Dallas's 4-1 victory masks a closer underlying picture. Expected goals split 4.35–3.66 in the Stars' favor, a 54–46 margin that barely reflects the scoreboard's lopsided finish.

The gap came down to goaltending. C. DeSmith stopped 18 of 19 shots and finished +2.66 above expected — he saved two-and-a-half goals the metrics say he shouldn't have. J. Daccord faced tougher volume (32 shots on goal) but couldn't overcome that deficit. Seattle underperformed its xG by 2.7 goals, a sign that quality chances don't convert without finishing precision.

The Stars held 63% of shot attempts (32–19) but struggled to generate Grade-A looks. Both teams finished 0–8 on high-danger chances — a rarity that usually evens out over time. Dallas created the edge in secondary scoring: rebound opportunities gave them the advantage. T. Harley picked up two assists in 22:38. W. Johnston scored twice in 19:50. M. Duchene added a goal and an assist in 17:57.

For Seattle, the story was wasted opportunity. Eight high-danger attempts should have yielded more than one goal, but precision abandoned them when it mattered. That's not randomness — it's the kind of game a playoff team replays when margins narrow.

DeSmith's performance was the fulcrum. He stopped every high-danger look and made the difficult routine. Against a team that generated as many dangerous chances as Seattle did, that's the difference between a competitive game and a rout. The Stars' goaltender didn't dominate in volume; he dominated in the one stat that counts.

The model favors Dallas at 56% win probability, projecting a 2.2–2.5 goal game in the Stars' direction. The edge is modest but rooted in where it matters: territorial control.

Dallas owns the 5v5 expected goals battle, 51.0% to 46.0%. This isn't just noise—the Stars generate cleaner looks and suppress Seattle's chances at an alarming rate. Dallas also edges Corsi at 47.0% to 45.0%, a smaller margin that trails the xG advantage but confirms possession isn't the anomaly here. Seattle is playing from behind more often, and the model sees that compounding.

But there's a caveat baked into Dallas's ask: The Stars are running a 1.024 PDO—three points above league average in puck luck. Seattle sits at 1.011. That gap is the difference between "Dallas is better" and "Dallas is luckier." Regression is a blunt tool, but asking Dallas to sustain 1.024 is asking a lot. A tick downward tightens the margin.

The personnel matchup sharpens the forecast. Jason Robertson anchors Dallas's attack at 0.96 xG/60 and carries a 25G–21A line. He plays the kind of volume game—12.4% shooting—that benefits when your team owns the volume. Michael Bunting complements at 0.83 xG/60 (8G, 14A, 7.6% shooting), a secondary weapon that inflates Dallas's offensive floor. Seattle counters with Shane Wright (0.77 xG/60, 8G, 10A, 10.4% shooting) and Berkly Catton (0.68 xG/60, 6G, 8A, 8.8% shooting). Both are shooting efficiently, but volume is the hidden tax: Wright and Catton generate fewer looks, which means fewer chances to exceed expectation.

There's an eerie symmetry in the tactical picture. Both teams sit at 0% high-danger goal rate and 0% rebound goal rate—the two most unsustainable outcomes in hockey. One or both will regress hard. Dallas's 1.024 PDO predicts they'll survive longer. Seattle's 1.011 suggests they won't.

The defensive vulnerabilities are specific. Chandler Stephenson is Seattle's leak; Tyler Myers fills that role for Dallas. Both will face pressure in transition if their team's underlying possession crumbles. That's where PDO regression hits hardest—not in a single unlucky bounce, but in a cascade when volume shrinks.

The Stars are the tighter case: better at 5v5, safer on the regression curve, and positioned to punish bad luck if it comes. Seattle has the talent to overcome variance, but they're betting against the model's read of the underlying game.