Quoting: A_K
The image I shared was from stat regressions from 2015-2018; Pat Maroon was not on the Blues. Also, this season Pat Maroon's on-ice xGF at 5v5 (13.89) is almost identical to his actual GF (14), so he's not really a poster child for a guy with misleading expected stats.
The graph is not just xGF. It is RAPM, Regularized Adjusted Plus-Minus. These are interpreted as a skater's offensive and defensive contribution (per 60 minutes) to the league scoring rate (Goals, xG, or Corsi). The entire purpose is to reduce all of the data collected in hockey games into coefficients that compare a single player's shifts to the league.
Besides, xGF has more than one variable. Shot distance is the most heavily-weighted, but shot type, angle, time after previous shot, and a whole mess of other variables are in play. It's not a perfect stat but it means a lot more than, say, "unblocked shot attempts - close".
Who is the one spouting off nonsense?
I believe in goals, shots, shot attempts, and the idea of trying to get a database to put them together into something bite sized. I don't think XGF is there, yet. Nor do I believe in point shares, or whatever. I do believe Parayko is currently stronger than Bowey and better at clearing the crease on the PK. Hence the offer.
Can your badly labeled graph tell me why Madison Bowey's p/60 is ahead of Parayko's this year? Can it tell whether Bowey's back to back 60 pt seasons in juniors give him a higher offensive ceiling?
I think there's really no way to accurately predict when Bowey becomes better than Parayko in similar minutes, but a lot of little stats make me think it should happen, eventually, as his endurance improves. I also think he could be better on PK. Eventually. The O'Reilly pick is top ten protected, so keeping the better defenseman right now doesn't seem like a logical priority.
If I'm Armstrong, I figure out the best prospect I can flip Eller for. This morning nobody batted an eye at a one for one for Puljujarvi (which Washington would not do; Eller is one of their best penalty killers).
That way you get three good prospects and cap space to add a bargain UFA in the summer, and you get a high lottery pick now, and then try to focus on winning, again, right after. Selling before LA and Florida should give the best odds at Hughes, so it's just a matter of finding a playoff team that could use some defensemen, and committing to a plan.