None of my posts are even remotely spammy. No URLs, or anything close to a URL--just a normal response. Yet I keep hitting the stupid CAPTCHA. Makes me want to go away.

Why isn't the code smarter? Why can't it at least look for URLs, repeated words, use a Bayesian filter like SpamAssassin does... It should be obvious that my posts are not spam.

Argh, even THIS post made me do one! CAPTCHAs are ruining the Internet.

  • Are you using a proxy ? blocking cookies ? This site hates my proxy -> lots of captchas, using the chat is all but impossible. – Panther Mar 7 '12 at 1:26
  • See I just want to know why a moderator (aka me) got a captcha. ;) – jrg Mar 7 '12 at 1:30
  • 1
    You might be posting too fast, see: meta.stackexchange.com/questions/1343/… – Jorge Castro Mar 7 '12 at 3:04
  • Ugh, that's...suboptimal. If an account is known to be good, to not be a spammer, then why pester the user? Unless he posts something that is clearly questionable, it shouldn't bring up a CAPTCHA. It shouldn't be so hard to tell when a post is obviously not spam. And no, no proxies and not blocking cookies for SE. – user41313 Mar 7 '12 at 5:56
  • 5
    Are you saying that a logged in user cannot be a bot? What if your account is highjacked and turned in to a bot (or what you want to replace it with), should it be allowed to do anything because it was trusted till now? Even people that visit the site everyday and are trustful users get captchas once on in a blue moon, dont be surprised, just prove your a human and move on, nothing to see here. ;) – Bruno Pereira Mar 7 '12 at 6:41
  • 1
    Accounts can be hijacked, yes, but if that happens it should be relatively obvious from the type of content posted. A spambot is typically going to post a URL of some kind, whether it has "http://" or not. If my post doesn't contain anything like "domain.tld", and doesn't contain repeated words, phrases, or lines, it should be pretty obvious that it's not spam. At the very least, if a post is flagged for a CAPTCHA, it could have a Bayesian filter run on it before making the final decision. Such a filter could trivially avoid false positives, which would prevent wasting people's time. – user41313 Mar 8 '12 at 1:19
  • @blujay not since one of the last big spam waves... they were spamming with almost "normal" content. – Mateo Feb 5 '15 at 17:37

We have improved our CAPTCHA system. Hopefully you have seen improvements, but feel free to let me know if there is more we can do.

You must log in to answer this question.