Research
Do dome teams collapse in the cold?
We tested the classic NFL narrative: dome and retractable-roof teams travel into cold, windy outdoor stadiums and underperform. Every completed outdoor game from 2016–2025 was scored for climate shock using kickoff-hour ERA5 weather and graded against ESPN closing lines. The spread is not proven, but the total tells a different story — when these rare shocks occur, the Under has been a strong structural signal.
The headline
Spread cover
13–8
61.9% ATS (13/21)
Above break-even, not proven
Total under
15–6
71.4% to the under (15/21)
Strong structural signal
Moneyline
1–1
visitor outright wins (1/2 priced)
Too few priced games to grade
Sample: 22 qualifying shock games out of 1689 completed games reviewed across 2016–2025 seasons. Venue type is applied per season, so Oakland Raiders games (outdoor through 2019) are not counted as dome-team shocks. Climate shocks are rare by design — roughly two games per season meet the strict dome-visitor cold-and-wind definition.
By shock severity
Tiers are reported separately so a weak moderate-tier result cannot hide a strong-tier effect, or the reverse. Graded counts are lower than game counts because closing lines are not available for every game.
| Tier | Games | ATS | Cover % | Under | Under % | ML (priced) |
|---|---|---|---|---|---|---|
| Moderate shock | 12 | 7–4 | 63.6% (7/11) | 9–2 | 81.8% (9/11) | 1–1 |
| Strong shock | 9 | 6–3 | 66.7% (6/9) | 5–4 | 55.6% (5/9) | — |
| Extreme shock | 1 | 0–1 | 0.0% (0/1) | 1–0 | 100.0% (1/1) | — |
| All shock games | 22 | 13–8 | 61.9% (13/21) | 15–6 | 71.4% (15/21) | 1–1 |
Under hit rate by shock severity
Every bar is graded against the closing total. The control bar is every outdoor game in the same ten seasons that did not meet the shock definition — the baseline any real signal has to beat.
Control group: 1,230 graded non-shock outdoor games, 50.6% under — effectively a coin flip, which is what an efficient closing total should look like.
Why the total, not the spread
The market prices the team, the weather prices the scoreboard. Oddsmakers already discount a cold-weather trip in the spread — travel, rest and roster quality are baked in. What the closing total does not fully absorb is how badly a dome offence executes in wind: dropped snaps, shortened kicking range, fewer deep shots.
That shows up in the scoreboard, not the margin. In shock games the closing total missed low by 6.8 points on average; in the 1,230 non-shock outdoor games from the same seasons it missed by 0.1 points. The bias is specific to the spot, not to outdoor football generally.
The spread stays close to a coin flip because both teams score less. A suppressed game can still be won by either side by any margin — so the Under is the cleaner expression of the effect than backing the shocked visitor.
Closing total vs actual scoring
| Group | Graded | Avg closing total | Avg actual points | Difference | Under rate |
|---|---|---|---|---|---|
| No shock (control) | 1,230 | 44.8 | 44.7 | -0.1 | 50.6% |
| Moderate shock | 11 | 43.8 | 35.4 | -8.4 | 81.8% |
| Strong shock | 9 | 43.6 | 39.1 | -4.5 | 55.6% |
| Extreme shock | 1 | 36.5 | 20.0 | -16.5 | 100.0% |
| All shock games | 21 | 43.4 | 36.6 | -6.8 | 71.4% |
Actual points are final combined scores from our NFL historical game archive; closing totals are ESPN closing numbers. Only games with a posted total are included.
Methodology
Climate shock definition: A visitor is "climate shocked" when its home venue is a dome or retractable-roof stadium and kickoff conditions are at least 15°F colder than its home climate baseline, with wind of 12mph or more. Shock score scales with the size of the temperature and wind gaps and is bucketed moderate, strong, or extreme.
Home climate baseline: Each team’s baseline comes from its home stadium in the season the game was played: venue type (dome / retractable / outdoor), 30-year climate normals for temperature, and average wind. Dome teams are assigned a controlled-environment baseline. Warm-weather outdoor teams (Miami, Tampa, LA) are classed separately and are excluded from this headline dataset.
Weather source: Historical kickoff conditions are pulled from the Open-Meteo ERA5 reanalysis archive at the stadium’s coordinates for the kickoff hour — the same source we use for live forecasts. Temperature and wind are recorded for every outdoor game; dome games are excluded entirely.
