The Stack Exchange reputation system: What's working? What's not?

Question

Update

Thank you everyone for the feedback you have provided. At this time, we will work through the feedback we have received to help us determine where to take our research on reputation next.

Please remember that this is just research and does not impact any product decision now or in the immediate future. We look forward to returning to the community to discuss more related and adjacent topics before embarking on any product development.

---

Having incentive systems to participate in online communities (like Stack Overflow) is not a new concept. Basically every Q&A site or forum has had some type of points or karma system, and they all work to varying degrees.

But as we’ve seen, our system and community in its current state has a high barrier to entry. Asking or answering a question back in 2009 was a lot easier as we were building the library of knowledge that exists today. There were a lot more questions yet to be asked!

We still want users to be able to participate on our site and feel as though they are a part of this community. But unfortunately, we’ve heard from so many of our users, particularly newer ones, that the learning curve is steep and it’s difficult to find ways to engage on the site.

In order to combat some of that, we’ve been experimenting with the Staging Ground work and with better onboarding for new users (currently only on Stack Overflow). These are just a few early attempts, and we know there could be additional solutions out there.

However, as the network has grown and changed, we’ve started thinking about our reputation and privilege system and how we might want to evolve it. Our Research and Community teams internally have been doing a lot of work to understand incentive systems more broadly, but we also wanted to hear from you all about what you think works really well with Stack Exchange’s reputation/privilege systems, as well as what is broken with reputation. We’d also love to hear about other rep/karma/incentive systems from other online communities that you think are exemplary that we could learn something from.

We don’t have any concrete plans yet. This is really the first part of a long process to see how we could potentially rethink reputation and privileges on this site and our larger network. We’ll be doing several rounds of research as a part of this larger exploration, and will update you on our findings as we go, so please be on the lookout for future updates.

So, for now, our questions for you are:

• What is really great about our reputation & privilege system? What do you think makes it great, and why is it worth keeping?
• What is broken about it, and why? Are there any solutions to it?
• What other systems work really well that we could learn from?

Since there are many answers already written now, if you'd like to add one but want to check first that someone hasn't already said what you want to say, try using the searchbar with inquestion:387356 is:answer ....

Answer 1

What is really great about our reputation & privilege system? What do you think makes it great, and why is it worth keeping?

It allows people who have authored very good answers or questions to edit other questions to improve them, close bad questions, etc. I think that stuff stuff should be kept.

What is broken about it, and why? Are there any solutions to it?

A lot. Just because somebody knows how to make good questions or answers doesn't mean they know what should be deleted. Make it so rep is not attached to privileges and instead actions. For example, Codidact (another Q/A site) gives privileges based on relevant actions. For example, editing freely once getting a lot of suggested edits approved. You can also take a look at roles the community would like.

What other systems work really well that we could learn from?

Like I said before, you can learn from Codidact and Discourse does with trust levels. You can lose TL3 (trust level 3) (which lets you edit categories and tags) if you are not active enough in the past 100 day by default.

Answer 2

What is working:

The rep cap, as I explained in 2012: it incentivizes posting not just an answer (for votes which are capped anyway), but the answer (the one chosen by the OP).
It is at the core of what makes Stack Overflow unique: a Q&A site where finding the right answer is actively promoted.

It is not perfect, as others explained in their answers to this question, with the "passive income" issue. But the main goal remains important: more points above rep cap means posting the right answers.

What is not working:

Most of everything else.
As Mithical explained, votes represent too many things (expertise, site familiarity, gamification, but also curation).
See also Mithical's answer on this page.

In particular, I would like a better way to assess expertise: I have 2000+ rep on scala on more than 200 answers (and a gold badge!), and yet, I used those answers to learn enough of that language to understand it was not a good fit for me (as I mentioned in "How to earn a million reputation on Stack Overflow: be of service to others").
Or: I have 1300+ rep on ruby (and a silver badge), but I know next to nothing about the language: those were questions about Git or Docker, involving Ruby.

The main other pain point is "what does a downvote mean?":

Take a legitimate question, validated by Staging Ground, about programming, showing research and what has been tried.
But what has been tried is a bit "dumb" (because you are a beginner on that topic): chances are, the question will be downvoted (strike 1).
And, after a few of those (again, good questions, but showing a poor understanding of the topic, as a beginner), you won't be able to ask other questions (strike 2: "We are no longer accepting questions/answers from this account")
The experience is quite dreadful for newcomers.
Don't take my word for it: this video "Uncovering Stack Overflow's TOXICITY" from Alex Ziskind illustrates the point nicely.

I do not dispute the merit of downvotes, as they help curate the site and prevent "bad" questions (vague, without research), and I hope the staging ground new review process will help mitigate the current questions "downvoted/closed in mere minutes".

But I would like to see a way to express the difference between:

• your question is valid, but expresses an incorrect idea/comprehension (should not be downvoted)
• your question is bad (off-topic, not enough focus, ...) (can be downvoted and/or closed)

Answer 3

It is hard to answer uniformly for every site because the cultures differ. For example, on the physics site answers are closed if they don't show research or if they are homework like. This is valid. We don't want to do people's homework for them. On the other hand, this is the prime use of Stack Overflow. I am a professional programmer. That is where I go for documentation or to get an easy answer to a problem. And this is valid.

Reputation keeps me coming back to answer questions at the physics site. I do it because I like to be helpful. Reputation is what makes me feel helpful.

It also motivates me to write good answers that get lots of votes. It is disappointing when number of votes often does not relate to the quality of the answer. That's life. On the average, better answers get more votes.

Downvotes discourage new posters. New posters don't know our rules about what we answer and what we close. They don't know our preferences to how we like questions to be posed. The wrong kind of question often gets downvoted and closed.

We justify downvotes as a way to encourage better questions. But a new poster doesn't know this. The user feels unwelcome and doesn't come back. I often feel that a courteous explanation of why the question doesn't fit our site and restraint on downvotes would improve this.

Here is an example question of a homework-like question that was closed with downvotes. I don't feel this poster should necessarily be discouraged. He should be encouraged to reframe his question so it fits the physics site better. That might not fit his needs - he might be looking for someone to solve the problem for him. If so, he is free to look elsewhere. Or he might ask again in a way that we accept. Unitary operator corresponding to Galilean transformations

Duplicate answers are a problem. Closing duplicates does keep us from having lots of duplicates. This does have benefits.

But answers more than a few days old don't get votes. I don't believe many people read them. When I add an answer to an old question it rarely ever gets a vote. When I link an old question to a new one (because it's a duplicate or related or relevant), the linked question seldom gets votes.

When a new poster asks a question and it is closed as a duplicate, it can be discouraging. Particularly when the duplicate answers at an expert level, and the OP is a beginner.

I don't know how to best handle this.

Downvoting and closing questions that could be answered. This is similar to duplicates. Some questions get closed too hastily. There are some common reasons for this.

