Where does spam come from?

As you may have noticed, the amount of spam scrolling down my sidebar has increased dramatically of late.  This morning, for example, I deleted 22 comments consisting of:

• how to get young guy to do [something in/with the Cyrillic alphabet]
• very good teenagers that cut themselves
• cannoli ["margarina vegetale, farina di manitoba, acqua, sale, malto, zucchero"]
• I like iTouch myself
• help them ebony bikinis
• tiny cartoon porn videos
• very interesting monster of cocks
• fables to funny qwerty dreams

Given the uptick, I thought I'd try to see where the spam came from.  I tried to correlate the time the comment was posted with the reference in my site stats, but the result was meaningless.  Take the cannoli comment, posted at 9:59 p.m. last night.  Accounting for random lag (five minutes earlier and later), the people who could have posted that spam came from:

• a syllabus linking to my Dark Knight posts [which, by the way, very cool]
• moria's post on experimentation in seminar papers to my homepage
• my 271 notes post to my Five Year Rule post here
• urbino's link to My Morning
• a sundry links post livejournal to my post on that kerfuffle
• another link post via livejournal to DISADVENTURE! et al.
• a post on Watchmen to my posts on Watchmen
• Scott Madin's linking to the same at Shakesvilles
• a couple of random pilgrims worshipping at the Shrine of Hello Kitty

All of those links and search results seem legitimate.  (Unless I'm wrong about the livejournalers, but they seem like actual people.)  Moreover, the comment itself was on this post.  How did it get there without registering in my stats?  I see no hits to that post in the entire hour before or after.  Where did it come from that it can do that? 

Comments

[The Modesto Kid]

Do you access your stats via a typepad or third party front end? The spam comments are generally from bots which depending on how the front end filters the access log, might not show up (since they would generally only be a single line in the access log.) I get about 100-200 attempts to spam my blog every day but they have not cracked my Javascript trap yet -- nearly 18 months old now. I have only once had a human post a spam comment to my sight -- it would not pay to hire people to post spam.

[Vance Maverick]

Also, your whole blog archive is in Google, so it wouldn't be too hard for a spambot to browse it without hitting your site at all.

[SEK]

My stats are from TypePad, so I'm not exactly sure what they show me. I'm guessing the bots access the pages, but TypePad doesn't show them? According to the official spam-filter, it's blocked 12,361 attempts to spam in the past 30 days, so I suppose the increase of 20 or so every day should be mitigated by the fact that 412 don't make it through. I can see why TypePad would delete the hits from the 412 it caught, but why would it delete the ones from the 20 it didn't?

Good point, Vance, the corollary to which would be---in addition to well-nigh unanswerable way---what attracts them to the pages they're attracted to? The other two pages that regularly attract attention are this one and this one. Are there, I don't know, literarily-inclined bots prowling my Google archives?

[SEK]

Ha! Now I'm reading the CT thread on comments to find Rich writing this. I normally don't, I swear!

[todd.]

My guess would be that it's possible to send a POST to the correct commenting script without actually loading any of your pages, and that the spambots already know the URL of that script. But I don't know as much about this stuff as I should.

[JPool]

I feel like "very interesting monster of cocks" should lead to muppets blogging, EotAW-style. See what you can do.

[G C]

Why are we dancing around the obvious? SEK is spamming his own blog and then doctoring the logs to hide the evidence.

[sharon]

I've just looked at my spam stats and they've gone a bit crazy this month - about 1100 in February has suddenly leapt to 7600 in March. Crikey. Fortunately, Akismet rocks.