Islanders shut out the Devils, 9-0

The model can't pick a winner because there isn't one to pick. New Jersey and the Islanders project to 2.6 goals apiece, with the Islanders narrowly favored at 51% win probability—but that's the furthest thing from confidence.

The underlying numbers explain why. At even strength, New Jersey's expected goal share sits at 50.0% to the Islanders' 48.0%—a dead heat. Corsi splits are equally muddled: 51.0% for the Devils, 49.0% for New York. You're splitting hairs to declare an edge either way.

The real divergence lies in what's actually gone in the net. New Jersey's PDO sits at 0.971, meaning the team has run cold relative to its underlying play. The Islanders are at 0.993, a sign they've converted at a clip closer to their chances. For a team like the Devils feeling snake-bitten, regression cuts one way: upward. That's built into tonight's odds.

Timo Meier is the first name to track for New Jersey. He's carrying 1.08 expected goals per 60 at even strength—among the highest on the team—and has converted at a credible 6.7% rate while collecting 13 goals and 15 assists. More immediately dangerous: he's scored 3 high-danger chances, the kind of shot that ends up in the back of the net regardless of goaltending variance. Jack Hughes runs just behind at 0.88 xG/60 with 18 goals, but his 11.5% shooting percentage is the real red flag for New York's defense. Hughes isn't just creating; he's finishing at an elevated clip.

The Islanders counter with Anders Lee, who edges out every other forward in expected goal generation at 1.20 xG/60. He's collected 9 goals and 20 assists, and he's gotten 2 high-danger chances home, which is efficiency. Bo Horvat adds another layer of risk. His 0.87 xG/60 looks modest until you note the 12.8% shooting rate, the highest on the Islanders' lineup. Both of these guys can bury half-chances.

The defensive liabilities cut both ways. New Jersey's Johnathan Kovacevic will be exposed on the ice. So will the Islanders' Alexander Romanov. One team gets lucky pinning the weakness against the other's hot hand. Neither team is manufacturing an unusual volume of rebounds or high-danger sequences in their own end—both sit at 0% for rebound-goal rate—suggesting goaltending stability. That means mistakes punish hard.

This is a game where the team that executes cleanest in the final minutes probably wins. The models can't separate them, and neither should you.