10

Sorry if this has been asked before, and I'm sure it must have been by now, but I couldn't find a previous question on this.

Alright, so I've finally got myself into "answer mode" - I've spent a lot of time on SO for research, asking questions, but I always assumed that I wasn't good enough to answer hardly any of the questions asked here. I finally started answering some CSS questions, and have now moved on to also answering some of the easier PHP and MySQL questions, and I was ecstatic when I finally hit the rep cap for the first time - a goal I never thought possible for me.

Something happened after that, however. Once the cap was met, I suddenly lost the urge to watch for new questions to answer. Don't get me wrong, I like helping folks with their issues, but once the cap was met I just lost interest.

I know there are issues with gaming, etc. that need to be dealt with, but I think it would be beneficial if the rules for the rep cap were changed. I think we need to move to an answer-based cap rather than just a hard global limit. This way, we're still limiting rep-inflation, but with the benefit of motivating answers to keep answering which is definitely something we want to encourage.

I propose the following:

  • Remove the global cap
  • Award reputation on answers, but limit it to five upvotes per answer.
  • Bounties and Accepts are exempt from all limits, as usual
  • Set limits on question-based reputation:
    • Questions can still gain you reputation
    • Limited to five upvotes per question (like answers are)
    • If your total rep for the day exceeds some number (keep it at 200?), question rep cannot be awarded, and new answer reputation will eat existing question rep

That last point might not make any sense, so let me give an example:

Let's say I've answered 4 questions. On three of them, I've hit my 50rep/5upvote limit (total 150 rep), and on one more I've got 4 upvotes (40 rep). My total for the day is now 190 rep.

Let's also say that I've asked a question that was upvoted twice, gaining me 10 reputation. I've now earned 200 reputation for the day. If I were to have an answer upvoted, the 10 reputation would nullify the two question upvotes, leaving me still at 200 rep. If I was then to answer a question and gain an upvote, it would still count as normal, giving me 210 rep on the day.

I think this is what I'm trying to say:

 Qc = N+MIN(MAX((200-N),0),Qe)

Where:

  • Qc = Question Reputation Counted (Rep applied to user total)
  • Qe = Question Reputation Earned (Raw upvotes)
  • N = Non-Question-Related Reputation

I think this system: - Keeps people answering questions, since answering more questions = gaining more rep - Limits the affect of a single good answer, since it's not really fair that someone can answer one trivial but often viewed question and get the same rep as someone who answers many seldom-viewed, hard questions. - Limits the affect of asking lots of questions, demotivating people who try to gain lots of rep by asking lots of inane questions. - Doesn't affect rewards, so bounties still get eyes on the really hard questions, and accepted answers get people to write really well thought out answers.

Does anyone see any major problems with my plan? I suppose Mortarboard/Epic/Legendary badges could be tricky, balancing legacy with future calculations. I don't know.

Thoughts?

3
  • you have tought it out, but wouldn't this make it overly complicated? and would this 5votes per answer be daily or permanent?
    – Andy
    Commented Dec 14, 2010 at 11:05
  • 1
    @Andy: See Jon's answer below, in retrospect I agree that the weird question limitations would be, well, weird. The caps would be permanent. The reason is that as the site grows and grows, long term users will have such a large back-catalog of answers that they'd otherwise gain huge rep everyday without actually having to do anything, because they'd gain so many upvotes from old questions. I think I remember Jon once saying that he (sometimes|often) hits the cap before he answers a single question simply because his old answers are so often voted. You'd have to ask him about that, though.
    – anon
    Commented Dec 14, 2010 at 21:14
  • I kind of agree with this systems... I've felt on an user recently who has 10k+ rep who do have a lot of answers but mainly 12k rep from 3 times the exact same answer on three duplicate questions. Commented Mar 5, 2013 at 3:00

1 Answer 1

8

For simplicity I think I'd ditch the "question rep" distinction. Just make it capped on a per question/answer basis. I'd suggest that the cap for both should be in terms of rep rather than votes - so you could get X rep for either a question or an answer, rather than X for an answer and X/2 for a question.

I'd also suggest a somewhat higher cap - 100 rather than 50, perhaps. But that's just a gut feeling.

Overall, I'd certainly prefer a per-answer cap than the current time one. But then I would, as the primary "victim" of the current system :)

One slight clarification I'd suggest: make this a cap on rep for all posts by a particular user on a particular question. So if I post 10 answers to a single question, they can't each hit the rep cap.

1
  • 1
    +1 For the questions, I was trying to address the "1000 question noob" issue, but I suppose setting any limit on question rep helps that, without the convoluted system I proposed. Excellent point on people who answer a question more than once. I hadn't considered that, I think because the only time I've ever seen that was on CW questions. On normal questions, I'd expect a single answer would usually suffice. I suppose you would know better than I would on that though.
    – anon
    Commented Dec 14, 2010 at 21:10

You must log in to answer this question.