4-2: Final (BOS). 2.12-1.86: xG (BOS-MIN). 0.35: BOS 4G total xG. -1.88: Hensley GSAx Boston scored four goals on shots the model priced at 0.10, 0.13, 0.11, and 0.03 xG . Combined: 0.35 xG of total expectation, four actual goals . The highest-xG attempt of the Fleet's four scoring shots was a 0.13 xG look. None of these were the kind of in-tight high-danger one-timers the model considers near-coin-flip to score; they were mid-quality wristers and snap shots that all happened to find space on the same Saturday night. Per-shot finishing rate worked out to 1.89 times expected . Across a season that ratio sits near 1.0; in a single game it can run wild in either direction, and on this night it ran in Boston's favour. Minnesota's two-minute tying run Down 2-0 to start the second period, Minnesota came back with a quick pair of goals. Lee Stecklein opened the period at P2 0:23 with a 0.11 xG snap shot on the 5-on-4 power play, then Taylor Heise tied it at P2 9:19 with a 0.13 xG backhand. Two scoring shots in the same range as Boston's, both finding space. The model graded both as normal mid-quality conversions, and for nine minutes of game-clock the Frost looked like a team that was about to take the game over. Boston's response, 2:09 later The 2-2 tie lasted 2 minutes and 9 seconds of game-clock. Liz Schepers retook the lead at P2 11:28 with another 0.11 xG wrister at 5-on-5. The pattern from the first period reasserted itself: Boston finishing on shots that don't normally find the net. Sophie Shirley added the insurance early in the third on a 0.03 xG wrister, the lowest-quality scoring shot of either team's night. Three Boston goals in 27 minutes of game-clock, none of them on chances the model thought were particularly dangerous. Goaltending, decomposed Nicole Hensley faced 30 live shots worth 2.12 xG and conceded four. That is -1.88 GSAx , which is bad on the model's accounting, but the bulk of that number is the gap between the model's expectation on those four shots and the goals they became. Three of the four were under 0.13 xG, one was under 0.04. A goalie isn't responsible for whether a 0.03 xG shot finds the net any more than she is for whether a 0.97 xG empty-net attempt scores. The save process underneath the GSAx number was much closer to expected than the headline suggests. Amanda Thiele on the other end stopped 21 of 23 live shots worth 1.86 xG, finishing at -0.14 GSAx . The goalie duel came out essentially even on process; the difference in result is at the other end of the rink, with shots that beat their own probability. The game in cumulative xG The scoring timeline P1 9:11 · Haley Winn · xG 0.10 · BOS 1, MIN 0. P1 16:45 · Sophie Shirley · xG 0.13 · BOS 2, MIN 0. P2 0:23 · Lee Stecklein (PP) · xG 0.10 · BOS 2, MIN 1. P2 9:19 · Taylor Heise · xG 0.12 · BOS 2, MIN 2. P2 11:28 · Liz Schepers · xG 0.10 · BOS 3, MIN 2. P3 5:16 · Sophie Shirley · xG 0.02 · BOS 4, MIN 2 Minnesota's looks above 0.08 xG Every Frost attempt the model graded above the 8% scoring-probability threshold. Two of them went in (the Stecklein PP and the Heise tying goal); the other entries are the chances that didn't beat Thiele on a night where Boston's mid-quality looks all did at the other end. P3 7:43 · Claire Butorac · xG 0.28. P2 15:37 · Grace Zumwinkle · xG 0.20. P3 5:51 · Kelly Pannek · xG 0.19. P2 4:09 · Britta Curl-Salemme · xG 0.15. P2 9:19 · Taylor Heise (goal) · xG 0.12. P2 0:23 · Lee Stecklein (goal) · xG 0.10 What this read isn't Not a claim Minnesota deserved a different result. Boston scored four goals; the Frost scored two. The point of the analytical read is that the underlying process was much closer than the score, which has implications for the next time these teams meet, not for what happened tonight.. Not a claim Nicole Hensley played poorly in a structural way. A -1.8 GSAx night where every goal against was under 0.13 xG is mostly variance landing on her side of the ledger. Single-game GSAx has wide variance on small samples, especially when low-percentage shots find space.. Not a claim Boston should expect this finishing rate again. 4 goals on 0.36 xG is roughly 11 times expected. Single-game finishing variance is enormous; the season-long shooting profile against expected is closer to 1.0 by construction.. Not a model failure. A near-even xG game ending 4-2 is well within the variance of single-game outcomes, especially when one team's finishing run lands in a single 27-minute stretch. 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/92 .