NHL Season Projections
2026-27: the rest of the regular season, played out 10,000 times by our game model. Updated 2026-10-02.
The Avalanche sit on top at 107 projected points, and they take the Presidents' Trophy in 21% of our runs. The East's coin flip is the Blue Jackets: 49% to make it, on 94 points. The West's coin flip is the Kings: 49% to make it, on 92 points. Three clubs get in at least nine times in ten. The Blackhawks bring up the rear at 74. It's early — 16 of 1,344 games are in the books — so this is mostly the ratings talking, not the standings.
Eastern Conference
Atlantic
| Team | Proj | Playoffs | Range | Div | Pres. | Proj rec | GP | PTS |
|---|---|---|---|---|---|---|---|---|
 Senators | 100.5 | 80% | 90–111 | 33% | 5% | 46-29-9 | 0 | 0 |
 Panthers | 96.9 | 65% | 86–108 | 18% | 2% | 44-31-10 | 2 | 3 |
 Lightning | 96.4 | 64% | 85–107 | 17% | 2% | 44-31-9 | 1 | 0 |
 Sabres | 95.5 | 59% | 84–106 | 14% | 1% | 43-32-9 | 1 | 0 |
 Bruins | 91.4 | 40% | 80–102 | 7% | 1% | 41-34-9 | 1 | 2 |
 Red Wings | 91.2 | 40% | 80–102 | 6% | <1% | 41-34-9 | 0 | 0 |
 Canadiens | 90.4 | 35% | 80–101 | 5% | <1% | 41-34-9 | 1 | 2 |
 Maple Leafs | 83.9 | 13% | 73–95 | 1% | <1% | 38-38-9 | 2 | 2 |
Metropolitan
| Team | Proj | Playoffs | Range | Div | Pres. | Proj rec | GP | PTS |
|---|---|---|---|---|---|---|---|---|
 Penguins | 107.3 | 95% | 97–118 | 53% | 20% | 50-26-8 | 1 | 2 |
 Hurricanes | 101.7 | 83% | 91–113 | 23% | 7% | 46-28-10 | 1 | 1 |
 Devils | 97.1 | 66% | 86–108 | 10% | 2% | 44-31-9 | 1 | 2 |
 Blue Jackets | 93.6 | 49% | 83–104 | 5% | 1% | 42-33-9 | 1 | 2 |
 Capitals | 92.3 | 43% | 81–103 | 4% | 1% | 42-33-9 | 0 | 0 |
 Rangers | 91.4 | 39% | 81–102 | 3% | <1% | 41-34-9 | 2 | 2 |
 Islanders | 86.3 | 19% | 76–97 | 1% | <1% | 39-36-9 | 1 | 0 |
 Flyers | 83.6 | 11% | 73–94 | <1% | <1% | 37-37-10 | 2 | 1 |
Western Conference
Central
| Team | Proj | Playoffs | Range | Div | Pres. | Proj rec | GP | PTS |
|---|---|---|---|---|---|---|---|---|
 Avalanche | 107.4 | 96% | 97–118 | 49% | 21% | 50-26-8 | 1 | 2 |
 Mammoth | 100.8 | 83% | 90–112 | 17% | 6% | 46-29-9 | 1 | 2 |
 Wild | 99.7 | 80% | 89–110 | 14% | 4% | 45-30-9 | 1 | 2 |
 Stars | 99.0 | 77% | 88–110 | 12% | 4% | 45-30-9 | 0 | 0 |
 Blues | 94.5 | 58% | 84–106 | 5% | 1% | 43-32-9 | 0 | 0 |
 Jets | 89.5 | 33% | 79–100 | 1% | <1% | 40-35-9 | 0 | 0 |
 Predators | 86.4 | 22% | 75–97 | 1% | <1% | 39-36-9 | 1 | 0 |
 Blackhawks | 73.8 | 1% | 63–85 | <1% | <1% | 33-43-9 | 2 | 0 |
Pacific
| Team | Proj | Playoffs | Range | Div | Pres. | Proj rec | GP | PTS |
|---|---|---|---|---|---|---|---|---|
 Golden Knights | 106.3 | 96% | 96–117 | 64% | 17% | 49-27-9 | 1 | 2 |
 Oilers | 97.5 | 76% | 87–108 | 18% | 2% | 44-30-9 | 2 | 3 |
 Kraken | 92.1 | 52% | 81–103 | 7% | 1% | 42-33-9 | 1 | 2 |
 Kings | 91.8 | 49% | 81–103 | 6% | <1% | 41-34-9 | 1 | 0 |
 Ducks | 89.2 | 37% | 78–100 | 4% | <1% | 40-35-9 | 0 | 0 |
 Sharks | 85.7 | 23% | 75–97 | 1% | <1% | 38-37-9 | 1 | 2 |
 Flames | 80.8 | 10% | 70–92 | 1% | <1% | 36-39-9 | 1 | 0 |
 Canucks | 80.0 | 8% | 69–91 | <1% | <1% | 36-40-9 | 2 | 2 |
How we build it
It's the same model that prices every game on the matchup board, run forward. Each club gets an even-strength rating built mostly on its expected goals for and against, adjusted for who it played, plus a power-play and penalty-kill rating and how often it takes and draws penalties. For every game left on the schedule, the model turns the two ratings into a full score distribution: who wins in regulation, how often it's tied after 60, and who takes it in overtime or the shootout. Rest counts too — a team on the second night of a back-to-back plays a little worse, and the model knows the schedule.
Then we play the season. Every remaining game is settled one at a time: a regulation win is two points, an overtime or shootout loss is one, a regulation loss is nothing. Do that 10,000 times and each club ends up with 10,000 final point totals. Proj is the average. Range runs from the 10th to the 90th percentile, so eight seasons in ten land inside it. The record column is the average wins, regulation losses and overtime losses, rounded, over a 84-game schedule.
How the playoff race is settled
Sixteen clubs get in: the top three in each division, then the next two best records in each conference as wild cards. A team can finish fourth in a stacked division and still get in on a wild card, which is why the cut line under the third row isn't the end of the story. Ties break on points, then regulation wins, then regulation-plus-overtime wins, then total wins. The NHL goes on to head-to-head points and goal differential; the simulation doesn't track goals, so anything still level after total wins is a coin flip. That only decides a sliver of seasons, but it's an approximation and we'd rather say so. Div is the share of seasons a club finishes first in its division. Pres. is the share it finishes with the best record in the league.
What it can't see
This is the game model run forward, not a separately tested season model. We grade the game model against real results every night; nobody has graded a season simulation built on top of it yet. The ratings are held fixed for every simulated game, so the spread you see is game-to-game luck and nothing else. It doesn't know about a trade in February, a starting goalie going down, or a team that simply turns out to be better than its numbers. Real seasons are wider than this. Treat a 95% as very likely, not done, and a 3% as alive. Nothing here is a lock, and it isn't betting advice.
Early in the year the ratings lean on last season, so the table mostly reflects what each club was. As games land, the ratings move and so do these numbers. The real table to date lives on the standings page, and every club's page has what's left on its schedule.