C-Command Software Forum

Changes to level of spamminess

I have been using SpamSieve for quite a few years. It is a fine product and has worked very well, so I have not spent much time thinking about how it works. But lately, with an increase in the number of spam messages, I have reason to fine tune it, if possible.

The Spam-catching strategy is set to Aggressive, and that has not caused any problems. I use your AppleScript program to delete messages from the Spam folder once I check for false positives. Since I see very few false positives, I would like to automate the deletion of the Blue and Gray spam messages. Those are the most common colors that I see in the Spam folder, but I also see a few messages with other colors that should never produce a false positive. I would like to include those selected messages having colors other than Blue and Gray in the automated deletion process.

First question: Once a message is defined as spam and assigned a color, will that color change and will the message be considered spammier as SpamSieve finds more messages like it? Will an Orange or Red spam classification for a message eventually change to a Grey or Blue spam classification?

Second question: If SpamSieve does not change the color classification of a message type as additional messages of that type are received, is there a way of me to manually increase the spamminess (along with the color) of those message types?

No, SpamSieve doesn’t go back and look at old messages again (unless you tell it to by manually re-applying the rules). However, if you get more messages like that in the future, SpamSieve may decide that they are grey or blue.

You could train them as spam. I would not do this for all caught spam messages—it’s generally better to only train mistakes. But if there are certain ones that SpamSieve thinks are closer to the border line that you want to make sure it thinks are spam, you could train them.

Thank you, Michael.

I guess my first question was not clear, but you gave me a good answer anyway. I now understand better what SpamSieve is doing.

I take it from your answer to my second question that manual training of SpamSieve is possible in this case, but that it might be more trouble than it is worth.

I will watch the Spam folder more closely now to get a better idea of what percentage of the spam messages are not colored Blue or Gray. If I see enough messages coming though that are less spammy than that, I might give training a try.