And to stay on topic, of the 52,180 spams tagged or deleted by my mail server in the last 2 weeks, 41,505 hit one or more SURBL for a nice round 80% spam hit rate. Compare that to the following hit rates and it seems SURBL are doing their jobs quite nicely. Thanks and congratulations to everyone involved. Note that I score Bayes_99 at 8 points whereas the network tests below get 3-5 points so Bayes gets a bit of an advantage from that. However, SURBL are unmatched at pushing scores towards my delete threshold of 25. Since the inception of SURBL I delete a much higher percentage of spam. Less spam delivered makes me and my users happy.
48,528 Bayes_99 93% 38,102 Razor 73% 29,434 DCC 56% 28,157 Spamcop RBL 54% 23,666 XBL RBL 45%
Daniel
Jeff Chan wrote:
Please add a rule for the JP SURBL list. It catches as much spam as WS or OB:
http://www.surbl.org/quickstart.html
jp - jwSpamSpy + Prolocation data source Joe Wein's jwSpamSpy program is used both by Joe's own systems and also Raymond Dijkxhoorn and his colleagues at Prolocation to process more than 300,000 likely spams per day. The resulting list has a very good spam detection rate around 80% and a very low false positive rate below 0.02%. This data is only available in the combined list multi.surbl.org.
An SA 3.0.1 and later rule and score using URIBL's urirhssub looks like this:
urirhssub URIBL_JP_SURBL multi.surbl.org. A 64 body URIBL_JP_SURBL eval:check_uridnsbl('URIBL_JP_SURBL') describe URIBL_JP_SURBL Has URI in JP at http://www.surbl.org/lists.html tflags URIBL_JP_SURBL net
score URIBL_JP_SURBL 4.0
An SA 2.63 and 2.64 rule and score using SpamCopURI 0.22 or later looks like this:
uri JP_URI_RBL eval:check_spamcop_uri_rbl('multi.surbl.org','127.0.0.0+64') describe JP_URI_RBL Has URI in JP at http://www.surbl.org/lists.html tflags JP_URI_RBL net
score JP_URI_RBL 4.0
Jeff C.
