The model projects Pittsburgh 2.7 to Boston 2.6, with the Penguins at 51% win probability—as close to a pick'em as hockey analytics produces.
Pittsburgh owns the underlying play. The Penguins generate 51.0% of expected goals at 5v5 versus Boston's 47.0%, and their Corsi sits at 50.0% compared to the Bruins' 49.0%. Consistent advantage in shot quality and volume. Boston, meanwhile, is running hot: a PDO of 1.023 means finishing above sustainable levels. Regression is the price of that pace. Pittsburgh's PDO sits at 1.012—already pricing in correction.
The execution gap is real. Rickard Rakell carries 13 goals on a 12.1% shooting rate with three high-danger chances—his production is sustainable. Filip Hallander was generating 1.28 xG per 60 minutes, but he's likely out (IR, undergoing further evaluation). That vacancy hurts Pittsburgh's depth.
Boston counters with Viktor Arvidsson: 18 goals at 1.04 xG/60 and two high-danger scores. He's both prolific and dangerous. Tanner Jeannot adds secondary punch at 0.75 xG/60 with six goals and an 8.3% shooting rate.
The defensive weak points cut both ways. Pittsburgh's Caleb Jones is a liability. Boston's Elias Lindholm is similarly exposed. Both teams are at 0% on high-danger goal rate and rebound goal rate—neither is capitalizing on prime chances, which means tonight likely hinges on sustained possession and which team's goaltending stays sharp longer.
Pittsburgh's edge is process. The Penguins generate better looks and control play. Boston's edge is heat. The Bruins are converting at rates that won't last. The model splits it because both are true and neither overwhelms. Fifty-one percent is thin, earned through work, not revelation.
The lever is Boston's regression risk. The Bruins cannot sustain 1.023 PDO; Pittsburgh doesn't need to improve to win, only hold. That slight process advantage—paired with Boston's unsustainable finishing—makes the Penguins the marginally safer pick.
