KHOP_RCVD_UNTRUST
After testing the khop rules for a few days, I noticed one oddity. TOP HAM RULES FIRED RANKRULE NAME COUNT %OFRULES %OFMAIL %OFSPAM %OFHAM 8KHOP_RCVD_UNTRUST 410 3.27 11.400.20 19.39 This is a 1-point rule which is hitting 19% of my ham and almost no spam. Should this rule be removed, or at least scored lower? -- Bowie
Re: KHOP_RCVD_UNTRUST
On 10/19/2009 10:11 AM, Bowie Bailey wrote: After testing the khop rules for a few days, I noticed one oddity. TOP HAM RULES FIRED RANKRULE NAME COUNT %OFRULES %OFMAIL %OFSPAM %OFHAM 8KHOP_RCVD_UNTRUST 410 3.27 11.400.20 19.39 This is a 1-point rule which is hitting 19% of my ham and almost no spam. Should this rule be removed, or at least scored lower? KHOP rules contained some useful ideas, but many appeared to be suspect to me so I didn't use it myself. They need to be tested in nightly masscheck to determine their true safety and efficacy. Warren Togami wtog...@redhat.com
Re: KHOP_RCVD_UNTRUST
Bowie Bailey wrote: After testing the khop rules for a few days, I noticed one oddity. TOP HAM RULES FIRED - RANKRULE NAME COUNT %OFRULES %OFMAIL %OFSPAM %OFHAM - 8KHOP_RCVD_UNTRUST 410 3.27 11.400.20 19.39 - This is a 1-point rule which is hitting 19% of my ham and almost no spam. Should this rule be removed, or at least scored lower? It fires only on mail passing through third-party-whitelisted relays like HostKarma-W and DNSWL. That one point is merely limiting the 2+ negative points assigned by the relays, so the net is still negative. However, I've been noticing those third-party-whitelisting relays steadily improve over time. My numbers don't lean quite as favorably towards ham as yours, but they've moved quite a bit from the original ratio.