Goal Differential in the Post-Season

I wouldn’t say it’s been consistent, there was that game against SJ with the back to back to back 5 on 3

Actually what’s happening to Vancouver seems to be the reverse of “clutch”… the occasional “meltdown” game where they give up a truckload of negative goal-differential.

Boston’s case is equally interesting, they seem to have run into a couple goaltending meltdowns of their opponents (Philly mostly, but Tampa as well) and collected the entirety of their high GD based on the Philly series and games 3-4 of the SCF… and otherwise managed to win two series they were outscored in.

Just very weird.

I certainly would not called VAN fluked by CHI. In fact, I think CHI was extremely lucky to have made it past 6 games.

I think a lot has to do with the toughness of competition. By my eyes, in the 4 rounds, VAN had to face the 3 best teams in the entire playoffs. The only so-called weakling is a NSH which is a top defensive team that features arguably the best defensive pairing in the league, a Vezina candiate, and a mindset that is determined to play it tight for the sake of survival. VAN never had a breather against weak goaltending (i.e. like BOS vs PHI). In fact, they’ve had to face 2 of the 3 Vezina candidates, 2 of the 3 Norris candidates – that’s not a recipe for running up the score.

I think VAN probably had one of the toughest route to the Cup in recent memory. And if you compare the aggregate team GD faced by the recent cup winners since the lockout, that seems to support this observation:

2011 VAN 144
2010 CHI 110
2009 PIT 117
2008 DET 79 <— no wonder these guys are at the top…
2007 ANA 186
2006 CAR 56

Anyone have the data over the last couple of seasons for how many stretches of 20 games where a team went -4 GD and still won 15 games? Are the playoffs so ‘unique’ that you can’t simply use regular season numbers to get a proper comparison?

Also, to what extent is everyone anomaly hunting in the data at this point? With better and better statistical tools and more data, do we actually have enough data historically to say much one way or the other in some of these cases? I think we can all take it for pretty much given that 90% of the MSM stat crap is anomaly hunting and small sample sizes, which all make for great (and weak) narrative. At the same time, while I’m pretty comfortable with possession statistics and chance counting, historical data mining does give me a fair bit of pause, in particular because it’s often used to change or remove context, or otherwise justify a narrative for the author.

My problem is that, as someone with a rooting interest, I constantly have to check my narratives at the door when reading posts like this, and it becomes increasingly difficult to harmonize disparate narratives that seem to conflict.