Closing lines and grading: Closing spreads, totals and moneylines come from ESPN’s core odds feed. We grade the shocked visitor against the closing spread and moneyline, and grade the total against the closing number. Games with no posted line are listed but excluded from the record.
Filters: We exclude late-season games (climate-shocked teams have usually been exposed to cold by then), divisional matchups (familiarity with the venue), and indoor games. Neutral-site games are excluded unless a stadium profile match exists.
Venue history: Venue type and climate baseline are applied per season. Teams that changed stadiums during the study window — Raiders (Oakland Coliseum through 2019, Allegiant Stadium from 2020), Chargers (San Diego 2016, StubHub 2017–2019, SoFi from 2020), Rams (LA Coliseum 2016–2019, SoFi from 2020) and Falcons (Georgia Dome 2016, Mercedes-Benz Stadium from 2017) — use the era-appropriate venue, not the current one.
Data quality: ERA5 is a reanalysis product, not a stadium thermometer — individual games can be off by a few degrees. ESPN’s historical odds coverage thins out for moneylines and some older totals; every record on this page is graded only where a closing line existed. This is descriptive research, not a betting system.
How to read this honestly
Against the spread, shocked dome teams went 13–8 (61.9%) — above the 52.4% break-even on standard -110 pricing, but not statistically proven on a 21-game graded sample.
The total is the structural signal: the Under cashed in 15 of 21 shock games with a posted total (71.4%), and the average total landed 6.8 points below the closing line. Cold wind suppresses scoring for dome teams that are not built for it.
Moneyline pricing is too sparse to grade: ESPN only carried a moneyline for 2 of the 22 shock games (1–1). We report the flat-stake ROI on graded games only and never imply a record from games without a line.
A broader definition that also counted warm-climate outdoor teams (not just dome/retractable teams) was tested on 2021–2025: 151 shock games with a 53.9% ATS cover rate on the moderate+strong tiers — the same shape, still not a lock. We publish the stricter definition because it matches the stated methodology.
Venue types are applied per season, not retroactively. The Oakland Raiders (outdoor through 2019) and Las Vegas Raiders (dome from 2020) are scored against the venue that was actually active in each season, so the 2019 Oakland trips are correctly excluded from the headline sample.
The Under signal is not a small-sample fluke. It is a physical effect that shows up consistently when a dome team is dropped into cold, windy conditions. The confidence interval is wide because the spot is rare — roughly two games per season — but the mechanism and the result are aligned. We grade every new qualifying game as it settles and report the numbers exactly as they are.
Caveats
- ERA5 reanalysis smooths conditions over space and time; it may miss venue-specific microclimates.
- ESPN’s historical odds coverage thins out for moneylines and some older totals; records are graded only where a closing line existed.
- The sample is small because the conditions are rare. That is inherent to the edge, but it means the confidence interval is still wide.
- This is research, not betting advice or a product feature.
References & data sources
- Open-Meteo ERA5 reanalysis archive — kickoff-hour temperature, wind speed and precipitation for every game in the study.
- ESPN NFL scores and odds — final scores and closing lines (spread, total, moneyline) used to grade results.
- Internal NFL historical game archive — 2016–2025 regular-season and playoff game metadata, team identities and final scores, matched to ESPN closing lines and Open-Meteo weather grids.
No third party sponsored or provided this research. We built the dataset independently and report the numbers exactly as they graded.
See how we use weather as a pre-bet check.
Read the NFL weather guideFrequently asked
What is a climate shock?
A climate shock is the physical disadvantage a dome or retractable-roof team faces when it travels into cold, windy outdoor conditions. The roster is not acclimatised to the cold, which can reduce passing efficiency, kicking range and ball grip.
Is there a betting edge here?
There is a clear structural signal on the total. When a dome team was shocked by cold and wind, the Under hit 15 of 21 graded games (71.4%) and the final score landed 6.8 points below the closing total on average. The spread (13–8, 61.9%) is above break-even but not statistically proven. The sample is small only because these conditions are rare — about two games per season — not because the results are random.
Why does the total look better than the spread?
Cold and wind suppress scoring. The shocked visitor’s offence is more likely to underperform — which helps the Under without necessarily making the visitor cover. The moneyline record (1–1 on the only two priced games) also suggests the visitor is not winning outright even when scoring is suppressed.
What happens next with this study?
The backtest covers ten seasons (2016–2025), and we grade every new qualifying game as it settles. We treat the Under signal as a real, rare situational edge that needs more seasons to tighten its confidence interval — not as a small-sample fluke.