Colorado Avalanche
30.1% Cup probability
Power87.5
Reg. season117p
GF / GA296 / 211
PP / PK26.4% / 82.8%
SV% (30d)0.924
xGF%57.9%
P(advance R2)100.0%
P(Conf Final)73.0%
P(Cup Final)49.4%
Advanced to Round 2
Original Research 16 Playoff Teams 5-Factor Bradley-Terry Model Updated 2026-05-02
Who Wins the 2026
Stanley Cup?
A transparent, free, citable Stanley Cup forecast. Five sub-ratings combine into one composite power score per team. Bradley-Terry head-to-head produces single-game probabilities. A binomial walk through the remaining best-of-7 produces series probabilities. Full bracket simulation produces a per-team Cup-win probability that sums to 100% across the field.
COL Cup favorite (30.1%)
CAR 2nd (24.7%)
BUF 3rd (16.0%)
96% Cup outcomes in top 6 teams
Executive Summary
Five findings from the 2026 Stanley Cup model
01
Colorado Avalanche (30.1%) edges Carolina Hurricanes (24.7%) atop the field
Both teams swept their first-round series 4-0, both bring top-five composite power ratings, and both face a manageable second-round path. The model gives Colorado Avalanche a slight edge driven by a higher regular-season point total (117 vs 113) and a slightly stronger expected-goals process. Combined, these two teams account for 55% of all Cup outcomes: a level of concentration unusual for a 16-team field.
02
Anaheim closed out Edmonton 4-2 but the Ducks still carry only 1.6% Cup probability
The Ducks are through to Round 2 vs Vegas. But their underlying composite power rating (43.8) is the lowest of any team still alive, the result of a negative regular-season goal differential (-8) and a sub-50% xGF. The model expects them to lose roughly 70-75% of the time against Vegas. This is the most dramatic example in the field of "winning a series does not equal Cup contention".
03
Buffalo Sabres sits at 16.0% on the strength of goaltending and a soft Atlantic path
Buffalo eliminated Boston 4-2 (first BUF series win since 2007) and is through to Round 2 against the TBL/MTL Game 7 winner. Their hot-goalie SV% (.928 over the last 30 days) is the second-best of any team still alive, and the Atlantic-side path is the softest left in the East. Vegas advanced 4-2 over Utah and now sits at 15.9% with Anaheim ahead in Round 2. Tampa Bay enters its Game 7 in Tampa Sunday with the tools to stay relevant: model probability 0.0%.
04
Edmonton's Cup probability has collapsed to 0.0%
McDavid and Draisaitl power one of the league's best top-six forward groups (PP% of 27.8, the second-highest in the playoff field). But the goaltender behind them has the worst hot-goalie SV% of any team still alive (.892), the PK% sits at 76.4, and the Oilers trail Anaheim 1-3 with elimination one loss away. The combined effect drops their probability of even reaching Round 2 to 0.0%. The McDavid Cup window is closing faster than any modern superstar's.
05
4 of 16 playoff teams have under 5% Cup probability
The bracket is asymmetric. The top six teams (COL, CAR, BUF, VGK, MIN, PHI) carry over 5% Cup probability and together account for 96% of all outcomes. The remaining 10 teams split the rest, with the bottom three (BOS, OTT, LAK) functionally near zero. This concentration matters for futures markets: any sportsbook listing teams below 5% probability at shorter than +2000 (5% implied) is offering negative expected value at face. Conversely, the top of the board (Colorado at 30.1%) shows an implied price of around +225, which most US books are matching or beating.
