Aaron, It took me a while to figure it all out too, especially since refbacks are so ungodly simple compared to the other notification types (at least from a plumbing perspective). There’s also a missing UI bit in the administrative fields of your comments admin screen and without it you may think they’re just duplicates of your webmentions, which in some sense they are, just with less data. I used the Admin Columns plugin to add a few additional columns to my comments admin UI so I can tell what kind of mention (Webmention, Pingback, Trackback, or Refback) I’ve received. My comments UI now looks roughly like this:

Additional columns at the end of my comments admin UI have been added to indicate the notification type as well as the semantic type. It also makes deleting duplicate/unnecessary refbacks easier.

In simplest terms , when someone clicks on a link on any website that directs to your website, your site logs that fact and indicates which site it came from. It doesn’t provide as much information as a Webmention, Pingback, or Trackback, and it relies on the link on a particular site actually being clicked on. Depending on how much traffic the “sending” site gets, it may take minutes or hours to months or even years for you to get a refback from such a page, so they’re definitely not as realtime a technology as something like Webmention can be.

As an example, the ProfHacker blog mentioned one of my posts in an article at https://www.chronicle.com/blogs/profhacker/weekend-reading-rediscovering-blogging-edition/65671, but since the Chronicle doesn’t support Webmention, pingbacks, or trackbacks nothing happened when they published their article. Thus I didn’t get a notification about it until someone actually clicked on that link to me several days later. As soon as they did click on it, I got a refback notifying me that the post existed. Without refbacks (or diligently pouring through my sites’ web traffic–and who wants to do that?) I would never have known that Lee wrote about and linked to me. Useful refbacks like this are infrequent, but when they happen, they can be quite useful.

Because Webmentions are prettier than refbacks, generally when I get a “useful” refback, I’ll go to the originating site and copy the URL and put it into the manual webmention box on my post’s comment section and use that to send myself a manual Webmention which then looks a lot better and richer than the refback version. I then delete the original refeback to get rid of the duplicate.

The tough part is that I also get a lot of self-refbacks, so anytime someone clicks on one of my own links back to myself, I get a refback for that too. I’ve tweaked my comment settings so all of those typically go to either my spam folder or trash. Given that I self-link a lot and have some reasonable traffic, I’ll get 300+ self-refbacks a day which is annoying, but worth cleaning out once a week for the couple of useful refbacks I get otherwise. Another side benefit is that it gives me a visual/physical indication of how popular particular posts are within a week. Maybe we could talk David into adding some code to mitigate self-refbacks?