Re: [Asrg] NOT E-Postage - Filtering Header Draft Posted

#0=spam- ; 10=spam+ ; 1-9 could be statistical confidence, or could 
#instead be like what Meng Wong recently described as classes of mail.

I'd rather see a score that has a natural neutral zero point, and then
real number values up and down from there. Using just ten levels seems
rather discrete and linear when the phenomena in question is probably
more Gaussian in nature.

You could map that distribution to a cumulative probability distribution
function, but I'd rather see a raw value used instead.

Regards,

Joe