What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

Now aren't these some trivial questions:

And all the usual suspects (including myself) are here to write an answer, because they can easily get lots of upvotes and lots of score reaching the higher levels of "k" (40k, 50k, 200k) in reputation. But that's not what Stack Exchange aims to be, from what I understand: A big collection of trivial questions.

Because there are still many very good questions left unanswered, e.g. this one:

Transaction activity running total in SQL

...simply because they're not trivial. Which means:

  1. You need more effort to write an answer
  2. There are less people proficient enough to actually answer the question
  3. Activity seems to have a positive feedback on itself on Stack Exchange. The more active a question, the "even more active" it becomes, because it ranks high in the "Top Questions" list. So the "good question" is even less interesting because of the "trivial ones" always being on top.
  4. So ultimately: You get less upvotes, lower ROI

In capitalist economics, it would be silly to answer the good questions. So maybe (as it's done in politics), we could shift interests a bit towards the "good questions" or rather, away from the "trivial ones"? First there would be need to distinguish "good questions" from "trivial ones". Maybe using a separate upvote/downvote system available only after 5000 or 10000 reputation. Then maybe, the trivial questions could be adapted such that upvotes will only count 1/2.

Maybe also closing a question as duplicate is some sign that it might've been "trivial". But these are just some ideas.

In the end, this goes into the same direction as the Chess rating system. If a grandmaster beats a newbie, then he only gets little score (and the newbie loses little), because it was a trivial game. But if a newbie beats a grandmaster, the change of score is massive because that means the newbie might've had an excellent game. On Stack Exchange, this would translate to:

who's asking -> || newbie      | grandmaster    |
who's answering ||             |                |
      v         ||             |                |
================++=============+================+
newbie          || as today    | answer gets    |
                ||             | lots of points |
----------------++-------------+----------------+
grandmaster     || answer gets | as today       |
                || few points  |                |

I've also posted this table to this similar question: The bikeshed problem and Stack Exchange

Of course, if grandmaster/newbie were the criteria for "good" and "trivial" questions, then answering the "good question" (Transaction activity running total in SQL) still wouldn't give a "grandmaster" many points because it was asked by a "newbie".

Another idea:

There could be some sort of distribution of points across answerers. If 10 people can provide a good answer to a question, then the question was rather "trivial" and those 10 people will get less points for every upvote they get. Or the other way round, being "the only one" who knows is rewarded.

share|improve this question
1  
This issue has been discussed before. e.g. The Bike Shed problem. –  Martin Smith Aug 30 '11 at 13:47
    
Hmm, true. That was a long time ago and it seems still that it is an active problem... I'll post my chess rating system idea there, and you guys may decide to close this as duplicate, if you think it should be closed... –  Lukas Eder Aug 30 '11 at 13:53
    
Related: meta.stackexchange.com/questions/3811/… –  Margaret Aug 30 '11 at 14:10
    
@Margaret: Yes, that's related, especially to the chess example. But it's not the same because it's not about giving an incentive to answer "good questions" rather than "trivial" ones. It's more about a general categorisation –  Lukas Eder Aug 30 '11 at 14:15
9  
An analogous suggestion was made some time ago and it was pointed out that ranking grandmasters and newbies is hard because a Perl guru might know nothing about Java. –  Pops Aug 30 '11 at 14:34
    
@Popular Demand, good point. Clearly this chess rating system won't be up for the job in SO, also because of the problem I mentioned at the end. newbie/trivial questions don't always correlate, because a newbie is only a newbie on SO, not necessarily in the field –  Lukas Eder Aug 30 '11 at 14:37
2  
@Martin, I think the claim here is a little more than just complaints about points from Bike Shed question, but rather lots of points for repeatedly answering the same Bike Shed questions. –  dmckee Aug 30 '11 at 16:23
    
