5-1: Final (BOS). 59%: Vancouver Corsi. 2.87-1.76: xG (BOS-VAN). -2.55: Maschmeyer GSAx Vancouver carried play. The Goldeneyes finished the night with 59% of the shot attempts , more zone time than their opponent, and the kind of territorial profile that usually correlates with a positive goal differential. Their reward was a single goal. Boston out-chanced them in high-danger looks 6 to 3 despite trailing in raw shot attempts, then converted a remarkable five times on 2.87 xG of total expectation. The model graded this match-up as roughly 60-40 in Boston's favour on chance quality. The actual scoreline was 5-to-1, which is what happens when a 60-40 process meets a finishing run on one side and the worst single-game GSAx of the year on the other. Five goals, 2.99 xG Boston's five goals came on shots the model priced at 0.06, 0.29, 0.22, 0.13, and 0.14 xG . Combined: roughly 0.83 xG of total goal expectation across the five shots that scored, against actual output of five goals. The single highest-xG shot of the night for Boston ( Megan Keller 's 0.29 power-play one-timer at the end of the first period) was also the most predictable goal of the bunch. The other four were the kind of mid-quality looks that the average PWHL shooter buries 6 to 22 percent of the time. Boston's per-shot finishing rate was 1.74 times expected . Across a season that ratio sits near 1.0 by construction; in a single game it can run wild in either direction, and on this night it ran in Boston's favour. Both sides of the special-teams ledger Boston scored both a shorthanded goal and a power-play goal. Megan Keller 's shorthanded marker at P1 19:14 came down a skater on a Vancouver power play; the model graded it at 0.29 xG, the highest-quality scoring chance of the night for either team. Three minutes and 50 seconds of game-clock later, with Vancouver now on the penalty kill, Jessie Eldridge 's second of the game cashed Boston's 5-on-4 at 0.22 xG. Going PPG-and-SHG in the same game is a rare line item; over the course of a PWHL season, fewer than 10% of team-game observations feature both, and most of those are blow-out games where the score is no longer in doubt by the second goal. Here it was the first two goals of a 5-1 night. Vancouver's six power-play opportunities produced 0 goals . Boston's five generated the one. The differential on special teams (one PPG plus one SHG for Boston, zero PP goals for Vancouver) is two full goals out of the five-goal margin, accounting for nearly half the spread by itself. Goaltending, decomposed The Vancouver crease tells the story most directly. Emerance Maschmeyer faced 23 shots worth 2.45 xG and conceded 5. That is -2.55 goals saved above expectation in a single game, the kind of number that lives in the bottom 1% of single-game starts across a PWHL season. Kristen Campbell came on for the rest of the game and stopped all 6 shots she faced, which suggests the change happened after the score was no longer in doubt rather than as a tactical adjustment. On the other end, Aerin Frankel faced 29 shots worth 1.76 xG and gave up one. Her GSAx was +0.76 , which is to say she played a normal good game. The story of the night isn't her save total; it's the gap between the two goaltenders' nights against similar workloads. The game in cumulative xG The scoring timeline P1 8:05 · Jessie Eldridge · xG 0.05 · BOS 1, VAN 0. P1 19:14 · Megan Keller (SH) · xG 0.27 · BOS 2, VAN 0. P2 3:04 · Jessie Eldridge (PP) · xG 0.22 · BOS 3, VAN 0. P2 15:46 · Susanna Tapani · xG 0.14 · BOS 4, VAN 0. P2 16:36 · Jamie Lee Rattray · xG 0.15 · BOS 5, VAN 0. P3 5:44 · Sophie Jaques · xG 0.09 · BOS 5, VAN 1 Vancouver's best non-goals The chances Vancouver generated that didn't beat Frankel. Each of these is a shot the average PWHL shooter converts at least 7% of the time. The model graded none of them as outliers in either direction; this was a Vancouver attack that produced volume without the in-tight one-time look that breaks a game open. P3 6:45 · Sarah Nurse · xG 0.18. P1 18:17 · Jenn Gardiner · xG 0.15. P2 17:18 · Sophie Jaques · xG 0.14. P3 6:29 · Sarah Nurse · xG 0.13. P3 4:33 · Tereza Vanišová · xG 0.11. P2 11:56 · Hannah Miller · xG 0.09 What this read isn't Not a claim Vancouver was the better team and got robbed. They were not the better team. Boston controlled the high-danger battle, won special teams in both directions, and beat the underlying chance-quality model 60-40. Vancouver controlled territory but produced a softer overall shot profile.. Not a claim Emerance Maschmeyer played poorly in a clean way. A -2.5 GSAx night is bad on the model's accounting, but five goals on 2.5 xG of expectation is the convolution of the shooter's run AND the goalie's night. Some of that GSAx is finishing variance landing on her side of the ledger; the rest is the goalie performance the number measures.. Not a claim Boston should expect this finishing rate again. 5 goals on 2.99 xG is a 1.67x finishing rate. Single-game sample. The right diagnostic for whether this is repeatable is the season-long shooting rate against expected, and that number sits much closer to 1.0 for both teams.. Not a model failure. A 60-40 favourite winning 5-1 is well within the variance of single-game outcomes. The model's job is to predict probability over many shots, not to forecast which mid-quality look beats a goalie on which Tuesday night. 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/101 .