Sunday morning. Sixteen games on the sheet, sixteen confidence points to hand out, and the group chat is already arguing about whether a 3-point road favorite deserves your 14.

Most people rank by gut. Some rank by spread. Almost nobody knows how often each size of favorite actually wins.

We counted. Every regular-season game from 2006 through 2025, matched to its point spread and final score. Favorites won 66.4% of the time.

That is the average. The average hides almost everything that matters when you rank a sheet.

Key findings on win rates by spread

Twenty seasons, one steep curve

By the numbers

  • 66.4% — Straight-up win rate for all NFL favorites (3,447 of 5,195 games) (nflverse games file, 2006-2025 NFL regular seasons)
  • 54.9% — Win rate for favorites of 3 points or less (nflverse games file, 2006-2025 NFL regular seasons)
  • 80.7% — Win rate for favorites of 7 points or more (nflverse games file, 2006-2025 NFL regular seasons)
  • 92.0% — Win rate for favorites of 14 or more (160 of 174) (nflverse games file, 2006-2025 NFL regular seasons)
  • 85.9% — Win rate for the biggest favorite of the week (345 weeks) (nflverse games file, 2006-2025 NFL regular seasons)
  • 5.0 — Underdogs who won outright in an average week (nflverse games file, 2006-2025 NFL regular seasons)

If you build your NFL confidence picks off the spread, here is what 20 seasons of results say:

  • Favorites of a field goal or less won 54.9% of the time. Barely better than a coin flip.
  • Favorites of 3.5 to 6.5 points won 65.7%.
  • Favorites of 7 or more won 80.7%, and still lost 298 times.
  • One half point, from 3 to 3.5, moved the win rate from 56.9% to 62.8%.
  • Ranking every favorite by spread earned 87.5 points in an average week. The same picks in random order would have averaged 80.8.
  • Only 1 of 345 weeks ended with every favorite winning.

How we built the sample

Real spreads, real scores, nothing estimated

The data is the nflverse games file (nflverse/nfldata on GitHub), a public dataset of NFL schedules, final scores and point spreads. We downloaded it on October 30, 2026. The script and the full game list are saved with this post, so every number can be rerun.

The sample is every regular-season game from 2006 through 2025 with both a final score and a point spread: 5,199 games. Four of them closed as pick'em games with no favorite, which leaves 5,195. Playoffs are out. So is the 2026 season, which is still being played.

A few definitions:

  • Spread: the file's spread_line field. nflverse's data dictionary describes it as the closing spread, sourced from Pro-Football-Reference.
  • Favorite: the team the spread favored, home or road.
  • Win rate: outright wins divided by all games. The 14 ties count as games the favorite did not win.
  • Against the spread: a favorite covers when it wins by more than the spread. Pushes (132 games) are reported separately and left out of cover rates.
  • Biggest favorite of the week: when two games shared the top spread, they split that slot's result. Same rule for every other slot.

Lines differ a little from archive to archive, so we checked. Against Sports Odds History, a second public line archive, the two sources matched exactly on 61% of these games, and most of the rest were a half point apart. Run on that archive's lines, the overall favorite win rate was the same 66.4%, and each of the three spread groups moved by less than one percentage point.

Spot checks against Pro-Football-Reference also turned up a few 1-point games where the file has the wrong team favored. Fixing those moves none of the headline figures by more than half a point.

Seasons ran 16 games per team through 2020 and 17 since 2021, which is why recent years carry more games. We also reran the three spread groups on 2016-2025 alone. Each one landed within about one percentage point of the 20-season figure.

If you want a refresher on reading numbers like these, start with statistics for smarter NFL picks. For team-by-team records, with splits like home, away and underdog, TeamRankings keeps NFL win trends data.

Win rates by NFL point spread

Every half point buys a little more trust

Here is the full NFL point spread ladder. Read it top to bottom and the pattern is obvious: the bigger the NFL spread, the more often the favorite wins. Read it closely and you see the curve is not smooth.

