Prevent questions on the Hot List from being upvoted by casual visitors (only rep is from association bonus)

Question

I have experienced this on a couple of sites (and I'm not the only one):

• You look through the question list. You see a pretty bad (or at least not-so-good) question heavily upvoted.

• Or even worse, a very poor answer to a good question - upvoted to the stratosphere.

• Or even worse, a very poor/incorrect answer upvoted high up - and a later, correct answer, languishing with a couple of upvotes (since the flood of views is no longer there).

• Upon checking, you see tons of views, which came from SE Hot List inclusion.

So, people see an "interesting" "fun" question, head on over in mass numbers (especially if the destination is a smaller site), and due to association bonus, can - and do - upvote any random "fun" crap that's "hot" yet to an expert on the topic is at best not worth being upvoted (and at worst should be downvoted).

As per Thursday's comment, this is made even worse because of the asymmetry of hot list effect: those coming from Hot Questions may recognize the question is crap, and perhaps 50% of them would downvote it under normal circumstances. But having only 101 rep, they can't. The 5% who want to upvote (for reasons that are difficult to fathom) can and do. The result: a vote count that does not represent the opinion of either the particular community, nor the network at large.

• UPDATE: The above effect just got a spectacular real world confirmation. As of this moment, the bad answer has only 7 downvotes (and 20+ upvotes), despite the fact that my comment which indicates it's a bad answer in need of downvotes has 23 upvotes - 3x number of people agreed that the answer is bad, than actually downvoted

One solution presented to this dilemma - which seems a bit too heavy-handed - is to prevent the association bonus rep from counting towards the vote-up privilege.

My suggestion is a much more surgical and localized:

Do not allow voting in the following case:

• The question is currently in the SE hot list
• The user's reputation on the site not counting association bonus is not enough to upvote.

Note that an association-only user can still upvote any other content on the site - or even questions that used to be on hot list, but aren't anymore.

---

_{As an example, an answer posted by myself recently was upvoted to 100+. It wasn't a bad answer, but if I'm being honest with myself, it clearly wasn't great enough to deserve 100+ upvotes that it got thanks to the hotlist.}

Answer 1

I like the ideas, and I see three options:

1. You need 15 in-site rep (i.e. association bonus doesn't count) or being a member for 24 hours to vote. Downside: I've got an account on RPG for a while without any activity but voting, with this system I would already generate this bad noise (and I certainly do, to some extent, without the intention to disturb the site).

2. You need 15 in-site rep to vote. Makes sense, but it's quite too strict I think. However, if this was the option, I would suggest lowering the downvote privilege limit from 125 to 115 so that you get both privileges at the same time when you get the assoc. bonus. Downside: It's a bit too strict to have to get +3 on a question to be able to vote, once you know the system from another SE site.

3. You need 15 in-site rep and make a post with a positive score to vote. IMHO the most sensible way to go.

At any case, even if no upvote restriction is added, we should change the downvote privilege limit to 100 so that people with only assoc. bonus can downvote instantly.

Answer 2

I pretty much agree.

The way it currently works, HNQ undermines the whole way the reputation and badge system works.

As a recent example, on the site at which I'm the most active, my most upvoted answer (at the time when this answer was originally posted) is a rather trivial, short, basic answer that I put nearly no effort into. I typed it on my phone with one finger while half asleep, lying in bed. While the daily rep-cap did do its job and limited the reputation earned from it, I did earn a silver "Guru" badge. (Certainly interesting, given the triviality of the answer.)

So, the asker earned (so far) 182 rep, a "Nice Question" badge and a "Popular Question" badge for a question that is (while on-topic and alright) quite basic and didn't feature much research.

I earned 265 rep, "Nice Answer", "Good answer" and "Guru" badges, all for an answer that (while correct and to the point) featured nearly no effort, and could have been made in seconds by most members on the site.

