Hello Brad,

Monday, January 5, 2004, 8:58:25 PM, you wrote:

BK> I've been getting a bunch of messages either squeaking by SA 2.61 or 
BK> nearly so (bigevil is nice), and added some rules to make them less 
BK> likely (I couldn't come up with a better name than glop, sorry):

BK> describe        glop_15 15 or more alphas in sequence
BK> body            glop_15 /[a-zA-Z]{15}/
BK> score           glop_15 1.0

BK> describe        glop_20 20 or more alphas in sequence
BK> body            glop_20 /[a-zA-Z]{20}/
BK> score           glop_20 4.0

BK> descibe         glop_25 25 or more alphas in sequence
BK> body            glop_25 /[a-zA-Z]{25}/
BK> score           glop_25 5.0

glop_15 -- 39425s/4835h of 85945 corpus (70035s/15910h)

Some words matched by this rule:
# glop_15="recommendations"
# glop_15="disqualificatio" n
# glop_15="misunderstandin" g
# glop_15="extraordinarily"
# glop_15="interpretations"
# glop_15="Congratulations"
and a whole bunch of PGP signatures

Need I say more?

glop_20 -- 18578s/1376h of 85945 corpus (70035s/15910h)
glop_20="contractorswarehouse" .com (my domain)
glop_20="XXXXXOOOOOXOXOXOXOXO"      (a boundary between sections of email)
glop_20="internationalization"
glop_20="NeuschwabenlandTimes"
glop_20="anthropomorphization"
glop_20="electrophysiological"
and a whole bunch of PGP signatures

glop_25 -- 5497s/205h of 85945 corpus (70035s/15910h)
glop_25="MMMMMMMMMMMMMMMMMMMMMMMMM"
glop_25="communityservicesincorpor"
glop_25="Bcccccccccccccccccccccccc"
glop_25="wwwwwwwwwwwwwwwwwwwwwwwww"
glop_25="XXXXXXXXXXXXXXXXXXXXXXXXX"
glop_25="futureofamericandemocracy"
glop_25="AlbertSherbertOrangehalfg"
glop_25="ZZZZZZZZZZZZZZZZZZZZZZZZZ"
glop_25="villagemarketHamBrownSuga"
glop_25="abcdefghijklmnopqrstuvwxy"
and a whole bunch of PGP signatures.

Most telling, I run with a 9.0 required hits threshold.  Your three rules
resulted in 205 false positives (all ham that matched glop_25).

OVERALL%   SPAM%     HAM%     S/O    RANK   SCORE  NAME
  85945    70035    15910    0.815   0.00    0.00  (all messages)
100.000  81.4882  18.5118    0.815   0.00    0.00  (all messages as %)

  6.634   7.8489   1.2885    0.859   0.59    5.00  glop_25
 23.217  26.5267   8.6486    0.754   0.39    4.00  glop_20
 51.498  56.2933  30.3897    0.649   0.27    1.00  glop_15


Bob Menschel





-------------------------------------------------------
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills.  Sign up for IBM's
Free Linux Tutorials.  Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to