I have just now realised that previewing Spotify playlists can lead these previews to be counted as listens (though it may only affect LB if the track is short).
I was thinking it weird that a track I’ve not heard in a while seemed to be rather popular within my past several listens. When I checked ListenBrainz, indeed. It appears the track was not only detected, but correctly tagged (i.e. not a LB link mistake), when I had not, in fact, heard it. Looking into my recent activity on Spotify, not only that track, but many, that I’d previewed in playlists but not heard, were logged.
I suspect the reason only this track was tracked by ListenBrainz is on account of the track being a 2-second “interjection” on an album, and as such a short preview would count as over half the song played. This could also explain why a Daily Drive opener “track” (“Today is Sunday, and this is your Daily Drive”) also appears in my past 1000 listens, when I’ve quit using that playlist entirely.
I don’t know what to do with this information, other than perhaps ceasing to use said feature. I don’t think there’s anything LB could do here, as the information probably appears no different than a regular listen. But I thought I’d share this discovery, as looking it up online doesn’t seem to bring anything up.
This is rather similar to another topic I’ve opened in the past, Is there a tool to dedup listens off ListenBrainz?, regarding the Mixing feature. In that case, I’ve solved the issue by manually removing extra listens (clearing out repeat listens that occur between song start and its eventual end by length, e.g. 3-minute song heard 14 times in five minutes) and not using the feature anymore. I reckon it’d be far more difficult doing something similar here, as I’ve been using this feature for way longer, and I’m not sure at what point, if ever, it started behaving this way. And it isn’t as easy to detect amongst my listens.