The question contains a mistake or misconception. Or the poster is too confused or new to physics to pose his question clearly.

This is often interpreted as the poster being too ignorant or hostile to physics to understand the answer. There is no point to answering.

This is a valid concern. We get many such posters. Some are religious. Some genuinely confuse physics with movie special effects. Some have their own homemade physics theories and cannot be convinced that anything else is right.

But other posters are just confused. A clear explanation would be of real benefit.

Some posters just pose their question wrong. We only address questions about mainstream physics. A question like "Why can't I go faster than light?" may well get an answer. But "Suppose I go faster than light. Would time go backwards?" smacks of a personal theory and/or poor understanding. It is likely to get closed.

More thought before downvoting and closing these would help.

Update on questions that could be answered

Here are some questions that illustrate what I mean.

If you could zero out your momentum, could you travel faster than the speed of light

This is obviously one of the lowest quality questions. The poster had heard something ridiculous and wanted an explanation. Most of the responses were in line with that. But though the poster was not familiar with physics, he was a reasonable person and was willing to listen. Giving a straightforward polite response that this is nonsense was helpful. He learned from it.

Special Relativity and Simultaneity and Movement

This poster has a better understanding of physics. He is learning special relativity (SR). Like many students, he is finding the concepts confusing. He sees what appears to be an absurd implication of SR and asks why it isn't so. This is a perfectly reasonable question right up our alley. If he had worded it differently, it would have been answered and maybe even up voted. People were hasty to down vote because they thought the poster expected silly implications. They were hasty to close as an inappropriate duplicate.

Is the Multiverse just a new religion?

This one is dubious. Many religious posters have the idea that truth comes from God, and physics disagrees with this truth. They are often unwilling to listen to ungodly answers. Such posts are downvoted and closed almost as a knee jerk.

You can see that this poster is somewhat sceptical of physics and has some misconceptions. Still he doesn't sound utterly convinced that physics is hogwash. I would have been willing to explain that the many-worlds interpretation is a speculative thought, something that might be true. If we find evidence, we will be more convinced. This makes it different from religion. This means physics isn't as unreasonable as he had thought.

He might have listened, and he might not. But as is, he deleted his question. I would be surprised if we ever see him again. His take surely must be that he asked what he considered a reasonable question. He was shouted down. Physicists are unreasonable and physics is hogwash.

The common thread is that a more tolerant response benefits the posters. It fulfills the mission of the site. It counteracts the reputation the physics site has built of being elitist.

Are time crystals real?

Here is one more, a video that shows how Don Lincoln of Fermilab handled one such issue. I wouldn't necessarily answer like this. But it is a lot of fun and clearly explains the misconceptions and the real ideas.

Answer 4

40 answers and counting, so I won't rehash what's been said, and sorry if my point has already been covered and I missed it.

There's a reputation problem on many small sites, but it isn't due to the design of the reputation system. For reputation to serve its purpose, people need to vote. The majority of the benefit of the Q&A knowledge, and the voting that sorts it, is for people finding the information in a search. They can't vote.

All of the voting is done by site users who have at least minimal reputation, generally meaning they have done some posting that was well-received by the community.

That's all great in theory. But many small sites have very little participation by users with voting privileges. Few of these people visit regularly. Of those who do, few bother to vote.

It becomes a vicious circle, where sites get fewer answers because investing the time to write good answers isn't rewarding on those sites. Lack of interest in writing answers and fewer answers by others means less reason for regulars to visit the site often.

For new contributors, it takes so long, and so many posts, to build the reputation to fully participate, that few continue to do so. Of the few who do, even fewer build the rep for community moderation. So small sites atrophy as participation by experienced users wanes and they are not replaced by involved new users.

Reputation plays a key role, but it doesn't work on many small sites because so few people vote. Tweaking the algorithms won't fix it, and radically changing it corrupts its purpose. It's a user behavior problem, not a reputation system problem.

Answer 5

I'm a long-time member of Stack Overflow and many other sites on the Stack Exchange Network. In fact, I think I joined before even the trilogy, when Stack Overflow stood alone. And I've never cared a jot about rep.

I like seeing the little green flag which means that someone somewhere found something I wrote useful. I always click to check which post received an upvote. (It's often something I wrote yonks ago off the top of my head with no effort involved.) It gives me warm fuzzy feelings, seeing that green flag. The actual number attached to it is completely irrelevant. Certainly I have no idea what my reputation on any given site is without checking, and I'd like it to be made less obvious.

(I do care about having enough rep to edit posts, because I do a lot of editing. I actively enjoy proofreading.)

What is really great about our reputation & privilege system? What do you think makes it great, and why is it worth keeping?

Allowing basic moderation abilities on the basis of experience with the system and value provided to the community makes sense, and rep is a reasonable proxy for that. Perhaps some people also find it rewarding in some way? I cannot claim to understand that.

What is broken about it, and why? Are there any solutions to it?

It's too big a deal. It's in our face all the time, and people are encouraged to chase it. See The problem with extrinsic motivation for more. I cannot imagine that any of the truely great contributions to Stack Exchange (https://stackoverflow.com/a/11227902 comes to mind) were made by someone chasing rep points.

Answer 6

This is quite a late answer, sorry, but I think the initial premise is wrong; the question is wrong (at least, for Stack Overflow, which is still, I understand, the biggest site in the SE network -- bigger than all the others added together).

SO shouldn't be trying to gain new users. It should be trying to grow the number of programmers on the site. I realise that this is a very similar thing, but I think it's an important distinction and recognising it may throw some light on the problem.

First, there's the 'growth' word. Attracting new users is part of growth, but it also recognises that retaining existing users is also important.

Then there's the 'Programmers' part. We don't want new users -- half the posts I close are off topic for SO. We specifically want to attract programmers. And, mostly, we want (trigger alert) professional programmers.

I've heard it repeated often that SO is unfriendly to new users. I think that anyone working in this community is probably not a snowflake and a couple of rude comments aren't enough to dissuade people. And, the learning curve is ok for programmers because we're programmers.

Having said that, I think the privilege system is broken. The system currently encourages answering duplicates and off-topic questions rather than flagging or downvoting them. That should change. And, I agree with previous posts that reputation doesn't demonstrate good moderation capabilities. Though, length of tenure might. I've been here ten years, and I know how to moderate, but I can't edit a tag wiki without having it reviewed first, which is a bit silly.

I'd remove rep penalties for downvotes, and I'd add a rep penalty for answering a question that is later marked as a dupe, and double that for answering an off-topic question.

I'd give more moderation rights to people based on tenure, but only if requested. We can decide for ourselves if we want that privilege.

Answer 7

Users without up/downvoting privileges should be able to vote up and down on answers to their own questions

I find it odd that users who haven't yet unlocked the up/downvote privileges—typically new users—has no voting ability on answers to their own question, but has to earn reputation in order to do this. I have no problem in even allowing downvotes in this specific case (to mark an answer seeming to be not helpful).

