Nashville Predators at Colorado Avalanche: score, goals, shots, period-by-period stats and the HabsMetrics prediction made before puck drop, from the NHL’s official data. Analytics for educational purposes only — not gambling advice.
How the win probability moved
COL’s chance to win
Before the game
60%
Low
41%25′
High
64%56′
COL
COL goal
NSH goal
Line: COL’s chance to win, minute by minute, as the HabsMetrics model saw it during the game.
For educational/analytics purposes only — not gambling advice
Pre-game call vs result
What the models said before puck drop, set against how the game ended.
Model miss: COL was the 60% pick but lost (NSH 3, COL 2).
Pre-game
Result
Winner
COL 60%
NSH won NSH 3, COL 2
✗Miss
Score
Expected NSH 3.0 – COL 3.7Likeliest final 2–3
NSH 3 – COL 2
Total goals
6.7 goals
5 goals
Overtime
22% chance
Went to OT/SO
By model
Win model · Elo onlyCOL 68%✗Miss
Goal model · PoissonCOL 60%✗Miss
HabsMetrics Analytics — For educational/analytics purposes only — not gambling advice
NSHaway
40%
at
COLhome
60%
Win probability20-pt gap
Model pick
COL · 60% to win
What’s driving it
3 of the top factors favour COL
← NSHCOL →
Team Offense (Goal Model)COLCOL scoring +17% vs league average, NSH +0% (time-weighted goal model).
Market Consensus BlendCOLBlended 35% toward the consensus market.
Home IceCOLLeague-wide home scoring edge of +6% applied to COL.
Team Defense (Goal Model)NSHCOL allowing +2% vs league average, NSH -3%.
Projected GoaltendingNSHProjected starters vs an average goalie, shrunk by shots faced: Mackenzie Blackwood (COL) -0.02 g/gm; Juuse Saros (NSH) +0.03 g/gm.
For educational/analytics purposes only — not gambling advice
Several models, several questions. Here is where they agree and where they do not.
The models pick the same winner (COL) but not with the same conviction: 60% to 68%.
Win model · Elo onlyWho wins?
NSH 32%COL 68%
Team strength ratings only.
Goal model · PoissonHow many goals?
Headline number
NSH 40%COL 60%
Projects each team’s goals, then derives a win chance from the score.
Expected goals NSH 3.0 – COL 3.7
MarketReference point
NSH 40%COL 60%
Bookmaker consensus with the margin removed. A benchmark, left out of the average.
They answer different questions, so their numbers can differ: the win model estimates who wins; the goal model estimates how many goals each team scores, then derives a win chance from that.
HabsMetrics Analytics — for educational purposes. Lines are inferred from shared five-skater ice time in NHL shift data; projections are estimates.
For educational/analytics purposes only — not gambling advice
Over the last 66 graded games, the win model and the goal model picked the same winner 57 times (61% correct).
HabsMetrics Analytics — For educational/analytics purposes only — not gambling advice
Projected score
Goal model
NSH
3.0
–
COL
3.6
Most likely final: COL 3–2 NSH · Actual: 2–3
Goes to OT/SO
22%
Likeliest scorelines
COL 3 – 2 NSH8%
COL 4 – 3 NSH7%
COL 2 – 3 NSH6%
COL 3 – 4 NSH6%
COL 5 – 4 NSH4%
Goal totals & margins
6+ total goals
63%
7+ total goals
53%
COL by 2+
34%
NSH by 2+
18%
Key factors (goal margin, + favors home)
Team OffenseCOL scoring +17% vs league average, NSH +0% (time-weighted goal model).+0.47
Market ConsensusBlended 35% toward the consensus market.+0.20
Home IceLeague-wide home scoring edge of +6% applied to COL.+0.18
Team DefenseCOL allowing +2% vs league average, NSH -3%.-0.14
Projected GoaltendingProjected starters vs an average goalie, shrunk by shots faced: Mackenzie Blackwood (COL) -0.02 g/gm; Juuse Saros (NSH) +0.03 g/gm.-0.05
InjuriesCOL without Josh Manson; NSH without Jonathan Marchessault.-0.02
Expected goals come from a separate goal model, so its likeliest score can lean differently from the win probability above. The model report explains how the two relate.
HabsMetrics Analytics — for educational/analytics purposes only — not gambling advice