Scott Williams , Area4 wrote:
Does anyone have a filter that would catch this one below?
Unfortunately, my filter even learned it a ham ! UGh!
With my existing rules, it scored a 2.1. Not flagged, but at least not auto-trained:
X-Spam-Status: No, hits=2.1 required=5.0 tests=BAYES_90 autolearn=no version=2.6
Looks like my previous bayes training got close, but not quite.
After training SA bayes on that specific message:
X-Spam-Status: Yes, hits=5.4 required=5.0 tests=BAYES_99 autolearn=no version=2.63
Better, but 1st one slips through until trained.
After adding: http://www.rulesemporium.com/rules/99_sare_adult.cf
X-Spam-Status: Yes, hits=8.7 required=5.0 tests=BAYES_99,SARE_ADULT2,SARE_BETTERORG autolearn=no version=2.63
I've been avoiding those content-specific rules, but it looks like they may be called for to help nudge bayes along, at least initially.
- Bob
