How Does Cold Weather Impact NFL Scoring?
During the season, it’s a question I get often on Sunday mornings. Fantasy owners want to know if they should bench a certain player due to extreme cold.
The short answer: No, cold temperatures have very little (if any) bearing on the total points scored in a game. Since actual (real-world) points scored and fantasy points scored have a relatively high correlation, cold temperatures should not significantly impact fantasy production on Sundays.
The long answer: I gathered gametime temperature and total points scored data for every outdoor regular season game since 1994, or 3,935 games in all. This data was gathered from Pro-Football-Reference.com, an excellent site with a ton of searchable data.
The table below shows a summary of the total points scored in extreme cold and extreme heat.
| Temp. Range | # | Total Points |
|---|---|---|
| 10 or below | 10 | 43.3 |
| 20 or below | 39 | 42.8 |
| 32 or below | 278 | 42.6 |
| 32.1 to 79.9 | 3424 | 42.1 |
| 80 or above | 233 | 41.5 |
| 85 or above | 68 | 41.7 |
| average | 3935 | 42.1 |
When compared to the league average in total points scored since 1994 (42.1 points), games that have been held in extreme cold have actually been high scoring. Gametime temperatures below 20 degrees resulted in an average of 42.8 points scored, while temperatures below 10 degrees had an average of 43.3 points scored.
On the flip side, games held at high temperatures (80 degrees or above) were actually on the low side (41.5) compared to the league average, though the difference isn't big enough to dissuade owners from starting players in high-temperature games.
Since we’re working with such a small sample size, I also pulled playoff data for low temperature games. In 18 playoff games with gametime temperatures of 20 degrees or below, the average was 45.1 points scored. In six games with a gametime temperature of 10 degrees or below, the average was 37.7 points scored.
Here’s a chart of all 3,935 regular season games.
The trendline shows a very low correlation (R-squared = 0.0082), which means that X-axis (Gametime Temp) has very little effect on the values in the Y-axis (Total Points). Moreover, the trendline has a negative slope so, if anything, total points scored drops a bit as temperatures rise.
At 4for4, we do adjust our projections for high winds (20+ MPH), but generally ignore precipitation and temperature, unless the forecast calls for a blizzard or a downpour. Unfortunately, gametime wind speed and precipitation data are not available at PFR.
But when it's simply cold weather -- don't worry about it.
