I know this has been discussed before, and that there was a change in the rep system because of it. But I propose that the issue be re-opened.

The reputation system still doesn't work. Even if you "only" get half the reputation for getting your question upvoted, it still means you can easily gain a huge amount of rep by just asking tons of questions and not answering any.

Exhibit A

As an example, according to his reputation this user is among the 1500 most trusted members of our community. I don't know about you, but for me this actually means that reputation has no value at all. Someone who solely uses the site as a library and the users as his information manservants has no business being anywhere close to the top, no matter how good his questions are.

As I understand it, the whole point of the site is taking some and then giving something back. Shouldn't the reputation system then be specifically designed to encourage this?

The fact of the matter is, noone is going to not-ask-a-question because he won't be getting rep for it. If he has a question, he's going to ask. He wants an answer, so it's in his interest anyway, right?

So then why are we rewarding reputation for it?

Reputation should represent how much the community trusts you, and how much you have contributed to making this site as great as it is. The people that ask the questions are not the ones that make it great. It's all those that put in the time and effort to provide the good answers.

These are the guys that deserve our recognition, the ones that deserve all the reputation.

When I joined the site, I started with a huge amount of respect for anyone with more than a couple hundred rep. When I finally started gaining some myself, I took it with a certain amount of pride. I felt my help was valued by the community, which is exactly how reputation should work.

How 'bout we make it all about that again?

  • 1
    Good points. Related/similar: meta.stackexchange.com/questions/70991/… – Pekka Mar 13 '11 at 23:13
  • 7
    And you might like this: meta.stackexchange.com/questions/81929/… – Pekka Mar 13 '11 at 23:14
  • 6
    And +1 for a coherent, well-researched first post. – Pekka Mar 13 '11 at 23:24
  • 9
    Well, without questions there would be no answers, right? – NullUserException อ_อ Mar 13 '11 at 23:32
  • 2
    @Pekka Thank you for the links. The idea of a 500 cap seems like a good place to start and would probably fix the worst cases. But I think it could be taken further. In my opinion someone who doesn't give shouldn't get. I am referring to reputation alone here, not help. Reputation should only be rewarded if someone is doing right by his peers and also giving back, that's what makes them a valueable member of the community. – Max Dohme Mar 13 '11 at 23:44
  • 3
    @NullPointer True, but I believe questions aren't something that we need to encourage with rep. People should be asking questions because they want answers, not because they want reputation. Write a great question, get a great answer! Doesn't that seem like a fair trade to you? – Max Dohme Mar 13 '11 at 23:48
  • 1
    +1 I can see that some rep will encourage questioners to perhaps join the community more fully, so a cap may be a reasonable compromise. But in the 100s - no higher. – amelvin Mar 13 '11 at 23:54
  • 1
    This discussion is already continuing at meta.stackexchange.com/questions/81929/… so unless you are arguing for a specific change (ie, you have a proposal that's different) then this is really a duplicate. – Pollyanna Mar 14 '11 at 0:34
  • 1
    @Adam The main difference is that I propose not awarding questions with reputation at all if the user is not also posting answers. Do you think this should rather be an answer to Jeff's question? When I searched for similar topics I missed that one, or else I might have posted it there from the start. – Max Dohme Mar 14 '11 at 0:42
  • @EyeSeeEm A few of the answers on Jeff's question do tie question reputation gain to answer reputation gain - in other words, the user has to be posting answers that get reputation in order to get significant question reputation. So I'd say your proposal is pretty close to that. Either way, it's very likely that Stack Exchange, Inc will implement one of the suggestions there, since the question is more established, and since Jeff has placed a bounty on it (ending in two days). If you really think your proposal is better than those discussion there, then you should put it there as well. – Pollyanna Mar 14 '11 at 0:48
  • 2
    -1, the reputation system certainly has problems, but discouraging asking is not the right solution. Reducing the number of good questions will reduce the number of good answers. – Pops Mar 14 '11 at 14:48
  • 2
    @popular You aren't discouraging asking questions by not giving rep. Questions are and should be asked for the purpose of getting an answer. If your question is good, you are rewarded by good answers. It really is that easy. – Max Dohme Mar 14 '11 at 14:51
  • 2
    @Eye, that's simply not true; asking questions you already know the answers for is a well-established part of the site. In any case, what you're describing is the missing piece in the positive reinforcement/negative reinforcement/punishment system, so it will modify users' behavior whether it's intended to or not. – Pops Mar 14 '11 at 14:56
  • 1
    @popular Okay, but do you really need to reward those questions with reputation? Will not getting reputation really discourage people from posing those questions? – Max Dohme Mar 14 '11 at 15:43
  • 1
    Wow. 1838 questions. 9 answers. 20959 rep. This is possible? – TRiG Mar 24 '13 at 2:03

I can only agree - not sure why reputation is given for questions any more. It made sense in the past to help build up the community - by encouraging everyone who contributed.

But the number of questions (and unanswered ones) is ramping up ever faster. If unanswered questions are rising then people who can answer questions are now a relative (not absolute) rarity and should therefore be encouraged more than people who just ask questions.

