From what I've heard and read about analytics use around the league, we shouldn't be so quick to say that they are better than the public ones. A lot of them are based on a hockey persons perspective, they use stats that a hockey guy feels makes sense rather than one that has been shown to have predictive or descriptive value. Edmonton raved over Russell because he gets the puck out of the zone more than almost anybody in the league. They used stats that showed how good he is at chipping the puck out, while not actually looking at some measurement to see if this is effective. In the end, Russell's zone exit attempts result in the opposition getting the puck more than almost anybody, and that is more detrimental in the long run. I know some teams still use incredibly flawed measurements like scoring chance +/-.
Like
@Daisy Jane says, let's not discount that we are actually slumping as a team lately and that's why the numbers have been bad. How much of the numbers discrepancy is because we came out the gate in a frenzy but have had, for example, Matthews injured and Nylander slumping lately?
I still feel like the whole thing smells of overcompensation to a problem that got exaggerated because Andersen couldn't save a puck, so a solution was implemented that looks better than it is because Andersen could pretty much put two nets out there and still be a competitive goalie. But who knows really?
It's also a good point by
@Daisy Jane (and others I've seen raise this) that it's too simplistic to just look at good results in one system, bad in another. Teams adapt, and we need to adapt with them. We saw before the change how we ran into a brick wall when teams dropped the forecheck on us and set up in the neutral zone. We need a counter-punch when that happens.