@dmckee: Absolutely. I don't blame anyone who's out for the score. I mean, we're all trying to collect our stamps, and panini football pictures and stuff like that, and it's fun getting upvoted. But maybe upvotes should be more directed towards higher quality and less trivial redundancy... –  Lukas Eder Aug 30 '11 at 16:33
1  
see also meta.stackexchange.com/questions/37466/… –  Ian Ringrose Aug 30 '11 at 16:37
    
@Ian Ringrose: Yes, finding duplicates should be rewarded in the same way as answering duplicates-to-be should be "penalised" –  Lukas Eder Aug 30 '11 at 21:40
1  
About the more active a question, the "even more active" it becomes: I think that the activity often indeed is high, but also for a relatively short period. –  NGLN Sep 6 '11 at 22:27
2  
This is basically what Microsoft has done with their TrueSkill ranking system, which was used by Halo (more details here). The concepts are similar to the Elo rating system as you pointed out. –  NickC Sep 22 '11 at 0:16
    
@Renesis: "Elo", I was looking for that name. Nice links! –  Lukas Eder Sep 22 '11 at 17:54

4 Answers 4

A part of the problem is, you get upvotes for good answers, But I only know if an answer is a good one when I understand it and know it is correct.
This is a lot easier for trivial questions than for good ones. So I think trivial answers gets sometimes a lot quick upvotes, because a lot of people see very quickly that this is the correct answer.

What about trying to identify the "good" and the "trival" questions on the amount of upvotes to the question?

E.g.

   Question  |  Answer points
    Votes    |   per upvote
  ----------------------------
     < -1    |        5
     -1,0,1  |       10
     >  1    |       15
share|improve this answer
    
While I think your scale is very extreme, I had thought about this as well... It might be a more interesting correlation than the involved users' reputation as I originally suggested. That would have to be analysed, I guess... –  Lukas Eder Sep 6 '11 at 13:58
    
@Lukas its just an example could also be 8,10,12 ... –  stema Sep 6 '11 at 14:02
    
What would you think about removing some point per upvote if there are many equally good answers? If 10 people can answer a question well, then every one of those answers become less valuable by themselves. Or in other words, there's a reward for being "the only one" who knows –  Lukas Eder Sep 6 '11 at 21:52
    
@Lukas if there is really a large amount of upvoted answers (not the pure amount of answers), then this could be part of an answer. This could maybe solve the problem for the immense amount of upvotes some answers on popular poll questions get. –  stema Sep 7 '11 at 5:49
    
Sometimes that's correct. But in a high quality Q&A, the question should be so concise that there can be only one or two good answers. The use case where you have 10 good (and thus useful) answers is rather rare, I think. See also NGLN's answer –  Lukas Eder Sep 7 '11 at 6:13
    
@Lukas to be clear, I agree with you that 10 answers is already a high amount. I just wanted to avoid that for e.g. 3 good answers (maybe they worked together, added important anspects, ...) and another 2 answers with no upvotes, because they are "wrong", are together enough answers to make the question less valuable. But I still think, this can only be part of a solution, i.e. it does not solve the whole problem. There will be still be the trivial questions with only one answer, because later people see at a glance that its already perfectly answered. –  stema Sep 7 '11 at 6:33
1  
You are right. If one thing can be said, then it's the fact that establishing rules to make Q's and A's more or less valuable is very tough... –  Lukas Eder Sep 7 '11 at 6:35
    
Or even (answer points per upvote) = 10 + (upvotes on question). –  David Wallace Nov 11 '13 at 3:05

We should be very careful when we are speaking about trivial, "indeed", "100% stupid", "obviously inconsistent" and so on, and so on questions.

Major part of information which is find of stackoverflow, actually, can be found using google, not matter is it about left shift operator in java or sorting algorithms regarding to cache prediction. Stackoverflow is like a supermarket for food. Yep, you can get it another way, but here we have something like extending services (if we are talking in capitalistic terms ))). It is pretty nice to have an ability to find apples and wine in one place, even if you are not drinking at all.

Besides, there are lot of questions, and quite useful I should say, that actually can be answered merely by reading the source code. Should we consider them trivial or not?

