Montreal's 4–1 victory masked a fundamental imbalance in conversion. The teams generated equal high-danger chances (four each), but the Canadiens converted all four while Nashville scored once. Montreal controlled 57% of 5v5 expected goals (3.85–2.88) and 57% of shot attempts (32–24)—numbers that aligned with the scoreline in direction but not magnitude.

Variance explained the gap. The Predators underperformed their expected goals by 1.88 goals, a deficit that transformed a narrowly competitive contest into a rout. J. Fowler (MTL) saved 1.88 goals above expected on 23 of 24 shots (0.958 SV%), while J. Saros (NSH) finished with 28 of 32 saves (0.875 SV%).

Montreal's efficiency extended beyond goaltending. The Canadiens scored two goals on 34 rush attempts—a conversion rate that exposed Nashville's defensive vulnerabilities. Zachary L'Heureux's wrist shot from the slot at 16:26 of the third period marked the turning point.

N. Suzuki orchestrated the attack, posting two assists in 18:53 of ice time. I. Demidov (1G, 1A, 16:11) and O. Kapanen (1G, 1A, 16:05) proved the scoring depth extended beyond the top line.

The projection is damning in its indifference: Montreal at 2.5–2.7 expected goals, Nashville at 53% win probability. This is as close to an even split as the model will produce.

The underlying metrics explain why. Montreal owns 48.0% expected goal share against Nashville's 49.0%—a difference you could miss mid-sip. Corsi splits are nearly identical: 49.0% for the Canadiens, 50.0% for the Predators. Neither team has seized control. Neither team has surrendered it.

But layers matter. Montreal's PDO sits at 1.018—a sign the puck is bouncing their way more often than hockey permits. Nashville at 0.992 means the Predators are due some luck. When regression arrives for Montreal, it could arrive quickly.

J. Annunen deserves scrutiny here. Playing on back-to-back nights, he might yield to a backup, or he might take the ice tired. That wrinkle alone can tilt a coin flip.

Montreal's offensive load rests on Cole Caufield. He generates 0.97 expected goals per 60 minutes and has buried 33 goals on 16 assists, but his 20.0% shooting rate reveals the whole story: he's converting chances at an unsustainable clip, riding high-danger finishes (1 HD goal recorded). Oliver Kapanen does similar work at a lower volume, 0.83 xG/60 with 19 goals and a 17.8% conversion rate that also won't hold.

Nashville's scoring comes via different avenues. Filip Forsberg sits at 23 goals and 17 assists with 0.80 xG/60—steadier production, less dependent on shooting luck, though his 13.6% mark still outpaces what regression suggests. Zachary L'Heureux presents a smaller puzzle: 4 goals on 0.85 xG/60 shows he's underleveraging his chances at 16.0% shooting.

Both teams have generated zero high-danger rebounds and zero high-danger goals, a statistical oddity that won't persist. Defense enters the picture. Mike Matheson is Montreal's weak point; Jonathan Marchessault shoulders similar burden for Nashville. Exposing either one becomes the path to separation.

The model calls it a coin flip because it is one. Both teams generate nearly identical shot quality. Both teams suppress it at nearly identical rates. The difference lies in what happens in the margin—Caufield's shooting luck, Annunen's legs, a bounce or two on rebound play—the very things that make hockey maddening and predictions honest.