Cold doesn't kill football.
Wind and water do.
The Weather Football Index measures what conditions take from the game. This page asks the opposite question — what does football look like when the weather gives everything back — and answers it from 6,923 games scored against the nearest airport station, 2000–2025. Pick a sky.
Outdoor · dry · wind ≤8 mph · 45–75°F. Football at its most itself.
The market prices these correctly — no edge, just football.
KC @ LA, 2018-11-19 — 105 points (62°F · wind 1 mph · WFI 0)
85°F+ at kickoff; the muggy subset adds 60%+ RH.
Heat is a conditioning problem, not a yardage problem — scoring barely moves.
DEN @ MIA, 2023-09-24 — 90 points (88°F · wind 11 mph · WFI 7)
Sustained 15+ mph; the gale subset is 20+.
The one input that never misses. Deep passing and long kicking die first.
BUF @ CLE, 2007-12-16 — 8 points (29°F · wind 37 mph · WFI 66)
Measurable rain observed during the game window, no snow.
The market's blind spot: rain games have gone under at 61.6% — the largest weather lean in the archive.
MIA @ PIT, 2007-11-26 — 3 points (51°F · wind 11 mph · WFI 30)
Snow observed falling during the game window.
The market's overreaction: books cut snow totals so hard the games lean over.
SF @ CIN, 2003-12-14 — 79 points (33°F · wind 6 mph · WFI 13)
≤25°F with wind ≤8 and no precip.
Cold alone does not kill scoring — and the market over-fears it.
JAX @ PIT, 2018-01-14 — 87 points (14°F · wind 4 mph · WFI 13)
The ladder
Average total points by condition class. The dome is the ceiling — and it isn't weather, it's architecture. Open air at its best gives back all but 2.5 of those points.
Two curves that explain everything
Temperature is a plateau
avg total points · calm (≤8 mph) and dry outdoor games
From the 20s to the 70s the game barely notices. The fade only starts above 85°F — and even there it is a couple of points, not a collapse.
Wind is a tax
avg total points · dry outdoor games at 50–75°F
Monotone, no plateau, no refuge: every 4 mph costs. At 20+ the game loses 9.1 points off the calm baseline — before any rain falls.
What the market fears vs. what matters
🌧️ Rain is the blind spot
61.6% under505 rain games — the largest weather lean in the archive, and it grows with the soaking: ≥0.25" games went under 64.3%. Breakeven is 52.4%.
❄️ Snow is the overreaction
42.4% underBooks cut snow totals so hard that 125 snow games actually lean over. Snow slows the game less than the number assumes — it's the wind that usually comes with it.
🥶 Cold is the myth
44.5 ptsCalm, dry games at ≤25°F score like September (62 games) — and went under only 43.5% of the time. Frozen air alone never stopped an offense; the market keeps pricing it as if it did.
Method: outdoor NFL games 2000–2025 joined to the nearest qualifying ASOS station's observations over the game window (median 5.8 miles); precipitation from present-weather reports, not the gauge. Classes use observed conditions; Bad/Harsh use the archive's WFI. Under rates vs the closing total via nflverse; breakeven 52.4%. Passing and kicking splits from the site's play-by-play study and kicking study. Full engine details on the methodology page.