I am curious regarding the confidence level, which shows up on Fresh Releases. Higher means more confidence that the release matches the user’s taste (or so I think). But what’s the scale, here? Is that, like, a percentage? Is there a maximum value?
I, for one, see mine range from 10 to 233. Are those just, results of some calculation, without necessarily a scale? Seeing values above 100 make me feel like “damn, that’s, like, a super fave or something”, like it just match me above 100%.
Also, now that I think about it, does that value change over time? I mean, I see releases from months ago (regardless of what “Range” as a filter is set to). If I listen to these releases more and more, does confidence rise over time? I guess I could store values myself and test this out in practice, but if anyone knows this, it’d be simpler.
Wait, really? Wow, so simple. I thought there’d be some sort of fancy algorithm or something. Is it a general listen count (All time) or is there some cutoff, a recency limit?
How do fresh releases get recommended? Is it based on number of artist listens?
How does confidence work for releases with multiple artists? Does it sum up the individual artist listens (OR), or count the artists together as one (AND)? I’m guessing the former.
If you listened to an artist a minimum number of times and that artist as a new release, that release ends up on your fresh releases. No magic, just really simple counting.
I don’t know the details about how multiple artists are included, but my guess is that if any of the artists have been listened to a minimum number of times, it will end up on fresh releases.
If I stop listening to an artist altogether, does that have any effect on Fresh Releases? Like, say I listened to an artist a lot in 2019. Nowadays, though, years later, no more. Moved on, listening to something else. Would that artist still show up under Fresh Releases?
On one hand, I feel like considering only recent listens could lead to ignoring part of someone’s music taste that might not be as prevalent at a given point in time. For instance, if someone goes through a phase of listening to more pop and less electronic (swap genres as you wish).
On the other hand, though, if Spotify was still recommending me stuff I listened to years ago and have since stopped listening to entirely, that’d be annoying. Like it didn’t adapt along with my music taste, and I’m to be forever attached to a bygone era of my music taste.
Interesting to know as well now that importing listens is finally here (I already asked Spotify for the latest copy of my extended streaming history). Like, is ListenBrainz going to suddenly start recommending stuff I’ve gotten over years ago? Is the random stuff I apparently listened to in 2016 going to haunt me? (I don’t even remember using Spotify in 2016, but I got an end-of-year playlist for it in late 2019 (???). I don’t even consider that as part of my music taste history. It’s the odd one out)
Funnily, I do kinda see that with the automated daily LB playlists (songs I haven’t listened to in a while). I listen to songs I like, and then they disappear next time. But since I don’t listen to the ones I don’t like, they remain.