Well first of all... I'm not trying to defend Campbell, or Skinner for that matter, I've lost confidence in both... but that has everything to do with the fundamentals of
how they play the game, and nearly zero to do with their HDsv%.
Over the long haul / large n-size, I might be inclined to agree with you, that HDsv% could distinguish between one goalie and another... but we're talking about 5 games for Campbell and 6 games for Skinner where they were BOTH on the same team and BOTH near the bottom of the league in HDsv%.
Now all of a sudden,
something changed, and Skinner's HDsv% is spiking through the roof over the last three games. Did he improve? Should we trust him as a starter now?
And when goalies switch from one team to the other and their SPCT spikes? Or tanks?
Or how goalies tend to have better save percentages when they have 50 shots against vs 25?
All of my personal experience says otherwise, so I’m a skeptic. There is just so much that these stats don't/can't capture. How often have you seen a goal go in and it takes the play by play guys three replays to determine that the puck was tipped? Do you think the goalie didn't know that already at full speed? Of course he did, so did the shooter and so did the guy who's stick or skate or sock it brushed... everyone else? Random chance whether they saw it or not... hopefully they were busy tying up their man
Does defining "HD" take into account whether there's a guy camped out at the side of my crease? And whether or not a trustworthy defender was there to tie up his stick? I can see that out of the corner of my eye... and that's going to be the difference between me being able to push to the top of my crease and have a chance to cut the angle on the shooter... and whether it brushes off my shoulder over the glass for a stoppage (since I'm way to slow to stop it otherwise) or whether it sails freely into the top corner because I had to stay back and respect the pass.
This is where numbers meet science, where trends meet regression analysis and I just don't believe the capture resolution is meaningful enough yet. We aren't talking about a pitcher and a batter... there are way to many uncontrolled variables and we aren't measuring them yet.