SH Sean Harding/blog
Spam stats
Thursday, February 13th, 2003

I added a simple little widget in the right column of the front page to list the percentage of my email that was spam yesterday and the long-term average percentage of spam I’ve been receiving.

The definition of "spam" here is simple: it’s what SpamAssassin tags as spam. That means the count will include some false positives, and it’ll miss a few pieces of spam. But my experience with my configuration suggests that these numbers largely cancel each other out. I think it’s close enough for my purposes.

There are a couple of other factors here that may make the spam percentage lower than it otherwise would be. I have a large list of blacklisted domain names and netblocks in my Sendmail configuration. I’m not using anything like MAPS RBL here. I’ve just constructed a list based on where a lot of my spam seems to come from. It’s been pretty effective for me (it’s currently blocking an average of 20,000 spam messages a month). It’s hard to say how many of those messages would truly have gotten through to me; some of them are sure to be dictionary attacks or otherwise sent to invalid addresses. But I have no doubt that my spam percentages would be significantly higher without the blacklists there.

Also, I’ve used temporary autogenerated addresses for a lot of things like posting on Usenet. And my main web page currently has uniquely-generated email addresses on each page. Those are real addresses — mail sent to them will get through to me. But their nature makes it very easy for me to disable any that I find to have been harvested by spammers. While I’m not currently taking advantage of it, I do also have the capability to make those addresses work only for a short period of time. If the spam flood from them becomes too much of an nuisance, I may turn that on. It seems unlikely to seriously inconvenience legitimate users and if the time window were small enough, it would make the addresses useless for spammers.

Update 2003-04-30:

I added some information on my blacklist blocking to the spam stats box on the main page. Some of the blacklist information is discussed above. Since I wrote that, I’ve started using the Spamhaus SBL blacklist. At the time of this writing, it accounts for a very small percentage of my overall blacklist blocked messages.

blog comments powered by Disqus