His xGF is low simply because his possession numbers were poor. It's already been explained why that was the case.
And his TOI% QoC(a far more accurate gauge of QoC) was third on the team. He was playing on the second pairing against second pairing competition.
I'm curious if that original graphic was from a Dellow article. Really seems like his(and the other advanced stats bloggers) style of click bait analytics. Come to a conclusion, find some numbers, frame them so they seem more important than they are(i.e. stats from two years ago), provide the numbers that support it and ignore the context that disproves it.