A few opinions of mine:

1. Most of the regular users on the site don't use their limited amount of upvotes on such trivial questions or answers. They save them for questions and answers that show more research/effort/knowledge, and/or are more interesting.
2. The vast majority of these votes came from casual visitors from other sites who came through the HNQ.
3. Since the issue is so trivial, they could easily recognize that the answer was correct (without research or special knowledge) and therefore upvote with a clean conscience. (I don't blame them, I might have, too, in their position. It's not their fault.)
4. One of the reasons this question became (and stayed) a HNQ was because people wanted to know what "virgin trains" were.

All assumptions, I know, but I think they are reasonable.

My currently most upvoted answer on the same site is also rather basic common knowledge, I typed it in three minutes after 30 seconds of research, and I've earned roughly 1000 reputation from it, which is not at all negligible on that site. This is undoubtedly because of the HNQ. There are other thoroughly researched answers that I spent hours on that have earned me less than 100 reputation.

IMO the whole point of a reputation system should be to (at least roughly) represent the quality and quanity of users' contributions to the site. As I have illustrated with the examples and arguments above, HNQ currently works against this goal.

Also, since HNQs typically get a lot of views, it often gives the asker silver and gold badges merely for the views.

Moving on, the HNQ system encourages clickbait titles. (I'm not saying OP meant "Virgin Trains" to be clickbait, but I figure it sort of ended up to be.) Very interesting titles that make people curious are likely to end up in the HNQ. And since HNQ is a self-fulfilling prophecy, once a question has hit HNQ, it typically stays for a while.

The way it is now, very trivial or easy questions can earn people a lot of rep (especially if a question remains in the HNQ for a few days), and that undermines the way the reputation and badge system works.

I personally think it's fun when my question hits the HNQ. But I think the system itself is slightly flawed.

I also agree fully with OP that the assymetry caused by the fact that these HNQ visitors can't downvote also makes the problem considerably worse.

So, I totally agree with your suggestion. Casual visitors who come through the HNQ and have never participated at all on the site they are visiting should not be able to upvote. (But if they are browsing the site unrelated to the HNQ, it's fine, they should be able to upvote.) Reputation is the basis for a lot of things here (including moderation), so it's important that the reputation system itself is trustworthy.

Like @yo' suggested, a criterion for voting when coming through HNQ could be that one has previously participated on the site in some way, such as an answer/question with a positive score (or even 7 helpful flags or 7 edited posts.) Even that would drastically reduce the problem.

Answer 3

I support this. I've been on the receiving end of this a lot on The Workplace.

Sometimes I write a super innocuous and not that spectacular answer, and it ends up somewhere insane like +200 within a day. That's unhealthy, and it's not based on quality either. Usually the answer that's currently highest voted gets all the drive-by votes.

This phenomenon is invalidating a lot of what voting is supposed to be about, so we may need to put a stop to it.

Answer 4

Although I support the idea here, I'd like to add a point of caution. At least on PPCG, we often see someone find a trivial (and often ill-fitting) question via Hot Network Questions, post an answer (this is easy, as a trivial question on PPCG tends to allow a large number of trivial distinct answers and thus it's easy to find one that hasn't been given), and then get upvoted by other Hot Network Questions visitors. This tends to lead to a self-reinforcing cycle of reputation gain among people who are new to the site, which often gets well out of control and gives people privileges well before they have experience with how they work on the site (this is an extreme example, but the same thing happens to a lesser extent frequently). Based on the other comments here, I don't think this is an issue that's unique to PPCG either.

As such, any approach based on reputation should probably exclude not only the reputation from the association bonus, but also any reputation gained from the question itself. As written, the proposal is likely to help (because drive-by upvotes are, on most sites, more common than drive-by answers), but not to the extent that people are likely hoping.

Answer 5

I think that a much simpler solution would be to move the "right to upvote" up to 125 reputation points, commensurate with the "right to downvote".

