4-3: Final (VAN-MTL). 22-28: Shots (VAN-MTL). 1.86-2.08: xG (VAN-MTL). -3.07: Combined GSAx Vancouver led 4-0 at 45:06 of game time on 0.47 xG worth of scoring shots . Then Hayley Scamurra tipped home three Montréal goals in 164 seconds — P3 12:10, 13:26, 14:54 — each one with Sandra Abstreiter on the bench for an extra attacker, and the Goldeneyes closed out a 4-3 regulation win. Both goaltenders finished below expectation: Abstreiter at -2.14 GSAx on 1.86 xG faced, Campbell at -0.92 on 2.08. The scoring-shot totals were 0.47 xG for Vancouver's four goals and 0.48 xG for Scamurra's three . Seven goals came out of under one expected-goal of scoring-chance quality on each side. Chance creation went to Montréal The process numbers lean Victoire. Montréal out-shot Vancouver 28 to 22 and out-xG'd them 2.08 to 1.86 . The high-danger split was 8 to 7 . Per-shot quality was 0.074 MTL to 0.085 VAN , so the chance profile tilts slightly Montréal on average as well as in volume. Five-on-five Corsi came out 51.8% MTL against 48.2% VAN — tight, but pointing the same direction. On every volume-or-quality measure the model produces, the Victoire were the side that should have led on the shot clock. They did not, because they did not finish. Montréal generated 2.08 xG across 28 shots and put three past Campbell. Vancouver generated 1.86 xG across 22 shots and put four past Abstreiter. The scoreboard followed the shooter-quality side, not the process side. The 4-0 lead, in order Vancouver's opener came at P1 15:12: Sarah Nurse tipped home a 0.11 xG shot at 5-on-5. The second made it 2-0 on the power play at P2 9:25, Ashton Bell from the top of the circle at 5-on-4, priced at 0.16 xG. Tereza Vanišová added a 0.12 xG 3-0 at P3 1:31, and the four-goal cushion went up on a Claire Thompson slap shot at P3 5:06 that the model graded at 0.07 xG — a low-probability shot that beat Abstreiter anyway. Through forty-five minutes of game time the Goldeneyes had scored four goals on 0.47 xG of scoring-shot quality. That is the kind of conversion tail that, if rate-extrapolated, implies an 80%-plus shooting percentage; the reality is it is a single-game variance spike, not a sustainable line. 164 seconds: the Scamurra natural hat trick Abstreiter went to the bench at P3 11:49 down 4-0. Twenty-one seconds later Hayley Scamurra tipped a point shot past Campbell for 4-1 (0.23 xG — the highest-priced shot of the entire game). Abstreiter returned for the face-off, left again at 12:46, and at P3 13:26 Scamurra tipped another one home for 4-2 (0.12 xG). Same pattern at 14:24 and 14:54: Abstreiter off, Scamurra tips in her third, 4-3, 0.12 xG. Three goals in 164 seconds — all of them redirections, all of them with the extra attacker on the ice. The xG total across her three was 0.48 . On average a tip-in sequence at those prices cashes roughly one time in two; Scamurra went three-for-three. Nicole Gosling assisted on all three; Erin Ambrose on two of them. Both goalies finished below expectation Sandra Abstreiter faced 22 shots worth 1.86 xG and allowed four. Most of her workload was perimeter — Vancouver's top-three non-goals were a 0.27 xG Sarah Nurse backhand on the power play at P2 7:58 (saved), a 0.24 xG Mellissa Channell-Watkins backhand at P2 19:06 (saved), and a Hannah Miller tip at P2 19:04 (saved). Her -2.14 GSAx reflects that four of the chances she did concede were priced in the 0.07–0.16 range — shots a PWHL-average goalie turns aside roughly 85% of the time, and she stopped none of them. Kristen Campbell had a -0.92 GSAx line of her own: 28 shots worth 2.08 xG , three allowed. The three that beat her were all Scamurra tips at the net front with Abstreiter pulled, which is a chance type the model prices at ~12–23% depending on the initial-shot quality — not a goalie-talent indictment so much as the consequence of giving up three consecutive uncontested net-front redirections in under three minutes. The combined -3.07 GSAx on the night is a tail result; rolling season lines sit