I. Sorokin stole a game the Lightning should have owned. New York won 2-1, but expected goals said otherwise: Tampa Bay generated 4.54 to the Islanders' 3.96, a 53-47 split in the visitors' favor. Sorokin's 0.968 save percentage and +3.54 goals saved above expected explain the gap—he turned a likely loss into a narrow win.

The Islanders controlled possession where it mattered. Shot attempts ran 24-31 against them, but they packed more danger into their chances: 0.090 xG per shot versus Tampa's 0.061. That quality advantage should have closed the margin. It didn't—at least not fully. New York underperformed its xG by 1.96 goals, wasting the sharp looks Dominic James and the attack manufactured.

The Lightning's collapse was sharper. Tampa Bay underperformed its xG by 3.54 goals, a chasm that included a 0-for-5 night on high-danger chances. For a team that created nine rebound opportunities, that won't sting as much as the raw shot-creation metrics suggest. Still, this was a game Tampa Bay could have won if variance had broken even.

A. Vasilevskiy held up his end, posting a 0.917 save percentage and +1.96 goals saved. He kept the Lightning in it through the third. But I. Sorokin's 3.54 goals saved above expected proved the difference. That's a goaltender who didn't just stop the puck; he rewrote a story the underlying numbers had already written.

The model projects a 2.7-2.5 Lightning win with Tampa Bay at 53% probability—essentially a coin flip. This is what parity looks like: both teams have paths to victory with negligible separation.

Tampa Bay is the hotter team, running an 8-2-0 streak, though that form masks a structural advantage built on efficiency rather than volume. The Lightning own the underlying play at 54.0% expected goals to the Islanders' 48.0%, and their 53.0% Corsi versus New York's 49.0% confirms they're controlling territory. But this efficiency comes with a debt: Tampa Bay's 1.019 PDO is artificially inflated, while the Islanders sit at 0.993. Regression favors New York.

Brandon Hagel anchors the Lightning attack, generating 1.10 expected goals per 60 minutes across 23 goals and 26 assists, including 2 high-danger goals, while shooting at 15.8%. Anthony Cirelli provides secondary scoring at 0.95 xG/60 with 15 goals, 18 assists, 5 high-danger goals, and 17.0% shooting. The Islanders counter with Anders Lee, generating 1.20 expected goals per 60 minutes with 9 goals, 20 assists, 2 high-danger goals, and 6.4% shooting. Bo Horvat rounds out their scoring at 0.87 xG/60 with 18 goals, 10 assists, 3 high-danger goals, and 12.8% shooting.

The stranger constraint: both teams sit at 0% high-danger and rebound goal rates this season. Neither has developed the net-front havoc that overwhelms playoff goaltending. Victor Hedman represents Tampa Bay's defensive crack; Alexander Romanov shoulders the same burden for New York.

The models call this a toss. The data agrees. Flip it in the air.