1 v 8most likely in 5
- Per game, at home
- 71.1%
- Per game, away
- 63.8%
Season 2027
The road to the title
Sixteen teams, four rounds, a mirrored board — the shape the league publishes. No games have been played, so every seed here is a projection and the number beside it is the probability that team actually finishes there.
Projected bracket · 2026-27
20,000 simulationsFirst-round cards show the projected seeding and the exact best-of-seven probability for that matchup. Every later card names the likeliest occupant of the slot and its probability of reaching that round — a marginal from the simulation, not a winner advanced through a fixed bracket. Those are different numbers and only one of them survives being wrong about the seeding.
Eastern Conference · first round
4 v 5most likely in 7
- Per game, at home
- 55.7%
- Per game, away
- 47.6%
3 v 6most likely in 6
- Per game, at home
- 59.1%
- Per game, away
- 51.1%
2 v 7most likely in 6
- Per game, at home
- 60.2%
- Per game, away
- 52.2%
Western Conference · first round
1 v 8most likely in 5
- Per game, at home
- 69.4%
- Per game, away
- 61.9%
4 v 5most likely in 7
- Per game, at home
- 54.4%
- Per game, away
- 46.3%
3 v 6most likely in 7
- Per game, at home
- 57.5%
- Per game, away
- 49.4%
2 v 7most likely in 5
- Per game, at home
- 67.8%
- Per game, away
- 60.2%
Per-game and series odds differ on every card because a series is not a coin weighted once: the higher seed hosts four of the seven, and the format stretches a per-game edge into a wider series one.
How far each team gets
| Team | Playoffs | Semis | Conf. finals | Finals | Title |
|---|---|---|---|---|---|
 New York Knicks | 100% | 81% | 62% | 46% | 29.1% |
 San Antonio Spurs | 100% | 80% | 59% | 38% | 22.6% |
 Oklahoma City Thunder | 100% | 77% | 54% | 31% | 17.3% |
 Detroit Pistons | 95% | 56% | 28% | 12% | 5.3% |
 Boston Celtics | 93% | 54% | 27% | 12% | 4.8% |
 Charlotte Hornets | 91% | 49% | 23% | 9% | 3.7% |
 Denver Nuggets | 93% | 52% | 22% | 9% | 3.7% |
 Cleveland Cavaliers | 86% | 41% | 18% | 7% | 2.5% |
 Los Angeles Lakers | 86% | 40% | 15% | 5% | 2.1% |
 Houston Rockets | 89% | 42% | 16% | 6% | 1.9% |
 Minnesota Timberwolves | 81% | 35% | 12% | 4% | 1.4% |
 Orlando Magic | 74% | 30% | 11% | 4% | 1.1% |
 Toronto Raptors | 69% | 26% | 10% | 3% | 1.0% |
 Philadelphia 76ers | 63% | 22% | 8% | 2% | 0.7% |
 Atlanta Hawks | 62% | 21% | 7% | 2% | 0.7% |
 Phoenix Suns | 69% | 23% | 7% | 2% | 0.6% |
 LA Clippers | 68% | 23% | 7% | 2% | 0.6% |
 Miami Heat | 57% | 18% | 6% | 2% | 0.5% |
 Portland Trail Blazers | 64% | 21% | 6% | 2% | 0.4% |
 Golden State Warriors | 24% | 5% | 1% | 0% | 0.0% |
 New Orleans Pelicans | 15% | 2% | 0% | 0% | 0.0% |
 Dallas Mavericks | 5% | 1% | 0% | 0% | 0.0% |
 Milwaukee Bucks | 3% | 0% | 0% | 0% | 0.0% |
 Chicago Bulls | 3% | 0% | 0% | 0% | 0.0% |
 Indiana Pacers | 3% | 0% | 0% | 0% | 0.0% |
 Sacramento Kings | 3% | 0% | 0% | 0% | 0.0% |
 Utah Jazz | 1% | 0% | 0% | 0% | 0.0% |
Every column is a survival probability, so each one is no larger than the one to its left. Teams below a 1% playoff probability are omitted from this table — they are still in the simulation and still in the season projection, and the two totals will not match this view by design.
Brackets that actually happened
All seasonsThe drawn bracket is the MODAL first round: the most likely occupant of each seed, with the probability it actually lands there printed beside it. Series odds are exact enumerations over a best-of-seven, not simulated estimates. Everything past round one is reported as a marginal probability from the simulation, which integrates over every seeding — reading a fixed bracket four rounds forward would compound one seeding assumption into a championship number.
Evidence
Aug 24, 2026, 1:42 PM UTC| Forecaster | Brier | Gap to close |
|---|---|---|
| Market (closing line) | 0.2070 | — |
| This model | 0.2141 | +0.0071 |
- Games scored
- 14,600
- Walk-forward n
- 25,749
- Walk-forward Brier
- 0.2106
- Calibration error
- 0.0114
historical walk-forward; not a live published record — the closing line is better (paired bootstrap 0.00712, 95% CI [0.00573, 0.00849]), which is the expected result for a model that carries no market features. Why the market is the benchmark
model 2026.08.1+943bde8b · generated Aug 29, 2026, 3:26 PM UTC