3-2: Final (BOS win). 36-22: Shots (BOS-MIN). 2.93-1.06: xG (BOS-MIN). -0.07: Hensley GSAx Boston outshot Minnesota 36-22 , held 57.8% Corsi , and generated 2.93 xG to 1.06 . On process, this was a comfortable Fleet win. The curiosity is how close the scoreboard stayed: 3-2, with Minnesota within one goal until the 6:09 mark of the third. That gap between process and scoreboard traces directly to goaltending and low-probability conversions. The Eldridge-Muller connection Jessie Eldridge scored twice. The first came at 4:55 of the opening period at 5-on-5, assisted by Alina Muller, on a shot the model priced at 0.05 xG . That is a one-in-twenty look. The second arrived at 6:39 of the second on the power play, assisted by Alina Muller and Haley Winn, at 0.22 xG . That is a more reasonable conversion, roughly one in five. Alina Muller then scored the game-clinching goal at 6:09 of the third (0.24 xG, 5-on-5), assisted by Abby Newhook and Jessie Eldridge . The Eldridge-Muller axis produced all three Boston goals: two scored by Eldridge with Muller assisting, one scored by Muller with Eldridge assisting. The combined xG on those three shots was 0.51. Scoring three goals on 0.51 xG is the kind of finishing variance that usually regresses, but in a game where Boston generated 3.12 total xG, the margin was wide enough that it did not matter. Goaltending, decomposed Nicole Hensley faced 36 shots worth 2.93 xG and conceded three. That is -0.07 GSAx , a slight overperformance that kept the Frost alive. Without that positive margin, the expected scoreline would have been closer to 3-1 and the game would have felt less competitive than it was. Hensley was not spectacular, but she was good enough to paper over the territorial gap for most of the night. Aerin Frankel faced 22 shots worth 1.06 xG and gave up two. Her GSAx for the night was -0.94 . On a light workload, she conceded two goals on chances the model priced at 0.02 and 0.12. Both were low-probability conversions, and the negative GSAx reflects that Minnesota scored more than the model expected them to. It was an uncharacteristically soft line from Frankel, though the small shot volume makes single-game GSAx noisy. The game in cumulative xG The scoring timeline P1 4:55 · Jessie Eldridge · xG 0.05 · BOS 1, MIN 0. P2 6:39 · Jessie Eldridge (PP) · xG 0.19 · BOS 2, MIN 0. P3 4:01 · Kendall Cooper · xG 0.02 · BOS 2, MIN 1. P3 6:09 · Alina Müller · xG 0.23 · BOS 3, MIN 1. P3 18:24 · Britta Curl-Salemme · xG 0.13 · BOS 3, MIN 2 Boston's power play The Fleet went 1-for-2 on the power play, generating 1.04 xG . That is an excellent rate. Eldridge's second goal came on the man advantage, and the unit created more than half an expected goal per opportunity. When the power play generates that kind of volume, conversion follows naturally. This was a night where the PP did its job. Minnesota's best non-goals The chances the model graded as Minnesota's best looks at Aerin Frankel . Every entry here is a shot the average PWHL shooter converts at least 7% of the time. The list is short, which tells its own story about how little Minnesota generated. P3 17:18 · Mae Batherson · xG 0.15. P2 11:37 · Lee Stecklein · xG 0.14. P2 10:53 · Taylor Heise · xG 0.09. P3 15:59 · Taylor Heise · xG 0.07 What this read isn't Not a claim Minnesota was unlucky to lose. They were not. Boston held a nearly 2:1 xG advantage and deserved the win on process. The 3-2 scoreline was closer than the underlying numbers warranted.. Not a verdict on Aerin Frankel from one soft GSAx night. Two goals on 1.15 xG is a small-sample blip. Frankel's season-level numbers remain elite; one game of negative GSAx on 22 shots does not change the evaluation.. Not a claim the Frost were never in it. Both Minnesota goals came from low-xG chances that beat their probability. Cooper's 0.02 xG tally and Curl-Salemme's 0.12 xG goal were the kind of improbable conversions that keep games competitive even when the process tilts the other way.. Not a model failure. A 3-2 game where the winning team held 3.12 xG and the losing team held 1.15 xG is a clean alignment between process and outcome. The model and the scoreboard agree on the deserving winner. 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 or less), 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/107 .