About one NFL game in four ends with a margin of exactly 3 or exactly 7. Nothing else about football betting matters as much as that sentence. The spread is a bet on the margin, the margin piles up on two numbers, and a half point on either side of those numbers is worth more than any other half point on the board. If you bet NFL sides and do not shop around 3 and 7, you are leaving money on the table every week.
The Numbers From 7,276 Games
Every regular season and playoff game from 1999 through 2025 comes to 7,276 results. Here is how often the final margin landed on each number.
- 3 points: 15.1% of all games
- 7 points: 9.0%
- 6 points: 6.0%
- 10 points: 5.5%
- 4 points: 4.9%
- 14 points: 4.9%
- 1 point: 4.2%
- 2 points: 4.1%
- 8 points: 3.7%
- 5 points: 3.6%
- Ties: 0.2%
Three is the most common margin by a wide gap. It shows up more often than 7 and 6 put together. The drop from 3 to 7 is steep, and the drop from 7 to everything else is steeper. Those two numbers are the whole reason the phrase “key number” exists.
The share has moved a little over time. From 2013 through 2019 exactly 3 happened in 13.9% of games, and from 2020 through 2025 it was back up to 15.2%. Three has never stopped being the number.
Why a Field Goal and a Touchdown Keep Showing Up
The reason is how close games end. A team down by 3 in the last two minutes plays for the tying field goal, not the touchdown. A team up by 3 with the ball kneels. Overtime games end on a field goal more often than a touchdown. All of that funnels close results onto 3.
Seven is the same story one score up. A team down 7 drives for the touchdown and kicks the extra point to tie, so a game that was 7 stays 7 or goes to overtime. A team up 7 with a late lead rarely needs more.
The extra point moved back to a 33 yard kick in 2015, and more of them miss now. That is why 6 and 8 have picked up some share since then. It has not dented 3 or 7 enough to change how you should bet.
What a Half Point Is Worth at 3
A standard spread bet at -110 needs to win 52.4% of the time to break even. Now look at a favorite at -3. In 15.1% of games the favorite wins by exactly 3 and your ticket pushes. Move that same bet to -2.5 and every one of those pushes becomes a win. Move it to -3.5 and every one becomes a loss.
That is a swing of roughly 15% of all outcomes hanging on half a point. Compare that with moving from -4.5 to -5, where the number in between, 5, lands only 3.6% of the time. The half point at 3 is worth about four times the half point at 5. Books know this, which is why the price to buy through 3 is always steeper than the price to buy through any other number, and why some books will not sell it at all.
Seven works the same way at a smaller scale. Going from -7 to -6.5 turns 9% of outcomes from pushes into wins. Worth paying for. Going from -7.5 to -7 turns 9% of losses into pushes. Also worth paying for, at the right price.
Shopping Lines Around 3 and 7
The practical rule is simple. When you like a favorite, you want to be on the right side of the key number, which means -2.5 rather than -3 and -6.5 rather than -7. When you like an underdog, you want +3 rather than +2.5 and +7 rather than +6.5. A book that is offering the better side of 3 at -115 is usually a better bet than a book offering the worse side at -105, because the 15% of games that land on 3 are worth more than ten cents of juice.
The one place reduced juice does matter is when both books have the same number. A -3 favorite at -105 needs to win 51.2% of the time instead of 52.4%. SportsBetting.ag runs a standing -105 on NFL sides, which is the cheapest way we know of to bet a key number when you cannot get off it.
The other move is the moneyline. If a favorite sits at -3 at every book you use, check the moneyline price. On a 3 point favorite a win by 1 or 2 pays the moneyline in full and pushes the spread, and the moneyline is often a better bet than paying 20 or 25 cents to buy the half point.
Where Key Numbers Stop Mattering
Once a spread gets past 10, the margins spread out and the advantage of any one number fades. Ten lands 5.5% of the time and 14 lands 4.9%, so they matter, but a half point around 12 or 16 is worth about a third of what it is worth at 3.
Totals do not have a number like 3 at all. No total lands on one figure anywhere close to 15% of the time, so the half point on a total is cheap to buy and rarely worth it. Spend your line shopping effort on sides, where it pays.
The last caution is the push on the number itself. Betting -3 at -110 is a bet with a 15% chance of giving your money back, which lowers your variance but also lowers your edge if you were right. If you have a real opinion and the book is offering -2.5 somewhere, that is the number.
The Books That Make This Easier
Key number shopping only works if you hold accounts at more than one book, because the half point you need is never at the same place twice. These books sit offshore and outside US state regulation, so whether you can use them depends on where you live. Of the five we cover, SportsBetting.ag is the one to open for the -105 NFL sides, and BetOnline posts its lines early enough that you can often get a 2.5 or 6.5 before the market settles on the key number. Compare the rest in our sportsbook reviews before you fund anything.