Answer 8

Speaking about reputation without speaking about voting (the main source of reputation) will be the opposite of intelligence. All the woes caused by reputation that people complain about are caused by our voting culture. One of the most damaging behaviors of our voting culture is penalizing and discouraging downvotes. Users should be able to vote their votes without worries. Negative feedback is more important when there's so much noise, so let’s fix that by removing all disincentives on downvotes: downvoting isn't a separate privilege; it is a single one with upvoting, the voting privilege; upvoting cost nothing, so downvoting should also cost nothing; downvoting makes a post total score to be reduced by one. It should also happen with reputation; the amount of reputation gained by an upvote is the same than a downvote.

Once you implement the first two, track the behavior of the ratio between upvotes and downvotes between users with the voting privilege and users without it (include both anonymous and registered ones). That would address some of the most egregious behaviors of the voting culture, and help towards creating a better voting culture.

The third is mostly aimed towards creating actual disincentives that the voting culture caused: post as often as possible. These three in tandem will steer the communities towards more productive behaviors and reduce further damage.

Those actions however, do not do anything to address the damage already done. I have a couple of ideas:

• More pruning: questions that are asked frequently (by the number of questions that are closed as duplicate against it), should be soft limited to 10 and hard limited to 50. Add some sort of expert queue so that users with knowledge about the topic (make it a badge privilege) can evaluate why users aren't finding the target post and take actions to make it either, easier to find or harder to misinterpret. Allow those users to see the deleted linked posts even if they don't have 10k yet. The limits should be enforced by the roomba and all posts deleted with by this should have the same effect as if it was deleted before the 60 days rule on reputation. How will this fix reputation: this will disincentivize and revert back the damage done by answering duplicate questions.

• Questions locked that are also closed by unclear, too broad or opinion based should have the same treatment deleted posts. All reputation gains of the question and associated answers should be invalidated. Locked posts are usually reserved to posts that would otherwise be deleted, so let’s make sure they have the same treatment as a deleted post without the actual deletion.

Those are the most lowhanging fruits that would disincentivize certain undesired behaviors that are done for reputation points. Others will take more time or require a deeper change of culture, like incentivizing edits to older posts, reducing the effects of popularity on voting behavior, and incentivizing posts that demonstrated both deep knowledge about the topic and accessibility in the presentation.

Answer 9

What is broken about it, and why? Are there any solutions to it?

It's not broken in the sense "needs to be fixed", but something that could maybe be modified/extended. I'm talking about janitorial services. I participate in some stacks, where, for some of them, it's more about reading, and, for some others, about answering/reading. But, for most of them, I'd like to be able to clean what can be, and put some extra efforts in it.

In order to do that, I need to gain reputation, and I can't always do that on that particular stack. So, I see many things I could improve/correct without being a diamond mod or "trusted" user (of this stack). But I am a trusted one on others. If I can edit some posts, VTC/VTD because I'm trusted, why would that trust stop at the border of another stack?

I'd like to think and talk about moderation rights' portability. Not for all privileges, but some basic ones (to be determined). The ones you can use to clean some mess, to improve a post by correcting some spelling/typo/grammar. At the moment, on some stacks, I have to go through some process and steps. It's a fair and logical process. But doing so gives more work to the "cleaning team" in the review queue. Maybe we could gain in "productivity" if a trusted user on one (or more) stack(s) could gain access to more janitorial tools.

In accordance with procedures to be laid down by SE, for some duties/tools, it could be immediately, or after a kind of probation period, or a number of duties achieved or a lowered thresold for reputation, or even a combination of some of those.

Answer 10

There is currently no consideration for reputation across sites when performing cross-network actions, specifically, in terms of migration. Here's a concrete example:

• A user has lots of reputation on site X.
• This user also visits site Y.
• They often find questions on Y that needs to be posted on site X (it could be off-topic on Y, or just more on-topic on X).
• Voting to close questions with a migration path to site Y is treated similarly to other closure votes, often ending in closure without migration.

If you have lots of reputation on site X, there's an understanding that you're well-informed of what types of questions belong on X.

---

This type of situation mainly occurs with Stack Overflow (SO = Y), which acts like a catch-all dumping ground for questions. Many questions that were on topic in (say) 2012 now have dedicated sites to answer that currently just muddy the waters on SO. It's closure review queue is constantly packed (like most other review queues). Using this SEDE query to count closures on SO, this chart displays the monthly closures over time:

As of the most recent months, between 12k-14k questions are closed every month (400-500 every day). Some of these may be legitimate closures aiming to migrate questions to a more suitable site (possibly from high-rep users on the target site).

There currently exists a migration path when closing to TeX.SE (X), but closure votes due to lack-of-information or debugging-details cause new-user questions to die on SO, when they would have received better feedback/responses on TeX.SE (despite the question quality).

Reference: What privilege should 30k users get?

Answer 11

And one more answer:

For me reputation points are mostly gamification and that has only very limited value. It's something attached to a user and may incentivize users to do more of the "right" things and less of the "wrong" things. I mostly treat content as detached from its creator and try to value it based on its own merits. The score of the content gives me a hint how good a certain piece of content could be. That's all I need (want) to know.

However, as a means to control behavior it's working okayish. The following aspects could maybe be tweaked.

• Moderation privileges (maybe the rep thresholds are not optimal, you could play around with these thresholds if only you had a way to measure how well people are performing in moderation)
• Asking privileges (ultimately the capacity of the network to curate, improve and finally answer new content is limited, so the asking privilege might be tied to positive prior interactions like editing, commenting, ...)
• Rep on questions and rep on answers was made equal some years ago, but it was never evaluated if this was actually beneficial to quality or rather decreased the value of the content. There is no intrinsic requirement for this rep to be the same, so maybe play around more with it, if you have a meaningful quantification how it affects content quality.
• Reward people for finding duplicates (but I have given up hope because that idea is super-old and discussed at length without being even experimented upon.).

But all these things might also work without reputation. For example the editing privilege could simply depend on a certain number of approved edits, the review privilege could simply depend on a certain number of approved reviews, .... Reputation mingles all that together into a single number and that is in my eyes its greatest weakness. I would like to de-emphasize reputation more in the future and instead control on quality as the central topic of every action on the network. Always ask yourself: "how does this improve the overall quality of the content". That's the right mindset. Of course keeping a sufficient large user base is part of the endeavor. Somebody needs to ask those really good questions that others then answer.

One of the strongest points of reputation is that you can lose it (via downvotes on created content). That influences behavior a lot and should be kept additionally to the downvotes.

Maybe related: What does Stack Overflow Inc. actually see as purpose of the site. "We ... want users to be able to participate on our site ..." doesn't really say how users should be able to participate on the site. Would for example just reading be a kind of participation?

