On 2020-01-24 at 11:29 -0800, Brandon Long wrote:
> On Fri, Jan 24, 2020 at 5:32 AM Jaroslaw Rafa wrote:
>
> Dnia 24.01.2020 o godz. 12:44:56 M. Omer GOLGELI via mailop
> pisze:
> > Google usually displays why it thinks an email is spam when
> an email marked as spam is opened.
>
> Yes, and it's usually always the same reason: "The message is
> similar to
> others identified by our filters as spam". I've never seen a
> different
> explanation in Gmail. That doesn't say anything.
>
> Last I looked the enum had about 10 entries, though I don't know the
> distribution of which one's we
> show.
>
> Brandon
The problem is that it seems that Gmail is pulling your leg. A technical
explanation ("We marked it as spam because the sender says any mails
sent from that server are not from him [SPF and DMARC link]") would be
fine, as it would explain what the issue is.
Browsing on the spam folder I found another gmail reason:
«This message seems dangerous
Similar messages were used to steal people's personal information. Avoid
clicking links, downloading attachments, or replying with personal
information.»
(and it was right, it showed for a netflix phishing mail)
But when a personal message, manually written, sent to one recipient, is
junked with a reason like "It is similar to messages that were
identified as spam in the past"¹, while there clearly that was the first
time gmail would ever have seen that email, that makes you angry with
the filter nonsense.
I guess that message is a generic one for "marked as spam by the general
AI", and will thus be responsible for 98% of junking. I guess those
message somehow look similar to some spam message (email less than X
words, it contains 1 url, headers contain foo, useragnet is bar...) but
not for a "similar" in human sense.
Interestingly, even "do not send to spam" filters are bypassed sometimes
(while others are inboxed but show a "This message was not sent to Spam
because of a filter you created." message).
I used to consider that the training was done on the mails marked as
spam by users, but reading the fine print of the message, is it training
itself based on its own decisions?
By the way, is there a way to move out of the Spam folder *not*
signaling it as not spam? (like a scam message you want to keep, but
that was correctly identified as evil)
Best regards
¹ What I am seeing on Gmail inteface is slightly different than what
Rafa reported. Rafa, was yours not a literal copy?
_______________________________________________
mailop mailing list
mailop@mailop.org
https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop
