On 31 Jan 2003, Tony Earnshaw wrote: > This joker/subscriber set off my automatic smtp 550 mail refusal system > [...] > I've gzipped his offending mail and attached it, so that people can see > *why* it was refused (it got 9.1 points, trigger is 5.0).
A 5.0 trigger is much too low (IMO) for generating an SMTP-level refusal. Even though that's the default level for tagging as spam, SA is just not _that_ accurate -- particularly if, as appears to be the case from your other postings, you're applying Bayesian analysis site-wide. Everything I've read so far about Bayes scoring indicates that it's best applied at the individual user level. ------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