Answer 12

What works - it works extremely well in all the obvious ways that you already know about. It was developed as alternative to forums, where valuable information is buried under a huge tide of noise. As a general principle, it does what it's designed to do, which is foreground good questions and good answers to same, and occlude the bad ones.

Bear that in mind while I say this next bit.

What's broken - humans are animals. If other people like something, you are predisposed to like it too. If they dislike it, you are predisposed to dislike it. No amount of intelligence or education makes you more immune to this tendency than any other human - or animal.

There's a snowball effect. Often vaguely good posts get rewarded way out of proportion, especially if the user posting them already has high rep.

As I don't want to disparage the efforts of other users, I'll use one of my own answers as an example:

Question

Why, in old movies and TV series, do they always use such extremely exaggerated wheel turns when driving a car?

My answer

In old cars you had to turn the wheel a lot further, because there was no power steering. For a feel of that, push a car when the engine is dead and try to steer it without engine assistance (it's hard!).

Also the actors would have to act driving with no idea what the screen behind them was doing, and often still do.

There are other answers that are a lot better - go into more detail about old filming techniques, describe how power steering and old cars work more accurately, and even one where someone went and asked a living 1950s actor that they knew. (one of those answers was from a user with higher rep than me but answered 3 days later, so admittedly there are other factors at play).

This snowball effect is not a particularly big deal - there's no way a user will start to rack up points without making some posts which are actually good. The snowball won't pick up lots of snow unless it had already picked up enough to start rolling down the hill. I honestly believe everyone who has rep of 100k plus has that much because they know something useful (note: I am not in that category and don't ever expect to be).

However, the snowball effect is a big deal in the opposite direction. As the question observes, asking an acceptable question these days is really hard. I'm not a newbie on my main site (History SE) but I still consider myself one on SO, even though I made my first post there 9 years ago. If I manage to ask a question that isn't closed, it will still usually not receive any upvotes - and part of needing to ask a question is being unable to always recognise it when answers to other questions tell you what you needed to know.

Looking at other (real) new users' questions, they often get a lot of downvotes before their question is closed. In other words they get much less slack that someone with lots of rep would get. You'd better make your first question perfect, or you're screwed.

For that reason, I strongly believe that for new users (not me, real new users) should have a reputation 'floor' on their questions. I'm aware that a user's rep can't fall below zero. I am saying that a new user's questions should never receive a score below zero, and they shouldn't receive notifications telling them that people are downvoting their question. And if a new user actually manages to ask a question which isn't downvoted or closed, they should get an automatic upvote. They won't stick around if their only reward is a zero - better than minus whatever, but still.

If a question is bad enough that it needs to be closed, then there is a mechanism to do that. If a user is trolling, again there are mechanisms to handle that - flagging and deleting questions, destroying user accounts. Neither needs to involve a horrible negative number shoved in a new user's face. (Comments are also a problem with new user's questions, but probably out of scope for this topic).

Second problem with rep system - cultural biases. I went into this one at length here. Just to add - this is mainly a problem at non-technical stacks like History. This is part of the bigger problem that the non-technical stacks don't receive as much attention as they should.

Third problem - you are not a gadget. This part is too radical for SO to accept, but I'm going to say it anyway. SO is a company which relies on the free work of very talented people in order to make millions. The real problem with rep is, ultimately, that it's meaningless internet points - this is people's only reward for creating huge wealth for SO's owners and shareholders. If you really want to make it sustainable and better, reward those 100k-plus users - financially. They are the people who, by moral right, own Stack Overflow.

Answer 13

I'll offer my thoughts as a new-ish contributor, who's slowly contributed maybe a dozen or so Q's and A's over a span of several years.

What about reputation is working to make me a good contributor:

• The existence of a reputation system. When I've thought of an unasked question, or found a question I can answer, the carrot of increasing my rep does motivate me to do more research, pay attention to the approachability of my prose, etc.
• The +10/-2 up/down vote disparity. I don't have the rep to see downvote totals on any sites yet, but I can infer from seeing notifications like "+38 rep from [answer]" that even my good contributions get a fair few drive-by downvotes. The smaller rep impact of downvotes makes this a lot easier to not take personally, and doesn't change the fact that I'm still plenty incentivized to delete a net-negative voted contribution (which I have done before.)
• The association bonus. It's much easier to get comfortable with a new site when I'm able to ease into contributing there by leaving helpful comments. I do wish that comment votes generated notifications; I understand why they don't generate rep, but I'd like to get the feedback on whether I'm leaving helpful comments without having to remember to go back to the question to look later.

What about reputation is not working to make me a good contributor:

• Low rep gain from niche answers. I'm getting far enough in my life to have some expertise on some subjects, and I do occasionally go tag spelunking to find niche questions I can answer with that expertise. This is probably the most value I can offer SE as a contributor, and it's the thing I get the least rep gain from doing. It takes a lot of valuable time watching a tag to find an appropriate question, and if I write a good answer it might get an accept and 1 or 2 upvotes.
• The HNQ reputation effect. The vast majority of the rep I've earned on most sites has come from answering HNQ questions when I get lucky and catch one that doesn't already have the answer I'm thinking of. This incentivizes me to use SE casually, checking HNQ every so often like I'm scrolling social media. I just discovered, after writing this answer, that there is a "filtered questions" view on the homepage which partially addresses this problem from the question discoverability side; I'm certainly going to give that a try. (It's not exactly prominent, it appears above the self-promo banner on the page and is hidden when logged out. By the time I started logging in regularly, I'd already tuned this part of the page out as self-advertisement.)

Answer 14

Not working: minimum rep to add comments

This has been broken forever. The way people want to enter a complex system like Stack Overflow is by dipping a toe in the water: making a tiny contribution and seeing what happens. The obvious way to do that is to add a comment on a question or answer: "I have a tiny bit of information that may be of use".

But the current system prevents that. It basically says "The only way you can participate here is by taking big, bold steps and doing them right".

It sucks. And since everyone who participates in these discussions has long since passed that minimum threshold, we forget how bad it was.

Answer 15

TL;DR:

In general, the reputation system might be doing a good work as the large user base is using the system properly and most of those having a high reputation have fairly earned it. Unfortunately, there are a few that are taking advantage of some system cracks, like neglected tags or some sites with low participation or have found how to game the system.

The reputation system might be benefit from being analysed by using a systems approach, i.e., by using a framework / model like Viable System Model (VSM).

Suggestions:

Improve the contingency / correction measures to better palliate the perverse incentive effects, better help to avoid game the system, and avoid usage vices.

Add more visibility to helpful information for the user and community explicitly pointed to facilitate and require a standard (minimal) self-regulation / self-organization.

Better align the reputation system to the welcoming, helping goals, participation and community health goals.

---

Using the VSM, the reputation system might be seen as a concrete component that interacts with several conceptual systems of this recursive analysis model.

