E. Merzlikins' exceptional goaltending turned a competitive underlying matchup into a blowout. The Devils generated 2.57 expected goals but were blanked by a goalie playing 2.57 goals above expected — the game's decisive margin.

Columbus actually dominated the metrics. They controlled 57% of 5v5 xG (3.41–2.57), converted two rush chances into goals across 22 attempts, and generated higher-quality looks at 0.071 xG per shot versus New Jersey's 0.049. They had six high-danger chances to the Devils' two.

But New Jersey wasn't completely overrun by volume. The Devils held 48% of shot attempts (24–26) and created legitimate scoring opportunities — just none that found the net against Merzlikins.

M. Olivier's shot from the high slot early in the third period (P3, 16:00) served as the turning point, a moment when Columbus' control crystallized into separation on the scoreboard. Olivier finished with two goals in just over 14 minutes; Z. Werenski added an assist across 23:11; C. Coyle chipped in another setup in 16:35.

J. Markstrom stopped 23 of 25 shots and actually outperformed expected performance slightly (+1.41 vs xG), so the loss falls squarely on his team's reduced volume and quality, not goaltending. Columbus' 3-0 scoreline was real — but Merzlikins made it more decisive than the underlying play suggested.

The model can't separate them. It projects 2.6 goals for Columbus against 2.6 for New Jersey and gives the Devils exactly 50% to win—a genuine toss-up.

Columbus comes in hot with nine wins in their last ten games, but the underlying metrics tell a tighter story. The Blue Jackets' 52.0% expected-goal share inches past New Jersey's 50.0%, yet both teams sit at 51.0% on Corsi—puck control essentially identical. The gap emerges elsewhere: Columbus has run hot (PDO of 1.003, meaning their actual scoring beats the shot maps), while the Devils are running cold (0.971 PDO, primed for positive regression).

Boone Jenner powers Columbus's attack at 0.98 expected goals per 60 minutes, with 11 goals and 21 assists anchored by a sustainable 9.6% shooting rate and 2 high-danger conversions. Mathieu Olivier provides secondary impact—0.89 xG/60, 12 goals, and an elite 13.3% shooting rate. The weak link is defensive: Conor Garland.

New Jersey counters with Timo Meier at 1.08 xG/60, the highest volume here, plus 13 goals, 15 assists, and 3 high-danger goals. Jack Hughes trails in opportunity (0.88 xG/60) but leads in scoring: 18 goals and 17 assists on an 11.5% rate that will eventually normalize. Johnathan Kovacevic represents their defensive vulnerability.

The oddity: both teams have registered 0% of their goals from high-danger and rebound situations, meaning their offense lives entirely on top-end talent. When the model cannot split possession metrics, shot quality, or depth-scoring patterns, when luck is the only separator, and when both defenses show similar soft spots, you get a true coin flip.