OT: Chunking the cruft - random lettered words

[David Relson]

On Mon, 15 Mar 2004 17:43:20 -0500
Matt Garretson wrote:

> Eric Wood wrote:
> > Are you saying "random words" or "random lettered words".  There is
> > a difference.  Lots of unique words having no previous score would
> > help pass the email.
> 
> 
> 
> Is that true?  I always thought that if robx is within the min_dev
> window around .5, then those "unique words having no previous score"
> would fail to have any effect on the spamicity calculation.  Am i
> misunderstanding something?

With 0.415 those words are slightly hammish.  However with min_dev=0.1
they're ignored.