Dallas 7, Edmonton 2: a scoreline that inverts the underlying play. The Oilers controlled 55% of expected goals (4.36 to 3.54), yet the Stars flipped it into a five-goal rout.

Goaltending won this one. J. Oettinger was immaculate—30 saves on 32 shots, +2.36 against expectation, and perfect on high-danger chances (10-for-10). T. Jarry was not: 20 saves on 27 shots, -3.46 against expectation, a full three goals below his workload.

Edmonton created the better looks. Ten high-danger chances to Dallas's six, six rebound opportunities (zero converted). On the rush, Dallas punished theirs (6 goals on 24 attempts) while Edmonton couldn't (2 on 23).

The variance column tells the story: Dallas overperformed their xG by 3.5 goals, Edmonton underperformed theirs by 2.4. When it mattered, the Stars executed.

Matt Duchene's backhander from the doorstep at 18:57 of the third marked the game's turning point.

The models see no daylight here. Edmonton projects to 2.5 goals, Dallas to 2.5, with the Stars holding a 51% win probability—this is as close to a toss as playoff hockey gets.

The Stars have earned it through form. They're 9-0-1 in their last 10, running at a hot points pace that makes them the hotter team entering puck drop. Yet that hot hand conceals a structural question: their PDO sits at 1.024, meaning they're converting at elite rates off average chances. Reversion is coming.

Edmonton, by contrast, is running cold. A 0.982 PDO signals they're due for positive regression—they're generating chances without finishing them yet.

Both teams drive play effectively at five-on-five. Edmonton holds the possession edge at Corsi 50.0%, while Dallas sits at 47.0%. On expected goals, the gap narrows: Edmonton 52.0%, Dallas 51.0%. These marginal differences explain why the models treat this as a coin flip. Neither team is pulling away in underlying metrics.

The Oilers' strength lies in volume. Expect a territorial tilt favoring Edmonton if both teams play their system. Dallas will defend in structure and punish on transition.

Goaltending could be the tilt. J. Oettinger backs up Friday and may start Saturday on the second of consecutive games—fatigue or a backup appearance both merit monitoring. C. Ingram faces the same back-to-back burden for Dallas. Whoever gets the nod will play a consequential game in an evenly matched series.

The roster matchups sharpen the edge. Zach Hyman (14G, 13A, 1.23 xG/60) is Edmonton's most dangerous forward; he's scoring at a 13.0% clip, a sign of shot quality and finish. Connor McDavid (25G, 33A, 0.92 xG/60) runs at 13.7% shooting—elite volume and elite efficiency, though his xG/60 lags Hyman's, suggesting Hyman is getting the very best looks.

Dallas counters with Jason Robertson (25G, 21A, 0.96 xG/60) at 12.4% shooting—comparable finishing talent, slightly fewer high-danger chances. Michael Bunting (8G, 14A) shoots at 7.6%, a more modest rate that tracks with his 0.83 xG/60; he's a playmaker first, finisher second.

The defensive weak link for Edmonton is Jake Walman. Dallas leans on Tyler Myers as the vulnerable back-end anchor. Both teams risk exposure if those defenders face sustained pressure.

This game turns on who exploits the PDO variance. Stars regression plus Oilers positive regression could flip the narrative entirely. The 51% pick favors Dallas, barely, because recent form still counts. But the models have seen enough in the underlying numbers to call this a genuine toss.