Those stats usually work on a full season if not more. They can't explain three games.
It's like playing dice. Just because you have maybe 10% more chance than the other, does not mean you won't loose here and there.
This being said, if you look at the line up, it is too top down heavy. It's the gamble they took, and that means less depth, so players are more tired, especially if the game gets physical. It's important to be able to rely on all your lines in the playoffs. If you have only one line, it is easy to stop. Normally they would have had two, but with Tavares out, it meant Matthews playing almost 30 minutes a game, which is insane when the optimum is usually 18-20 for a first line.
Also, stats are useless in games where there is a skin in the game. Like when faced with elimination, you can't rely on above average chances. You need something more. Like in the last game, the Habs played defense the last period, chipping the puck away asap. It's not good stat wise, but it worked. They are a decent team five on five as well generally, but poor in PP usually. So not having a lot of penalties helped them quite a bit. Also shots from the perimeter. Price is perfect with those, not so much in the slot or of course deviations.
Some of those things can be seen in stats, some don't. But it's not something that can be predicted really. Not sure how useful those stats are. I certainly would not use them to build a team that wants to compete in the playoffs. Could work for the regular season though. If they truly followed stats though, I think they would find the cheapest players that bring the most at all positions, taking into account special teams. So superstars would probably be out. They ask too much money for what they bring. A real moneyball team would have four good lines that simply rotate no matter what at about the same pace. I don't see any team doing that.
