I have had results similar to Peter's. I'm currently running Pyzor as an additional spam qualifier, but I have disabled shortcircuiting of the PYZOR_CHECK rule, due to the number of FPs pyzor hits.
On my server I have shortcircuited the highest hitting rules, that also have the lowest level of FPs: URIBL_BLACK, URIBL_JP_SURBL, RCVD_IN_BL_SPAMCOP_NET, RAZOR2_CHECK I originally had BAYES_99 in this list, but it generates FPs form time to time, so I've removed it from the shortcircuit list. Using shortcircuit rules keeps down processing time, although this is not much of an issue on my server, because 99% of the spam is blocked by spamdyke before even accepting the mail with almost no cpu time. -- Felix Buenemann Am 14.10.2008 17:32 Uhr, Peter Nitschke schrieb: > I just added Pyzor to a server for the last 24 hours out of curiousity. > > All the spam it hit, was already well tagged as spam, eg scores in the 20+ > range, but it also hit a few hams which fortunately had enough good points > to not go above the threshold. > > This may well be a reflection on the effectiveness of my current setup, but > I am about to remove Pyzor from the server. > > Peter > > > *********** REPLY SEPARATOR *********** > > On 14/10/2008 at 9:46 AM Matus UHLAR - fantomas wrote: > >> On 13.10.08 08:25, William Taylor wrote: >>> Is Pyzor worth running these days? >>> Is it still effective? >>> Can anyone using it comment on it? >> works for me, however there are still some error messages. >> And it has FPs for some mailing lists monthly notices (and I have to >> register to be able to list/delist). >> >> -- >> Matus UHLAR - fantomas, [EMAIL PROTECTED] ; http://www.fantomas.sk/
