts discussing such stats.) $\endgroup$ Apr 29, 2018 at 13:43
  • 2
    $\begingroup$ Also something similar to your last paragraph (some kind of selection of "question of the week") was discussed a few times. Some of these questions are at least partial related: Star Ranking System?, Is it possible for this site to get some sort of “best of” feature to compile classic answers? and Proposal for promoting more interesting, deeper questions. $\endgroup$ Apr 29, 2018 at 13:52
  • 4
    $\begingroup$ BTW I think that both tags on this question seem are rather out of place. The tag allowed-question is supposed to be used for discussions what is on-topic on the main site. The tag policy is for discussion about various policies on this site. AFAICT, your question touches on neither of these two topics. $\endgroup$ Apr 29, 2018 at 13:59
  • 5
    $\begingroup$ What kinds of questions are you looking for? Sorting by votes isn't the only approach to searching for "popular" questions. $\endgroup$ Apr 29, 2018 at 14:24
  • 11
    $\begingroup$ Older Questions have naturally had time to garner a large number of votes, so I scarcely believe that is evidence of Math.SE "aging". Of course other factors (accessable content, for example) help to improve the chances for wider approval. Personally I always feel a bit of a thrill when one of my long forgotten Answers gets an unexpected upvote. $\endgroup$
    – hardmath
    Apr 29, 2018 at 16:06
  • 4
    $\begingroup$ Maths is quite a mature subject so the proportion of study which takes place at the forefront, of new things, is relatively small. That means the major topics in established fields have all been asked. If you contrast with Mathoverflow, many new questions there have high scores. $\endgroup$ Apr 29, 2018 at 17:30
  • $\begingroup$ This comparison is unfair: an old question has more time to receive upvotes than a new one. It's better to compare two posts by their score $T$ days after creation, where $T$ is a fixed duration. $\endgroup$ Apr 29, 2018 at 20:24
  • $\begingroup$ @ProducerofBS In MO you have exactly the same situation. Most highly upvoted questions are old. If you check the first 50 questions ordered by votes, more than 40 of them are asked 7 years ago. $\endgroup$
    – user99914
    Apr 29, 2018 at 22:01
  • $\begingroup$ @JohnMa agreed. $\endgroup$ Apr 30, 2018 at 8:23
  • 6
    $\begingroup$ Someone asked 6 years ago "Is $dy/dx$ a ratio?" and many many many others have asked it since; hopefully most of those repeats were closed as duplicates. $\endgroup$
    – GEdgar
    Apr 30, 2018 at 12:28
  • 7
    $\begingroup$ I'm pretty old. $\endgroup$
    – Will Jagy
    May 2, 2018 at 21:10
  • 4
    $\begingroup$ I dunno. I’m aging too and I think it’s better than the alternative. $\endgroup$
    – Randall
    May 2, 2018 at 23:24
  • $\begingroup$ I always sort questions with newest on top, so from what I see there are no questions more than an hour old on M.SE. $\endgroup$ May 13, 2018 at 4:39

1 Answer 1

14
$\begingroup$

If you read about the questions with most votes in MSE, you can find that until the $35$-th of question, the questions are all asked at least 4 years ago. Some of them are even about 8 years ago.

There is a positive feedback loop inherent in having a "most popular X" list on any website. It's not only that older questions dominate the "Most voted" list: older answers also have a score advantage over newer ones, even when the newer one is better. Unfortunately the only real solution to this is to remove the ability to sort by score, and the cure is probably worse than the disease.

Does this mean new questions can't receive enough attention?

No. I think it's safe to say that scarcely anyone uses the list of all questions sorted by votes to find questions to answer. Some people use the front page (all questions, sorted by "interestingness", which combines activity with votes); others use the full question list sorted by either activity or age; and others use tag lists sorted by either activity or age.

If you tag your question correctly, it is likely to receive attention. You can improve the chance of receiving attention by choosing an informative title and first line, because people will see either the title or both of those as the summary of the question.

One of the reasons of this phenomenon, is, I think, is about the 'hot topics':

HNQ has the same positive feedback loop mentioned above, plus an additional one: people who have 100 rep on another site can get an "association bonus" which gives them enough rep to upvote, but not enough rep to downvote. I and others have complained on the "mother meta" site that this has negative effects on the quality of some sites, but thus far we don't seem to have convinced the people who matter.

Does this mean often, receiving many votes and attentions really doesn't mean it is a good question?

I don't think I could justify "often", but I would certainly say that there are perverse mechanisms which mean that some questions get far more attention and upvotes than they deserve, including some extreme cases where they should really be deleted instead. (NB don't ask for examples on this site, because the specific ones I'm thinking of are on a sister site).

$\endgroup$
3
  • 4
    $\begingroup$ There is another "real solution" -- use the vote data more wisely. Here and here are some articles about how to do a better job of incorporating up and down votes. $\endgroup$
    – user14972
    May 2, 2018 at 21:26
  • 1
    $\begingroup$ @Hurkyl, the Evan Miller article is a nice bit of nostalgia for me (it relates to my only ever question on the stats stack), but I don't see how it is a "real solution" to the positive feedback loop giving a score advantage to older posts over newer posts. An older post at +60/-0 is still going to rank above a newer post at +5/-0. $\endgroup$ May 2, 2018 at 21:41
  • $\begingroup$ Adding to Hurkyl's comment, Quora seems to have a reasonably good answer sorting system. Even a simple formula based on votes divided by views would be pretty reasonable. $\endgroup$ May 11, 2018 at 16:28

You must log in to answer this question.

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