luckadjusted
NFL Week 9 · 2026

The NFL, adjusted for luck.

The scoreboard tells you who won. It doesn’t tell you how much of that was the bounces. So we replay every game 4,000 times, using the plays both teams actually ran, and count how often each side comes out on top. It’s a model, not a ruling. Everything that goes into it is on the methodology page.

2026 gamesAdjusted standings

The bar on each card shows how often the loser wins the replays. Fuller bar, more luck. An outline means the winner was behind in the replays.

Week 9, 2026

14 of 15 games scored so far (new games appear the morning after they’re played). The replays back every winner. Week 9: upsets and luck ranking

Week 8, 2026

2 of 14 winners were behind in the replays. Week 8: upsets and luck ranking

Week 7, 2026

The replays back every winner. Week 7: upsets and luck ranking

Week 6, 2026

3 of 14 winners were behind in the replays. Week 6: upsets and luck ranking

Week 5, 2026

2 of 14 winners were behind in the replays. Week 5: upsets and luck ranking

Week 4, 2026

2 of 16 winners were behind in the replays. Week 4: upsets and luck ranking

Week 3, 2026

4 of 16 winners were behind in the replays. Week 3: upsets and luck ranking

Week 2, 2026

2 of 16 winners were behind in the replays. Week 2: upsets and luck ranking

Week 1, 2026

1 of 16 winners was behind in the replays. Week 1: upsets and luck ranking

The adjusted standings, 2026

The standings adjusted for luck, if the season ended today. We re-roll every kick, fumble bounce, drop, conversion and ref call, and keep everything else as played. Think we count too much as luck, or too little? Move the sliders.

How much of each is luck?

0% treats it as skill, 100% as pure luck. The tick is our fit.

What counts

None by default: our tests measured no carryover after a lucky break, and counting some doesn’t make these standings forecast the rest of a season any better. Slide it up to count a break for more than its own swing. Why

Order by Group by

AFC

Adjusted for luck, the Browns and the Ravens would make the playoffs if the season ended today instead of the Broncos and the Patriots.

AdjustedNowTeamRecordAdjusted winsLuckSeed
1▲12Jaguars6–26.7−0.71
2▲24Chiefs6–26.2−0.22
33Bengals6–25.9+0.13
4▲26Steelers6–25.9+0.15
55Bills6–15.2+0.84
6▲28Browns5–44.2+0.86
7▲310Ravens4–54.2−0.27
8▼17Patriots5–44.1+0.9out
9▼81Broncos7–24.0+3.0out
10▲111Texans3–53.9−0.9
11▲112Jets2–73.2−1.2
12▼39Raiders4–53.1+0.9
13▲215Dolphins1–72.1−1.1
1414Titans1–71.8−0.8
15▼213Colts2–71.8+0.2
1616Chargers0–80.6−0.6

NFC

Adjusted for luck, the Cowboys, the Rams and the Seahawks would make the playoffs if the season ended today instead of the Giants, the Vikings and the Lions.

AdjustedNowTeamRecordAdjusted winsLuckSeed
1▲12Bears7–26.7+0.31
2▼1149ers7–16.7+0.32
3▲25Packers6–36.5−0.55
4▲48Rams5–46.4−1.46
5▲27Seahawks5–45.9−0.97
6▲39Cowboys4–55.5−1.53
7▼16Giants5–2–14.4+1.1out
8▲210Buccaneers3–64.3−1.34
9▼54Lions6–24.3+1.7out
10▲111Panthers3–54.2−1.2
11▼83Vikings6–13.9+2.1out
1212Saints3–53.4−0.4
1313Falcons3–63.3−0.3
14▲115Commanders2–5–12.7−0.2
15▼114Cardinals3–62.3+0.7
1616Eagles2–71.7+0.3

Adjusted wins: every kick, fumble bounce, drop, conversion and referee call re-rolled at your settings, everything else as played. Seed: playoff seed in the adjusted standings; 5 in only in the adjusted standings (luck kept them out), out is in by record only thanks to luck. Adjusted: rank in the conference by adjusted wins, ▲ places up because luck cost them, ▼ down because luck helped. Now: rank by record.

