Pittsburgh dominated the game on paper. The Penguins controlled 63% of expected goals at 5v5, outshooting Toronto 37–20, and generated three high-danger chances to Toronto's two. The scoreboard, though, said 4–3 Leafs. This was a heist.

The difference was goaltending and finishing. A. Stolarz stopped 34 of 37 Penguins attempts for a 0.919 save percentage, beating his expected mark by 0.84 goals. T. Jarry, meanwhile, surrendered 1.74 more goals than his underlying numbers predicted, allowing 4 goals on just 20 shots. Toronto overperformed its xG by 1.74 goals overall—either elite shot placement or lucky bounces, probably both.

But there was method to Toronto's night. The Leafs were clinical off the rush, converting 2 of their 20 rush chances. Pittsburgh matched that efficiency—2 goals from 29 rush opportunities—yet it wasn't enough. Bobby McMann's wrist shot from the slot at 13:43 of the third period proved decisive, a goal that broke through Pittsburgh's expected-goal advantage.

The Three Stars tell the story. W. Nylander led Toronto with 2 goals and an assist across 19:04 of ice time. B. Kindel answered for Pittsburgh with 2 goals in 22:23. A. Matthews added a goal and assist for Toronto in 21:58. None of it mattered quite as much as those saves Stolarz made when it counted.

Pittsburgh's dominance in underlying metrics—the shot volume, the controlled looks, the xG edge—won't show up in the standings. Toronto's win did.

Pittsburgh holds a narrow edge in a game the model has rated essentially even. The Penguins project to 52% win probability with an expected score of 2.7–2.5—the definition of a pick'em.

The advantage resides in territorial dominance. Pittsburgh runs Corsi at 50.0% to Toronto's 45.0%, meaning the Penguins control play location and generate 51.0% of expected goals at 5v5 compared to Toronto's 46.0%. Both teams sit near league-average conversion luck (Pittsburgh's PDO 1.012, Toronto's 1.002), so the underlying quality gap isn't masked by goaltending variance.

For Pittsburgh, Rickard Rakell carries the throughline. He averages 1.01 expected goals per 60 minutes, has converted 13 times, and maintains a 12.1% shooting rate on top of 3 high-danger goals—a combination that edges close games toward the puck-possessing team.

Toronto's counter is two elite finishers. Auston Matthews leads with 17 goals and 1.10 xG/60, shooting at 10.6%, while John Tavares has 15 goals on nearly identical 0.91 xG/60 but punches at 12.4% and has produced 3 high-danger goals. When these two get looks, they finish.

On paper, territorial advantage should tip the scales. Pittsburgh's possession edge is structural—it limits high-danger looks and forces lower-quality attempts. But Toronto's finisher caliber can collapse that gap in a single game, which is why the model won't pick a side. Pittsburgh wins if it enforces its play style and stays stingy on chances. Toronto wins if it forces puck chaos and executes on limited opportunities. The model sees both paths as equally probable.