NFL Against the Spread: Do Favourites or Underdogs Cover More?
Underdogs edge favourites against the closing spread. The real story sits in the key numbers, in road favourites, and in what each side actually paid.
Ask a room of NFL bettors whether to back favourites or underdogs and you will hear loyalty, not evidence. I wanted evidence. This study covers 7,245 games with a closing spread from 1999 to 2025, regular season and playoffs. It asks who covers, how that changes with the size of the line, and whether venue matters.
The headline answer is that underdogs covered 51.3% of games that did not push. That is a real lean, but it sits below the 52.4% you need at standard -110 pricing, so it does not make you money on its own. The more useful findings sit underneath, and they shape how I build my NFL betting tips page.
How the key numbers of three and seven swallow bets
This chart splits every game by the size of the closing spread. It shows the share the favourite covered, the share that pushed and the share the underdog covered. Across the whole sample 2.7% of games landed exactly on the number, but that average hides the shape: favourites of exactly -3 pushed 9.0% of the time and favourites of exactly -7 pushed 6.5%.
Away from those two numbers the underdog side edges ahead almost everywhere. Favourites of -1 to -2.5 covered 49.8% of non-push games. The biggest favourites, -14 or more, covered just 47.3% across 215 games.
View the data: cover, push or fail by size of spread
| Favourite covered | Push | Underdog covered | |
|---|---|---|---|
| -1 to -2.5 | 49.4% | 0.8% | 49.8% |
| -3 | 43.3% | 9.0% | 47.7% |
| -3.5 to -6.5 | 48.6% | 0.9% | 50.5% |
| -7 | 47.0% | 6.5% | 46.5% |
| -7.5 to -9.5 | 47.3% | 0.0% | 52.7% |
| -10 to -13.5 | 46.9% | 2.5% | 50.6% |
| -14 or more | 45.6% | 3.7% | 50.7% |
For me the lesson is about price as much as side. Half a point either side of 3 or 7 changes the result of a bet far more often than half a point anywhere else, so that is where I shop hardest.
Road favourites held up better than home favourites in mid-range spreads
Here the favourite cover rate is split by venue, with pushes excluded and neutral sites removed. Overall, home favourites covered 48.3% across 4,692 games and road favourites 49.5% across 2,459. That means home underdogs covered 50.5% and road underdogs 51.7%.
The gap opens most in the -3.5 to -6.5 band, where road favourites covered 51.4% against 47.9% for home favourites. At exactly -3 the split is starker still: 44.6% at home against 51.1% on the road. I read that as the market charging slightly too much for home field in short and medium spreads.
Neutral-site games excluded. Road-favourite rows at -10 or more are small samples.
View the data: home vs road favourites against the spread
| Home favourite | Road favourite | |
|---|---|---|
| -1 to -2.5 (776 home / 647 road) | 49.2% | 50.3% |
| -3 (647 home / 488 road) | 44.6% | 51.1% |
| -3.5 to -6.5 (1,593 home / 816 road) | 47.9% | 51.4% |
| -7 (332 home / 154 road) | 51.0% | 49.7% |
| -7.5 to -9.5 (610 home / 193 road) | 48.9% | 42.0% |
| -10 to -13.5 (555 home / 128 road) | 49.7% | 40.8% |
| -14 or more (179 home / 33 road) | 48.3% | 40.6% |
The large road-favourite rows drop to 40.8% and 40.6%, but they rest on 128 and 33 games. Several road bands sit under 200 games, so I treat gaps of a few points there as noise until the tape gives me a reason to believe them.
What backing each side actually paid at closing prices
Cover rates only matter once you put a price on them. This chart shows the return per 100 staked at the recorded closing spread odds from 2006, across 5,290 games. Backing every favourite returned โ5.1% and backing every underdog returned +0.2%. The sharpest contrast is at three: underdogs of +3 returned +6.2%, while -3 favourites posted the heaviest loss of any band on the chart.
