Many spammers are using badly broken bulk email generators, which ensure that a significant proportion of the spam that does arrive could be impossible to display or incomprehensible.
Richard Jowsey, the author of in-development spam filter Death2Spam, says almost 1% of spam messages come with “bad content type, bad encoding, bad structure, null content” or simply “malformed” notices. Jowsey arrives at this figure from accumulated data from D2S’s analysis of traffic to its alpha and early beta users.
“These people are not only greedy and rude, they’re obviously stupid enough to be using completely dodgy bulk-email software,” Jowsey tells his users.
One statistic might encourage the view that spam does only minor harm to the network: 58% of spam messages are less than 10kB in size, with another 28.1% falling between 10kB and 25kB. However, legitimate messages received by the sample are shorter, with 62.3% under 10kB.
“Notably, there are no spam messages larger than 250k (although that could change),” Jowsey writes cautiously.
Curiously, virus-infected messages are practically all of a similar length; 97.2% of them fall in the 10-25kB bracket.