In VSM, the "system 1" refers to autonomous recursive components. In the case of the Stack Exchange (SE) Network, each user could be seen as a system 1 who is part of a community that also could be seen as a system 1. Each community is part of the SE network, and that could be seen as the system 1 in focus.

The core concept of the reputation system is that reputation points are earned by each SE user as the result of their interactions with other SE users through the SE software. Unfortunately, many users and apparently many communities too, don't look to have the requisite variety to achieve the autoregulation "standards". In other words, the environment and the tasks to be done are too complex in relation to capacity installed understood as the skills, knowledge, being able to participate, willing to participate.

On the other hand, there is also an alignment problem as not all users are willing to participate according to the goals and policies, but they still found ways to earn reputation. Sometimes this is referred to as "game the system".

When a user is suspended, their reputation is temporarily set to 1, but once the suspension finish the reputation is restored completely. The suspension records aren't public.

There are automatic and manual applied bans to restrict specific privileges. The ban records aren't public.

The access to higher privileges is merely based on reputation points, and it doesn't consider suspensions and bans.

There are some pages that show aggregated data about users, i.e., new users, editors, voters, and moderators, and there are leaderboards, but there isn't any clear guidance, and apparently there is a lack of management knowledge and skills (requisite variety), i.e., about how can users as individuals and as community might use the reputation system to take control and policy decisions as individual (self management) and as community (self organization).

Reputation system points

IMO, ideally the reputation system, besides tracking the current points, might do a better representation of trustiness than nowadays do, if it also tracks an indicator about give the community parameter about how the user behavior is aligned to the community values, let’s call it the "karma" indicator. The karma indicator might be calculated based on user participation in tasks that doesn't give reputation points, but contributes to the community like comments, votes, etc. will make earn karma, by the other hand getting spam flags, abuse flags, suspensions, that some way go against the community will make that the user loose karma.

The karma points might be used for instead of the reputation privileges, i.e., to participate in the per-site meta and in the a site chatroom for new users. Also at a certain level it might earn some "one-day pass" that can be given to new users to allow them participate in the per-site meta and in a site chatroom for helping new users to learn the ropes of the site.

Exceptions

Flag for Moderator Attention and CM escalation process

The automatic algedonic (pleasure and pain) channel is responsible of handling the notifications about the good and bad expections. The reputation system triggers inbox notifications or emails about achieving milestones that are related to earning a priveling or badge.

The replies to flags and suspensions are other notifitications of the overall system that should be more clearly related to the reputation system. At this time only the user and the moderators of the correspondig sites might see them in the user profile but they aren't not clearly linked to the reputation system. The closer might be user community moderator process that have a nominee score card showing some metrics and that might block the participation of users for having suspensions.

The reputation systems doesn't have warnings levels and there are important things that are not throwing notifications that help / guide users and communities to better self regulate / organize.

Incentives

After the site graduation process, the reputation system doesn't track / recognize commitment.

The reputation system incentives to interact with the users from other communities and with the environment are minimal. I think that the only incentive in the reputation system are badges earned for views through shared links.

The reputation system incentives have shown causing undesired effects. Probably it might be qualified as perverse incentive system as in a certain grade, it has caused problems like:

• very low quality posts being answered and upvoted instead of being downvoted, closed and deleted when they aren't improved,
• "fastest gun in the west" (be the first to answer to take advantage of the chronological sorting, being seen first),
• "tactical voting" (to make a post to have a better ranking to another for other reasons than the post content)
• "balance voting" (to compensate an "unfair" vote)
• "welcome voting" (to allow the user to post in the per-site meta and or use the chat)...

Vices

Sometimes good-intentioned users take advantage of the flexibility of the system to deviate its usage in order to help a user that "deserve it". The recurrence of this might relax the attention and eventually that might opened opportunities for gaming the system.

Meta System Relative to the whole SE Network

The reputation system aggregates reputation across all the sites only in the chat and in combined-flairs.

https://stackexchange.com/leagues shows the leaderboard for each site showing only the reputation points.

When your fellow users vote up your questions and answers on a Stack Exchange site, you generate reputation. Reputation is a rough measure of: how much the community trusts you your communication skills the quality and relevancy of your questions and answers These friendly reputation leagues are an informal way of tracking your reputation within the community on a particular Stack Exchange. Reputation is capped at 200 per day, but remember that bounty awards and accepted answers are immune to this daily reputation cap.

It might be helpful that also shows a leaderboard of participation (comments, votes, reviews, participation in per-site meta, participation in the site main chatroom)

Network Wide Policies

All users starts in Meta Stack Exchange with reputation 1 or with the association bonus. That might be fine for certain type of posts, but for posts discussing stuff that might affect all the network it might be convenient to have a better idea of the contribution done across the network without having to individually navigate to the network profile of each user.

Thinking loudly, posts with discussion might show the regular user card (reputation earned in Meta SE only). On posts with feature-request might be shown a user card similar to combined-flair showing the total network earned reputation and a selection of the sites where the user have being participating recently more frequently.

Related answers

Perverse incentives

• user3840170's answer from March 9

Karma

• Mithical's answer from March 11

Answer 16

There are already some suggestions to handle voting differently depending on the number of votes. We should consider also handling it differently for new users. A current oddity revolves around the timing of downvotes for users hitting the 1-rep minimum. Here's an example:

• User with (say) 1 rep asks a question.
• They receive 4 downvotes ▼ and one ▲.
• Depending on the timing of the votes cast on the question (+1/-4), they could end up with the following reputation:
  • 11 (▼ = 0, ▼ = 0, ▼ = 0, ▼ = 0, ▲ = +10)
  • 10 (▼ = 0, ▼ = 0, ▼ = 0, ▲ = +10, ▼ = -1)
  • 9 (▼ = 0, ▼ = 0, ▲ = +10, ▼ = -1, ▼ = -1)
  • 8 (▼ = 0, ▲ = +10, ▼ = -1, ▼ = -1, ▼ = -1)
  • 7 (▲ = +10, ▼ = -1, ▼ = -1, ▼ = -1, ▼ = -1)

In all of these voting scenarios, the consensus seems to point to a poor question (1 upvote + 4 downvotes). In all scenarios, they end up with higher rep than they started, but it varies because downvotes at 1 rep has no impact.

I agree that the user should receive some reputation, but that it's variable, depending on when the votes are cast, seems out-of-place. In the above example, the order of counting should be upvotes first, then downvotes, leaving the OP with 7 rep, regardless of the vote order. In the above case, the new user could circumvent the new user restrictions at 10 reputation, even though it may be doubtful that this should be allowed.

Answer 17

It is one of the parts needed to make community moderation work, and that is a really important part. It's also flawed enough that only using reputation for access to mod tools is problematic. It's useful enough as a very, very rough measure of how much someone has participated in a positive way.