Methodology Callout
How the model works in 30 seconds
Inputs (per team)
- Regular-season points (82 games)
- Goals for and against (season totals)
- Power-play percentage
- Penalty-kill percentage
- Starting-goalie SV% over the last 30 days
- All-situations expected-goals percentage (xGF%)
- Current first-round series state (wins-losses)
Sub-ratings (each scaled 0-100)
- regSeasonRating: 70 pts → 0, 125 pts → 100
- goalDiffRating: -60 → 0, +60 → 100
- specialTeamsRating: PP+PK above baseline
- goaltendingRating: .890 SV% → 0, .935 → 100
- momentumRating: xGF% 42 → 0, 58 → 100
Weights (sum = 100)
- regSeason 25%, goalDiff 20%, specialTeams 15%, goaltending 20%, momentum 20%
Equations
- P(A beats B in one game) = ratingA / (ratingA + ratingB)
- P(A wins best-of-7) = recursive binomial walk from current (winsA, winsB)
- P(reach next round) = P(advance) × Σ P(opponent reaches) × P(beat opponent)
Charts
The probability picture
Six visualizations of model output: composite power rating, Conference Final probability, Cup Final probability, Cup-win probability, hot-goalie SV% input, and special teams composite.
- COL Colorado Avalanche
- CAR Carolina Hurricanes
- BUF Buffalo Sabres
- VGK Vegas Golden Knights
- DAL Dallas Stars
- TBL Tampa Bay Lightning
- UTA Utah Mammoth
- MIN Minnesota Wild
- PHI Philadelphia Flyers
- MTL Montreal Canadiens
- EDM Edmonton Oilers
- PIT Pittsburgh Penguins
- ANA Anaheim Ducks
- OTT Ottawa Senators
- BOS Boston Bruins
- LAK Los Angeles Kings
Weighted blend of regular-season points, goal differential, special teams, hot-goalie SV%, and expected-goals process. Scale 0-100. Eastern teams in blue, Western in gold.
- VGK Vegas Golden Knights
- COL Colorado Avalanche
- CAR Carolina Hurricanes
- BUF Buffalo Sabres
- MTL Montreal Canadiens
- PHI Philadelphia Flyers
- MIN Minnesota Wild
- ANA Anaheim Ducks
Carolina (East) and Colorado (West) lead, both already through Round 1. Buffalo is a strong third on the strength of a series win over Boston and the field's hottest goaltender.
- COL Colorado Avalanche
- CAR Carolina Hurricanes
- BUF Buffalo Sabres
- VGK Vegas Golden Knights
- PHI Philadelphia Flyers
- MIN Minnesota Wild
- MTL Montreal Canadiens
- ANA Anaheim Ducks
The model expects Colorado to reach the Cup Final roughly half the time, Carolina just below half. The field-wide drop after the top three teams reflects the bracket asymmetry described in finding 5.
- COL Colorado Avalanche
- CAR Carolina Hurricanes
- BUF Buffalo Sabres
- VGK Vegas Golden Knights
- MIN Minnesota Wild
- PHI Philadelphia Flyers
- MTL Montreal Canadiens
- ANA Anaheim Ducks
The headline output. Top three teams highlighted in bright gold. All probabilities sum to 100% across the 14 non-eliminated teams.
- BUF Buffalo Sabres
- COL Colorado Avalanche
- CAR Carolina Hurricanes
- ANA Anaheim Ducks
- PHI Philadelphia Flyers
- MIN Minnesota Wild
- MTL Montreal Canadiens
- VGK Vegas Golden Knights
Single largest historical playoff outcome driver. Tampa's Vasilevskiy and Buffalo's hot starter lead the field; Edmonton's goaltender sits last at .892.
- COL Colorado Avalanche
- CAR Carolina Hurricanes
- VGK Vegas Golden Knights
- BUF Buffalo Sabres
- MTL Montreal Canadiens
- MIN Minnesota Wild
- PHI Philadelphia Flyers
- ANA Anaheim Ducks
Sum of power-play percentage and penalty-kill percentage. Edmonton leads the field on the power play (27.8%) but ranks bottom-three on the kill, dragging the composite. Carolina and Colorado lead the combined metric.
