>Some other "negative" scores that I find odd: > >SPAM: LOW_PRICE (-1.2 points) BODY: Lowest Price >SPAM: HTML_FONT_COLOR_RED (-1.2 points) BODY: HTML font color is red >SPAM: BIG_FONT (-0.4 points) BODY: FONT Size +2 and up or 3 and up
I also saw these curious negatives while evaluating the SpamAssassin reports in "caught spam" as well as the "near misses" (things that got through). They seemed counterintuitive when I looked them up on the website's lists of tests and I felt that most should in fact be positive since they were signs of spam. I understand that the scores are generated by a genetic algorithm that scans a test archive of spams and derives the scores -- but that doesn't mean that a little seat-of-the-pants intuition by the administrator can't come into play. :-) I grepped for all the negatives and overrode them with positive scores in my local.cf file. In each case I tried to pick a score that "made sense" although I freely admit that I could be off base in many cases. Truth is I'm still in the process of tuning them so that I don't get false positives. For what it's worth, here is the list that I inserted in my local.cf -- take these with as many grains of salt as you wish. They seem to be working fairly well for me so far, although it's become clear that some of the point scores are a little overzealously high and could be lowered. (I think I assigned these point values during a bout of especially high anti-spam fever, so...) The point is, these are the ones that were negative in the stock scores that really seem like they should be positive. # Overrides on scores that GA made nonsensically negative score AUTO_EMAIL_REMOVAL 3.0 score BIG_BUCKS 1.5 score BIG_FONT 1.5 score BUGGY_CGI 5.0 score CABLE_CONVERTER 3.5 score CALL_FREE 0.5 score CONGRATULATIONS 1.0 score COUPON 3.5 score CREDITORS_CALLING 2.5 score CREDIT_BUREAU 2.5 score DEAR_SOMEBODY 1.0 score DEAR_SOMETHING 1.0 score DISCLAIMER_LEGALESE 2.5 score DONT_DELETE 5.0 score EXCUSE_14 3.0 score EXCUSE_15 3.0 score EXCUSE_16 3.0 score EXCUSE_17 2.0 score FOR_JUST_SOME_AMT 2.5 score FREE_ACCESS 3.5 score FREE_DVD 4.5 score FREE_QUOTE 4.5 score FREE_WEBSITE 4.5 score FROM_HAS_MIXED_NUMS 2.5 score FROM_US_PHONE 2.5 score GET_PAID 2.5 score GET_STARTED_NOW 3.5 score HIDDEN_ASSETS 4.5 score HTML_50_70 1.0 score HTML_COMMENT_UNIQUE_ID 4.5 score HTML_FONT_COLOR_GRAY 0.5 score HTML_FONT_COLOR_RED 0.5 score HTML_FONT_COLOR_UNSAFE 1.0 score HTML_FONT_FACE_ODD 0.5 score HTML_WIN_OPEN 4.5 score HTTP_NUMBER_WORD 2.5 score INSTANT_ACCESS 2.5 score JAVASCRIPT 2.0 score JAVASCRIPT_UNSAFE 4.5 score JAVASCRIPT_URI 4.0 score LIMITED_TIME_ONLY 3.0 score LINES_OF_YELLING_2 1.0 score LINES_OF_YELLING_3 1.0 score LOW_PRICE 2.5 score MAILTO_LINK 0.5 score MIME_BOUND_DIGITS_2 2.5 score MIME_MISSING_BOUNDARY 3.0 score NO_CATCH 4.5 score NO_REAL_NAME 1.0 score PLEASE_READ 4.0 score PROFITS 2.0 score RATWARE_MBOMBER 5.5 score RATWARE_SCREWUP_2 5.5 score RESERVES_RIGHT 2.0 score SAVINGS 2.0 score SPAM_PHRASE_00_01 0.5 score SPAM_PHRASE_01_02 0.75 score SPAM_PHRASE_02_03 1.0 score SPAM_PHRASE_03_05 1.5 score SPAM_PHRASE_05_08 2.0 score SPAM_PHRASE_08_13 2.5 score SPAM_PHRASE_13_21 3.0 score SPAM_PHRASE_21_34 3.5 score SPAM_PHRASE_34_55 4.0 score SPAM_PHRASE_55_XX 5.0 score SUBJ_ENDS_IN_SPACE 2.0 score SUPERLONG_LINE 2.5 score SUSPECT_LIST_HEADERS 3.5 score TONER 6.0 score TO_EMPTY 6.0 score TO_MALFORMED 6.0 score UNIVERSITY_DIPLOMAS 6.0 score US_DOLLARS_2 6.0 score WINNER 2.0 score WINNER_CAP 2.0 score WRINKLES 5.0 score WWW_REMOVEYOU_COM 5.0 ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
