Why spam blocking is so difficult

This is probably at the root of many complaints. "I have this spam blocker software, why am I still getting spam?". In the introduction we touched on the impression that users have when a spam filter is put in place. No matter if users are just you and your family, or an entire organization for which you are responsible for spam blocking, the result can be the same. Sometimes people are unhappy about receiving 3 spams even though 300 spams were blocked. There's nothing this site can do about that, except to give you the ammo to help people 'see the light'.

Spam blocking is never as simple as a boolean equation - a go-no-go decision. The holy grail of machinery, and the somewhat worn out notion of artificial intelligence must be applied. A machine is being asked to make a decision for us - is this email a spam or a ham. Simple enough for a human to decide - sometimes. Often times we are fooled into opening an email to see if it is indeed a ham. And yet some expect a dumb silicon machine to make better decisions than us humans.

In my daily trials and tribulations as a specialist in this subject matter, I often hear from people "I got this spam, and anyone can see that it's a spam. Your spam blocker doesn't work". Requests like these are offered salt to help improve the taste crow they're about to eat. The most rudimentary analysis of an emails raw source code shows that it's not quite so simple for a machine. Besides - if it's so obvious, why bother complaining? Just delete the damn thing. 

Having vented a little, let me state up front, that there are many many reasons why any spam gets through your spam blocker. Some analysis of samples will usually indicate which knob to turn - tune. This site will cover each and every knob that can be tuned - so that you can figure out this stuff out for yourself. Sites that can tolerate no spam at all require constant tuning, tweaking and adjusting.

Employees of companies with such technically incompetent management have few options except spending their weekends and holidays searching spam pits for false positives. Indeed, most dinosaurs do not suffer learning too well. But the vast majority of sites tolerate a 95% effective spam blocker quite well. As stated above, some simple analysis of sample spams tells the tail - what can be done to improve the filtering.

Again, all of this subject matter will be covered in both high level detail, and excruciating detail, but as you are aware of by now, it is no simple subject. It can be very difficult to get a dumb machine to make smart decisions that agree with ours. If you digest the information presented here, it is by no means unreasonable to expect 95% of spams be blocked - with a very minimal amount of false positives.