Venue shows up again. Home underdogs returned โ1.7% and road underdogs +1.4%, while underdogs of +10 to +13.5 returned โ1.5%.
Pushes refunded. Historical return, not a forecast.
View the data: what the spread paid
| Back the favourite | Back the underdog | |
|---|---|---|
| -1 to -2.5 | โ5.2% | +0.3% |
| -3 | โ10.2% | +6.2% |
| -3.5 to -6.5 | โ4.2% | โ0.8% |
| -7 | โ0.1% | โ4.6% |
| -7.5 to -9.5 | โ4.3% | โ1.0% |
| -10 to -13.5 | โ3.5% | โ1.5% |
| -14 or more | โ6.9% | +1.6% |
| All games | โ5.1% | +0.2% |
Now the honest caveat. The recorded consensus prices carry a thinner margin than most bookmakers charged before 2023, which flatters these returns. Had every underdog since 1999 been backed at a flat -110, the return would have been โ2.1%, so the +0.2% is break-even at a friendly price, not an edge to rely on.
Season by season, the underdog lean is real but streaky
The final chart tracks the underdog cover rate each season. Underdogs covered more than half of non-push games in 20 of the 27 seasons, which is consistent rather than dramatic. The range is what stands out: a high of 56.4% in 2006 came straight after a low of 41.3% in 2005.
With a 95% range of ยฑ1.2 points on the overall rate, single seasons swinging that far show how much noise one campaign carries.
View the data: underdog cover rate by season
| Underdog covered | |
|---|---|
| 1999 | 53.1% |
| 2000 | 52.6% |
| 2001 | 52.9% |
| 2002 | 55.8% |
| 2003 | 50.2% |
| 2004 | 50.2% |
| 2005 | 41.3% |
| 2006 | 56.4% |
| 2007 | 49.2% |
| 2008 | 51.0% |
| 2009 | 51.4% |
| 2010 | 50.8% |
| 2011 | 51.2% |
| 2012 | 51.9% |
| 2013 | 46.7% |
| 2014 | 52.1% |
| 2015 | 54.1% |
| 2016 | 47.5% |
| 2017 | 48.0% |
| 2018 | 54.3% |
| 2019 | 53.3% |
| 2020 | 55.8% |
| 2021 | 53.4% |
| 2022 | 54.7% |
| 2023 | 48.0% |
| 2024 | 47.3% |
| 2025 | 51.4% |
This is why I never overreact to a hot or cold start. One season can run several points either way, and the long-run lean is narrower than any single year suggests.
What I take from this
- Respect three and seven. Favourites of -3 pushed 9.0% of the time and -7 favourites 6.5%. Line shopping around those numbers matters more than picking a side blindly.
- Lean road, not home. Road underdogs covered 51.7% and returned +1.4%, while home underdogs returned โ1.7%. Home field looks slightly overpriced in short spreads.
- Blind underdogs are not enough. Backing every underdog returned +0.2% at friendly consensus prices but โ2.1% at a flat -110. The lean is a starting point for matchup work, not a system.
Questions from readers
Do underdogs cover the spread more than favourites in the NFL?
Slightly. Across 7,245 games from 1999 to 2025, underdogs covered 51.3% of non-push games, which is still below the 52.4% needed to break even at -110.
How often does an NFL game push on the spread?
Across all games, 2.7% landed exactly on the closing spread. The rate jumps at the key numbers: 9.0% for favourites of -3 and 6.5% for favourites of -7.
Is backing road underdogs profitable against the spread?
At recorded closing prices from 2006, road underdogs returned +1.4% and home underdogs โ1.7%. Those consensus prices carry a lower margin than typical bookmaker odds, so real-world returns would likely be lower.
I'm Tariq Al-Harbi, based in Riyadh, and I write the NFL bets of the week at htftpredictions.com โ breaking down spreads, totals and the matchups that settle them.
View profile โ
