Quantcast
Channel: Untangle Forums
Viewing all articles
Browse latest Browse all 5188

Scoring...

$
0
0
I have an untangle box thats been running a week or so with Spam Blocker Lite on it.

Spam Blocker Lite is working and stopping some spam.

My issue is that 99.9% of the legitimate mail has a very low score, usually under 2 and never over 3 - so far so good. The filer is set to 3.5 and super spam turned on.

The rub is that most of the spam is either 0 or -ve so gets through - one viagra type mail scored -62. So there's very little point in turning the filter up to high.

DNS is configured to the ISPs DNS and I'm not getting any "can't resolve DNSBL queries" alerts.

So a couple of questions arise:-

With this detection pattern, is it worth upgrading to the full version? - my initial reaction is not as, for my use, there is something about the detection mechanism that doesn't work for me.

Is there anyway I can maybe manually influence the detection to reduce the number of -ve scores?

Whilst I appreciate that a new system has to "learn", how long before the learning process start to achieve results?

Viewing all articles
Browse latest Browse all 5188