High-danger chances said the Blue Jackets should have won. They created two; the Islanders created one. The Islanders won anyway, 4-3 in overtime, because Ilya Sorokin played brilliantly and Joonas Greaves did not. Sorokin stopped 27 of 30 shots for a 0.900 save percentage. Greaves surrendered 23 of 27 for a 0.852 save percentage, 0.88 goals worse than his underlying volume predicted.

Simon Holmstrom settled it at 01:47 of overtime with a wrist shot from the doorstep. The game had already been decided.

Columbus commanded the underlying matrix. They out-attempted New York 30-27 and expected goals ran 2.82–3.12, a slight edge to the Islanders. The Blue Jackets were lethal on the rush, scoring two goals on 25 chances—a rate that should carry teams to victory. They attacked the net, generating seven rebound opportunities. None converted. The Islanders created six rebound chances and converted one, a trifle until you note this: the Islanders ran 0.88 goals above their expected total, and Greaves ran 0.88 below his. Same variance, opposite ends.

Z. Werenski led the Columbus assault with two assists over 28:02. M. Barzal answered for the Islanders with two assists in 20:52. S. Mayfield, the third star, produced a goal and an assist in just 14:19—production that underscores New York's clinical finish.

The underlying numbers promised a game too close to call. Sorokin made it look settled.

Columbus owns the underlying play here. The model projects a 2.5–2.8 Blue Jackets victory with Columbus holding 56% win probability—a lean backed by the season's 5v5 numbers. Columbus owns 52.0% expected goals and 51.0% Corsi at 5v5, against New York's 48.0% xG and 49.0% Corsi. These aren't blowout margins, but they're consistent across both volume and quality metrics.

Form favors the road team. Columbus is 8-2-0 in their last 10 games, translating to a points pace well above league average. New York is playing at a lower clip—one reason the model projects them as a secondary outcome.

Both starters face back-to-back pressure. E. Merzlikins handles the bulk of Columbus's workload, while I. Sorokin anchors New York. Either could see a backup pressed into duty depending on coaching decisions and recovery needs.

The player edge tilts toward Columbus's depth. Bo Horvat drives the Islanders' offense with 18 goals, though he posts just a 12.8% shooting rate on 0.87 xG/60—suggesting some luck in conversion. Anders Lee, New York's power-play driver, carries 1.20 xG/60 with 9 goals and 2 high-danger chances, but his 6.4% shooting is below typical benchmarks. Boone Jenner leads Columbus with 0.98 xG/60 and 21 assists against 11 goals, implying quality creation rate-limited by finishing. Mathieu Olivier adds 0.89 xG/60 with a 13.3% shooting rate—a secondary weapon at full efficiency.

Columbus's possession edge sits atop a small PDO advantage. Columbus's 1.003 PDO versus New York's 0.993 suggests marginal puck-luck benefits, but the underlying volume advantage is the real story. New York's defensive weak link, Alexander Romanov, faces another night in a compressed system. Columbus counters with Conor Garland, a forward who'll anchor secondary matchups.

Players to Watch

• Anders Lee (NYI) — 1.20 xG/60, 9G–20A, 2 HD goals, 6.4% shooting
• Bo Horvat (NYI) — 0.87 xG/60, 18G–10A, 3 HD goals, 12.8% shooting
• Boone Jenner (CBJ) — 0.98 xG/60, 11G–21A, 2 HD goals, 9.6% shooting
• Mathieu Olivier (CBJ) — 0.89 xG/60, 12G–9A, 1 HD goal, 13.3% shooting