Simulation Lab · Experimental
From each matchup to the Super Bowl.
Completed regular-season results stay fixed. Explore thousands of possible finishes using the latest saved team strengths.
Every fresh run draws new possible seasons. A fixed seed reproduces the same draws when the snapshot and game status are unchanged. Overall probabilities should stay fairly similar across large runs; a repeated champion alone does not mean a run is stuck.
Scroll inside each results table to compare every statistic. Your team, week, and example choices stay selected as you browse.
What the simulations say
These are frequencies under an experimental model, not guarantees or validated betting probabilities. A favorite can still lose in most simulated seasons.
Teams · game by game
Expected final scoring
Both teams' simulated average final score, scoring SD and middle-80% range. These are scoreboard statistics, not player yards or touchdown box scores. Final results show actual scores, not retrospectively fitted forecasts.
Whole-league outlook
Expected records & division races
W–L–T averages include real completed results plus simulated remaining games. Fractional records are averages across seasons; the range describes simulated wins. Division order below is by average wins, not an official tiebreaker ranking.
Qualification & championship chances
The playoff picture
Seven qualifiers per conference, four division winners, three wild cards, and one first-round bye. Playoff probabilities sum across all simulated seasons; they are not a fixed set of 14 predicted qualifiers.
One possible finish · explore this run
A complete simulated season & playoff bracket
Final records in this example season
All 272 regular-season game results in this example
We retain 20 complete examples spread through the run; the averages above use every trial. Each example keeps its own game results, records and playoff seeds together. Winners are sampled outcomes, not guaranteed favorites, and different examples can have the same champion.
What this model does—and doesn't—know
Each game uses the same paired historical forecast-error model as the head-to-head simulator. One shared score settles each matchup, so opponents cannot both win. Current team strengths are held fixed through the remaining season; games are independent conditional on those strengths. Injuries, starters, roster changes, coaching, weather and season-long changes in strength are not simulated.
Only completed regular-season scores in the saved feed are fixed. A game underway or awaiting a final result is simulated as unresolved, without live scores or in-game adjustments. The feed uses a conservative six-hour completion delay. Actual playoff results are not ingested yet: the postseason shown here is hypothetical.
Seeding uses overall winning percentage (a tie is half a win), head-to-head, division/conference/common-game results where applicable, strength of victory and schedule. Late tiebreakers are simplified to overall point differential and a seeded draw; official combined scoring ranks and net touchdowns are not reproduced. Therefore close seeds can differ from official NFL tiebreakers. Division winners occupy seeds 1–4; wild cards 5–7. No. 1 gets a bye, the divisional round reseeds, higher seeds host within the conference, and the Super Bowl is neutral.
Postseason sampling conditions on a decisive score: no playoff game ends tied. It is not a drive-by-drive or overtime-rule simulation. Mean ± SD and percentile ranges describe sampled variation, not calibrated confidence intervals.
Official NFL tiebreaking procedures ↗ · Head-to-head model details · Current saved betting lines