All — thanks so much for your time. Regarding your questions, I'll do my best to answer given my inexperience with SA.
Kevin A. McGrail wrote > We need more headers to see. What version of SpamAssassin does the > headers show? What rules does it show it hits when it does hit? Here's raw source of another example I received today, although I don't see an SA version in there. I also again included the contrasting output from an online check. The difference is 10 points. http://pastebin.com/HbtC6ETu How is Bayes being trained? I'm a stranger to Bayes principles, so I honestly don't know. Here's what the ISP provides me in terms of configuration: 1) A basic cPanel interface labeled "SA configuration" which is made up of form fields that add white list and black list entries and fields for reassigning test scores. http://i.imgur.com/5IrbCz4.png 2) Access to a few raw files, which contain Bayes info and user prefs, which I'm assuming coincide with the above. http://i.imgur.com/ZUagJzp.png That's the extent to which I can tweak SA. Anything beyond requires working with the ISP. They've mentioned editing the user pref file per the following: />Because of that you can train further Spam Assassin to catch more spam and to performed more test if >required. You can edit this configuration file to do the changes: >/home/xxxxxxx/.spamassassin/user_prefs >Different scores can be used, alongside further tests etc. to be carried out. The following are increased >scores for tests that you may consider to be spam catching: >score RCVD_IN_BL_SPAMCOP_NET 2.5 >score RCVD_IN_SBL 2.0 >score URIBL_SBL 1.5 >score URIBL_OB_SURBL 2.5 >score RAZOR2_CF_RANGE_E8_51_100 0.1/ Should this be necessary for a well-maintained instance of SA? Again, thanks very much and just to reiterate, what I'm hoping to understand is what's causing the discrepancy between the scores. Is it my user error (can I affect which tests are triggered? Should I be performing "learning" somehow to affect triggering tests?) or if this is poor maintenance from the ISP ( they're not updating the software, not adding common tests, etc). Given what you've seen, is this "shoddy" plug'n'play for an install of SA? -- View this message in context: http://spamassassin.1065346.n5.nabble.com/Irregular-Test-Reports-in-SA-tp115438p115481.html Sent from the SpamAssassin - Users mailing list archive at Nabble.com.