When people ask questions, they get a solution - that is their reward. Getting badges is bonus enough.

If there is a feeling that entirely turning off the question rep taps will cause the questions to dry up then put an absolute cap on the reputation that questions can yield, say 500 points.

The blankman account that the OP flags up should not be able to rip nearly 10k in reputation simply for asking an awful lot of questions. It takes a lot of answers to generate 10k in reputation in 2011. That merely asking 1000+ questions suggests someone is more trustworthy (or whatever rep definition) than all the 5k, 6k, 7k users out there who have answered hundreds of questions sticks in the throat.

  • 7
    Ahahahaha! I just noticed that the only strongly upvoted answer (6+) in Exhitbit A's account is to the question "Your personal, successful coding practices." I half expected the answer to say "Crowdsource questions on SO on every possible occasion" :D – Pekka Mar 13 '11 at 23:59
  • 2
    OT: "It takes a lot of answers to generate 10k in reputation in 2011" - while we are at it, why people who gained lots of rep on poll-like questions in the past (forbidden practice today) should keep that rep? Is it fair in regard to newcomers? That could be another topic to discuss (which I believe would be very controversial)... – Xoum Mar 14 '11 at 0:16
  • 2
    @Xoum They don't. As these questions are closed and deleted, and reputation recalcs done, the reputation gained from them is lost. If you see a question that doesn't belong on stackoverflow, please vote to close it. – Pollyanna Mar 14 '11 at 0:41
  • 2
    "It takes a lot of answers to generate 10k in reputation in 2011." I have a hard time believing this statement. Look at the people in the weekly top reputation league: stackoverflow.com/users?tab=reputation&filter=week I know a lot of people keep harping on the idea of how hard it is to generate 10k rep, but it's not just one or two people that hit 200 rep daily, and they aren't using special tricks to get their reputation. – Pollyanna Mar 14 '11 at 0:44
  • 1
    @Adam The average number of answer upvotes per question is steadily declining and the number of questions is exploding meta.stackexchange.com/questions/74271/… - so questions and their answers are now getting fewer and fewer views - and fewer and fewer opportunities to get get votes. I also think the electorate badge is a problem meta.stackexchange.com/questions/80339/…. Of course you can still hit 200 but it takes more answers than it did in 2008. – amelvin Mar 14 '11 at 9:48

I think there should be a not-insignificant incentive to ask good questions, at least at first.

I'm not opposed to a cap on reputation past a certain point so that high reputation has a greater meaning, but encouraging new users to write good questions is important as well.


A big problem I haven't seen mentioned in here nor in linked threads is that I think that reputation is kind of a "expertness" metric. If I see someone with 50k reputation I automatically assume, that his answer has more merit than someone's with 250 rep. And I did this EVEN THOUGH I know how reputation works.

Lately I've seen low quality answers from people with very high reputation and upon further examination I noticed a trend of asking simple questions. Second most top rated question on StackOverflow is How to undo last commit(s) in Git?. I'd argue that it did not take too much work to ask the question. Also there is nothing to be rewarded for in for example this question: What is the difference between 'git pull' and 'git fetch'?. It's very common, but the person asking did not do anything special to get reward for.

Question upvotes more often than not reflect the importance and commonness of the question, rather than quality.

Therefore it should be only loosely related to reputation. I think it would be great to give badges for great questions. That way we're saying "Thank you for your contribution, we've noticed that your question is very common and therefore helpful". But we should not devalue reputation just because someone asked a very common question.


Today I noticed that I'm as well having some reputational gains from the question upvoting.

There is this answer I've been working on for ~ 3 years improving it multiple times. It gained me 24 upvotes. But I gained much more upvotes WITHOUT doing any actual work in this and this. I just plain asked what I needed to know. Maybe the whole debate was too focused on people abusing it while it should be more about whether it reflects anything and if it actually motivates anyone.


(This question was asked in 2011 and was probably about SO. But it's here on Meta.SE now, where it would apply to all 160+ sites on the network, and most of them don't have SO's scale.)

A good, well-asked question benefits the community, not just the asker.1 Most importantly, it provides an entry for good answers, the pearls that we optimize for. Demonstrably, most people don't ask and self-answer most of the time, which they could do to share their knowledge; they need a prompt. Why shouldn't we reward good questions that provide that?

If somebody is getting lots of rep from bad questions, your community needs to fix that. Don't upvote questions that don't meet a reasonable level of usefulness! Downvote ones that are actively bad! You get the practices that you reward through votes.

1 Heck, I've sometimes asked questions that weren't even my questions, as a way of helping out somebody who can't ask directly.

  • 1
    this is especially important on small sites that are finding their way. – Kate Gregory Apr 2 '17 at 11:45

I think questions should only add to your points up to 500 points because all these users getting 50,000 reputation for asking popular elementary questions is just polluting the quality of the site.

( I was informed this is already implemented for edits because people were making minor whitespace edits and getting points )

That way the reward for asking a question would be getting an answer instead of points

It may help stackexchange fight the quality versus popularity conflict of interest.


You must log in to answer this question.

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