comments from a new user

[David Relson]

At 08:54 AM 5/10/03, Andrew Pimlott wrote:

.. [snip]...

>This is exactly backwards.  In the following line from -vvv,
>
>                                          n    pgood     pbad      fw     U
>     "andrew"                           785  0.685078  0.226744  0.248672 +
>
>fw is (at least roughly) the likelihood that a message containing
>"andrew" is spam.  Not pgood or pbad.

pgood and pbad are different animals than fw.  pgood and pbad are 
proportions.  The value of fw is a probability.  If memory serves me, it is 
equal to pbad/(pbad+pgood)

>Anyhow, the new wording is perfect.
>
>Andrew

Andrew, to you the thanks for spotting the poor wording.

Peter, to you the thanks for the great new wording that everyone likes!