For example: Why do people say there is modulo bias when using a random number generator?

That was quite a new question, yet, got a lot of views and votes very fast, then it just died all of sudden and nobody watched it anymore. I barely finished writing my comment when the traffic already died and is now completely dead, and nobody is going to read my comment anymore (which I would like to get an answer too).

So how did that happen? It just seems weird to me, the question itself wasn't that special, and the solution was simply disgusting and the 76 votes worth answer itself wasn't very clearly written. I've seen better questions/answers with less votes given, and worser questions with more traffic/discussion, here it just all died suddenly. So what happened? Just got lucky? Or cheating? Or I simply didn't catch the awesome point of that so awesome answer... Maybe someone can explain what is so good in that answer, to a dummy like me? It just isn't worth 76 upvotes IMO. I just don't get it.

Every now and then I see these high-voted questions, and I always wonder, how do they get votes so much. Is that question shared via social media or something?

I suspect that once the votes reach some limit, people will start upvoting without thinking more, and upvoting generates more upvoting and traffic, and so on, and once all the non-thinkers are spent, the traffic just dies straight to a wall. Are there any evidence of this happening? Just a theory, don't take it personally. Still, it's a mystery how a bad question can get its first N votes to get the chain reaction started.

  • 1
    On the other hand, there are other questions which start very slow (at one point, this had a -3 score), which get huge after an answer is posted (was at the hottest for the week for 5 straight days, I'm proud :) ) stackoverflow.com/questions/10289890/… – Richard J. Ross III Jun 13 '12 at 15:04
  • The user who posted the answer you commented on gets notified of your comment, so if they choose to respond, they will. However, asking questions in comments is for getting clarification on a question/answer, so if your question is more in-depth than that, or not 100% directly related to the answer, post it as its own question (keep in mind that if it is very closely related, it could potentially be closed as a duplicate, depending on who reads it and how they interpret it) – yoozer8 Jun 13 '12 at 15:13
  • 1
    "which I would like to get an answer too" If you wanted an answer to a comment, you should have posted a question. That's what we're here for: Q&A. – Nicol Bolas Jun 13 '12 at 15:13
  • @NicolBolas, im afraid the question is so stupid to ask about it; because it got damn 76 votes up, how could that ever be bad answer, right? Well, maybe i go ask it then. I expect it to be closed as offtopic, or just simply by "not real question" votes, though. I'll blame you then if that happens ;) – Rookie Jun 13 '12 at 15:41
  • @Rookie You might be interested in this: meta.stackexchange.com/questions/138176/… – Mysticial Jul 21 '12 at 0:46
  1. The question is not dead. It arrived just today at the very top of a Google search for help understanding the "modulo bias" issue.

  2. Questions can be upvoted for reasons other than being "special." I upvoted this one because it asked the same question I myself had, and asked it clearly enough to elicit just the answer I needed. (Well, I see now the questioner and the answerer were the same person, but so what.)

  3. The highly upvoted answer did include a solution, and I agree it was a very bad solution, but the solution wasn't really the point; the explanation of modulo bias was. And it was a great explanation, easier to understand than the one on Wikipedia.

  4. Armed with that explanation and the additional solution posted as an answer by Nick Dandoulakis, I not only came up with an efficient bias remover, but I actually understood what I was doing. It might be small beans to you, but it was an exciting day for me.

  5. Your comment is likely to appeal more to advanced coders, not people who need help understanding what gives rise to modulo bias. So it doesn't need to be bonded to this particular question. If you are keen to have a discussion about "xor128 RNG," I agree with Dave Newton that you should post it as your own question.


If you have a comment you'd like an answer to, ask a question!

High-traffic tags will get high view counts, leading to more potential votes.

High-traffic tags also get... high traffic: questions churn faster, pushing even relatively new questions off the radar quickly. Only questions with consistent, sustained activity will be prominent on the first page.

Unless someone is specifically watching, un-directed comments like yours may well go un-noticed.

It's possible there's some level of voting fraud. Self-promotion of a question or answer isn't a Bad Thing, so if that's the case, doesn't really seem like an issue, and it's not fraud.

Regarding that specific answer, if people think it answers the question in a reasonable way, they'll upvote. It does seem to answer the question, more completely than the other answer from yesterday. Only one newer answer remains, and traffic will be reduced now that it's an older question.

  • Tags itself cant be the reason. I tag with popular tags frequently (including c++) but the best i've ever had was maybe 15 upvotes, and I think that wasnt even c++ question :P Usually i get around 2-4 upvotes, and not so many views. – Rookie Jun 13 '12 at 15:36
  • @Rookie Anecdotal evidence and not really relevant. Of course tags aren't the only reason--question content matters. Things a reader never thought to ask (specifically mentioned in one of the answer's comments at least once) are much more likely to get an upvote as well, meaning "Ooo, huh, how 'bout that." – Dave Newton Jun 13 '12 at 15:39

You must log in to answer this question.

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