2-1: Final (MTL). 1.46-1.85: xG (MTL-MIN). 10-8: HD (MTL-MIN). +0.85: Desbiens GSAx Game 3 of the Montréal-Minnesota semifinal turned on a 24-second window in the second period . Down 1-0 after a Sidney Morin 0.03 xG point shot at P1 3:22, Montréal answered with Maggie Flaherty at P2 4:36 (0.10 xG, primary Hayley Scamurra ) and Hayley Scamurra herself at P2 5:00 (0.07 xG, unassisted). Two goals on a combined 0.17 xG of recorded shots, both from below the model's high-danger threshold, and both behind a Minnesota defence that had just held up under 96 minutes of siege two days earlier. Minnesota outshot the Victoire 28 to 19 and outchanced them 1.85 to 1.46 , but never got past Ann-Renée Desbiens ( +0.85 GSAx ) more than once. Montréal leads the best-of-five 2-1 with Game 4 back in St. Paul on Saturday. The 24-second window P2 4:36. Maggie Flaherty , 0.10 xG, 5v5. A Scamurra entry off the half wall fed Flaherty stepping into the slot from the blue line. Wrist shot through traffic, in. The model rates the look as a median high-percentage chance for a defender (Flaherty's 14th shot from inside 35 feet this postseason); Rooney was screened by two bodies in front. The Frost don't get the puck out of the zone on the next shift, and 24 seconds later: P2 5:00. Hayley Scamurra , 0.07 xG, 5v5. A second-chance recovery off a board scramble. Scamurra found space at the high slot and put a wrister into the same low-glove corner Flaherty had hit. Two goals, 24 seconds, no line change for either team in between. The combined 0.17 xG was below the median quality of the seven non-goal shots Minnesota had taken to that point, but both went in and the road team had a lead they would not relinquish. Minnesota's chance edge, neutralized For the home team this was the inverse of Game 2. Minnesota controlled the chance battle ( 59.7% Corsi , 1.85 xG to 1.46 , 8 high-danger looks to 10 ), and lost. Kelly Pannek 's third-period wrister at 19:56 (0.18 xG) with the goalie pulled was the Frost's best chance to tie. Taylor Heise added two looks above 0.12. Grace Zumwinkle had three. None went in. Ann-Renée Desbiens faced 28 live shots worth 1.85 xG and gave up only the Morin point shot, a 0.03 xG shot from above the circles that found a seam through traffic. Across the three-game series she's now at +5.18 GSAx on 90 shots faced. That's the difference in the series so far. Eleven penalties, zero PP goals Game 3 was chippy: 11 minor penalties (7 on Montréal, 4 on Minnesota) including a Nicole Gosling pair in the third period and roughing-fest at P1 14:26 that sent both Amanda Boulier and Taylor Heise off. Neither team converted on the special-teams chances. Montréal killed all four Frost power plays at zero shots against on three of them. Minnesota's penalty kill held Montréal to 0.31 xG total across seven power plays, well below league-average PP xG production (~1.1 xG per ~7 PP minutes). The officiating swung disproportionately Montréal's way in volume but the Victoire didn't pay for it on the scoreboard. Goaltending, decomposed Ann-Renée Desbiens : 27 of 28 on 1.85 xG faced , +0.85 GSAx . The lone goal she allowed was the Morin 0.03 xG point shot through traffic, the kind of shot a goalie at the 50th percentile would stop seven times in ten. Her three-game series totals: +5.18 GSAx, .974 SV%, 87 of 90 . Through three playoff games she's the highest-process goaltender in the postseason. Maddie Rooney : 17 of 19 on 1.46 xG faced , -0.54 GSAx . Both goals against were under 0.10 xG, and the Flaherty goal was effectively unsightable through the traffic. The third-game line is a small step back from her +3.36 GSAx Game 2, but the volume was much lower (19 vs 52 shots) and the workload doesn't tell as much. She didn't lose this game; the team in front gave up two recovery goals 24 seconds apart, and recovery goals under 0.10 xG don't reflect on the goaltender. The game in cumulative xG The scoring timeline P1 3:22 · Sidney Morin · xG 0.03 · MTL 0, MIN 1. P2 4:36 · Maggie Flaherty · xG 0.17 · MTL 1, MIN 1. P2 5:00 · Hayley Scamurra · xG 0.07 · MTL 2, MIN 1 Top scoring chances MTL P2 4:36 · Maggie Flaherty wrist · xG 0.17 (goal). MTL P1 3:08 · Laura Stacey snap · xG 0.15. MTL P2 13:37 · Catherine Dubois wrist · xG 0.14. MTL P2 5:53 · Maggie Flaherty wrist · xG 0.11. MTL P3 17:40 · Hayley Scamurra wrist · xG 0.10. MTL P2 4:36 · Maggie Flaherty wrist · xG 0.10. MIN P3 19:56 · Kelly Pannek wrist · xG 0.17. MIN P3 14:59 · Grace Zumwinkle tip · xG 0.12. MIN P2 17:50 · Taylor Heise wrist · xG 0.11. MIN P3 15:19 · Kelly Pannek wrist · xG 0.10. MIN P3 13:28 · Élizabeth Giguère tip · xG 0.09. MIN P3 2:42 · Grace Zumwinkle wrist · xG 0.09 What this read isn't Not a Minnesota process problem. A 1.88 xG home playoff game with 28 shots and 8 high-danger chances is solid offence, especially against a Desbiens who's running +5 GSAx through the series. The Frost have outshot Montréal cumulatively across the three games (109 to 99) and have been tied or ahead in xG across two of them. The shot mix isn't the issue.. Not a Rooney indictment, again. Two goals on shots worth a combined 0.17 xG, both with traffic in front, is a hard line to swing on. Across the three games Rooney's GSAx is +3.20 on 99 shots faced, a top-tier process line. She's not why this series is 2-1.. Not a series-deciding swing. The road team won. Game 4 is back in St. Paul on Saturday with the Frost in a 1-2 hole. PWHL home win rates this postseason are 71 percent. Minnesota at home is the highest-rate scenario for a series tie heading to a Montréal Game 5 at Place Bell.. Not a special-teams story. Eleven penalties, zero PP goals, zero PP xG above 0.05 by either team on any single power play. The game was decided 5v5. 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.. 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/312 . The series projection, including how Game 3 moves Minnesota's and Montréal's series-WP, lives at /playoffs/bracket .