Season Projections

We play out the rest of the NFL season 5,000 times on a computer and count how often each of you finishes first, in the top five, or dead last.

Each simulated season starts from where things actually stand — the points your six teams have already banked — and then plays every remaining game on the real NFL schedule, one at a time. For each game we need to know how likely the home team is to win, and we get that from DraftKings: their moneyline for that game, with the sportsbook's built-in cut removed, so the two sides add up to exactly 100%. A −150 favourite becomes roughly a 58% chance to win; the +130 underdog gets the other 42%. Then we flip a weighted coin at those odds and hand out 2 points to whoever owns the winner.

Because every game is played once per simulated season and shared by everyone, the results hang together the way real life does: if the Rams win, the 49ers lose, and everyone holding either team feels it in the same run. That's also why drafting the same team as three other people doesn't help you — you all rise and fall together.

Later in the year DraftKings will stop posting lines for games that are still weeks away. For those, we use a strength rating for each team that we work out from all the lines DraftKings has posted plus the results so far, with a home-field edge included. It's the market's opinion of the two teams applied to a matchup the market hasn't priced yet. The line under the timestamp tells you how many games are running on real DraftKings odds versus the rating.

At the end of each simulated season everyone is ranked by total points, with a projected Points For as the tiebreaker — your current Points For plus what DraftKings' spreads and totals imply your teams will score the rest of the way. Add up the 5,000 finishes and you get the percentages on this page.

The whole thing re-runs every morning after the lines refresh and again after every scoreboard update, so it always reflects today's odds and what's already been played, not a preseason guess.