How Our Playoff Odds Work
Every number on this site comes from the same three steps: rate each team from its results, turn ratings into game win probabilities, then simulate the rest of the season many thousands of times.
1. Team ratings
Each team gets one number: how many points (or goals) better or worse than an average team it has played, on a neutral field. We fit it to every final score this season with a regularized least-squares model: the predicted margin of a game is the home team's rating minus the away team's rating plus a home-field edge. Blowout margins are capped (24 points in the NFL, 25 in the NBA, 4 goals in the NHL) so garbage time doesn't distort the ratings, and every rating is pulled toward average by a few "phantom" games of zero margin, which keeps early-season ratings from overreacting.
2. Game win probabilities
A rating gap becomes a win probability through a normal distribution. The standard deviation of game margins is about 13.5 points in the NFL, 12 points in the NBA and 2.6 goals in the NHL. Home field is worth about 1.6 points in the NFL, 2.4 in the NBA and 0.18 goals in the NHL, and neutral-site games get no home edge. In NHL simulations about 23% of games go to overtime or a shootout, where the loser still earns a point.
3. Season simulations
We play out every remaining regular-season game 20,000 times for the NFL and 10,000 times for the NBA and NHL. Before each simulated season every team's rating is nudged by a random amount sized to how uncertain that rating is, which is why odds stay humble early in the season. Each simulated season is scored with the league's real rules:
- NFL: the full tiebreaking procedures for divisions and wild cards (head-to-head, division and conference record, common games, strength of victory, strength of schedule and net points), seven seeds per conference, a reseeded bracket and the full draft order, including the strength-of-schedule tiebreaker.
- NBA: the top six per conference plus the play-in tournament for seeds 7–10, then best-of-seven series with home court to the better record.
- NHL: the top three in each division plus two wild cards per conference, with regulation wins and the other tiebreakers, then the divisional playoff bracket.
A team's playoff odds are simply the share of simulated seasons in which it made the playoffs; the same goes for division titles, seeds, championships and draft picks.
College Football Playoff
The CFP picture is not simulated. It applies the 2026 selection rules to the latest College Football Playoff rankings (or the AP poll before the first CFP rankings): the four Power Four champions and the highest-ranked Group of Six champion get automatic bids, Notre Dame is in if it is ranked in the top 12, and the rest of the field is filled by ranking. Projected champions are the highest-ranked team in each conference.
Data
Schedules, final scores, standings and rankings come from public sports data feeds and are refreshed every few hours. If a feed is unavailable, the site keeps the last good snapshot and shows its update time.
Frequently asked questions
Why do the odds change after a game we didn't play in?
Every result changes the standings and the ratings of the two teams involved, and that ripples through tiebreakers and every simulated season. A rival losing can be worth as much as your own win.
Why might a team with a worse record have better odds?
Odds depend on the remaining schedule, tiebreakers already won or lost, division structure and team strength, not only the current record.
Do you account for injuries?
Not directly. Ratings come from scores only, so a key injury shows up only as results change. That keeps the model simple and transparent.