← Game board
The Edgework model · interrogate the forecast

What If?

Because team strength is built from the dressed players, you can take it apart. Scratch a star, swap the goalie, trade a player — the win probability moves, recomputed in your browser on the same model the forecasts use. This is the question Elo can't answer.

Goalie swapsOur goalie input is ~0.07 year-over-year reliable (vs ~0.85 for skater offence), so even a big goalie change moves the forecast far less than intuition expects — by design, not a bug.
COL win prob vs avg opponent
63.0%
Change from baseline
—
Baseline 63.0% uses the dressed lineup + projected (best) starter. Edit below to see the swing.
SkaterPosTOIValue (scratch Δ)
Cale MakarD25.7-4.0 pts
Devon ToewsD24.6-2.0 pts
Nathan MacKinnonC22.8-1.9 pts
Brent BurnsD21.0-1.4 pts
Samuel GirardD20.8-1.1 pts
Artturi LehkonenL20.4-1.5 pts
Valeri NichushkinR19.7-1.4 pts
Martin NecasC19.0-0.9 pts
Brock NelsonC18.9-1.7 pts
Josh MansonD18.0-0.5 pts
Wyatt AamodtD17.1-0.9 pts
Sam MalinskiD16.0-1.5 pts
Jack AhcanD15.8-0.8 pts
Ilya SolovyovD15.7-1.2 pts
Calvin de HaanD15.0-0.7 pts
Victor OlofssonL14.5-0.7 pts
Ross ColtonC14.4-1.1 pts
Logan O'ConnorR14.1-0.8 pts

“Value (scratch Δ)” is each skater’s win-probability swing if replaced by a replacement-level player — the exact validated-model arithmetic, run live in your browser. Trade out a roster slot and trade in anyone to see a swap; combine with scratches and a goalie change.