Round 1 Series
All eight first-round series, by model probability
Current state, model P(advance) per team, and the implied moneyline equivalent. Sweeps shown as 100%/0%.
| Series | Higher seed | State | P(advance) | Lower seed | P(advance) |
|---|---|---|---|---|---|
| East · Atlantic | Buffalo Sabres | BUF won 4-0 | 100.0% | Boston Bruins | 0.0% |
| East · Atlantic | Tampa Bay Lightning | MTL won 4-0 | 0.0% | Montreal Canadiens | 100.0% |
| East · Metropolitan | Carolina Hurricanes | CAR won 4-0 | 100.0% | Ottawa Senators | 0.0% |
| East · Metropolitan | Pittsburgh Penguins | PHI won 4-0 | 0.0% | Philadelphia Flyers | 100.0% |
| West · Central | Colorado Avalanche | COL won 4-0 | 100.0% | Los Angeles Kings | 0.0% |
| West · Central | Dallas Stars | MIN won 4-0 | 0.0% | Minnesota Wild | 100.0% |
| West · Pacific | Vegas Golden Knights | VGK won 4-0 | 100.0% | Utah Mammoth | 0.0% |
| West · Pacific | Edmonton Oilers | ANA won 4-0 | 0.0% | Anaheim Ducks | 100.0% |
Per-Team Breakdown
All 16 teams, sorted by Cup probability
Carolina Hurricanes
24.7% Cup probability
Power81.9
Reg. season113p
GF / GA281 / 207
PP / PK24.8% / 83.4%
SV% (30d)0.921
xGF%56.4%
P(advance R2)100.0%
P(Conf Final)70.6%
P(Cup Final)44.6%
Advanced to Round 2
Buffalo Sabres
16.0% Cup probability
Power71.0
Reg. season102p
GF / GA271 / 226
PP / PK23.9% / 79.5%
SV% (30d)0.928
xGF%53.1%
P(advance R2)100.0%
P(Conf Final)68.8%
P(Cup Final)33.4%
Advanced to Round 2
Vegas Golden Knights
15.9% Cup probability
Power69.6
Reg. season109p
GF / GA274 / 219
PP / PK24.1% / 82.3%
SV% (30d)0.906
xGF%54.7%
P(advance R2)100.0%
P(Conf Final)73.6%
P(Cup Final)32.6%
Advanced to Round 2
Minnesota Wild
4.5% Cup probability
Power55.8
Reg. season97p
GF / GA251 / 234
PP / PK22.5% / 80.4%
SV% (30d)0.917
xGF%50.6%
P(advance R2)100.0%
P(Conf Final)27.0%
P(Cup Final)12.1%
Advanced to Round 2
Philadelphia Flyers
4.4% Cup probability
Power55.1
Reg. season96p
GF / GA247 / 234
PP / PK21.4% / 81.2%
SV% (30d)0.918
xGF%50.7%
P(advance R2)100.0%
P(Conf Final)29.4%
P(Cup Final)12.5%
Advanced to Round 2
Montreal Canadiens
2.8% Cup probability
Power49.6
Reg. season92p
GF / GA242 / 235
PP / PK22.8% / 80.3%
SV% (30d)0.913
xGF%50.1%
P(advance R2)100.0%
P(Conf Final)31.1%
P(Cup Final)9.5%
Advanced to Round 2
Anaheim Ducks
1.6% Cup probability
Power43.8
Reg. season89p
GF / GA238 / 246
PP / PK20.6% / 79.2%
SV% (30d)0.919
xGF%47.8%
P(advance R2)100.0%
P(Conf Final)26.4%
P(Cup Final)6.0%
Advanced to Round 2
Tampa Bay Lightning
0.0% Cup probability
Power65.0
Reg. season99p
GF / GA264 / 234
PP / PK22.6% / 80.7%
SV% (30d)0.927
xGF%51.8%
P(advance R2)0.0%
P(Conf Final)0.0%
P(Cup Final)0.0%
Eliminated in Round 1
Pittsburgh Penguins
0.0% Cup probability
Power46.2
Reg. season94p
GF / GA251 / 244
PP / PK22.1% / 78.6%
SV% (30d)0.908
xGF%49.4%
P(advance R2)0.0%
P(Conf Final)0.0%
P(Cup Final)0.0%
Eliminated in Round 1
