On Wed, 1 May 2002, Daniel Rogers wrote: > I find this quite interesting, beucase it gives and example of how > ineffective some of the rules have become. (NIGERIAN_SCAM, for example)
There are a couple of things to note about this analysis: (1) It doesn't account for duplication. If you got the same spam 5000 times, whatever rules it triggered will show up proportionately in the counts. This doesn't accurately reflect the effectiveness of the rule in *identifying* the spam -- unless you've already de-duped your archive? (2) It doesn't account for rules that appear in combination. CLICK_BELOW doesn't score anywhere near 5 all by itself; it's only as effective as the other rules that matched with it. So about all you could say from just this analysis is that rules that were never hit could possibly be deleted. _______________________________________________________________ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk