Specific Sender Repeatedly Judged Spam

macOS 10.12.6, Mail 10.3, SS 2.9.29

Hi Michael,

For several years, I have had a regular email correspondence with a friend who is in my Contacts. SS has always correctly categorized his emails as good. This week, however, SS has judged them to be spam, and moved them to the spam folder. Each time SS makes this error, I train the false positive with Mail > Message > Spam Sieve - Train as Good.

After this happened a few times, I wondered if I’d mistakenly “trained as spam,” as correcting false negatives is a lot more common than correcting false positives; perhaps muscle memory and habit had controlled my response. So I went back and carefully re-trained those same recent messages as good.

Even after this correction, the error continues to recur; it’s happened twice this morning. SS is otherwise operating correctly; here’s the current Statistics window:

Screen Shot 2017-08-30 at 10.28.59 AM.jpg

The percentage correct number had been about 99.6% before this series of miscategorizations.

Any ideas? Would seeing the SS log file help?

Thanks.

David Brick
Santa Cruz CA

Whoops. I just saw this: https://c-command.com/spamsieve/help/why-is-spamsieve-not-ca

Let me work though that before I ask for your help.

I think the page you want is this one. If the address is in your contacts, it’s unlikely that SpamSieve is classifying the messages as spam. Perhaps you have another rule or server filter that’s doing that. If you need further assistance please send in a diagnostic report.

Hi Michael,

I think I’ve solved it.

I looked at the Spam Sieve log, and found my friend’s messages marked as both presumed spam and presumed good. Then I looked at the Spam Sieve whitelist and blocklist, and found my friend’s email in both, with some “apply rule” checkboxes checked in both lists, and some not. The entire combination seemed inconsistent, and made me think I’d fouled things up during training by selecting the wrong training commands — having in the past selected “Train as Spam” much more often than “Train as Good,” I probably selected the former rather than the latter through carelessness and/or muscle memory. My subsequent retraining on the same messages may have made the situation worse.

To start anew with this sender’s email, I erased all its whitelist and blocklist entries. Then I asked him to send me another email. This email was correctly assessed and appeared in the Inbox.

Thanks for the specific explanations and suggestions in the manual. I’m not a programmer, and the contents were very useful in explaining how the various pieces operate.

David Brick
Santa Cruz CA

Yes, if you train a good message as spam that will add it to the blocklist and also disable the whitelist rule. You can fix this by deleting the rules, as you did, or by enabling the one that you want to keep.

However, I think there may be a second problem. You said that the address was in your contacts, so if you had Use macOS Contacts checked in the preferences, SpamSieve should always classify the messages as good, regardless of the blocklist and whitelist. So I suggest that you make sure that this option is checked and that the address in question really does appear in the Contacts (and in an e-mail field, rather than a phone number field).

Thanks.

The “Use macOS Contacts” Preference was (and remains) checked, and Darryl’s email listings were (and remain) shown in my Contacts as emails and not in a phone number field:

I don’t know why Spam Sieve might have not registered his email. But I’m content to stop poking at it now.

Thanks again for your help.

Then you should be all set unless the address is excluded.

OK. If you wanted to look into this more, you could click this link to enable some additional debug logging of the address book, quit and relaunch SpamSieve, and send in a diagnostic report.

I just checked, both using Terminal and viewing the ME card, and his address is not excluded.

OK. If you wanted to look into this more, you could click this link to enable some additional debug logging of the address book, quit and relaunch SpamSieve, and send in a diagnostic report.

Well, I’m intrigued. I’ll do it.