|
Hi JD:
In my opinion the more tests we have the less false
positives we will have since we can reduce the high weight for some of our other
tests.
What I find interesting is our spam is no longer
failing with borderline weights. As we have added more and more tests a
spam fails with much higher weight and those that are borderline (we hold on 20)
have a higher probability to be false positives.
One example of a test that at first we found to be a
great test but over time we have reduced its weight to 5 is REVDNS. More
and more legitimate mail fails that test. So if we can come up with tests
that can trigger spam and yet with reverse tests negate the ones we know our
chances of success increases.
What triggered this was getting a porn email where HELO
was .Microsoft.com. Can we ever imagine eBay sending an email where HELO
is eBay but REVDNS is EarthLink or a DSL?
For example:
Domain REVDNS
HELO (perhaps even IP range)
this will work best with such companies as eBay,
Microsoft, Amazon, etc.
So if a REVDNS comes up that does not match the domain
or HELO then it gets a certain weight.
it should be easy to implement...
Regards,
Kami From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of J.D. Springer Sent: Saturday, October 11, 2003 11:08 PM To: [EMAIL PROTECTED] Subject: Re: [Declude.JunkMail] Integrity Checker I think this is a good idea in concept, but could produce a lot false positives if not used carefully. For example: When I travel, I have to use mail.earthlink.net as my SMTP server.J.D. Kami Razvan wrote: --- [This E-mail scanned for viruses by Declude Virus at MAILER.DB2Consulting.com] --- [This E-mail was scanned for viruses by Declude Virus (http://www.declude.com)] --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com. |
- Re: [Declude.JunkMail] Null Sender R. Scott Perry
- Re[2]: [Declude.JunkMail] Null ... Roger Heath
- [Declude.JunkMail] Null Sender ... Roger Heath
- Re: [Declude.JunkMail] Nul... R. Scott Perry
- RE: [Declude.JunkMail] Nul... John Tolmachoff \(Lists\)
- Re[2]: [Declude.JunkMa... Roger Heath
- Re: [Declude.JunkMail] Integrity Checker R. Scott Perry
- RE: [Declude.JunkMail] Integrity Checker Markus Gufler
- Re: [Declude.JunkMail] Integrity Checker J.D. Springer
- RE: [Declude.JunkMail] Integrity Checker Markus Gufler
- Kami Razvan
