S. Wedgewood was the difference. The COL goaltender stopped 1.8 goals above expected—the variance that turned a tight underlying game into a 3–1 final. Both teams underperformed their expected goals: BUF by 1.81, COL by 1.11. Wedgewood's +1.81 tipped the balance.

The underlying play favored COL cleanly. They controlled 59% of expected goals (4.11–2.81 xG) and generated shots worth 0.065 xG per attempt versus BUF's 0.043. On high-danger chances, the gap was smaller: COL 6, BUF 4. But COL was lethal off the rush, scoring three goals on 31 rush chances—a rate that matters when volume is high.

Nathan MacKinnon's wrist shot from the high slot in the second period was the game's turning point. It crystallized what the metrics already showed: COL was the team built to win this game. BUF never recovered.

BUF's offense disappeared. They mustered 28 shot attempts to COL's 37 (43% share) and generated a thin 41% expected-goals share. A. Lyon made 34 saves on 37 shots (0.919 SV%), a solid performance that masked the problem: his skaters didn't generate enough volume or quality. When a team underperforms its xG by 1.81 goals, chance creation isn't the issue—finishing is.

The Sabres' one goal came on a 34-save night for Lyon, which would win most games. COL's three came from a team that controlled shot quality, forced the rush, and had a goaltender who exceeded expectations. Wedgewood's save percentage (0.964) doesn't lie: he faced 28 shots and stopped 27, with an additional goal's worth of value baked in above the baseline.

COL won because they were better in every dimension—possession, quality, and goaltending.

Colorado 2.8 to Buffalo 2.5 represents a near-pick'em: the model projects 46% win probability for the Sabres, which is about as close to a true coin flip as you'll find in any matchup.

The Avalanche own the underlying play. At 5v5, they control 57.0% expected goals and 57.0% Corsi — metrics that translate to territorial dominance and higher-quality scoring chances. Buffalo's 50.0% xG and 49.0% Corsi show a team consistently outplayed in this regard. That's not happenstance. Colorado's recent record, 6-2-1 over their last 10, reflects an outfit firing on schedule.

Both teams are running hot offensively at an unsustainable clip. Colorado's PDO sits at 1.021; Buffalo's at 1.020. Those numbers indicate both sides are converting above their expected rates, which means regression looms. Colorado's advantage in underlying play keeps them slight favorites despite the luck factor.

The goaltending situation bears monitoring. S. Wedgewood faces a back-to-back, which may yield a backup start or a fatigued starter.

Colorado's attack centers on Artturi Lehkonen, whose 1.11 expected goals per 60 minutes and 18-goal, 22-assist line with a 14.9% shooting rate suggest he's finishing well above true talent. Ross Colton provides secondary depth at 1.10 xG/60, though his 4.9% shooting rate hints at positive variance that may not hold. Buffalo counters with Jason Zucker (1.07 xG/60, 14 goals, 15.1% shooting rate) and Josh Doan (1.01 xG/60, 15 goals, 11.5% shooting rate) — both contributors, neither dominant.

Nazem Kadri represents a defensive liability for Colorado; Jack Quinn fills a similar role for Buffalo.

Colorado's edge in shot quality and recent form outweighs the regression signals that apply to both teams. The model's marginal preference for the Avalanche is justified, even if the 46% win probability for Buffalo fairly captures how tight this matchup truly is.

Data: MoneyPuck 5v5 models, NHL API