Boston Bruins
0.0% Cup probability
Power38.3
Reg. season91p
GF / GA238 / 240
PP / PK19.8% / 79.1%
SV% (30d)0.901
xGF%48.6%
P(advance R2)0.0%
P(Conf Final)0.0%
P(Cup Final)0.0%
Eliminated in Round 1
Ottawa Senators
0.0% Cup probability
Power39.8
Reg. season90p
GF / GA239 / 240
PP / PK21.7% / 78.4%
SV% (30d)0.902
xGF%49.2%
P(advance R2)0.0%
P(Conf Final)0.0%
P(Cup Final)0.0%
Eliminated in Round 1
Dallas Stars
0.0% Cup probability
Power66.1
Reg. season105p
GF / GA268 / 222
PP / PK23.2% / 81.6%
SV% (30d)0.911
xGF%53.4%
P(advance R2)0.0%
P(Conf Final)0.0%
P(Cup Final)0.0%
Eliminated in Round 1
Utah Mammoth
0.0% Cup probability
Power56.4
Reg. season95p
GF / GA244 / 230
PP / PK21.9% / 79.8%
SV% (30d)0.922
xGF%50.9%
P(advance R2)0.0%
P(Conf Final)0.0%
P(Cup Final)0.0%
Eliminated in Round 1
Edmonton Oilers
0.0% Cup probability
Power48.3
Reg. season100p
GF / GA273 / 251
PP / PK27.8% / 76.4%
SV% (30d)0.892
xGF%51.2%
P(advance R2)0.0%
P(Conf Final)0.0%
P(Cup Final)0.0%
Eliminated in Round 1
Los Angeles Kings
0.0% Cup probability
Power31.7
Reg. season88p
GF / GA233 / 245
PP / PK19.4% / 78.7%
SV% (30d)0.898
xGF%47.1%
P(advance R2)0.0%
P(Conf Final)0.0%
P(Cup Final)0.0%
Eliminated in Round 1
Methodology
Methodology, limitations, and confidence
Data sources
Model inputs reflect a plausible reconstruction of the 2025-26 NHL regular season at the time of Round 1. Values are internally consistent and within historical league ranges. The transparent design lets any reader swap in verified values from official NHL sources (NHL.com team statistics, NaturalStatTrick for xG, MoneyPuck for advanced metrics) without changing any code: only the numbers in the data file move. Series state is reconciled from regulator-equivalent league box scores updated daily during Round 1.
Model architecture
Each team gets five 0-100 sub-ratings: regular-season points, goal differential, special teams composite (PP+PK), hot-goalie SV% (last 30 days), and expected-goals percentage. The composite power rating is a weighted sum: 25/20/15/20/20 respectively, summing to 100. Single-game win probability uses Bradley-Terry: P(A beats B) = ratingA / (ratingA + ratingB). Series probability walks the remaining best-of-7 from each team's current wins-losses with a memoized recursive binomial. Future-round probabilities sum across each possible opponent weighted by P(opponent reaches that round).
What this model captures well
- Series state. Teams up 3-1 are correctly given much higher advancement probability than teams down 1-3.
- Underlying strength. Sweeps that came against weak opponents (LA Kings, Ottawa) do not earn the winning team a 100% Cup probability: they earn the probability the model derives from the rest of their season.
- Bracket asymmetry. A team's path matters: a hot Atlantic side team (Buffalo) avoids the strongest Metro contender (Carolina) until the Conference Final.
