THE BETTING FOOL
By a Man Who Should Know Better
Here is the question I ask every football bettor sooner or later, usually right after he tells me how much he knows. Of all the NFL games ever played, in what percentage does the team that wins the game also cover the spread? Do not look it up. Pick a number and write it down. I will wait.
If you wrote down fifty, you are in respectable company. The reasoning goes that a point spread exists to make the two sides equal, so the team that wins the game should cover about as often as the team that loses it. It is a lovely theory. It is wrong by thirty-three points. Over twenty seasons the team that won the game also covered the spread 83.2 percent of the time. That is five games out of six.
I counted every regular-season game from 2006 through 2025 that had a posted spread and a winner: 5,181 of them, using the closing line. Pushes are set aside, which is the honest way to do it, and if you count a push as a miss instead the answer is 81.1 percent. Either way you are nowhere near fifty.
Here is what I like about this one: it is not really a statistical finding. It is arithmetic. A team that wins the game covers the spread in exactly two situations. Either it was the underdog, in which case it won the game and therefore beat any number of points it was getting. Or it was the favorite and won by more than the spread. There is precisely one way for a winner to fail to cover, and it is that the favorite wins, but by less than the number. That is the whole list.
Now put the league's numbers in. Favorites win about two games in three — 66.5 percent here — which means underdogs win the other 33.5, and every one of those is a cover, whether the dog was getting a field goal or three touchdowns. Of the favorite wins, 47.7 points of the total went on to cover, 16.3 did not, and 2.5 landed on the number and pushed. Add the underdog wins to the favorite covers and you get a hair over 81 percent of all games, and 83.2 once you take the pushes out of the pile. Nobody had to pick anything.
Figure 1 · Where every game lands
5,181 regular-season games, 2006–2025, closing point spread. The first, second and fourth segments are games where the winner covered or pushed. Only the red one is a winner who failed.
Winner covers = first two segments ÷ (all games − pushes) = 83.2%.
So how does a spread that splits the money fifty-fifty produce a winner who covers five times in six? Because the spread splits the bettors, not the winners. In this sample favorites covered 48.9 percent and underdogs 51.1, which is about as close to even as a market gets. But look at who is inside that underdog number. 67.2 percent of all underdog covers — two out of every three — were underdogs that won the game outright. The two groups overlap so heavily that the spread and the scoreboard are telling you the same story nearly every Sunday.
You might suspect I picked a lucky stretch. I did not. In the twenty seasons from 2006 to 2025 the number never dropped below 77.3 percent (in 2020, the season of empty stadiums and scrambled schedules) and never rose above 87.9 percent (in 2016). The typical season lands right around 83. And the current one is behaving: through Week 4 of 2026, the winner covered in 50 of 59 games that did not push, or 84.7 percent.
Here is where the question gets useful, because the 83.2 is an average over very different games. It depends almost entirely on how big the spread is. When the line is three points or less, the winner covers 95.7 percent of the time; there is hardly any room between “the favorite wins” and “the favorite covers.” When the line is ten or more, the winner covers only 62.0 percent. Favorites win 86.1 percent of those games and still fail to cover 36.9 percent of the time, because winning is the easy part and winning big is the part they charge you for.
Figure 2 · The bigger the spread, the less the winner covers
Share of games in which the outright winner also covered, pushes excluded.
| Spread | Games | Favorite wins | Fav. wins, no cover | Winner covers |
|---|---|---|---|---|
| 3 or less | 1,881 | 55.1% | 4.1% | 95.7% |
| 3.5 to 6 | 1,491 | 64.9% | 15.3% | 84.5% |
| 6.5 to 9.5 | 1,124 | 76.0% | 25.5% | 74.1% |
| 10 or more | 685 | 86.1% | 36.9% | 62.0% |
| All games | 5,181 | 66.5% | 16.3% | 83.2% |
Same 5,181 games. “Fav. wins, no cover” is a share of all games in the row, pushes included in the base.
The individual numbers on the board tell the same story. At a one-point spread, a winner covers every time the game does not push: there is nowhere for a favorite to win without covering except by exactly one, and that is a push. At three, the winner covers 93.1 percent of the time (781 games, the most common line in the sample). At a touchdown it is 76.8. At ten, 61.6. Each point you add to the spread opens up another margin of victory that is a win for the favorite and a loss for the bettor.
First, in a close game, you are picking a winner. If the line is three or less, the dog covers about as often as it wins or loses by a field goal at the worst, and the cover is almost never the point. In a pick-five pool, the question on a small spread is not “who covers” but “who wins,” and the football you need is the football that decides games: quarterbacks, turnovers, who can run the ball when everybody knows it is coming.
Second, the points are worth less than you think. Take a 2.5-point underdog. The winner covers 94.5 percent of those games, which means the only thing the 2.5 points buy you is the sliver of games in which the favorite wins by one or two. That is 5.5 percent of the time. The other ninety-four-odd percent, the dog covers because it won, or lost by more than the cushion. The cushion is real. It is just thin, and it is why I do not pretend that a +2.5 is anything other than a pick on the underdog to win.
Third, big favorites are a trap for a different reason. You can be right that the favorite wins, which is the easy part, and still lose the ticket. At ten points or more the favorite wins 86.1 percent and covers about 46. The gap between those two numbers is every game where the bettor said “sure thing” and the sure thing won by six.
Fourth, look at what the card is really asking. This week's column has three underdogs at +2.5 — the Bills, the 49ers and the Falcons — and in games with a number like that the winner covers about 94.5 percent of the time. So those three plays are really the question “can this team win the game,” with a little padding. Last week's five sides all covered, and they split three and two in this exact way: Jacksonville and Cleveland were underdogs that won the game outright, and San Francisco laid three and won by ten, while Tampa Bay lost by three and Miami lost by five and covered on the cushion. That is the normal mix, not a sign of anything.
This is not a system, and I would be a poor Fool if I let you walk out thinking it was. The 83.2 percent is not an edge. It is a description of what the spread is, and the market has already priced it. To make money you still have to pick the winner, or the cushion, better than a price that already knows all of this. Underdogs covered 51.1 percent in this sample, which sounds like a gift until you remember that a bet at minus-110 needs 52.4 to break even. Not a gift; a rounding error with a story attached.
What the number does give you is a way to read a bet. When you take a side, ask what has to happen. If the answer is “the favorite has to win by more than a touchdown,” you know you are asking for a lot. If the answer is “the underdog has to win, or lose by a hair,” you know you are mostly asking for a win. Call your shot accordingly.
How this was counted
Source: the nflverse games table. All regular-season games, 2006–2025 (5,199 games); 4 pick’em games (spread of zero) and 14 games that ended in a tie are excluded, leaving 5,181. The spread is the posted closing line. A game “pushes” when the favorite wins by exactly the spread; pushes are excluded from the headline figure and counted as misses in the alternate figure. Postseason games add 231 more games and move the answer by about six-hundredths of a point (83.19%). Season-to-date 2026 figures are Weeks 1–4.
“Bet less than you brag, learn more than you lose, and never trust a tout who smiles too much.”
— The Betting Fool
Disclaimer: The Betting Fool is for entertainment purposes only. Nothing on this site is financial advice, and past performance does not guarantee future results. Must be 21+ to wager where legal. If you or someone you know has a gambling problem, call 1-800-GAMBLER.