As you can see above, the dataset has over 14,000 rows, all of which include a game played, the home and away team, who was favored to win said game, weather data, etc.
Now, not every game has a stated over/under line, as you will see in the dataset. The reason for this is because over/under totals were not introduced until 1969. As a result, any game before 1969 will contain NA values on that column. To solve this problem, I cleaned the dataset to remove any games without an over/under line. Along with this, I got rid of any game with an NA temperature, wind value, and total points value. After that, we were left with about 10,603 rows of data to examine.
Part 2: Descriptive Analysis
How often does the favorite cover the spread, depending on the wind speed?
The bar chart visualization examines how often favorites cover the spread under different wind conditions, and the results show a clear decline (albeit not massive) as wind speed increases. In calm conditions (0–5 mph), favorites covered around 58.5% of the time, the highest rate in the dataset. That rate dipped slightly to 56.9% in moderate wind (5–10 mph), then fell down to 54% in the 10–15 mph range. In high‑wind games (15+ mph), favorites covered only 53% of the time. Although the decline is gradual rather than dramatic, the pattern is consistent: as wind increases, favorites cover less often. This supports the broader idea that wind disrupts offensive efficiency, especially passing, which reduces scoring margins and makes it harder for favorites to separate enough to beat the spread.
How often do favorites cover based on temperature?
This visualization examines how often favorites cover the spread across different temperature ranges, revealing a clear downward trend as conditions get colder. In warm games (70°+), favorites covered at the highest rate, followed by moderate‑temperature games (50–70°), where cover rates remained relatively strong. Once temperatures dropped into the 32–50° range, the cover rate declined noticeably, and in sub‑freezing games (<32°), favorites posted their lowest cover rate in the dataset. This pattern aligns with the broader scoring dynamics observed earlier: cold weather suppresses offensive efficiency, reducing explosive plays and lowering scoring margins. Because favorites rely on creating separation to beat the spread, colder conditions make it harder for them to win by enough points to cover. In short, temperature acts as a quiet but meaningful influence on favorite performance, especially in freezing conditions.
#3: How does temperature affects total scoring in NFL game?
My third visualization analyzes how total scoring varies across temperature ranges, and the results show a clear, if gradual, upward trend as conditions warm. Games played in sub‑freezing temperatures (<32°F) averaged a total of 41.32 points, the lowest of any bucket. Scoring increased slightly in the 32–50°F range, where games averaged 41.84 points, and rose again in the 50–70°F range to 42.37 points. The warmest games (70°+) produced the highest scoring environments, averaging 44.12 total points. While the differences between adjacent buckets are modest, the overall pattern is consistent: warmer conditions support higher scoring, likely due to improved passing efficiency, fewer drive‑stalling mistakes, and more explosive plays. This visualization reinforces the broader conclusion that temperature is a meaningful driver of total points, making it a key factor in whether a game is likely to go over or under the posted total.
#4: Is the over or under more likely to be covered, depending on the wind speed?
This next visualization examines how wind speed influences whether games go over or under the posted total, and the results show a clear downward trend in over frequency as wind increases. In calm conditions (0–5 mph), overs were hit 50.4% of the time, basically a coin flip. As wind increased to 5–10 mph, the over rate dipped slightly to 49.3%, still close to even, still a coin flip. The decline became more pronounced in the 10–15 mph range, where overs fell to 47.1%, reflecting the growing impact of wind on passing efficiency and explosive plays. The sharpest drop occurred in high‑wind games (15+ mph), where overs were hit only 40.5% of the time, making unders overwhelmingly more common. This pattern reinforces a central finding of football analytics: wind is the strongest single weather predictor of unders, as it disrupts deep passing, reduces the reliability of field goals, and slows offensive pace. The visualization makes this relationship unmistakable. As wind speed rises, games become increasingly likely to finish under the total.
#5: Do big favorites (-7 or more) underpreform in cold games?
Lastly, this visualization examines how spread size interacts with temperature to influence how often favorites cover, revealing a clear pattern: cold weather consistently depresses favorite performance across every spread bucket. As spreads increased into the 3–7 and 7–10 point ranges, the gap widened, with favorites covering considerably less often in colder conditions. The steepest drop occurred in the 10+ point bucket, where large favorites struggled the most in sub‑freezing or near‑freezing temperatures. This visualization reinforces the broader theme emerging throughout this analysis: weather affects more than scoring: it affects margin, making it harder for favorites to create the separation required to cover. Spread size and temperature interact in a predictable way: the larger the spread and the colder the game, the less likely the favorite is to cover.
Part 3: Secondary Source
To supplement the league‑wide analysis, I incorporated a secondary dataset containing detailed weather and scoring information for all 2024 Cincinnati Bengals home games. This dataset serves as a real‑world case study to validate whether the broader patterns observed across the NFL also appear at the team level. The Bengals’ home environment is particularly useful because Paycor Stadium is an outdoor venue exposed to the full range of Midwestern weather conditions, which makes it an ideal test case for examining how temperature and wind influence scoring and totals outcomes.
Across the Bengals’ 2024 home schedule, the same weather‑scoring relationships identified in the league‑wide dataset appeared clearly. Warmer games produced higher total points, while colder games clustered near the lower end of the scoring distribution. Wind showed an even stronger effect: games with wind speeds above roughly 12–15 mph consistently produced lower totals and were more likely to finish under the posted line. When comparing over/under outcomes directly, warm, low‑wind games leaned toward overs, while cold or windy games leaned toward unders, mirroring the national trend. This alignment between the Bengals’ localized data and the league‑wide results strengthens the overall conclusion: weather is a reliable and consistent determinant of scoring efficiency and totals outcomes, even when examined at the micro level of a single team’s season.
Conclusion + Key Findings
Across the descriptive and supplementary analysis, one theme emerged amongst them all: weather is the most consistent and influential determinant of whether a game goes over or under the posted betting total. Three key findings support this conclusion. First, wind speed showed the strongest single relationship with totals outcomes, with overs occurring 50.4% of the time in calm conditions but dropping sharply to 40.5% in games with winds above 15 mph, a clear indication that wind suppresses scoring by disrupting passing efficiency and reducing explosive plays. Second, temperature meaningfully shaped total scoring, with warm games (70°+) averaging 44.12 total points compared to just 41.32 in sub‑freezing conditions, demonstrating that cold weather acts as a drag on offensive production and pace. Third and finally, spread size interacted with temperature to influence favorite cover rates, as large favorites struggled disproportionately in colder games, reinforcing that weather affects not only scoring totals but also scoring margins.
These findings held true not only in the spreadspoke dataset, but also in the 2024 Cincinnati Bengals home‑game case study, where warm, low‑wind games consistently produced higher totals and leaned toward overs, while cold or windy games skewed toward unders. Taken together, the evidence shows that while team strength and spread size matter, they are secondary to the direct, predictable impact of environmental conditions. For analysts, bettors, and teams alike, understanding how weather shapes offensive efficiency is essential for anticipating game flow and totals performance and this analysis demonstrates that wind and temperature are the most reliable indicators of whether a game will finish over or under the posted line.