4-3: Final OT (BOS). 2.03-2.38: xG (BOS-MTL). 6-9: HD (BOS-MTL). 2/3: Montreal PP Montreal scored two power-play goals in the first period and added a third 5-on-5 marker with two seconds left in the second to take a 3-0 lead into the third . Boston had 45 minutes to find one goal worth of space against Ann-Renée Desbiens and produced none of it. Then in the third period, the Fleet scored three goals in 13:44 of game-clock , the last one with 15 seconds left in regulation , and Ella Huber won it 1:33 into overtime. Two Montreal power-play goals built the early lead Marie-Philip Poulin opened the scoring at P1 14:50 with a 0.05 xG tip on the 5-on-4. Kati Tabin added a 0.09 xG slap shot on the next Montreal power play 2:10 later at P1 17:00. Two PP goals on combined 0.14 xG of expectation, both in the first period. The Victoire's power play converted at 2 for 3 on the night, with Abby Roque adding a 0.13 xG 5-on-5 marker at P2 16:58 to take it to 3-0. Combined Montreal regulation xG on scoring shots: 0.26. Three goals on a quarter of expectation. Three Boston goals to tie it Loren Gabel broke the seal at P3 6:01 with the highest-xG attempt of either team's night, a 0.26 xG wrister. Megan Keller cut the lead to one at P3 17:26 on a 0.03 xG slap shot , the lowest-quality scoring shot of the night by a wide margin. With 15 seconds left in regulation, Susanna Tapani tied it at P3 19:45 on a 0.12 xG snap shot. Three goals on combined 0.40 xG across 13:44 of game-clock. The Keller goal in particular is the kind of low-percentage shot that tends to find space when a team is pressing hard and a goalie is starting to lose her structure. The OT winner Ella Huber took a Megan Keller feed and beat Desbiens 1:33 into overtime on a 0.07 xG wrister. The model rates that shot at roughly one in fourteen against an average goalie. In OT it didn't have to be a high-percentage chance to be the right one; the first shot to find space won the game. Goaltending, decomposed Ann-Renée Desbiens faced 26 live shots worth 2.03 xG and conceded four. GSAx came out at -1.97 . Two of the four goals against her were under 0.10 xG (Keller's 0.03 in the third and Huber's 0.07 in OT); two were in the 0.12 to 0.26 xG range. The single biggest swing was Gabel's 0.26 chance, the only Boston shot above 0.20 xG that found the net. The compressed timing of the third-period goals is the kind of late-game pressure pattern that produces this shape of below-expected save line; the underlying save process wasn't as far off as the headline GSAx number reads. Aerin Frankel stopped 27 of 30 live shots worth 2.38 xG, finishing at -0.62 GSAx . The two PP goals against her in the first were both under 0.10 xG; she gave up the Roque even-strength marker at 0.13. Above-expected by the model, even in a near-loss. The game in cumulative xG The scoring timeline P1 14:50 · Marie-Philip Poulin (PP) · xG 0.05 · BOS 0, MTL 1. P1 17:00 · Kati Tabin (PP) · xG 0.08 · BOS 0, MTL 2. P2 16:58 · Abby Roque · xG 0.11 · BOS 0, MTL 3. P3 6:01 · Loren Gabel · xG 0.25 · BOS 1, MTL 3. P3 17:26 · Megan Keller · xG 0.03 · BOS 2, MTL 3. P3 19:45 · Susanna Tapani · xG 0.12 · BOS 3, MTL 3. P4 1:33 · Ella Huber (OT) · xG 0.09 · BOS 4, MTL 3 Boston's looks above 0.07 xG The Fleet's chances the model graded above the 7% scoring-probability threshold. Four of them went in (Gabel, Tapani, Huber in OT, plus the marginal Keller if you include it). The shape of the distribution is consistent with a road team that took its chances when they came rather than dominating the chance creation. P3 6:01 · Loren Gabel (goal) · xG 0.25. P2 7:10 · Laura Kluge · xG 0.21. P3 10:57 · Jamie Lee Rattray · xG 0.18. P3 10:45 · Susanna Tapani · xG 0.15. P3 12:26 · Theresa Schafzahl · xG 0.15. P3 4:10 · Jill Saulnier · xG 0.14 What this read isn't Not a claim Boston dominated chance creation. The Fleet finished with lower xG and fewer high-danger chances than Montreal. The comeback was finishing variance compressed into 14 minutes, not a structural takeover.. Not a single-game indictment of Ann-Renée Desbiens . Two of the four goals against her were under 0.10 xG; the model's GSAx number reads worse than the underlying save process suggests. Late-game pressure with the team protecting a lead produces this shape of save line.. Not a claim Montreal collapsed structurally. Holding a 3-0 lead through 45 minutes against a top opponent and giving it up in the final 14 happens. The defensive shell didn't break in any structural way; mid-quality chances found space and Boston's scoring run lined up with Montreal's least-effective stretch of the night.. Not a claim either special-teams unit's profile is repeatable. 2-for-3 on PP for Montreal and 0-for-3 for Boston are small samples; the conversion rates fade on rolling samples and don't fold into expectations for the next meeting. 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. Overtime shots are included.. 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/76 .