It is a fundamentally flawed measure and it won't ever accurately measure knowledge nor moderator ability/qualification. But it's useful enough and I think it would be problematic to try and make it more accurate. I don't think you can get a really accurate measurement anyway, and additional restrictions would likely just make the whole thing more complex, harder to understand and more frustrating. It's useful enough as it is right now.

One huge problem is that on larger sites like SO new posts get so few votes on average that achieving the reputation necessary for participating in community moderation can seem impossible to achieve. A large number of questions and answers simply don't receive any votes at all. I'm not sure how to fix this, but this is a rather large barrier for new users compared to users that started much earlier.

My most controversial opinion on the voting system is that I don't think downvotes are useful enough on questions (please visit this meta post of mine if you want to downvote this idea). My impression is that new voters perceive downvotes and close votes as very hostile. Closing is necessary to keep the site working, but close votes at least carry some more constructive meaning and a chance to fix the question compared to downvotes which pretty much just tell new users that they suck. Downvotes on answers are necessary to ensure the good answers are sorted on top, they don't have the same purpose for questions.

Answer 18

Concur with NeMo in general outline, but I have a slightly different take. I spend 90% of my time on history, so my opinion may be skewed.

I think it might be more useful to group contributions (questions or answers) into a couple of categories where the reputation system works differently.

Bad contributions aren't really affected by reputation scoring - they are dealt with by votes for closure/deletion/moderation. So let's ignore bad contributions.

I think there is a weak correlation between good contributions and reputation. @NeMo identifies some edge cases where upvotes do not indicate optimal answer. I'm not sure what policy implications I take from the gap - I'm not sure I'd go to the effort of changing anything. (see below for additional thoughts)

I think the really interesting case is contributions that need refinement, most especially contributions by new users. I perceive the downvote spiral, and I'm not fond of it. Increasingly I'm finding new contributions by new users which have merit/worth/value, but which don't fully participate in the SE culture or the site culture. The reputation system does not serve this category well.

If I were to offer a suggestion, I'd wonder about broad categories of option (refinement welcome - these are outlines, not fully developed). For questions where "refinement" is the correct intervention, downvotes don't count (or are fractional value). Upvotes count normally. Users are encouraged to edit & contribute. Gamify this - upscale the badges & incentives that come from refinement of contributions that are on this margin.

How do we decide which contributions belong in the refinement category? Could be either:

• Users with <= 100 reputation stay in the refinement queue until they exceed 100 rep. That provides time to acclimate to the site culture.

• "refine" becomes a special label (like "protected") Normally users vote to apply the label. Mods can apply the tag. So long as the question has the refine label, downvotes are discounted, and positively voted edits garner reputation.

The goal is to incentivize refinement and community curation ,and to guide users towards meaningful contribution that aligns with the site culture.

(site culture can always be changed through appeal on Meta)

Answer 19

What is really great about our reputation & privilege system? What do you think makes it great, and why is it worth keeping?

There are lots of things that are great about our reputation and privilege system.

The bar for participation is really low on Stack Exchange, and that is a good thing. We are (by design) primarily a question-and-answer site, and every user, no matter his or her reputation, can post questions and answers.

When other users find your question and answer posts useful, upvotes will cause your reputation to climb quickly. The system encourages upvotes and subtly discourages downvotes, and upvotes impact your reputation score more than downvotes. This is a systemic way to avoid penalizing new users too much while they are learning the intricacies of the site and culture, also a great thing.

The most useful privileges are triggered at very low reputations, which is a great thing. Participating in meta, voting, commenting, flagging posts, using the chat rooms, and more are all available to you, in steady increments, with fewer than 10 upvotes on any of your posts. And 20 upvotes on any one site gets you expanded access to all Stack Exchange sites. This is a wonderful feature as well, eliminating a lot of noise from spammers and trolls, yet opening up the doors quickly to new users who are actually contributing on even a minimal level.

Community moderation privileges are granted with higher reputation scores, and this is a great design as well. Increased participation is rewarded, and it makes sense to trust the users who have proven their dedication to the site to be the ones who help moderate the site.

What is broken about it, and why? Are there any solutions to it?

In the Stack Overflow Blog post that is referenced in the Question, in the "What do you find most frustrating?" poll there are two items on the list that I want to discuss that might be considered related to this topic: "Unwelcoming community" (10.6%) and "Barrier to participation" (8.3%)

In my opinion, "Unwelcoming community" really has very little to do with the reputation system. Instead, new users are made to feel unwelcome by other actions that have little to do with the reputation and privilege system. Rude comments and questions that are downvoted and/or closed inappropriately or without adequate explanation in the comments are the biggest "unwelcoming" behaviors that I notice on the Stack Exchange sites I frequent. This would not be solved by revamping the reputation system. Instead, it needs to be solved by changing the culture of the site to be more inclusive and welcoming. Writing blog and meta posts to remind people to welcome new users and be more charitable in voting could be a solution, as well as moderators and high rep users modeling positive behavior.

"Barrier to participation" can be frustrating to some new users, but again, our bar for participation is already really low. We want to, first and foremost, encourage posting questions and answers, which all users can already do. And as I said earlier, after just a few upvotes, you have access to do all the important things on our sites.

---

In conclusion, I think the current reputation and privilege system works well as is, and does not need a massive overhaul.

Answer 20

What is broken about it, and why? Are there any solutions to it?

The Reputation point. The page literally says

Reputation is a rough measure of:

• how much the community trusts you

• your communication skills

• the quality and relevancy of your questions and answers

However, I have seen some long-time users who just asked a single, yet basic question, then disappeared forever, still gained thousands of points. In this case, it is a correct measure of "the quality and relevancy of your questions and answers", but I think it partially measures (or not at all) trust and communication skills. Can we safely say that that guy, who asked one famous question, has better communication skills and trust than a new user who continuously posts questions and answers, but can never achieve that points?

My proposed solution is to consider implementing a decay system. Kaggle ranking system and H-index have done this, and it is working pretty well: you must be active in order to maintain a "high rank". This thing is not perfect, but it's better than nothing.

(High rep users might downvote this because it is not favoring them. This system is not targeting you, this one targets users who post one question and disappear forever.)

EDIT: add a related paper

Answer 21

Solution: consider network reputation for comment privileges

I am a very low-rep user who has used SE for a long time, mostly as a reader. The "problem" part of this answer has more-or-less already been posted, but since I have never posted on Meta before I cannot comment.

"Not working: minimum rep to add comments"

https://meta.stackexchange.com/a/387552/1029183

I understand that needing a minimum reputation to comment prevents some unwanted behaviours.

However, I think that maybe you could consider network reputation when deciding to allow comments. Specifically, if a user earns comment privileges on one sub-site, allow them to comment on all sub-sites, or at least a group of related sub-sites (e.g. computing). I believe that transferable comment privileges would probably increase quality. This is speaking from personal experience based on the number of times I would have left a helpful comment but just ended up closing the tab instead.