NFL favorites' straight-up record by point spread, 2006-2025 regular seasons

SpreadGamesW-L-TWin rate
1 to 2.51,104591-509-453.5%
3782445-336-156.9%
3.5505317-187-162.8%
4 to 6.51,228822-404-266.9%
7331253-75-376.4%
7.5 to 9.5559429-128-276.7%
10 to 13.5512430-81-184.0%
14 or more174160-14-092.0%

Favorite win rate by point spread

Favorite win rate by point spreadValue
1 to 2.553.5%
356.9%
3.562.8%
4 to 6.566.9%
776.4%
7.5 to 9.576.7%
10 to 13.584.0%
14 or more92.0%

Source: nflverse games file, 2006-2025 NFL regular seasons

Small favorites

A coin with a slight lean

Favorites of 1 to 3 points went 1,036-845-5. That is a 54.9% win rate across 1,886 games.

One-point favorites won 52.1% of 466 games. Exactly-3 favorites, the most common line in the sample, won 56.9% of 782.

A field-goal favorite feels like a favorite. Over 20 seasons it played like a coin that lands heads a little more often than tails.

For your sheet, that means these games belong at the bottom. Missing one costs you a few points. Missing one you ranked 14th costs you the week.

Mid-range favorites

The 3.5 line changes the math

Favorites of 3.5 to 6.5 points went 1,139-591-3, a 65.7% win rate across 1,733 games.

The first half point does a lot of work. Three-point favorites won 56.9%. At 3.5 the rate climbed to 62.8%, across 505 games. By 6.5 it reached 72.0%.

This is the crowded middle of every confidence sheet. These teams should win about two of every three. Treat them like it, not like sure things and not like toss-ups.

Heavy favorites

Strong, never safe

Favorites of 7 or more went 1,272-298-6. That is an 80.7% win rate across 1,576 games.

The rate kept climbing with the number. Seven-point favorites won 76.4%. Favorites of 10 to 13.5 won 84.0%. Favorites of 14 or more won 92.0%, losing 14 times in 174 games.

Read that last one again. Even the biggest favorites in football dropped about one game in twelve. The week's biggest favorite failed to win 14.1% of the time in our sample. Across an 18-week season, that is two or three Sundays where your top pick scores zero.

What win rates mean for confidence points

Spread order beats gut order most weeks

A confidence pool pays you the points you assigned to every pick you get right. So the value of a pick is its win rate times its points.

Put 16 on a 14-point favorite and history says you collect about 14.7 points on average (92.0% of 16). Put the same 16 on a 3-point favorite and the average drops to about 9.1 (56.9% of 16). Same slot, same team you think will win, very different expected payoff.

That is the whole case for ranking by spread. With nfl straight up picks you only need the winner. The spread tells you how hard to lean on it.

We tested it. For every week from 2006 to 2025, we picked the favorite in every game and gave the most confidence points to the biggest spread, down to 1 point for the smallest. Then we compared that order with the same picks in random order.

Picking every favorite, three ways to order the sheet (average week, 2006-2025)

OrderingPoints per weekShare of max
By spread, biggest favorite on top87.571.7%
Random order, same picks80.866.3%
Perfect hindsight, same picks104.585.8%

Spread order beat the random-order average in 275 of 345 weeks. It earned about 6.7 more points per week, with the exact same winners picked.

The slot-by-slot record shows why. The week's biggest favorite won 85.9% of the time. Second biggest, 81.1%. Third, 77.5%. At the other end, the pick in the smallest-spread game won 51.9%.

The hindsight row is the ceiling. Nobody gets there. The 17-point gap between spread order and hindsight is where your own read lives: injuries, weather, a team resting starters. Tools like power rankings in confidence pools help you decide which games to move up or down from the spread's starting order.

NFL favorites vs the field in a confidence pool

Same winners, different order, different week

NFL favorites win about two of every three games. That sounds like a plan. It is also everyone else's plan.

In an average week of our sample, about 15 games had a favorite. Favorites won 10.0 of them. Underdogs won 5.0.

