Bunkum Splotched, Spatula Seethed And The Creativity Of Spammer Naming
from the it's-that-time-of-the-year dept
Yes, we're reaching that time of the year when the news slows down and we get a combination of predictions (again?) and bizarre filler stories. Count this next one among the latter. Jack Kapica of the Globe and Mail apparently has been keeping track of the bizarre (yet creative) names that spammers choose in emailing him. If you've never noticed, it seems that spammers tend to come up with some of the most bizarre names: "Ovaries Secreter, Emm Zcacsog, Bella Pxolc, Candida Outlaw, Capote Dogie, Macon Expel, Exhibitionism Phoneys, Tillman Unscrew, Nuptials Overgenerous, Letdowns Gastritis, Dionysius Swindall, Slugged Shindig, Concessions Burgles, Fikriyya Gurney and Shea Snay." It makes you wonder why they bother. Do these names get through spam filters better? Are they more convincing to make people buy from them? Or are the spammers just bored out of their minds, and this is an outlet? Or, most likely of all, they just use some program to grab two words at random off the internet somewhere.
Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.
Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.
While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.
–The Techdirt Team
Reader Comments
Subscribe: RSS
View by: Time | Thread
Spammers suck...
Seriously though, before I started using GMail, I used HoTMaiL. Every single day I'd clear out at least 10-20 spam emails. Even AFTER I'd check them as spam, they would STILL get through. I'm sure M$ sells @hotmail.com accounts to spammers. I'm never using hotmail again...(well, except for my X360 account...)
[ link to this | view in chronology ]
Spam filters
[ link to this | view in chronology ]
No Subject Given
[ link to this | view in chronology ]
Looks familiar
[ link to this | view in chronology ]
No Subject Given
[ link to this | view in chronology ]
No Subject Given
From: "Gladwyn Orick" gladwyna@khnp.co.kr
To: "Jarvis Jacox" [a fake name with no relation to the e-mail adddress]
Subject: Re: runin cornflower
www.--------.com
=20
Vl
Xa
Va
Cl
A
So
Le
AGRA (30)
nax (30)
LlUM (30)
ALlS (30)
mbien (30)
ma (30)
vitra (30)
- $135
- $124
- $86
- $170
- $120
- $76
- $166
=20
David, what happened? Alex mentioned Medusa- Its a mess and hes right,
he has to go higher up with it. Him, not us. We stay out. Far away out.
What happened? repeated Marie. Whats the old Medusa got to do with
anything? Theres a new Medusa-an extension of the old one, actually-and
its big and ugly and it kills, they kill. I saw that tonight; one of
their guns tried to kill me after thinking hed killed Cactus and
murdering two innocent men. Good God! Alex told me about Cactus when he
called me back, but nothing else. How is your Uncle Remus? Hell make it.
The Agency doctor came out and took him and the last brother away.
[ link to this | view in chronology ]
Re: No Subject Given
Step 1: Take a corpus of spam and legitimate mail.
Step 2: Break the entire corpus into tokens, breaking along whitespace. Include every part of a message, including headers, into the token list.
Step 3: Analyze the relative frequency of each word, and assign it a probably of appearing in a spam mail vs. legitimate mail.
Step 4: Whenever a new mail comes in, tokenize it and assign probabilities to each mail based on the corpus data.
Step 5: Grab the 20 or so 'most interesting' tokens, defined as the tokens that are most strongly spam or non-spam.
Step 6: Average the probabilities of the interesting tokens, and if it's above a certain limit, label the mail as spam.
Step 7: Update the spam probabilities based on this new mail.
There you go! Spam trigger words become very highly indicative of spam (a probability very near 1) while words that often appear in your legitimate mail become very indicative of the opposite (probability very near 0). Since only the 'interesting' words are taken into consideration, trying to dilute the probability with white noise won't have any effect (and will only make the words used in white noise be rated as more likely spammy). Weird spellings and such that spammers use to try to get around blacklists are even better, because they'll almost certainly not appear in legitimate mail, and become an automatic flag for spam after you see them once.
Trials with the process show a highly effective filter that produces virtually zero false positives.
Given this, that mail you just posted would have been caught automatically.
[ link to this | view in chronology ]