I would have been able to post this as a comment on the question I linked if you had considered my network reputation. But then, I wouldn't have needed to post this as a comment.

Edit:

Per the comments, this solution is already implemented as the association bonus. It sounds fine to me.

If when I had tried to comment on the other answer the denial message had told me about the association bonus, I would not have written this answer.

Revised solution: advertise the association bonus when blocking actions where having the bonus would make a difference

For example, instead of "You must have 5 reputation to comment", Meta could tell me "You must have 5 reputation on Meta to comment, or 200 reputation on any other Stack Exchange site".

Answer 22

Maybe I'm late, but I just read some answers and I didn't see an important point: Reward for reviews.

On Stack Overflow, I made over 13k reviews and the only things I received was badges. It's cool to have more gold badges than most of people with 3k rep, but if at a time I stopped, it would also be because of the lack of reward.

Making review takes time

Making reviews isn't a job, but it takes a lot of time as a job. Even while being quick, it can take 45 minutes - 1 hour to complete 40 reviews in the same queue. That's long, and after coming back for 25 days and making 40 reviews for each days, you earn one badge. 50 days for normal queue.

People wants, or not

I will take me as an example, but maybe I'm not representative of people

When just reach reputation to be able to review, we made some of them as we feel useful for the website. We will not talk about the issue. But after a few times, when realizing you just pass 1 hour to do them, or just when you receive the silver badge, you are more able to stop review.

Adding a reward will encourage people to review. A few reputation points will give a good reason for some people to do more reviews than now.

Review are necessary

If no one makes reviews, SE will stop working: question will not so much be closed, beginners not helped, etc. It's required to be sure this part of the system works. And always keep more than 1k post waiting for review on Close Vote is -for me- not a good thing. By giving an advantage, it will help people participate and so, make the system safer.

Robot and farming

As told in comment, as I quickly think about this, it's possible some people try to farm reputation. To prevent this, there are multiple ways:

• Cap reputation to not so high value. Robot will reach the cap every time and not so much get rep.
• Fix the review audits by better selecting them, make them less obvious, etc.
• Check for others things: time of vote, which one is chosen, the "agreement percent" (if the person always choose something different than the final solution)
• Let moderators more check for recurrent people. For example, people like me will be more able to be checked for bot than people that just do 5 reviews
• Be rude when seeing a bot to discourage this

Another way is to ask for moderators or very high reputation people (100k & more) to check like one time each month if people could get reputation, a sort of checking for reality of people. It will add a new step, but could fix this issue too.

Also, personally, while the bot is doing good review, it's not an issue. I'm sure review could be partially automated. I already proposed (10k+ link) to add automatic management for some post. So, even if I understand the issue, if it's good review, it's not an issue.

The actual reward

It's only one bronze, one silver and then a gold badge per 1000 reputations. I also reached TOP 1 of reviewers all time on First question (yes, let me flex :p) and I didn't received anything, just the satisfaction to be first on something that help the community.

Reward advantages

The objective is not to give like 5rep/review, but one of those:

1. Like 1 rep per review given each days, capped to 25. If not capped, it could be "too easy" like just spam in all review queue and in one day you reach next privilege.
2. Review given for each gold badge obtain. It's easier to make 0.5rep/review for example, and add "step"
3. Each days, when a review queue is complete (40/40, 20/20...), you obtain some. It will help people to do all review for one queue, and not necessarily to wait until 1000 review (which is big)

Failing review audit could cancel reputation gain for today or something.

As I told, maybe it's possible to give special reward at 10k review, 100k review, for TOP 1 of month or something. Always give more reason to people to participate, and reward them for their efforts.

Notes

• It will also help people that doesn't answer/doesn't receive upvote but participate.
• Some mod that comes from another website could get reputation everywhere (and will not be mod with 200 rep :p).
• I totally agree with most of others posts, so I didn't repeat what they explained as not capped post, rep on closed post etc.
• I'm happy to make all of my reviews, but the lack of reward (and sometimes lack of adding feature) make it sad.

Answer 23

Remove reputation gain from accept votes.

Question askers are the least suitable people to determine the best answer. Yet, the accepted answer feels to a lot of users like it's the be-all and end-all. Answerers are begging and harassing users for accept votes. New users flag posts for moderator's attention asking to change the "correct answer" mark. Users viewing solutions consider the one with a green tick mark as the one and only true solution.

The first step to fixing this mess is to remove the reputation gained from accept votes. And I know I would lose a lot of gained reputation, but that's an acceptable sacrifice. Reputation should be gained from upvotes, not from the whimsical decision of the question asker. This alone would fix a lot of quality issues on our sites.

The end goal could be to completely remove accept votes, or just leave it as a placebo to the question asker and show it only to logged-in users.

Answer 24

My main interest in SO are expert level niche answers to rare edge cases.

What's going to happen if you remove the downvote penalty (following Dharman's proposal which I've discussed with him before) is that the irresponsible downvoters will immediately (in the short term) outnumber the experts who do know how to evaluate intricate solutions to rare problems (in the long term).

I think Dharman's perspective can make sense looking at the massive flow of low quality questions on the PHP tag, but I have to remind of Shog9's argument: "not destroying value". Dharman once asked me: "if you think that question is so good why don't you upvote it?" I'll repeat my answer here: "because understanding the question and running some tests takes a minimum of 2 hours, that's how hard it is to make a judgement on the upvote". (While downvoting something you don't have the slightest clue about at zero cost would take no effort at all.)

Now, if you couple Dharman's proposal of removing the downvote cost with Nij's proposal of increasing the downvote penalty for the poster, you've just empowered reckless irresponsible downvoting at the cost of leaving the niche Q&A posters (seeking to solve hard problems) outnumbered and outgunned. You won't be filtering for pearls any longer, nor promoting the production of those rarer pearls.

(It's happened too many times that I've seen downvotes and critiques in comments from experienced engineers who completely failed to understand the question or the answer they were recklessly voting on.)

Answer 25

What is great about the reputation system.

Over time, really common questions and trusted users start to stand out. This becomes valuable for new users or people who are just learning about a topic and want to know similar information.

What is broken about it, and why?

New users have a difficult time contributing new questions or information to the site. Users that posted a common question years ago and no longer contribute to the site are still gaining lots of reputation. There should be a balance that incentivizes people to add missing information to already existing pages other than answering a question from 10 years ago. If someone provides a better answer to a common question asked many years ago, does that answer really get recognized?

One way that new users can contribute is by answering their own questions, but I am not sure how this type of contribution would be received by current moderators. It may be more intuitive to format the forum as a guide that people can add new technical information to over time and the Q&A portion becomes more specific to different parts of the guide.

What other systems work really well?

