Wins by season
Career NHL wins, ending with the projection at a 60-start workload (violet).
Pool value
Projected full-season fantasy worth.
why 84?
Every fantasy goalie category, scored on the season model's projections at a 60-start workload:
- proj wins32 × 4.5+144
- proj shutouts2.9 × 3+8.7
- SV% 0.895 vs .900-0.005 × 250-1.3
- GAA 3.03 vs 2.70-0.33 × 50-16.5
- total135 pool points · an elite goalie season scores 160135 = 84/100
Weights: a win is worth 4.5 (the heaviest goalie category in every format), a shutout 3, SV% pays 250 per full point above .900 and GAA pays 50 per hundredth below 2.70. 100 = a Vezina-tier season (~26 W, 6 SO, .918, 2.45). Projections shrink SV% toward career and league averages because save percentage repeats poorly year to year; wins lean on team quality for the same reason.
Season stats — goaltending
NHL regular-season goaltending numbers from the league's goalie feed — wins, GAA, save% and shutouts, season by season. Skater columns (G/A/P) would be all zeros for a goalie, so they're gone. The 60-start proj line is a model, not a pace: save% is projected with empirical-Bayes shrinkage toward career and league (save% is weakly repeatable — Schuckers 2016; a James–Stein study on NHL save% found shrinkage beats the raw estimate 2.2× in error), GAA is rebuilt from its SA/60 × (1−SV%) identity, shutouts are the Poisson zero probability at the projected goals-against rate (Buttrey & Washburn, JQAS 2011), and wins are regressed toward the league per-start rate because wins are mostly a team stat (Berri & Brook 2010). Backtested on 330 goalie seasons from our own data: 26% lower save% error and 16% lower GAA error than naive last-year pace. Hockey's honest truth: save% barely repeats, so this model shrinks hard on purpose.
| Season | Team | GP | GS | W | L | OT | GAA | SV% | SO |
|---|---|---|---|---|---|---|---|---|---|
| 17-18 | WSH | 37 | 30 | 15 | 11 | 3 | 2.47 | 91.9% | 3 |
| 18-19 | COL | 49 | 45 | 25 | 14 | 5 | 2.55 | 91.9% | 4 |
| 19-20 | COL | 43 | 43 | 23 | 12 | 5 | 2.51 | 91.6% | 3 |
| 20-21 | COL | 50 | 49 | 36 | 13 | 1 | 2.08 | 92.0% | 7 |
| 21-22 | SEA | 55 | 54 | 18 | 31 | 5 | 3.16 | 88.9% | 2 |
| 22-23 | SEA | 53 | 50 | 24 | 21 | 4 | 2.89 | 89.7% | 0 |
| 23-24 | SEA | 36 | 34 | 14 | 16 | 2 | 2.85 | 89.9% | 2 |
| 24-25 | SEA | 26 | 26 | 8 | 17 | 1 | 3.49 | 87.5% | 0 |
| 25-26 | SEA | 32 | 28 | 13 | 12 | 4 | 2.65 | 90.9% | 0 |
| 26-27 | SEA | 10 | 8 | 5 | 3 | 1 | 3.08 | 89.0% | 0 |
| 60-start proj | SEA | 60 | 60 | 32 | 24 | 6 | 3.03 | 89.5% | 2.9 |
Data via the public NHL API · 2026 Pro Hockey Stats