How Does Weather Affect NFL Betting?

Weather affects NFL betting mainly through wind, and mainly through totals. Strong wind makes long passes and kicks harder, and games played in it have tended to score less. Cold on its own shows much less effect than its reputation suggests, and rain and snow are harder to measure because the public data rarely records them well.

The larger point is that sportsbooks see the same forecasts you do. The market lowers totals for wind before kickoff. A bettor who knows it will be windy knows what the book knows; the opportunity, when there is one, is in acting on a forecast before the price catches up, or in judging the forecast better than the market does.

What the data shows about wind

We took every completed NFL game from 1999 through the 2025 season in the nflverse results file that was played outdoors or with the roof open and had a wind speed recorded: 5,205 games. For each wind band we compared the total line recorded in the file with the actual combined score.

Recorded wind	Games	Average total line	Average final total	Final minus line	Under rate
0–4 mph	1,061	43.4	44.8	+1.4	50.0%
5–9 mph	2,248	42.9	43.9	+1.0	48.2%
10–14 mph	1,212	42.4	42.5	+0.1	53.8%
15–19 mph	497	42.1	40.5	−1.6	57.3%
20+ mph	187	40.9	39.9	−1.1	56.0%

Under rates exclude pushes. Three things to read from it.

Scoring drops with wind. Games with 15 to 19 mph recorded averaged 40.5 points against 44.8 in near-calm conditions, a gap of more than four points.

The market already moves. The average line fell from 43.4 to 42.1 to 40.9 as the wind rose. The book is not ignoring the weather; it is adjusting, by about 1.3 points between calm and 15-plus mph and about 2.5 points at 20-plus.

Historically it did not move far enough. At 15 mph and above the under won 57.3% and 56.0% of the time, above the 52.4% needed at −110. That held in both halves of the file: 57.2% at 15 mph or more from 1999 to 2014 (456 games) and 56.4% from 2015 to 2025 (228 games).

Before taking that as a system, weigh the caveats. They are serious.

This is recorded wind, not forecast wind. The file has one wind figure per game and does not say when it was measured. It describes conditions that happened. When you bet, you have a forecast, and forecasts miss. Part of the gap in the table is simply the market not knowing on Friday what the wind would be on Sunday, and you will not know either.

The samples are small. 497 games at 57.3% carries a sampling error of about two percentage points, and 187 games at 56.0% about four. The direction is consistent; the precise size is not settled.

The recorded line is not your line. The data dictionary does not say whether the total is an opening or closing number. Across all outdoor games the final score beat the line by about half a point on average, so the useful comparison is between wind bands, not against zero: calm games ran about 1.4 points over and 15 to 19 mph games 1.6 under, a three-point swing.

Coverage has gaps. Wind is missing for some outdoor games, most visibly in 2022, when only 107 of 198 outdoor games in the file have a wind value.

Cold and indoor games

Temperature on its own shows little in the same file:

Recorded temperature	Games	Under rate
20°F or below	96	47.9%
21–32°F	288	46.6%
33–50°F	1,318	52.3%
51–70°F	2,174	51.0%
71°F and above	1,329	51.0%

The coldest games went under less often than average, not more, although 96 games is too few to say much. Cold matters when it comes with wind. The National Weather Service's wind chill formula makes the point: 20°F with a 15 mph wind feels like about 6°F. If you are reacting to a cold forecast, check whether it is a windy one; that is the part the numbers support.

Domes and closed roofs are the control group. In the 1,815 games under a roof, the final total averaged 46.9 against a recorded line of 45.7, and the under won 48.5%. Weather-related handicapping for a game in a closed stadium starts and ends with checking that the roof is actually closed.

Rain and snow

The public data we used does not record precipitation, so we have no number to give you. Mechanically, a wet ball raises the risk of fumbles and dropped passes and a snow-covered field slows everyone down, but bad footing hampers offence and defence alike, and we are not aware of a reliable public dataset that separates them. If you see a confident percentage for "snow games", ask where the precipitation data came from.

How to use weather without fooling yourself

Get the forecast from the source. The National Weather Service publishes a free forecast API that needs no key, only a User-Agent header. Look up a stadium by latitude and longitude with the /points endpoint and read the hourly forecast, which runs seven days ahead. A forecast made days out is less reliable than one made the morning of the game, and hourly detail around kickoff is more useful than a daily summary.

Move before the market, or not at all. If a forecast for 20 mph winds has been public since Wednesday, the total has probably already fallen. The table above says the average line has adjusted by one to two and a half points for strong wind. Your edge, if any, is in betting a total before a new forecast is priced, or in holding a view about the forecast that differs from the consensus for a reason.

Watch totals and passing props together. Wind that affects the total also affects long passes, field goals and quarterback yardage lines. The player props guide covers how those lines respond, and the Academy's over/under guide covers how totals are priced.

Know the rules for weather delays. DraftKings' filed football rules void bets on a regular-season game that has not started and is moved more than 7 days from its scheduled date. A game stopped with more than 10 minutes of regulation left and not finished within 72 hours is void unless a market is already decided; one stopped with 10 or fewer minutes left, or in overtime, and not finished within 72 hours is settled on the score at the stoppage. Playoff games have their own rule, and all bets include overtime.

Size for variance. Even at 57%, an under loses more than four times in ten. A weather angle is a small edge in a noisy market; the Academy's guide to variance shows how long a run of losses a 57% bet can produce.

Sports betting is legal only in some states and only for adults 21 and over in most. Historical rates describe past games, not future ones.