False Positive Scanning

When I scan through Akismet’s spam log, looking for false positives, I basically just flick my scroll wheel as fast as I can, and “absorb” the spams as they float past. False positives stick out like a sore thumb… but I have no idea why. I can scan through a page with 100 spams in about 3 or 4 seconds, and be positive that they’re all spam. I can’t scan through legitimate comments looking for spam quite so quickly, but it’s still pretty fast. I have no real point to make here… I just thought it was fascinating that I can tell spam apart from ham quickly, but without really knowing how I’m doing it. I doubt that’s unique to me. Spammers are doing a really crappy job at making their spam blend in.

I’ve encountered only two comments where I wasn’t sure if they were spam or not. In one case, I just removed the URL and let the comment stand, because it looked like maybe it was actually a legitimate business user who got overzealous with a link to her furniture store. And I’ve personally dealt with hundreds of thousands of spams, across various sites. They’re going to have to do better if they want to survive.

2 thoughts on “False Positive Scanning

  1. i don’t think the spammers are out to fool us anymore. i think they’re purely trying to confuse the mechanical means of spamm prevention.

    confusing non-spam leads me closer to agreeing with lorelle about deleting any comment, human or not, that doesn’t make a real contribution to the discussion. if your comment could be generated by a machine, why bother? y’know.

    i tracked down one guy who posted a fishy url with a real comment, and found out that he was just a guy, who was kind of overzealous with his hosting link. i let the link stand, since it was a reasonably valid comment.

    i’m generally a fan of respecting the effort of commenting by giving real links to people who do. unfortunately the wp-dofollow plugin doesn’t work so well on my server (IIS)

Comments are closed.