Pictured is a torrent that has stalled for me. It is a very large torrent 100GB+ and I had only selected a portion of the contents to download. It wasn't intuitively clear to me what the 'Relevance' column was actually showing me. My expectation was that it would show what percentage of the pieces I still needed were possessed by a given peer. Obviously this is not the case. It took quite a bit longer than I'd like to admit, but eventually it occurred to me that the column must be showing the percentage not of the pieces I still needed but of all the pieces I had chosen to download.
This explains the numbers I'm seeing (multiple peers with the same relevance percentage as my own completed percentage). But if I'm correct about how the relevance number is being calculated, it isn't a very useful datum. There's no way no way of knowing if any given peer has the pieces I actually need (ie the pieces I want but don't yet have). A peer that shows 30% relevance might have all the pieces I need to complete the torrent or none of them.
A quick poke around the internet led me here and subsequently here. A comment here reads: "Relevance", "i.e: How relevant this peer is to us. How many pieces it has that we don't."—which is exactly what I would have expected the column to provide.
I'm not a programmer or familiar enough with github to determine what the code there actually does or even if that code actually made it into qBittorrent. But I think it is fair to say that the Relevance column in qBittorrent (I'm running 3.3.7 from the Debian Stable repo) tell me what I actually want to know. So I'd love confirmation of how the number is actually calculated and also to request that the information displayed be altered to more clearly show what portion of the pieces I still need (to complete my download) are possessed by a given peer. If that cannot be calculated from the information available (I don't know enough about the torrent protocol to know for sure), then it would be nice if what is shown was more clearly described.