Frequently Asked Questions
How are NFL playoff odds calculated?
BlitzChaser simulates the rest of the regular season 100,000 times. Every unplayed game is decided by a win probability from each team's Elo rating (with a small home-field edge). At the end of each simulated season the 14 playoff seeds are set using the official NFL tiebreaking procedure, and the whole postseason bracket is played out. A team's playoff odds are simply the share of simulated seasons in which it made the field.
Why do the odds jump so much from one week to the next?
Because 17 games is a very short season: a single result moves a team's expected win total by close to a full win, and one win often separates a #5 seed from the couch. But a swing of 20 or 30 points after one game is not what a well-calibrated model produces — we checked, by replaying every week of 2021–2025 and comparing the odds shown with who actually made it. The model now carries an explicit uncertainty about each team's strength into the simulation, so a September loss costs a team roughly 8–14 points of playoff odds, not 20–30, and no team is written off or crowned after Week 1. The swings shrink as the sample grows.
Do you handle ties and tiebreakers?
Yes. Ties count as half a win, simulated games can end tied, and playoff seeding in every simulated season uses the league's own cascade: head-to-head, division record, common games, conference record, strength of victory and strength of schedule (with the sweep rule and the minimum-four-common-games rule for wild cards). For today's real standings the points-based steps are applied too. Fed the complete 2025 season, the code reproduced all 14 real 2025 playoff seeds. See the methodology for exactly what is and is not implemented.
What does the Elo rating mean?
A single number for team strength. Each team starts the season at last season's final rating regressed halfway toward the average, then moves after every game — up for a win, down for a loss, by more when the result was a surprise. The K-factor (40), home-field bonus (+30 Elo) and regression were tuned jointly and out of sample on every game from 2018 to 2025, and the simulation treats each rating as an estimate with a measured uncertainty (100 Elo) rather than a known fact. There is no margin-of-victory or quarterback adjustment yet. Details on the methodology page.
What are the game scenarios and the Butterfly Effect?
"If [Team] wins" percentages are read straight from the same 100,000 seasons: the playoff odds of every team, restricted to the simulations where that game went a certain way. The Butterfly Effect stacks those shifts as you pick winners for the week's slate.
When does a team show as clinched or eliminated?
Only when it is mathematically certain from the standings, or when the league's own standings flag it. A 99.9% simulation result is not a clinch and is never shown as one.
How often is the data updated?
Weekly during the season, after the week's games. Each update re-runs the full 100,000-season pool and locks a snapshot so the Odds Movement chart grows one point per week.
Is this for betting?
No. BlitzChaser is a hobby project for entertainment and informational purposes only. The odds are model output, not betting lines. Not affiliated with the NFL.