2019-09-19 Spam

I’m looking to migrate my mail away from Gmail to Migadu. And I’m noticing the spam again. Right! That was the problem and I had forgotten. Back in 2007 I said that my Spam box contained 4065 spam mails. These days I don’t have that many. My Spam folder had 739 mails in it dating from August 18 to September 18, address, so about 25 per day.

I’ve enabled a mailbox for alexschroeder.ch, and I’m forwarding my old emacswiki.org mail, and campaignwiki.org mail, all via an alias, plus all the RFC accounts (abuse, admin, hostmaster, postmaster) via an alias... and so now I’m looking at my Junk folder:

Sep 1626
Sep 1719
Sep 1819
Sep 1924
Sep 2023
Sep 2123
Sep 2230
Sep 2363
Sep 2461
Sep 2527
Sep 2659
Sep 2738

Looks like numbers in the same ballpark. I definitely don’t want to use email without a spam filter. Many of these don’t get caught automatically, unfortunately. Perhaps I have to up my spam filter settings.

Also notice the point at which I started redirecting alex@gnu.org to the same address, doubling my spam input. I wonder how that worked. I used to redirect these to Gmail, so now I’m wondering how this can call add up. It seems to me that more data is required.

For the moment, I haven’t had any false positives (no legit email ended up in my Junk folder) but I have many false negatives (spam ending up in my Inbox which I then manually move to my Junk folder).

Apparently Migadu uses rspamd (based on the headers I saw) and I noticed that you can set the “filter strength”. It was set to “high” and I’ve changed that to “aggressive” on Sep 25. Let’s see whether that helps.

A few days ago I also set my DMARC policy to “reject” in order to help protect other people from spam sent in my name.

Update: I’ve set the “filter strength” to “paranoid” on Sep 28.

Update: I’ve collected a bit more data... False negatives is spam that slipped through. I marked it as “read” before moving it into the Junk folder, allowing me to better quantify how well the filter is doing. False positives is real mail that got classified as spam. I’ve only found them as I looked through the spam folder. Those are the real problematic decisions!

DateSpamFalse NegativesFalse Positives
Sep 2839160
Sep 294600
Sep 304550
Oct 15201
Oct 25941
Oct 35110
Oct 45340
Oct 54730
Oct 65200
Oct 77011
Oct 88310
Oct 97141

The false positives all came from the same political site giving updates on a campaign, thus something I’d be unsubscribing anyway, I think. 🙂



Please make sure you contribute only your own work, or work licensed under the GNU Free Documentation License. Note: in order to facilitate peer review and fight vandalism, we will store your IP number for a number of days. See Privacy Policy for more information. See Info for text formatting rules. You can edit the comment page if you need to fix typos. You can subscribe to new comments by email without leaving a comment.

To save this page you must answer this question:

Just say HELLO