> Benny Pedersen wrote: > > On tir 27 okt 2009 18:44:28 CET, John Hardin wrote > > > >>> 0.000 0 112532 0 non-token data: nspam > >>> 0.000 0 844 0 non-token data: nham > > > > try to get them more equal numbered in your trains > > > >> reflect your actual raw spam/ham ratio, but yours is a little strongly > >> skewed towards spammy tokens.. > > > > result of not scanning outgoing mails will be sign of this
On 27.10.09 18:14, Adam Katz wrote: > I disagree. I see no reason to scan outbound mail, and this > particular aspect of it is more harmful than helpful. I see good reason to scan (and even refuse over some score) outbound mail: If your customers start spreading spam, you can save yourself from problems. Of course, the rejecton limit should be safely high and you should be carefull when choosing 3rd party rules, e.g. botnet. > Bayes examines both bodies and headers of messages; if you scan your > outbound mail to even out your ham:spam ratio, you are watering down > your bayes db. Basically, it will learn from the headers that all > outbound mail is ham and all inbound mail is spam. Outbound mail must not be automatically marked as ham, mostly because of the reason above. This is just another reason to be careful about that. -- Matus UHLAR - fantomas, uh...@fantomas.sk ; http://www.fantomas.sk/ Warning: I wish NOT to receive e-mail advertising to this address. Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu. I just got lost in thought. It was unfamiliar territory.
