What −3.5 actually asks a team to do, why both sides cost −110, and how pushes, hooks and line moves decide spread bets.
The spread is a handicap, not a prediction
A point spread takes a mismatched game and manufactures a coin flip. When the Chiefs are −3.5 against the Raiders, the market is not predicting a 3.5-point game — it is saying that if you subtract 3.5 points from Kansas City's final score, the two teams become equally likely to "win" the adjusted game. Bet the favourite at −3.5 and they must win by 4 or more to cover. Bet the underdog at +3.5 and you win if they lose by 3 or fewer — or win outright.
That last clause matters more than newcomers expect. A spread bet on the dog is two bets in one: the upset, and the respectable defeat. It is why underdog backers celebrate meaningless late touchdowns and why a garbage-time field goal can be worth exactly as much as the game's opening drive.
Reading a line: −3.5 (−110)
Every spread quote carries two numbers, and they answer different questions. The first — the spread itself — is the handicap. The second, usually −110, is the price: what you must risk to win 100. The spread tells you what has to happen; the price tells you what it costs to find out.
At −110 on both sides, you need to win 52.4% of your spread bets just to break even — that gap between 50% and 52.4% is the bookmaker's vig. The two numbers also move independently: a book that wants more underdog money might improve the price to −105 before it touches the spread. Half the craft of betting spreads is noticing which number moved.
Pushes and the hook
Whole-number spreads can tie. Lay −3 in a game the favourite wins by exactly three and the bet pushes — stakes returned, nobody paid. Books avoid this outcome, and sharper bettors court it: a push is a free look at a bet that might have lost.
The half point that prevents the push is called the hook, and in the NFL it is not decoration. The difference between −2.5 and −3.5 spans the single most common margin of victory in the sport. Two bets that look nearly identical on a betslip can have meaningfully different win probabilities — which is why the next guide in this series is entirely about the numbers 3 and 7.
Why the spread exists at all
A moneyline on a heavy favourite is a dull product: risk 400 to win 100 bores most bettors. The spread turns every game into a near even-money proposition, which is better entertainment and better business. The book's ideal world is balanced action at −110 on both sides, collecting the vig with no opinion of its own.
Reality is messier — books carry positions constantly — but the principle explains line movement. A spread moves when the market's money and the sharpest accounts disagree with it, not because an algorithm changed its mind about the teams. By Sunday morning the number has absorbed a week of information: injuries, weather, and the opinions of bettors whose limits are far higher than yours. That closing number is the most accurate public forecast of an NFL game that exists.
What a spread bettor actually shops for
Because covering is priced like a coin flip, small edges are the whole game. The same match might be −3.5 (−110) at one book and −3 (−108) at another — the second quote is simply a better bet on the favourite, no analysis required. Line shopping across books, and knowing when a half point is worth paying for, compound quietly over a season.
Our model plays the same game from the other side: it projects each match as a score distribution, converts that into cover probabilities, and compares them with the de-vigged market number. When the two disagree by enough, the gap shows up as an edge on Best Bets — a lean against the market's coin flip, never a lock.
Put it to work
- Odds Board Every book's spread side by side, best line highlighted.
- Game Projections The model's score, spread and total for every game.
- Best Bets Where the model's cover probability disagrees with the market.