The Capitals had the underlying advantage and lost 3-1 anyway. Expected goals split 4.13 to 3.66 in their favor, high-danger chances nearly even at 8–7—yet the Islanders got the win. That gap traces directly to Sorokin, who finished at 0.957 save percentage and stopped 2.66 goals above expected.

The game itself stayed tactically balanced: 23 shot attempts each, 15 combined high-danger chances. Washington pressed effectively, generating five rebound opportunities without converting any—a rare miss for a net-crashing team. L. Thompson faced 22 shots and saved 2.13 goals above expected, a strong line, but the Capitals still couldn't overcome the deficit.

The Islanders' edge came through transition. Two goals on 25 rush chances—a conversion rate that neutralized Washington's underlying advantage. Mathew Barzal's snap shot from the point at 18:44 of the second set the tone the Capitals never answered.

Washington finished 2.66 goals below expected. When a team carries better underlying metrics to a three-goal loss, goaltending tilts the result. Thompson performed above his expected line. Sorokin performed farther above his. One save percentage at 0.957, one at 0.909—the difference decided an otherwise even game.

The Three Stars:

• #3 M. Roy (WSH) — 1A in 21:21
• #14 B. Horvat (NYI) — 1G in 20:34
• #6 J. Chychrun (WSH) — 1A in 20:21

The model projects a 2.5–2.8 goal spread with Washington at 54% win probability—a coin flip with slight Capitals lean.

New York has slipped to -5 goal differential over its last ten games. At 48.0% expected goals, the Islanders are underperforming expectation in shot quality; Washington's 50.0% xG% sits modestly above, though both sit at 49.0% Corsis.

Washington's real edge is managerial: the Capitals are running a 1.018 PDO, cashing chances at rates that won't hold. Regression is near. New York's 0.993 sits at league-expected levels, suggesting their recent skid reflects actual play drift rather than variance luck.

The goaltending axis tilts toward uncertainty. D. Rittich enters on the second half of a back-to-back, a scenario that often yields either a backup start or a fatigued starter—neither helps Washington's cause.

Offensively, Anders Lee drives New York's attack at 1.20 expected goals per 60 minutes, converting at a bloated 6.4% rate (9 goals, 20 assists) that obscures genuine quality—he's posted two high-danger goals. Bo Horvat's 12.8% conversion is more sustainable: 18 goals on 0.87 xG per 60 yields real upside from his three high-danger tallies.

Washington counters with Alex Ovechkin at 0.97 xG per 60, a dip from his career average. His 13.8% shooting (19 goals, 17 assists) is inflated; two high-danger goals offer some shelter, but expect regression. Anthony Beauvillier's steadier profile—0.89 xG per 60, 14 goals at a realistic 10.4%, anchored by two high-danger conversions—poses the greater long-term threat.

Defensively, Alexander Romanov and Declan Chisholm both present weak links, though neither team generates rebounds or high-danger chances at rates that fully exploit those gaps. The defensive vulnerabilities exist; the volume to punish them does not.

This is as close as it gets.