Hi, On Thu, Oct 19, 2017 at 10:54 AM, Reindl Harald <h.rei...@thelounge.net> wrote: > Am 19.10.2017 um 16:50 schrieb Alex: >> >> My bayes is trained such that most marketing emails are bayes99. I've >> also now removed mcsv.net from the whitelist and see it resulted in 70 >> messages from mcsv.net being caught today, all of which were from >> marketing@ or news@ or similar accounts from sites like >> news@firma.agency > > well, your users will be grateful when you use a biased bayes and reject > their subscribed newsletters - "bayes is trained such that most marketing > emails are bayes99" is idiotic - the only question for traing is SPAM OR NOT > SPAM and if you are not 100% sure don#t train a sample at all
And yet, it would have stopped the email in question. It also relies on other ham rules to subtract points or trusted senders to be whitelisted. I'm also not convinced all of these are opt-in in the first place.