I have been a casual user of Stack Overflow for only about a year and my experience using the site is not going to be deep enough to comment on the intricacies what the reputation system was originally designed for. However, I still want to post this answer in case others think it meaningful.

One thing that I have wanted to mention since I began using Stack is that I think questions and answers could be categorized in a way that acts more like tutorial pages or guides rather than an overwhelming thread of Q&A. This could play into the reputation system by providing users an incentive to build and modify guides or technical information and then additional incentive by answering questions posted to those guides. To me, this seems like a more intuitive interface to show users how a task could be conducted or how technical knowledge should be categorized and written (perhaps a more technical and gamified version of a wiki with a Q&A feature).

When I first joined Stack Overflow, it was overwhelming to try to avoid duplicate questions by searching for pre-existing questions that were similar to mine because there were so many questions and not enough tags to filter to what I really wanted to know. It was also frustrating to see answers be accepted that were kind of close but didn’t fit my specific use case. I was worried that if I posted a new question that wasn’t different enough from existing questions, someone would downvote or close the thread for being too similar to a question asked 5 or more years ago that I didn’t see. Which would be fair to downvote if it decreased quality of content.

I use Stack Overflow mostly to help with programming in R. I also mostly use the ggplot package of R in my typical workflow. I can use the tags to filter [r][ggplot][geom-point] to search for questions that have already been asked on the topic of the “geom_point” function, but what if Stack Overflow had an incentive for people to contribute to a guide on this topic rather than a list of questions? Instead of seeing a list of random questions about using the geom_point function in R (usually R studio), I would prefer to see a guide on this topic that has been contributed to by many users over the years. If the existing guide or guides do not answer my question or provide knowledge I am looking for. I can ask a question and post a bounty for someone to add this missing information to the guide. If I ask a question that has clearly already been answered through the guide or other questions on that guide, I could be downvoted and a moderator could remove that redundant question. On the other hand, if I am proactive, I could add to the guide with my own information and get upvoted for adding new and useful information. In a way, this would make Stack websites become community supported guides for technical information in addition to Q&A forums. Think something similar to how Steam community pages allow users to post guides on specific topics for videogames or how Genius.com allows users to comment on song lyrics to provide additional information and sources for what may have inspired those lyrics.

During my journey learning R programming for the past year, I have wanted to post what I have been learning to a website like Stack Overflow so that people going through a similar journey won’t have to spend as much time figuring out the same things I am. For people who want internet points for doing this, I think the reputation system on stack would be a decent incentive for them to start or modify a guide. There is already a prompt at the bottom of questions that tells users they can answer their own questions. I think the reputation system would work better if answering your own questions or creating guides became a main feature of Stack websites.

This would be very valuable to programmers because every new software and software feature could receive a new or updated guide. It would also be valuable to professional experts that want to publicly share their knowledge of old and new topics in their field (think of science, math, photography, design, fitness, and all other Stack websites attracting expert knowledge in this way).

I apologize for a redundant answer if this feature is already available. If it is available, I have not noticed it.

Answer 26

Reputation for tags (1500) vs. edits (2000)

As someone that's been around for a long time (since the private beta), I'm at the stage where reputation is just a target number for earning the privs to what I want to do.

I like answering questions (and asking them), but over time, helping with triage as been great. I had to do some work to make that happen.

Now I'm at about 1250, so access to tags is next. Frankly, I'm not interested that.

Helping with edits is far away (+800), and I see many questions and answers where I'd like to help with cleanup.

If my rep ever gets to 2001, I won't want to increase for the sake of the number, I'll just want to keep questioning, answering, triaging, and editing.

Answer 27

I wonder how much you've looked into the question of culture and internationalisation? Because from my perspective, the whole points system feels very American. Here in the UK, small children were awarded gold stars and other such badges for effort and achievement, but we were expected to grow out of that, and as we became older, to value things such as knowledge sharing for their own sake. I still feel that when SO gives me points and badges, they are treating me like a child.

Answer 28

What is really great about our reputation & privilege system? What do you think makes it great, and why is it worth keeping?

One thing is that users get the same amount of rep for each upvote, regardless of whether it is a question or an answer (+10). This encourages people to post more questions to contribute to both the Q&A database more quickly because the answers will not appear if there are no questions.

What is broken about it, and why? Are there any solutions to it?

Same as above. Making the same worth of each vote clears the difference between a good answerer and a good questioner. Before that, only looking at the user rep amount was the indicator of power of knowledge for this user on the site. But not anymore. Just look at me, for instance: I mostly ask questions, but rarely give answers. And this leads us to another problem...

... many answers that I posted here were actually not answers, but the posts posted on the answer sections since the OP (staff) asked to post in the answer section just for the case of not creating many other related questions on the newly announced feature on site. Here are examples:

• Some changes to the profile while we make it responsive
• Network-wide HTTPS: It's time
• New post summary designs on site home pages and greatest hits now; everywhere else eventually
• Tag synonym dashboard 2.0

and so on.

Despite the fact that the number of reputation gained per vote is the same for a question and for an answer, only answer votes contribute towards gold tag badges (that helps close/reopen questions without other user votes). I think that posting good questions should give users a similar ability.

The rest of the issues are related to Meta sites:

• On MSE some questions (mostly tagged bug) can be answered only by developers (staff), so the only restricted kind of users may actually answer. Mostly, but not always, since some bugs can be resolved in case of a workaround, or the answer may contain the possible solution that can be nevertheless only implemented by developers.

• There is no ability to make a bounty on per-site metas, so there's no regular way to raise additional attention to a question rather than bumping it with meaningless edits. Here's a feature request for this issue, but on ruSO.meta: Конкурсный вопрос на Мете

What other systems work really well that we could learn from?

IDK, I mostly use only SE now. Unfortunately :)

Answer 29

One thing that I do not find that is working right is that there is any rep gained from making trivial, meaningless edits to posts. Like changing SQLServer to SqlServer or its to it's. It is a cheap way for some to get an extra 1K rep, because aside from me, how many others are really going to reject that edit?

If someone wants to spend their free time fixing contractions that are missing apostrophes or casing for products to match a brand name (which are things that a script or bot could fix), by all means let them waste their time, but dont reward them for the kind of activity that is supposed to be discouraged

Answer 30

What's broken 1: not everything is Stack Overflow / Mathematics. Privileges and voting rights are much harder to gain on some networks. I know many on DSP.SE who are qualified to close vote, but can't.

What's broken 2: 1-rep can't vote. Some users, especially experts, have nothing to ask or don't wish to participate in Q&A's - yet their votes remain valuable. "Non-participants shouldn't get to vote" - it takes 3 upvotes on a question to change this; good as nothing.

Two in one post because they're related, and emphasis on "not all is SO".

Solution 1: allow sites to decide their own privilege thresholds.

Solution 2: a verification system that addresses the abuse concerns of letting any new account vote. Phone + e-mail, mod review, are examples.