Spam protection Aggressive setting blocks legitimate emails!

[jkc054]

Because of the phishing emails I have been receiving fraudulently claiming to be from Fastmail support I created a ticket and Fastmail support changed my Spam Protection from Normal to Aggressive. I created an Uber account but did not receive the confirmation email either in my Inbox or my Spam folder. I changed the spam protection from Aggressive to Basic and requested Uber to resend the confirmation email. I received it right away in my Inbox.

My concern is that when I have my spam protection set to Aggressive that legitimate emails labeled as spam do not land in my Spam folder and are not even delivered!

I cannot afford to have legitimate emails labeled as spam and not even delivered so I can read them.

Why did this happen and should it have happened?

Has anyone else experienced this? I cannot afford to have legitimate emails blocked from view so I guess I will have to move the spam protection back to Normal.

[n5bb]

I'm surprised that the Uber message had a spam score of 10 or greater, which is the discard threshold for the Aggressive setting. If you expect to receive messages with such high spam scores, you should never use the Aggressive spam setting unless you know that all messages you receive are whitelisted (sent from people in your online Address Book). Look at the Spam/Virus Protection page description of Aggressive protection:

Aggressive

• Known insecure email hosts/relays are always blocked (xbl)
• Possible spam is placed in your 'Junk Mail' folder
• Extremely probable spam is discarded without you seeing it
• Email from people in your address book is definite not spam

If the normal spam settings are not aggressive enough for you, here is what I suggest so you can customize your spam settings:

1. Change your Spam Protection setting to Normal and click Save Settings. It is very important that you click Save Settings!
2. Next change your Spam Protection setting to Custom and again Save Settings. Your Advanced Spam Protection Settings will now be defaulted to the Normal setup.
3. In Address Book, leave the setting at "Always accept".
4. In Define Spam, do not use the "then discard" feature.
5. In Probable Spam you can lower the setting from the default for normal (5.0) to a lower value. No messages will deleted - spammy messages will just be filed into your Junk Mail folder (which is renamed Spam in the current normal Fastmail web interface main page).
6. In Possible Spam, if you are using the current normal Fastmail web interface to read email you can uncheck the box to change the subject.
7. I recommend setting the Backscatter Action to Junk Mail.
8. Now click Save Changes.

More suggestions:

• In the normal current Fastmail web interface, messages from Fastmail staff about your account will be marked with a green checkmark in the From area. If you don't see the green checkmark it's not from Fastmail support. You should never click a link in any email or reply to the email unless you are sure it is from a proper sender who is not a scammer. Fastmail will never ask you your login password, and you should never enter that password unless you go to to the Fastmail website in the normal fashion - never log in after clicking a link in an email!
• You can check the spam rating of the fake messages you received to determine how low to set your Probable Spam setting.
• If the fake support messages are similar to each other, you may be able to file them automatically in your Junk Mail (spam) folder using the rules system. If they are coming from a non-Fastmail From address you can discard them in the rules if you are very careful.
• I strongly suggest that you keep your online Fastmail Address Book updated with your desired correspondents. This will keep their emails from accidentally going to the Spam (Junk Mail) folder.

Bill

[BritTim]

One thing Bill does not mention is checking whether you are using a personal Bayes database. Generally speaking, if you are using settings that train your database effectively, it will be better to be using a personal Bayes database as it will learn the characteristics of the spam you are receiving and react accordingly. It can, however, happen that an improperly trained personal Bayes database can actually lead to worse results than the default global Bayes database. It is worth checking which database is in use and considering whether you are receiving the expected results.

[n5bb]

That's an excellent point, BritTim! The things which in the long term I find to be most useful in managing spam are:

• Use the user Bayes database appropriately.
• Keep all of your normal correspondents in your online Address Book and use whitelisting.
• You should always check the Spam (Junk Mail) folder periodically and use the Not Spam button to move good messages back to your Inbox and mark them as good. Do not manually move these messages.
• Use the Report Spam button to move bad messages to your Spam folder and report them. Do not manually move these messages.
• If you make a mistake in reporting spam, don't worry - just fix the problem using the buttons described above. Your user Bayes filter will correct the mistakes. If you use an email client there are also ways to use the User Bayes filter.
• If you use these techniques (address book whitelisting and user Bayes filter), you can lower your spam "file to Junk Mail" threshold and you will see very few good messages ("ham") in your spam folder. Look at the spam scores reported in the current web interface for messages in the Spam folder and you can see how best to set the spam filing threshold.

Bill