As a way of avoiding work, I check the comments on this blog and decide which to approve and which to send to the spam folder. (Lots of stuff gets sent directly to spam; these are almost 100% classified correctly and I basically never need to check there.)
There are different kinds of spam, but I can typically spot it by being close to content-free and with a link to a site that is selling something. I don’t mind if you’re a statistical consultant and you link to your consulting site, but, no, if you submit a comment with a link to some discount DVD site or whatever, yes, you’re going straight to the spam fliter.
Today, though, I got a new kinds of spam: it looked just like the usual stuff but there was no URL, either in the mssage or in the regular URL field. I can’t figure out why somebody would bother to do this.
Perhaps its a mistake? Spammers' code likely has bugs as frequently as anybody's.
I've seen this a few times myself. I think they're either trying to prime the spam filters by putting what looks like valid content, or they're trying to game us by pretending to be a normal user. Or it's a bug and it didn't drop a URL properly : )
Keyword loading? You have a pretty well established site, so it's possible that spammers could use that weight to promote terms. I don't know the end goal, but it seems plausible.
Nate Silver once got some spam in the comments of one of his posts. It was advertising sex toys in Chinese. I mentioned this to Nate in a comment, expecting Nate to remove it, but it never went away. It is one of the most peculiar spams I've seen; I wonder who the target audience was!
Did you try Googling for the precise text of the comment, in quotes? It may be a way of both testing the spam filter and marking with a peculiar and somewhat uncommon phrase combination as a way of marking unmoderated targets for later (this might be particularly effective in systems where comments by the same name/email that are moderated as non-spam once are automatically marked as clean on subsequent tries – this is how Google Groups filtering works).
I received one of those in one of my blogs. It sounded unfocused but legitimate (my blog is in Portuguese, the comment was in English), and I accepted it. Over the next two days, I received some attempts at real spam.
My hypothesis: that kind of comment is a test to see if it passes. Then they may use a spider to track if the comment was accepted. If it was, the traditional troops move in.
yeah, i'm pretty sure adam's right. it's to screw with spam filters and classifiers, so that somehow another wave of spam gets through.
What I don't understand at all is how the spam filters miss the most obvious things… it's a pretty sad commentary on data mining/machine learning/etc. etc. when the spam filter won't notice that anyone whose name includes the word "viagra" is not a real person. I stopped counting how many such comments I have removed from my blog. So goes for anyone whose name includes the name of a pharmaceutical drug.
Yeah, I bet it's probably meant to mess with a (likely Bayesian) spam filter…priming it for future hits.
One of the reasons you get a lot of spam here is because you are a .edu site. In the eyes of the search engines .edu sites are worth a lot more. So, you'll probably always have to deal with it. It sucks, I know.
And I agree with the first poster, basic text spam comments are either screwups or sometimes they are "tests" to see if you are paying attention.