First, the correct interpretation from these charts - or pretty much any advanced stats - is that they had similar impact overall last season. You're operating within inherent margin of errors of the statistical methods used.
Secondly, it's demonstrable that a combination of actual production AND expected production (and IIRC there's more weight on actual production) at Season N is better at predicting future production in Season N+1 than just either alone. You're putting a lot of weight on "expected" while ignoring the "actual", and if you follow development within the field, emphasis on "finishing" has increased. You can't ignore a player's ability at creating high % scoring chance and exploiting their teammate's strength when it's done repeatedly. Suzuki is a really good finisher.
Purely relative stats (and % too) are comparable for similar volume and context, which isn't exactly the case here for the latter. And, anyways, when you look at their difference in xGF±, it's a difference of 4 xG over their entire season. That is pretty much negligible and external factors that xG doesn't account for could explain a lot. Relative stats are also very sensible to hot streaks and cold streaks of teammates.
Lastly, I voted Zegras because he's fancier and funnier to watch, and at roughly equal impact, give me the more entertaining player - since I watch hockey to be entertained first and foremost.