Maybe it would help to get a deeper understanding of Bayesian Filtering: Since it's a probabilistic estimation it isn't necessary to train it with the exact same patterns again and again.

Having trouble understanding and wonering if doing right thing with spamhalter I have some 200+ spam messages in a spam folder Here's what I have done1] I've created A SPAM FOLDER

2] when a spam message comes in, I will right click mouse and look for spamhalter classification and then click "train message as spam"and then proceed to move said message  to that spam folder I've created.

I know I've read where it says it takes several messages before not receiving again  but I'm not getting always same messages as I'm getting various  and lord knows when enough of each will meet the "training" aspect. Is there something I'm not doing correctly?

Thanks in advance