The model sees Islanders over Canucks as a 55% proposition in a game it projects at 2.5-2.8 expected goals total. Form backs the lean: the Islanders are 6-3-1 over their last ten, while the Canucks sit at minus-5 goal differential in the same stretch.
New York's edge runs deep into underlying play. The Islanders command 48.0% xG versus Vancouver's 44.0%, and lead 49.0% to 46.0% in Corsi percentage. These margins are modest but consistent—the kind that usually holds.
The structural advantage: Vancouver is running a 0.974 PDO, which is cold and screaming regression. They're due to catch some bounces. The Islanders sit at 0.993, neither particularly lucky nor unlucky.
What's this gap? Largely unsustainable variance colliding with a team that doesn't have enough talent to overcome it. Vancouver generated enough good chances to sustain higher production earlier in the season. Now they're at the whim of puck luck.
Individually, Filip Chytil carries Vancouver's offense at 1.03 xG/60 but has just 3 goals on a 15.0% shooting rate—he's overdue for positive regression too, though with less predictive power than team-level PDO. Drew O'Connor is working with 13 goals and 11 assists, quieter than his high-volume peer but efficient at 11.8% on shot.
For the Islanders, Anders Lee owns 1.20 xG/60 with 9 goals and 20 assists, operating as both finisher and orchestrator at a sustainable 6.4% shooting rate. Bo Horvat brings 18 goals and 10 assists on just 0.87 xG/60, converting at 12.8%—he's exceeding his expected output by a margin that rarely holds.
Defensively, Victor Mancini is the weak link for Vancouver, while Alexander Romanov holds that distinction in New York. These aren't game-breaking liabilities, but they're asymmetries worth noting when the game tightens.
The Islanders should win this. The model agrees. The only variable is how quickly Vancouver stops playing like a team running in concrete.