Clean sweeps almost never happen. Only one week in 345 ended with every favorite winning: 2017 Week 15, when favorites went 16 for 16. At the other extreme, underdogs won 11 of 14 games in 2015 Week 10. In 204 of 345 weeks, five or more underdogs won outright.

Underdog wins per week, number of weeks (2006-2025)

Underdog wins per week, number of weeks (2006-2025)Value
0 upsets1
1 upset6
2 upsets22
3 upsets44
4 upsets68
5 upsets64
6 upsets70
7 upsets40
8 upsets14
9 upsets13
10+ upsets3

Source: nflverse games file, 2006-2025 NFL regular seasons

Here is what that means in an NFL confidence pool. If most of your group takes the favorites, picking the same winners keeps you level with them. It does not move you ahead.

You gain ground in two places. The first is the order, since the record above shows spread order beats a random one most weeks. The second is the bottom of the sheet, where small favorites win barely more than half the time and an underdog pick risks only a few points.

That second spot is where NFL team analysis pays off. A 1-point favorite won 52.1% of the time over 20 seasons. Call it a coin flip. If you have a real reason to like the other side, the low-point slot is the cheap place to act on it.

Use the data in your confidence pool picks

Rank by spread, then make it yours

Turn the numbers into confidence pool picks with a short routine:

  • Start from the spread. Sort the week from biggest favorite to smallest and use that as your first draft.
  • Top points to double-digit favorites. Favorites of 10 or more won 84.0% or better. They earn the slots that hurt most to miss.
  • Respect the 3.5 line. A 3.5-point favorite won 62.8%, a 3-point favorite 56.9%. Rank them that way.
  • Keep 3-and-under at the bottom. These are near coin flips. Low points limit the damage.
  • Pick your spots. If you go against a favorite, do it in a low-point game, not at 15 or 16.
  • Check the line before your deadline. Spreads move during the week. Rank off the latest number you can see.

None of this replaces knowing the teams. It gives your nfl confidence pool picks a floor, so your gut only has to move a few games instead of building the whole sheet.

A good confidence pool strategy is boring at the top and brave at the bottom.

Tip

Straight-up pools reward this data most. Against the spread, favorites covered only 48.8% of the time in our sample (pushes excluded), so the spread's edge disappears once it becomes the scoring line.

Want to run your own? A Splash Pick'Em game lets you pick every matchup straight up or against the spread, with spreads pulled by Splash or entered by the commissioner. Splash has also announced a Confidence format for Pick'em, where you rank every pick from most to least confident and a correct pick scores its rank value.

You play against your friends and other fans, not the house. Entries, prizes and standings are handled for you, and live scoring shows exactly where your points went. Availability varies by state and province, so check where you can play first.

Sort by spread. Move a few games. Send your sheet.

Frequently asked questions

Quick answers on spreads & confidence points

In NFL regular seasons from 2006 to 2025, favorites of 1 to 3 points won 54.9% of the time (1,036 of 1,886 games, with 5 ties). Exactly-3 favorites won 56.9%, and 1-point favorites won 52.1%. That is why these games usually belong near the bottom of a confidence sheet.

Usually, yes. Across 345 weeks from 2006 to 2025, the week's biggest favorite won 85.9% of the time, and favorites of 14 or more won 92.0%. They still lose sometimes. The top slot holds the most likely winner on the sheet, never a sure one.

They do different jobs. Straight-up picks decide who you take, and the spread tells you how much to trust each one. In our 2006-2025 sample, picking every favorite and ranking by spread earned 87.5 points a week, versus 80.8 for the same picks in random order. Spread order beat the random-order average in 275 of 345 weeks.

Slightly. From 2006 to 2025, home favorites won 67.2% of 3,306 games and road favorites won 64.8% of 1,889. The spread already accounts for where the game is played, so the bigger signal is still the size of the number, not the location.

About five. In the average regular-season week from 2006 to 2025, underdogs won 5.0 games outright and favorites won 10.0. Only one week in 345, 2017 Week 15, ended with every favorite winning, and in 2015 Week 10 underdogs won 11 of 14 games.

More from the Splash blog