4-3: Final (MIN). 2.04-1.86: xG (MIN-NY). 2/6: NY power play. -1.96: Osborne GSAx This game spent 32:32 of game-clock looking like a comfortable Minnesota home win, then almost stopped being one. The Frost scored four times by the middle of the second period, all on shots the model graded between 0.08 and 0.14 xG. New York responded in the third with three goals in 4:25 of game-clock on a combined 0.42 xG , two of them on the power play. The Sirens couldn't generate a fourth in the closing minutes; Minnesota held on 4-3. Four Minnesota goals on 0.48 combined xG The Frost's four scoring shots were 0.14, 0.14, 0.12, and 0.08 xG . Combined: 0.48 xG of expectation against four actual goals. Grace Zumwinkle accounted for two of those personally; Taylor Heise and Britta Curl-Salemme added the other two. None of the four was a high-danger one-timer. Mid-quality wristers and snap shots all finding the net inside a 32-minute window is a finishing run that doesn't repeat on demand; on this night it was the Minnesota lead. The almost-comeback Anna Bargman broke the seal at P3 12:34 with a 0.22 xG slap shot on a 4-on-3 power play (the Sirens' two-skater advantage off back-to-back Minnesota minors). 76 seconds later, Paetyn Levis added a 0.17 xG snap shot at 5-on-5 to make it 4-2. Allyson Simpson converted again on a 5-on-4 power play at P3 16:59, this time on a 0.04 xG attempt. Three NY goals in 4:25 of game-clock, two of them on the power play. The pace of the comeback was the kind that sets up a tying goal in the final minute, but Minnesota's defensive shell held the line and the Sirens couldn't generate a comparable fourth chance. Goaltending, decomposed Nicole Hensley stopped 23 of 26 live shots worth 1.86 xG, finishing at -1.14 GSAx . The three goals against her were on shots in the 0.04 to 0.22 xG range; the Bargman 4-on-3 was the highest-xG look she faced all night and the only one that beat her on a true high-quality chance. Kayle Osborne faced 28 live shots worth 2.04 xG and conceded four. GSAx came out at -1.96 . All four goals against her were under 0.15 xG. The save process was much closer to expected than the headline GSAx number suggests; the team in front of her gave up a tight cluster of mid-quality chances and four of them found space. The game in cumulative xG The scoring timeline P1 10:04 · Grace Zumwinkle · xG 0.11 · MIN 1, NY 0. P2 0:22 · Taylor Heise · xG 0.11 · MIN 2, NY 0. P2 8:22 · Britta Curl-Salemme · xG 0.12 · MIN 3, NY 0. P2 12:32 · Grace Zumwinkle · xG 0.09 · MIN 4, NY 0. P3 12:34 · Anna Bargman (PP) · xG 0.22 · MIN 4, NY 1. P3 13:50 · Paetyn Levis · xG 0.16 · MIN 4, NY 2. P3 16:59 · Allyson Simpson (PP) · xG 0.04 · MIN 4, NY 3 New York's looks above 0.07 xG The Sirens' chances the model graded above the 7% scoring-probability threshold. Three of them went in (Bargman, Levis, Simpson). The remaining list is the shots Hensley turned away or that missed; nothing outsized, just a reasonable distribution of mid-quality looks. P3 12:34 · Anna Bargman (goal) · xG 0.22. P2 2:20 · Savannah Norcross · xG 0.17. P3 13:50 · Paetyn Levis (goal) · xG 0.16. P1 19:02 · Micah Zandee-Hart · xG 0.12. P1 14:14 · Anna Bargman · xG 0.10. P2 9:43 · Sarah Fillier · xG 0.09 What this read isn't Not a claim Minnesota's finishing is repeatable. Four goals on 0.48 xG is roughly eight times expected on a per-shot basis. The chance-creation profile is what to keep; the conversion rate fades on rolling samples.. Not a single-game indictment of Kayle Osborne . Four regulation goals against on shots all under 0.15 xG is mostly variance landing on her side. The save process was much closer to expected than the score suggests.. Not a claim New York's PP is on a hot streak. 2-for-6 on a small sample is noise around an average power play. Both PP goals also came in the third when Minnesota was protecting a four-goal lead; the kill structure context matters.. Not a claim Minnesota collapsed. Holding a 4-3 lead through three minutes against a team in full push mode is a different process than a structural breakdown. The defensive shell did its job; the lead held. Methodology xG model : PWHL-calibrated logistic regression on 16,709 non-EN shots. Same fit as the methodology post . Features: distance, angle, rebound (≤3 s), penalty shot, power play, overtime, scorer-tagged quality chance, seven shot-type dummies. Empty-net shots are priced outside the main model (MoneyPuck / NST convention).. GSAx : summed xG on shots faced minus goals allowed, computed per-goalie from the PBP feed. Empty-net shots and goals are excluded from both sides of the ledger.. Numbers regenerate on every build. A PBP re-sync or xG recalibration reshapes every figure in this post on the next static export. Full play-by-play, shift overlays, and the per-player game log for this game are at /games/78 .