And that's why you have to watch the games, and look at all the stats in context. It would be expected that your line with by far the highest offensive zone start percentage has more attempts on goal. It would also be expected from being sheltered against top offensive threats and defensive matchups. You would also know that Corsi is not a perfect representation of actual possession. Perimeter players who make a large number of trivial shots look great according to shot attempts. High danger scoring chances often don't result in a shot attempt at all. This would explain why Marner has 2 even strength goals in 37 games, and Bozak has 6 in 36.
So why is the lowest scoring player on our first line on pace for 3x the number of even strength goals as the lowest scoring player on the Bozak line? Despite much more difficult matchups and roughly 10% higher o-zone start percentage? Because our best possession line is actually the Matthews line. The Bozak line is terrible at even strength and everybody can see it. Their goal differential compared to the matchups and zone starts they get is bad, and it's not despite Marner.
Except they also grade out well by xGF% or even HDCF%. Marner also leads all forwards in HDCF% at 5v5. So, when a guy grades out well by all metrics that capture possession and chances generated, it would mean he is not spending that much time in his own end. You are overvaluing zone-starts, considering a majority of shifts take place on the fly. So this isn't a case of having a high amount of shots from the perimeter inflating their possession stats, which seems to be the conclusion you have reached to dismiss Corsi.
Marner's line has dictated play when he is on the ice. They aren't constantly trapped in their own end. They also don't see a massive split in their favor for zone-starts. Given how little zone-starts actually mean to possession numbers. They still dominate possession and chances generated when you adjust for score, venue, and zone.
I watch the games, the Bozak line isn't our best line but the main difference between Matthews line and Marner's line in regards to the goal differential between Marner and Hyman at 5v5 goals is driven by one line having an 11.5% oiSH%. Part of that is no one on the Marner line is as good a shooter as Matthews, and Nylander is better than everyone not named JVR. I'd also add, Nylander only has 3 5v5 goals to Marner's 2.
Also, a 55% GF% rate is fine, which is what the JVR-Bozak-Marner line is currently doing. I'd like to see them do better, but I do think Marner's struggles have been over-exaggerated. The great Stamkos-Kucherov line is only at a 60% GF ratio, while Scheifele/Wheeler/Connor is below them.