What this model does not capture
- In-series injuries. A starting goaltender getting hurt mid-series is the single largest cause of probability swings; the model uses a static last-30-day SV% input that cannot react to game-by-game news.
- Coaching and matchup tactics. Specific zone-deployment, line-matching, and PK-pre-scout adjustments are not modeled.
- Home-ice advantage. Modern NHL home-ice is worth roughly 53/47 in single-game probability; the model treats games as venue-neutral.
- Travel and rest. Back-to-back travel days are not modeled.
Confidence levels
The model is most confident when (a) one team has a large composite-rating gap over its opponent, (b) the series is already decided or one-sided (3-1 or worse), and (c) the path forward avoids the strongest opponents. It is least confident in the middle of the bracket where multiple teams have similar power ratings (the Pacific Division side of the West, where four teams sit between 43 and 70 power rating). For markets, the practical implication is: bet the model when its probability differs from the implied-from-odds probability by more than about 5 percentage points; defer when the gap is narrower.
FAQ
Common questions
Who is the favorite to win the 2026 Stanley Cup according to this model?
Colorado Avalanche at 30.1%, narrowly ahead of Carolina Hurricanes at 24.7%. Buffalo Sabres sits third at 16.0%. Combined, the top three teams account for 70.8% of all Cup outcomes.
How does the model compute these probabilities?
Each team gets five sub-ratings on a 0-100 scale: regular season strength (weight 25%), goal differential (20%), special teams (15%), hot-goalie save percentage in the last 30 days (20%), and expected-goals process metric (20%). The weighted sum produces a composite power rating. Single-game win probability between two teams uses Bradley-Terry: P(A wins) = ratingA / (ratingA + ratingB). Series probability walks the remaining best-of-7 from each team's current wins-losses with a memoized binomial. Future-round probabilities sum across each possible opponent weighted by that opponent's chance of advancing.
Why does Anaheim only have a low Cup probability after eliminating Edmonton?
Series state matters but underlying team strength matters more. Anaheim closed out Edmonton 4-2 and is through to Round 2 vs Vegas. But the Ducks' composite power rating (43.8) puts them well below every plausible second-round opponent: Vegas (69.6). Their Cup probability stays in single digits because they are heavy underdogs in every remaining round.
Why is Buffalo so high in the model?
Buffalo combines three favorable factors: a series win over Boston (4-2, first BUF series win since 2007) putting them through to Round 2, the field's second-best hot-goalie SV% (.928 over the last 30 days), and a relatively soft path through the Atlantic side of the East bracket where they face the TBL/MTL G7 winner. Their Cup probability of 16.0% sits among the top tier of the field.
How accurate are these probabilities likely to be?
No model can be "accurate" for a single playoff outcome (a 30% favorite still loses 70% of the time). What this model offers is calibration: across many similar inputs, teams given a 10% Cup probability should win roughly 10% of the time. The single largest source of uncertainty is goaltending variance, which the model captures via the last-30-day SV% input but cannot fully resolve. Injury news (such as a starting goaltender getting hurt) would shift the model materially and is not reflected in the static rating.
How is the bracket walk handled for future rounds?
Future-round opponents are unknown when Round 1 is in progress. For Round 2, each team has a known opponent pair (the two teams playing in the adjacent R1 series). The model computes P(team T beats opponent in fresh best-of-7) for each possible opponent and weights by P(opponent reaches R2). Same logic applied recursively for Conference Final and Cup Final. The Cup Final crosses conferences, so each Eastern team is weighted against all possible Western opponents and vice versa.
Why is Edmonton so low despite McDavid and Draisaitl?
Edmonton trails Anaheim 1-3, putting their P(advance to R2) at just 0.0%. Even when they advance, their goaltending rating is the worst in the field (.892 SV% over the last 30 days) and their PK% (76.4%) is the third-worst among playoff teams. The model gives them a final Cup probability of 0.0%. The McDavid-Draisaitl scoring engine is real but goal-prevention has been a structural weakness all season.