Luckiest so far

  1. Broncos7–2+3.9 wins
  2. Vikings6–1+2.0 wins
  3. Giants5–2–1+1.6 wins
  4. Packers6–3+1.5 wins
  5. Browns5–4+1.4 wins

Unluckiest so far

  1. Cowboys4–5−2.3 wins
  2. Rams5–4−2.0 wins
  3. Chargers0–8−1.7 wins
  4. Panthers3–5−1.7 wins
  5. Jets2–7−1.6 wins

All 32 teams, after Week 9 · Power rankings: who is best, adjusted for luck

The biggest heists of 2026 so far

Games the winner almost never wins in the replays. The biggest since 2019

  1. Buccaneers 24, Cowboys 16 · Week 5 · the Buccaneers win 4% of replays
  2. Broncos 30, Rams 26 · Week 3 · the Broncos win 7% of replays
  3. Browns 21, Panthers 18 · Week 3 · the Browns win 8% of replays
  4. Giants 28, Cowboys 20 · Week 1 · the Giants win 15% of replays
  5. Broncos 17, Seahawks 14 · Week 6 · the Broncos win 23% of replays

How it works

Here’s the idea. We take every snap each team ran in a game and deal them out again into new drives, matched by down and distance. Then we play the game over. And over. 4,000 times. Kicks, fumble bounces and turnovers happen at their normal rates, not the way they happened that day.

How often a team wins those replays is roughly what its play was worth. The gap between that and the result is what we call luck. Lose a game you’d win 9 times out of 10 in the replays, and that’s −90 on our luck index. We also go play by play, with each kicker and receiver measured against his own record, so you can see where the luck came from.

Does it predict anything? Rematches, about as well as the betting line. Who covers next week? No. The market’s already on that. Methodology and every test

Common objections

Do you rate players?

No. The replays reuse the plays both teams actually ran in that game, matched by down and distance. So if the pass rush ate that quarterback alive, that’s already in the plays. The only things we reset are the coin flips, using rates from every game since 2019. Who falls on a fumble (the team that fumbled keeps it 52% of the time). Whether a kick from that distance goes in. Whether a defender holds on to a pass he should have picked.

A 62% kick went in. Why is that luck?

Mostly, it isn’t. A make on a 62% kick counts as +0.38 over what you’d expect. A miss would have been −0.62. A kicker who hits exactly his share comes out at zero over a season. And the verdict on a game comes from replaying the whole game, not from any one kick.

Isn’t recovering a fumble a skill?

Forcing fumbles and holding on to the ball? Sure, and the replays keep both as they happened. Who ends up on the loose ball is a different story. A team’s recovery rate one year predicts the next at a correlation of 0.09 (192 team-seasons). A coin would give you zero. Turnover luck by team

Doesn’t the game situation matter?

It does, and we only handle part of it. Plays are matched by down and distance, so a 3rd-and-20 only draws from that team’s long 3rd downs. Teams that are behind late go for it on 4th down. But field position, the score and the clock don’t change which plays get drawn. That’s a real gap. It still holds up, though: replaying the first game of a division series picks the rematch winner as well as the closing betting line (log loss 0.666 vs 0.668, 336 rematches).

A receiver who drops a lot isn’t unlucky. He’s just bad.

Agreed. That’s why each drop is measured against that receiver’s own drop rate. A guy who drops a lot is expected to drop more, so he doesn’t get much “bad luck” for it. Kickers get the same treatment (their own record from that distance), and so do defenses (how often they actually catch the passes they should pick).

Does any of this predict anything?

We check it against what happened next. Teams the replays gave a 50%–75% chance won 61% of the time (we said 63%). And records adjusted for luck pick the winners of a team’s next games better than its actual record does. How we test it, with every number

Every season since 2019

20262025202420232022202120202019