Analyzing over/unders with pace, efficiency, and environmental factors
Totals betting offers several advantages over spread betting:
You don't care who wins. You're betting on the combined output of both teams - a fundamentally different analysis than picking sides.
Pace determines how many possessions each team gets. More possessions = more scoring opportunities.
Efficiency determines how well teams convert possessions into points.
| Metric | Good Offense | Average | Poor Offense |
|---|---|---|---|
| EPA/play | +0.10+ | 0.00 | -0.10- |
| Points/drive | 2.5+ | 2.0 | 1.5- |
| Success rate | 48%+ | 44% | 40%- |
Teams that score TDs vs FGs in the red zone directly impact totals.
Teams converting 60%+ of red zone trips to TDs boost totals. Each TD vs FG adds 4 points.
Teams settling for FGs (under 50% TD rate) depress totals. Multiple FG drives lower combined scores.
Weather is one of the most exploitable factors in totals betting. The public often underestimates its impact.
| Condition | Adjustment | Impact |
|---|---|---|
| Wind 15-20 mph | -3 to -5 points | Affects passing, field goals |
| Wind 20+ mph | -5 to -8 points | Significant passing impact |
| Temp under 20°F | -2 to -4 points | Ball handling, kicking affected |
| Rain (light) | -1 to -2 points | Slight passing impact |
| Rain (heavy) | -3 to -5 points | Run-heavy games |
| Snow | Varies | Can increase or decrease scoring |
Lines are often set before weather forecasts are finalized. Check forecasts Sunday morning for late adjustments - the market may not have fully priced in weather changes.
When a heavy favorite (-10+) jumps ahead, they often run the ball and kill clock. This can push games under even when both teams are high-scoring.
The public has consistent biases in totals betting:
Use opponent-adjusted scoring rates, not raw PPG
Team A scores 24 vs avg D, Team B allows 21 = 22.5 for Team A
Fast-fast matchup +3, slow-slow -3
Per the table above
If your number differs by 3+ points, investigate
Strategy talk means little without the base rates. From the nflverse game log bundled with this site — all 7,276 completed games, 1999–2025 — the over went 48.7%, the under 49.8%, and 1.5% pushed. Excluding pushes, that is 49.4% over / 50.6% under: the market centers totals almost perfectly, with a whisper of an under lean that is far too small to beat the vig. It stays that way at every price level:
| Closing total | Games | Over hit (pushes excluded) | Landed exactly (push) |
|---|---|---|---|
| Under 40 | 1,634 | 50.3% | 1.1% |
| 40 to 43.5 | 2,112 | 50.4% | 1.8% |
| 44 to 47.5 | 2,159 | 48.7% | 1.5% |
| 48 to 51.5 | 980 | 48.0% | 1.2% |
| 52 and higher | 391 | 48.6% | 1.5% |
Data: nflverse games.csv bundled with this site (7,276 completed games, 1999–2025), computed by the author.
Note the mild tilt: high totals lean slightly under, low totals slightly over. If you want a folk rule that survives the data, it is “the market stretches a little at the extremes” — but 48–50.4% everywhere means no row clears the 52.4% break-even. The famous “public loves overs so bet unders” play would have won 50.6% of decisions since 1999: real direction, still a losing bet after juice.
The adjustment tables above give rules of thumb; the game log lets us check the two big ones directly:
| Condition | Games | Avg closing total | Avg actual points | Over hit |
|---|---|---|---|---|
| Outdoors, wind ≥15 mph | 684 | 41.8 | 40.4 | 43.0% |
| Outdoors, wind <10 mph | 3,309 | 43.1 | 44.2 | 51.2% |
| Dome (closed, fixed) | 1,194 | 45.7 | 47.0 | 52.0% |
| Outdoors, all | 5,333 | 42.8 | 43.3 | 48.7% |
Data: nflverse games.csv bundled with this site (completed games with recorded roof/wind, 1999–2025), computed by the author. Over rates exclude pushes.
Two things are true at once. Wind really does kill scoring — windy games produced 3.8 fewer points than calm ones — and the market already knows, pricing windy totals 1.3 points lower before kickoff. Historically that discount wasn't quite enough (unders hit 57.0% in the windy sample), making wind the strongest environmental lean in this dataset; but 684 games is a modest sample, forecasts are public, and modern books shade harder than they did in 2005. Domes run the other way: about a point more scoring than the line expected, and overs at 52.0% — right at the edge of noise.
Spreads have 3 and 7; totals have a softer set. The most common combined scores since 1999, with the share of all games landing exactly there:
| Final total | 37 | 41 | 51 | 44 | 43 | 40 | 47 | 33 |
|---|---|---|---|---|---|---|---|---|
| Share of games | 3.9% | 3.8% | 3.8% | 3.7% | 3.5% | 3.3% | 3.2% | 3.0% |
Data: nflverse games.csv bundled with this site (7,276 completed games, 1999–2025), computed by the author.
The peaks are flat — nothing like the 15.1% tower at a 3-point margin — which is why buying half-points on totals is rarely worth a surcharge. Whole-number closing totals landed exactly (pushed) just 2.9% of the time, versus 4.8% for whole-number spreads. Scoring context matters too: league scoring rose from 41.9 points per game in 1999–2004 to 45.8 in 2017–2022 and eased to 45.1 since 2023, so any “historical” totals rule needs to survive an environment shift of four points.
Keep reading: how accurate is the over/under, weather & scoring, the league scoring trend, and stadium scoring.
Nothing here is betting advice, and no number on this page predicts any single game. Sports betting is legal only in some jurisdictions and only for adults (21+ in most U.S. states). If betting stops being entertainment, call or text 1-800-GAMBLER. Read our full disclaimer.