Well, I'm not going to join the Johnson bashing, but as you said GAR is an estimate. You should always use it with grain of salt. Like every other adv stat out there.
Hockey is a very complicated sport and it is played with so little margins that you just can't extract all the decisive factors in to numbers. There have been plenty of very good players, elite or generational even, who are or were of utmost importance to their teams and wouldn't have gotten near the McDavids GAR numbers.
Does that mean that McDavid is on another level compared to them? And might I ask what that "level" is and how important that "level" is for a game called ice hockey?
I don't mean GAR doesn't work. I'm just saying GAR is just one stat that might help you to see something in players you might not notice without it.
For example if you'd pick the 20 best GAR players of their position to an one team, you'd probably get one helluva team, but it might not even get you in to the playoffs. It certainly would not guarantee any of your players or your team the Rocket, Hart or any of the "important stuff". And strangely most of the "top GAR" team players would see their GAR drop singnifically after one season together. Would that mean that they have gotten so much worse with all the other "top GARs"? Or just imply the shortcomings of that individual stat...
To keep this on topic; even as Laine would never get near McDavids GAR numbers he might still one day be discussed as one of the greatest players. Or might not. Only the future will tell (not the GAR
).