We read all 7 NWSL games ESPN's scoreboard listed last weekend: the winning goalkeeper had a higher save rate in 6 of them

In 6 of the 7 NWSL games played September 19-20 — the full set the ESPN endpoint listed as Final — the goalkeeper on the winning side recorded a higher save percentage than the goalkeeper who ended the night with a loss. Three winning goalkeepers stopped every shot they faced. Washington Spirit's keeper made 5 saves on 5 shots on target in a 3-0 road win over Chicago Stars. Bay FC's keeper made 2 on 2 in a 2-0 home win over Racing Louisville. Angel City FC's keeper made 2 on 2 in a 3-0 road win at Utah Royals. (Utah finished with 2 red cards and 25.6% possession.) Houston Dash's goalkeeper made 10 saves from 11 shots on target in a 2-1 road win at Boston; Houston itself put only 6 shots on target. Combined across all 7 games: winning goalkeepers faced 33 shots on target and saved 29 (87.9%). Losing goalkeepers faced 55 shots on target and saved 36 (65.5%). (ESPN's shots-on-target count includes shots that strike the woodwork: one of those 55 — a Bay FC shot — hit the frame, neither saved nor scored, so saves and goals do not sum to shots faced.) In Washington-Chicago the split was stark — winning keeper 100% (5 of 5), losing keeper 25% (1 of 4). The one exception: Denver Summit FC beat Seattle Reign FC 2-1 while Seattle's goalkeeper stopped 8 of 10 (80%), more efficient than Denver's 3 of 4 (75%). Denver put 10 shots on target to Seattle's 4. Conversion rates by result across the seven games: winning teams turned 18 of 55 shots on target into goals (32.7%); losing teams converted 4 of 33 (12.1%). Boston Legacy FC put 11 shots on target and scored once (9.1%). Washington Spirit scored 3 from 4 shots on target (75%); Portland Thorns scored 4 from 13 (30.8%).

The following is analysis, not fact. NWSL matches receive little of the box-score analytics coverage that the men's game generates, which means the full picture from any weekend usually goes unread. Read across all 7 games — every save, every shot on target — goalkeeper efficiency and finishing conversion each aligned with the result in 6 of the 7 games. ^mmodeled Possession also tracked the result in 6 of 7; total shots was less reliable at 4 of 7, with Boston (22 total shots, loss), NC Courage (14 total shots, loss), and Orlando Pride (19 total shots, loss) all recording more shots than the team that beat them. The direction of the save-rate gap held in 6 of 7 games. That pattern is not independent of context — teams that score first may compact defensively and absorb shots from a trailing side pressing for an equalizer ^mmodeled, and it is not possible from this data to separate goalkeeper skill from game-state management. But even under that caveat, the gap is wide: three winning keepers did not allow a single goal on the shots they faced. Two numbers frame a stark shot-volume-versus-result split. Boston put 11 shots on target and scored once (9.1%). Houston's goalkeeper, on the other side of that game, faced all 11 of those shots and stopped 10 (90.9%). The team with 11 shots on target lost; the team with 6 won. Shot volume and result pointed opposite ways here; save rate and conversion pointed with the result. ^mmodeled The Denver-Seattle exception marks the pattern's limit. Seattle's goalkeeper stopped 8 of 10 shots (80%) and still lost; Denver created 10 shots on target to Seattle's 4. In that one game the higher-volume side beat the more efficient keeper — a plausible reading is that volume outweighed efficiency there ^mmodeled. It was the only such game.

Seven games is not a sufficient sample to characterize NWSL save rates or draw conclusions about the league as a whole. The 6-of-7 pattern is consistent with a game-state explanation — teams protecting leads may allow lower-quality shots in the second half, which would systematically inflate the winning keeper's numbers and deflate the conversion rate of trailing teams. Utah's two red cards make that game's comparison particularly unreliable as a standalone data point. The Denver-Seattle result is consistent with volume overriding efficiency when the shot gap is large enough.