much closer to zero on both sides. The game in cumulative xG The scoring timeline P1 15:12 · Sarah Nurse · xG 0.12 · VAN 1, MTL 0. P2 9:25 · Ashton Bell (PP) · xG 0.15 · VAN 2, MTL 0. P3 1:31 · Tereza Vanišová · xG 0.12 · VAN 3, MTL 0. P3 5:06 · Claire Thompson · xG 0.08 · VAN 4, MTL 0. P3 12:10 · Hayley Scamurra · xG 0.24 · VAN 4, MTL 1. P3 13:26 · Hayley Scamurra · xG 0.12 · VAN 4, MTL 2. P3 14:54 · Hayley Scamurra · xG 0.12 · VAN 4, MTL 3 Special teams: the two-goal cushion, briefly Montréal drew 3 power-play opportunities and generated 5 shots for 0.38 xG , most of it on a single second-period chance above 0.12 xG. They scored none of it. Vancouver's 2 opportunities produced 4 shots for 0.54 xG and the Bell 2-0 goal at P2 9:25. Net special-teams contribution was +1 goal Vancouver on the night. In a game decided by one goal, the 1-for-2 versus 0-for-3 split is the clearest place to point for the result. The penalty timeline P1 16:01 · VAN · Sophie Jaques · Delay of Game (2 min). P2 5:04 · VAN · Sarah Nurse · Tripping (2 min). P2 7:36 · MTL · Jessica DiGirolamo · Tripping (2 min). P2 12:14 · MTL · Shiann Darkangelo · Slashing (2 min). P3 9:35 · VAN · Claire Thompson · Cross Checking (2 min) Montréal's looks above 0.08 xG Every Victoire attempt priced at 0.08 xG or higher. The three Scamurra tips show up as goals; the top non-goals are a Dara Greig attempt at 0.11 xG and two Catherine Dubois chances on the second-period power play. Campbell stopped everything outside the Abstreiter-pulled window. P3 12:10 · Hayley Scamurra (goal) · xG 0.24. P3 3:46 · Shiann Darkangelo · xG 0.15. P1 13:48 · Shiann Darkangelo · xG 0.14. P3 12:10 · Lina Ljungblom · xG 0.13. P1 16:16 · Catherine Dubois · xG 0.13. P3 14:54 · Hayley Scamurra (goal) · xG 0.12. P3 13:26 · Hayley Scamurra (goal) · xG 0.12. P1 16:31 · Lina Ljungblom · xG 0.12. P3 15:17 · Dara Greig · xG 0.11. P2 14:45 · Shiann Darkangelo · xG 0.08 Vancouver's looks above 0.08 xG P2 7:58 · Sarah Nurse · xG 0.27. P2 19:06 · Mellissa Channell-Watkins · xG 0.23. P2 9:25 · Ashton Bell (goal) · xG 0.15. P1 15:12 · Sarah Nurse (goal) · xG 0.12. P3 1:31 · Tereza Vanišová (goal) · xG 0.12. P2 19:04 · Hannah Miller · xG 0.11. P1 10:25 · Abby Boreen · xG 0.10. P1 13:28 · Jenn Gardiner · xG 0.09 What this read isn't Not a claim Vancouver outplayed Montréal. They did not on any process measure — shots, xG, HD, Corsi all leaned Victoire. The Goldeneyes won because they finished at a rate that is not sustainable game-over-game, not because they generated the better chance profile.. Not an indictment of either goalie's season. A -2.14 / -0.93 pair is a tail outcome. Abstreiter's four allowed goals came on shots priced 0.07–0.16; Campbell's three came on uncontested net-front tips. Single-game GSAx has wide variance; the right test for goalie talent is the rolling sv%-vs-expected.. Not a claim the model underrated Scamurra's tips. 0.23 / 0.12 / 0.12 are exactly the prices the model gives tip-ins from those locations — she converted three in a row at rates that are individually normal and jointly uncommon. Three hat-trick goals priced below 0.25 xG each is a finishing night, not a mispriced one.. Not a claim pulling the goalie was wrong. Down 4-0 with the clock in the third, the win-probability math says pull — the downside is small because you are already losing, and Montréal's three goals in 164 seconds almost brought it back. The tactic worked; they still needed a fourth shot to beat Campbell and did not get it. 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.. GSAx : summed xG on shots faced minus goals allowed, computed per-goalie from the PBP feed. Empty-net shots and goals are excluded on both sides; Scamurra's three goals are counted against Campbell because they went into a defended Vancouver net (not an empty one).. Corsi and high-danger splits : computed at five-on-five only. Power-play shots land in the PP xG columns, not the Corsi or high-danger columns.. 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/114 .