Bayes scoring question

[Theodore Heise <th...@heise.nu>]

Hi all,

This may have been addressed previously, but I couldn't find it in
the list archives.

I was looking over scores of my newly installed 3.0.0-rc2 and
noticed that for fourth column[1] the BAYES_95 score is higher than
BAYES_99.

	score BAYES_00 0 0 -1.665 -2.599
	score BAYES_05 0 0 -0.925 -0.413
	score BAYES_20 0 0 -0.730 -1.951
	score BAYES_40 0 0 -0.276 -1.096
	score BAYES_50 0 0 1.567 0.001
	score BAYES_60 0 0 3.515 0.372
	score BAYES_80 0 0 3.608 2.087
	score BAYES_95 0 0 3.514 2.063
	score BAYES_99 0 0 4.070 1.886

This seems counterintuitive to me, based on my understanding of
probability and statistics (which is admitedly just enough to be
dangerous).  Is this a result of some interaction?  For example a
message that meets BAYES_99 is also more likely to trigger some
network tests, so a lower score is needed to reduce FPs?

Thanks for any patient explanations!

Ted

[1] Thanks as well for the great explanation yesterday of what the
four columns mean--I was wondering myself.

-- 
Theodore (Ted) Heise     <th...@heise.nu>     Bloomington, IN, USA

Re: Bayes scoring question

[Tom Meunier <to...@mvps.org>]

Theodore Heise wrote:

>This seems counterintuitive to me, based on my understanding of
>probability and statistics (which is admitedly just enough to be
>dangerous).  Is this a result of some interaction?  For example a
>message that meets BAYES_99 is also more likely to trigger some
>network tests, so a lower score is needed to reduce FPs?
>  
>
http://wiki.apache.org/spamassassin/HowScoresAreAssigned