Question-related badges are too easy to get. Can we change the Popular, Notable and Famous Question badge requirements?

Question

Question-related badges are far too easy to obtain. Badges awarded to those who ask questions greatly outweigh those which are awarded for answers. Of course, it doesn't help that there are more question-related badges, however the main culprit is the set of badges which are based on question views: Popular Question (1,000 views), Notable Question (2,500 views) and Famous Question (10,000 views).

---

Firstly, lets take a look at the top three badges for each category (Bronze, Silver and Gold):

Gold

1. Famous Question (62,192)
2. Fanatic (8,076)
3. Great Answer (6,928)

The Famous Question badge has been awarded more times than all other Gold badges combined.

Silver

1. Notable Question (345,834)
2. Yearling (231,080)
3. Necromancer (72,875)

There are more Notable Question badges than the second-most and third-most Silver badges combined.

Bronze

1. Popular Question (878,622)
2. Student (606,325)
3. Editor (477,659)

As you can see, the view-related badges dominate all 3 badge categories by a long way.

---

One of the most viewed questions on StackOverflow is "How to undo last commit(s) in Git?" with an impressive 573,219 views, 2,410 upvotes and 1,143 favourites.

This is a famous question.

The three most recent Famous Question badges were awarded for "How to read XML file in android", "Scp command syntax for copying a folder from local machine to a remote server" and "What's the CSS style I need to make a readonly textbox control look disabled".

Combined these questions amount to 30,000 views, 6 upvotes and 0 favourites. One of which is closed!

These simply aren't famous questions.

---

Proposed Change

I think a more fitting system would be:

• Popular Question : Ask a question with 5,000 views, at least 5 upvotes and 1 favourite.
• Notable Question : Ask a question with 15,000 views, at least 10 upvotes and 2 favourites.
• Famous Question : Ask a question with 75,000 views, at least 30 upvotes and 5 favourites.

Answer 1

I disagree with changing the criteria for what counts as Popular, Notable, and Famous question badges.

StackOverflow has been around for a while, and the user base has increased substantially over the years. Many of the questions you see with those badges are older questions that have had much more time to gain views.

As a quick test, I ran this Data.SE query to find out how many famous questions badges were actually asked recently:

2008    10261
2009    23210
2010    16504
2011    11514
2012    2250
2013    22

As you can see, the majority of questions considered "Famous" were ones that were asked years ago. I'm sure many of them are also simple/easy questions that hadn't been asked before, which now frequently come up in Google searches.

Traffic also plays a big role in this.

To compare, check out a smaller SE site, such as Programmers, which is one of the larger SE 2.0 sites that has been around for a few years. Currently they have 246 Famous Question badges, and that's actually lower than the other gold badges you mentioned; 250 Fanatic badges and 404 Great Answer badges.

Even if we did change the criteria though, then what happens in 5 years? 10 years? Do we keep raising the bar based on how much time has passed and how much traffic the site gets?

So to summarize, don't change the criteria for what counts as "Popular", "Notable", and "Famous" just based on site traffic or age. Those should be static and achievable goals for anyone posting a question today.

And if your question survives for a few years and happens to end up being viewed and possibly helping over 10,000 people over the course if it's lifetime, then I'd say that's great and worthy of a gold badge!

Answer 2

I sort-of agree with your argument but in changing the numbers you're going far too far. There are only 2,260 questions with 75,000 views, or more.

If you want to make the system fairer then I would stick with just views being considered (there are already badges for everything else mentioned) but increase the number of views. Make,

• Gold 25,000 views; there's 15,933 questions
• Silver 10,000 views; there's 61,688 questions
• Bronze 2,000 views; there's 424,297 questions

Famous, as a definition means well-known, not upvoted a lot. Just because it's been potentially helpful to tens of thousands of people doesn't mean that they all upvoted it; they might not have Stack Overflow accounts. If votes are taken into account then it should also include the "helpful" votes that anonymous users can provide.

Also, you, and I, are only looking at Stack Overflow, which has a huge number of visitors. Don't forget that there are plenty of other sites in the network where a change in the number views that gives you the badge almost certainly isn't warranted.

Lastly, if this is made retroactive, there'll be riots as people's badges are revoked... is it worth it? If you don't make it retroactive then you're discriminating against new questions.

Answer 3

I agree with the general gist of the issue. Here's my reasoning.

Gold badges are supposed to be the most coveted, the most important, and therefore the rarest and most difficult to get. And we know this because of how few of them are acquired.

Here is the list of all badges on SO. Gold badges are by far the fewest.

Unless you look at Famous Question. Here are some fun facts:

• There are more Gold Badge Famous Questions than there are Silver Badge Good Answer.
• There are almost as many Famous Question badges as Necromancer (Silver).
• There are almost twice as many Famous Question badges as there are Bronze Tag badges. Across all Stack Overflow tags. That's right, for ever person who's gotten a cumulative answer score of 100 on a tag, there are twice as many questions that are considered "Famous".
• There are fewer people on all of SO who have completed at least one review task (Custodian) than there are Famous Question badges.
• If you add up all of the other Gold badge tallies, all of them, you get less than half of the Famous Question count. That means 2 out of 3 Gold Badges are Famous Questions.

To provide counterpoint to Rachel's answer:

And if your question survives for a few years and happens to end up being viewed and possibly helping over 10,000 people over the course if it's lifetime, then I'd say that's great and worthy of a gold badge!

Questions don't die (at least, not answered ones). Eventually, given sufficient time, every question will become a "Famous Question". And when every question is famous, how will you tell the difference between the exceptional famous questions and the "famous due to the passage of time"?

Rather than rollback tens of thousands of badges, I say we simply retire the badges (like Analytical), and declare a new set of them. For these, you get them if your question gets X unique views in a fixed time period since it was posted. Maybe 6 months.

That way, you're actually rewarding people who deliberately spread the word on their questions, rather than rewarding people who do nothing at all.

Famous Question should not be easy to get. This should be true for any Gold badge.

Answer 4

Stack Overflow is not like other sites. The big three are huge and old; they have lots of questions, lots of votes, and lots of visiblity. The current forms of the badges are already difficult to earn on some smaller sites; changing them to make them more elite for SO would make them useless everywhere else.

Perhaps the answer is to make the thresholds adjustable per-site. That would probably entail a fair bit of work, since the badge-calculation code almost certainly wasn't written with configurability in mind. It could be changed, but it doesn't seem important enough.

On any given site, some badges are "too easy" and some are "too hard". That's just life. So SO gives out "too many" question-view badges -- is that really a problem? It's probably also a lot easier to get Nice/Good/Great questions/answer there (we have no Greats yet on Mi Yodeya); so what? Where does it end if you go down this badge-changing path?

The thresholds are what they are; no number is right for all sites, but in the absence of per-site configurability, it's better to leave it alone than try to adjust in favor of one site at the expense of others.