High-danger chances split 8-8 in a 6-3 Islanders win—a complete tie in the most dangerous areas, yet the Islanders pulled away by three. That efficiency came from two sources: shot quality and goaltending.

The Islanders generated 0.076 expected goals per shot versus Colorado's 0.058, a quality edge that mattered more than volume. On the rush, where both teams had 26 chances, the Islanders scored three to Colorado's two. NYI converted 2 of 8 rebound opportunities; Colorado converted zero.

I. Sorokin (NYI) made 35 saves on 38 shots (0.921 SV%), finishing 1.26 goals above expected. M. Blackwood (COL) saved 36 of 41 (0.878 SV%), a gap in performance that decided the game.

The underlying expected goals split 4.94–4.26 in NYI's favor, a modest edge for a three-goal margin. The story isn't that one team outplayed the other—it's that the Islanders made their few advantages count where it mattered: shot quality, rush conversions, and goaltending.

Colorado's projection holds firm: 2.9 goals to 2.5, with a 56% win probability edge. The Avalanche have earned it on form, not luck alone.

The form is obvious. Colorado has run 9-0-1 over their last 10, stacking recent wins into an edge New York hasn't built.

The underlying play reinforces the projection. Colorado dominates at 5v5, generating 57.0% of expected goals while allowing 48.0%, and they control territory at 57.0% in Corsi—a territorial tilt that puts the Islanders on their heels. This is the core edge.

But here's the drag on confidence: Colorado's PDO sits at 1.021, running hot. The Islanders are cooler at 0.993. Regression will come for both, and when it does, the gap narrows. Factor in variance, and that 56% conviction softens.

The goaltending setup is murky. I. Sorokin faces a back-to-back, as does S. Wedgewood. Both could see backup time or tired starters—a wild card that cuts both directions.

The matchup details add texture. Colorado leans on Artturi Lehkonen at 1.11 xG/60 with 18G and 22A, shooting at 14.9%, and Ross Colton at 1.10 xG/60 with 7G and 15A at just 4.9%—a shooter waiting to heat up. New York counters with Anders Lee at 1.20 xG/60 (9G, 20A) and Bo Horvat at 0.87 xG/60 (18G, 10A, 12.8% shooting). Lee's per-60 threat is sharp; Horvat's volume offsets his efficiency gap.

The defensive vulnerabilities cut both ways. Colorado must account for Nazem Kadri's exposure; New York has Alexander Romanov as a weak link. These aren't hidden—both teams know who to target.

Take Colorado at 56%. The model's lean holds on pure play, even with PDO regression factored in. The Avalanche own the territorial battle badly enough that mean reversion in luck closes but doesn't flip the edge. The form, combined with underlying control, is too much for a team running cooler on both luck and efficiency.