Besides, I believe each developer could confess, something, which seems to you to be a hard topic for now, just few months later can be trivial.

Those are very subtle issues. I don't believe we can found here something better than human votes. Event if we try to produce some empirical rules which supposed to help people make correct decisions and more likely penalize trivial questions.

share|improve this answer
    
My point here is not about "trivial" questions, which are perfectly ok. But one should not be able to get +100 upvote points on something that almost anyone can answer, whereas a tough answer on a tough topic only gets you +25 points –  Lukas Eder Sep 6 '11 at 21:29
2  
@Lukas, it's looks like it is nevertheless about trivial and non-trivial questions :) because, once again, it is very difficult to separate one from other. Say, If we'd talked not about difficult/trivial question but about megapopular/"mundane" questions, I would fully agree with you. Megapopular questions are like black wholes with it's gravitation wave. –  shabunc Sep 6 '11 at 21:32
1  
Subjectively, it's not difficult at all. "Trivial" questions have a very high probability of being closed as duplicates, or being answered and accepted in a very short time. If someone can answer a question very easily (not because of their skills but because they're just fast enough), then there should be slightly less reward. That will remove some of the "gravitation" from these questions, over to other ones who deserve more attention... –  Lukas Eder Sep 6 '11 at 21:50

Although the suggestion looks valuable at first sight, I actually think you are proposing to beat psychology.

There already is a good system for questions nominating trivial or good and that is the question vote count. And there already is a good system too for limiting the vote count on answers and that is the ability not to upvote.

So why are answers given by high rep users sometimes heavily upvoted? Maybe because:

  • they were (one of) the first to answer and get the most readers,
  • high rep users attract upvotes more, because the answers from low rep users are more carefully upvoted (the answer might be wrong or incomplete or is not well formed),
  • badge hunters knowing nothing about the subject suspect the high rep user knows what he/she talks about and consider the answer valid anyway,
  • all of the given answers to this question.

In all these cases, the upvoter simply often does not get what an upvote should mean: being useful. Though maybe "being useful" is worth a discussion on its own (, because the answer should be useful to whom?), an average nice answer on a trivial question with possibly a high number of duplicates simply is not that useful.

Maybe this should or could be more pointed out in the faq? (That an answer only deserves an upvote when it is useful). There is a help topic on how to ask questions, but there isn't one for how to vote.

share|improve this answer
    
Nicely put. In a high quality Q&A, a good and concise question will have 1 or 2 good answers. If a question gets 10 answers, then it was either too broad (-> not all answers get upvotes, that's ok too), or too trivial (-> all answers get upvotes, then they're not so useful as you put it...) –  Lukas Eder Sep 7 '11 at 6:15

This feels a bit like you are proposing a "technology" fix to what is essentially a "people" problem (To wit: folks who should know better keep answering duplicates).

Worse it would strip what little transparency remains in the SE voting system (I mean...acceptance is +15 (+2), up-votes are +5 on questions, +10 on answers while down-votes are -2 (-1) on answers and -2 (0) on questions unless the post is CW then votes have no effect on rep, also unless the rep cap has been reached (but the rep cap neglects acceptance and bounties), and there are some points for approved edits.... ARGHHHH!!!)

Other solutions that I recall being suggested include

Obviously all of these suggests also have problems.


On a side note you seem to think that sky high reps are significantly founded on these kinds of simple answer, and I am not sure that is the case. Certainly they are helped by them, but that is not the same thing.

share|improve this answer
    
I agree with you. None of these ideas will permanently solve a problem that is not really such a big problem. But I think the voting system should be used like taxes, to encourage one behaviour, and discourage another. It should be more interesting to answer "good questions" than "trivial ones". If it wasn't for a sophisticated voting system, then SO would never be as good as it is. Reputation is a currency that buys you good answers when you finally need them (e.g. with bounties). And if you give a bad answer, you have to pay because you're downvoted. –  Lukas Eder Aug 30 '11 at 16:28

You must log in to answer this question.

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