RE: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-10 Thread Benjamin Tomhave
gt; Sent: Wednesday, July 09, 2003 11:49 PM > To: [EMAIL PROTECTED] > Subject: RE: [SAtalk] Wow, spam with scores as low as 2.1?? > > > simply out of interest, may i ask what you define as a 'false negative'. > > -Original Message- > From: Benjamin Tomh

RE: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread mwestern
simply out of interest, may i ask what you define as a 'false negative'. -Original Message- From: Benjamin Tomhave [mailto:[EMAIL PROTECTED] Sent: Thursday, July 10, 2003 1:51 PM To: [EMAIL PROTECTED] Cc: Matt Kettler; Simon Byrnand Subject: RE: [SAtalk] Wow, spam with scores

RE: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread Benjamin Tomhave
> I'd say that with all checks (bayes, rbl's, razor, dcc etc) enabled, that > 5.0 is pretty good for "straight laced" email acounts, and 7 is reasonable > for "most people". Any higher than 8 and you start missing a lot of spam. > Given the commentary in this thread and the fact that my bayes seems

RE: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread Simon Byrnand
At 21:36 9/07/03 -0400, Matt Kettler wrote: At 10:46 AM 7/10/2003 +1200, Simon Byrnand wrote: >0.5 (Severe), Egads, > 1.5 (High), Ouch, > 4.0 (Medium), Oww. These three settings are going to give you false positives galore. 4.0 isn't all that bad.. according to STATISTICS.txt 4.0 should g

RE: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread Matt Kettler
At 10:46 AM 7/10/2003 +1200, Simon Byrnand wrote: >0.5 (Severe), Egads, > 1.5 (High), Ouch, > 4.0 (Medium), Oww. These three settings are going to give you false positives galore. 4.0 isn't all that bad.. according to STATISTICS.txt 4.0 should give you 0.44% FP rate in v 2.54, which is a

RE: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread Benjamin Tomhave
> These three settings are going to give you false positives galore. > > Our default "suggested" setting is 7.0 and customers can't see it > any lower > than 5.0 > Actually, under our system, I tested for over a month with different levels and, given the underlying scoring system(s), there are very

RE: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread Simon Byrnand
At 12:28 9/07/03 -0600, Benjamin Tomhave wrote: I don't know that I'm configured correctly on my system, but our relative spam sensitivity scores (required hits) that are pre-set for customers are: 0.5 (Severe), Egads, 1.5 (High), Ouch, 4.0 (Medium), Oww. These three settings are going to g

Re: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread Jim Ford
On Wed, Jul 09, 2003 at 12:28:49PM -0600, Benjamin Tomhave wrote: > 0.5 (Severe), 1.5 (High), 4.0 (Medium), 10.0 (Low). The majority of spam I > see tagged by SA w/ DCC and basic checks enabled (no RBL) tends to fall in > the 2-7 range. I'm always amazed to hear about scores in the teens and > t

Re: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread Jim Ford
On Wed, Jul 09, 2003 at 08:19:03PM +0200, Tony Earnshaw wrote: > See my site, http://www.rimblesister.com Tried this link, but it didn't work - shure there's not a typo? Now I'll never know what a 'rimblesister' is! ;^} Regards: Jim Ford -- Spam poison - don't use! ---> [EMAIL PROTECTED] <---

RE: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread Benjamin Tomhave
I don't know that I'm configured correctly on my system, but our relative spam sensitivity scores (required hits) that are pre-set for customers are: 0.5 (Severe), 1.5 (High), 4.0 (Medium), 10.0 (Low). The majority of spam I see tagged by SA w/ DCC and basic checks enabled (no RBL) tends to fall i

Re: [SAtalk] Wow, spam with scores as low as 2.1??

2003-07-09 Thread Tony Earnshaw
Dragoncrest wrote: I just saw something weird today. I'm running SA 2.55 and I hit a spam message that scored as low as 2.1 and I still have yet to figure out how. It was a blatently obvious spam, but it scored very low. I know that a number of members talked about this not too long back, bu