Currently, a user requires 15 reputation to upvote questions and answers, and 125 reputation to downvote. I suspect that I understand why things were set up this way—downvoting is, in some sense, more open to abuse than upvoting, hence there is a higher bar. Unfortunately, the association bonus makes it possible for users to upvote immediately, but requires more work if they want to downvote. This means that if a question makes it onto the Hot Network Questions (HNQ) list, there is likely to be a surge of interactions from users who may upvote, but not downvote. As has been noted repeatedly in other answers and comments here, this is profoundly distorting, and not healthy for the ecosystem.

Heck, I would even support changing the reputation requirement for upvoting up to 102 points (universally). Right now, the issue is that there is literally no bar for entry for upvoting. You don't have to demonstrate any kind of attachment to a site (beyond your first) before you can vote everywhere. It is very simple to create a new account and start upvoting with your association bonus. If a user has to post a positively received question or answer, or even have an edit suggested, then they are clearing a higher bar, and are demonstrating an interest in (and commitment to) the new site.

In any event, I believe that the goal should to prevent drive-by voting of any kind (up or down). Aligning the upvote privilege with downvote privilege seems like a reasonable solution to me.

Answer 6

This effectively removes half of the feedback that Hnq visitors can send. I could explain why, but Jeff Atwood does it better. To quote the relevant parts:

The advantage of this system is that nobody gets downvoted, but at a steep cost: we’ve lost half the potential information. If a post has zero upvotes, does that mean it’s bad? incorrect? uninteresting? mediocre? There’s no way to tell, because zero has multiple meanings.

And when a question (from communities that aren’t massive) hits hnq, votes from the hnq make up a significant amount of votes on that question and it’s answers. This means when a question has a low score, it becomes even more difficult to tell what it means.

Normally, a post with a low score can mean one of two things:

• uninteresting or complicated
• Controversial

You can tell thick one of these it is with the view vote counts or an userscript. However, when a question hits the hnq, another option is added:

• people who actively use the site know this post sucks, but hnq voters upvote it for no reason

Even with a ton of rep and plenty of userscripts, the only way to tell wether this is true, or a post is controversial is to see who’s casting votes, and that’s only available to employees (and moderators in a more limited fashion, but I don’t know how that works).

So, other than the ridiculous amounts of rep you can gain by posting wrong answers, this is a problem because it makes

it harder to know what a post’s score means.

Answer 7

I’ve marked this as ‘status-declined’ for now as we have no plans in the near-term to make updates to hot network questions. Related, we are doing some research around voting as issues with voting is one of the top themes we hear about.

Answer 8

As an alternative to Xander's answer perhaps rethink what an association bonus means.

Privileges on a graduated site are currently as follows:

Reputation	Privilege	Description
100	edit community wiki	Collaborate on the editing and improvement of wiki posts
100	create chat rooms	Create new chat rooms
75	set bounties	Offer some of your reputation as bounty on a question
50	comment everywhere	Leave comments on other people's posts
20	talk in chat	Participate in this site's chat rooms
15	flag posts	Bring content to the attention of the community via flags
15	vote up	Indicate when questions and answers are useful
10	remove new user restrictions	Post more links, answer protected questions
10	create wiki posts	Create answers that can be easily edited by most users
5	participate in meta	Discuss the site itself: bugs, feedback, and governance
1	create posts	Ask a question or contribute an answer

You would get the effect which the OP asks for, if you reduced the association bonus to 10 instead of 100.

That would be not enough to vote, but is enough to "remove new user restrictions" which sounds about right.

The one other thing that experienced users (i.e. with an association bonus) maybe ought to be able to do is flag (and chat?) -- so perhaps reduce that privilege from 15 to 10.

Anyway, that would mean you don't get to vote unless you earn some reputation on the site.

However that would be a problem when bootstrapping a new Beta site, which maybe depends on some of the first beta users starting with an association bonus, which lets them cast the site's first votes.