2-1: Final (OTT). 1.30-4.23: xG (OTT-BOS). 3-9: HD (OTT-BOS). +3.23: Philips GSAx Game 3 of the Boston-Ottawa semifinal was Game 2 in louder colours. Boston outshot the Charge 37 to 22 , outchanced them 4.23 xG to 1.30 , won the high-danger battle 9 to 3 , and lost in front of a Canadian Tire Centre crowd that had no business feeling comfortable until Ronja Savolainen 's point shot at P3 19:31 found space behind Aerin Frankel . Ottawa scored both goals on shots the model priced under 0.03 xG . Gwyneth Philips finished +3.23 GSAx on 37 shots faced , her second straight start above +2.0 GSAx in this series. Ottawa leads the best-of-five 2-1 with a chance to close it out at home on Sunday. The Savolainen winner P3 19:31. Ronja Savolainen , 0.03 xG, 5v5, assisted by Michela Cava and Fanuza Kadirova . A Cava recovery off the wall pushed the puck back to Savolainen at the inside edge of the blue line, then a long wrister through a crowded slot. The model priced the look at 0.03 xG , a sightless point shot from outside scoring distance. It found the short side, low. Boston pulled Frankel in the closing seconds for the extra attacker and couldn't generate a clean equalizing look before the buzzer. Ottawa won a playoff game on what the model says was one of the two least likely shots of the night to find net. Boston's chance edge, neutralized The Fleet built another championship-level chance profile and lost again. The second period alone was 2.50 xG of Boston offence against 0.30 from the Charge. Abby Newhook put two wristers on Philips in the dying seconds of the period, graded 0.42 and 0.25 xG , both turned away. Alina Müller added 2 different looks worth more than 0.20 xG across the second and third periods. Liz Schepers 's tying goal at P2 5:11 graded 0.24 xG , a wrister off a Shay Maloney entry that found space between Philips's glove and the post. Boston put 9 high-danger looks on Philips (5v5 shots from inside 50 feet) and beat her on the Schepers goal. That is the difference in the series so far. Goaltending, decomposed Gwyneth Philips : 36 of 37 on 4.23 xG faced , +3.23 GSAx . The single goal she allowed was Schepers's wrister, a 0.24 xG look the model rates as roughly a coin flip for a starter at her process tier. Three-game series totals: +5.87 GSAx, 0.958 SV%, 92 of 96 . The Fleet have generated 9.87 xG across the three games to the Charge's 3.86 ; the series is 2-1 Ottawa because Philips keeps stopping the high-danger looks one tier above what a replacement goalie would. Aerin Frankel : 20 of 22 on 1.30 xG faced , -0.70 GSAx . Both goals she allowed graded under 0.03 xG . Ottawa managed only 22 shots and 1.30 xG tonight, but two of them found net on looks the model says she stops well over nine times in ten. Three-game series GSAx: -1.14 across 60 shots faced . She has not been the problem; Boston keeps losing the variance side of the ledger. The game in cumulative xG The scoring timeline P1 13:33 · Fanuza Kadirova · xG 0.03 · OTT 1, BOS 0. P2 5:11 · Liz Schepers · xG 0.24 · OTT 1, BOS 1. P3 19:31 · Ronja Savolainen · xG 0.03 · OTT 2, BOS 1 Top scoring chances OTT P3 7:57 · Emily Clark default · xG 0.18. OTT P3 3:17 · Taylor House default · xG 0.13. OTT P2 1:00 · Gabbie Hughes default · xG 0.09. OTT P2 15:15 · Brianne Jenner default · xG 0.09. OTT P1 2:41 · Gabbie Hughes default · xG 0.08. OTT P1 15:30 · Rebecca Leslie default · xG 0.08. BOS P2 19:22 · Abby Newhook default · xG 0.42. BOS P2 10:16 · Alina Müller default · xG 0.26. BOS P2 19:21 · Abby Newhook default · xG 0.25. BOS P2 5:11 · Liz Schepers default · xG 0.24 (goal). BOS P2 19:54 · Alina Müller default · xG 0.22. BOS P2 13:02 · Ella Huber default · xG 0.17 What this read isn't Not a Boston tactics problem. A 4.23 xG playoff road game with 37 shots and 9 high-danger chances is the kind of process line that wins three out of four games against a league-average goaltender. The Fleet have run 9.87 xG across the three games and finished 4.2% on 96 shots.. Not a Frankel cratering. Two goals on shots worth a combined 0.06 xG is variance, not technique. Across the series she's at -1.14 GSAx on 60 shots, roughly +1.1 goals short of breakeven.. Not a sustainable run for Philips. +5.87 GSAx through three games is well above the regular-season top decile, and not a process line any goaltender holds across a full series. The base rate says she regresses toward her career mean in Game 4; Boston needs that to happen before they go on summer vacation.. Not a series-clinching read. Ottawa leads the best-of-five 2-1 with Game 4 at home on Sunday. PWHL home win rates this postseason are 71 percent; the simple base case is a Charge clincher. Boston has driven the chance battle in all three games, but they need either a Frankel night above their goaltender's recent median or for the model's bottom-decile finishing luck to start regressing. 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, including the series cumulatives across games 308, 310, and 313. Full play-by-play, shift overlays, and the per-player game log for this game are at /games/313 . The series projection, including how Game 3 moves Ottawa's and Boston's series-WP, lives at /playoffs/bracket .