Home-Field Advantage Is Shrinking. The Point Spread Noticed.

Home-field advantage is one of the oldest assumptions in football, and the number attached to it is usually three points. Look at 24 seasons of results and two things are true at once: home field is worth noticeably less than it used to be, and the point spread has already adjusted for it.

There is one exception, and it sits exactly at that traditional three-point number.

Everything below uses the regular season only, with neutral-site games removed: the 62 international games plus the Bills' six Toronto games (2008-2013), 68 in all. A game in London has a home team on paper and not in the stands.

Home teams win less often than they used to

Start with the plainest measure: how often the home team won the game outright.

For most of the 2000s and 2010s, home teams won about 57% of the time and by about two and a half points on average. Since 2021 they have won 54% of the time, by about two points.

2020 is the one natural experiment football has ever run on this. With crowds absent or capped for most of the season, home teams went 127-128-1 straight up and won by an average of five hundredths of a point. Whatever home field is made of, the crowd is a large part of it.

Crowds came back in 2021. The full advantage did not.

The spread moved with it

The point spread carries a market estimate of home field inside every number. Across a full season every team plays at home and on the road, so team quality largely washes out of the average, and what is left is roughly the home-field credit the market is giving.

The credit fell from about two and a half points to about one and three-quarters, and the home team's record against the spread barely moved: between 48% and 50% in every era. In 2020 the market cut the credit to just over a point within the season itself, and home teams still landed at 49.8% against the spread.

That is what an efficient number looks like. The underlying advantage shrank by roughly half a point and the market took almost a full point out of its estimate, but the outcome it is trying to predict, whether the home team beats the number, stayed near a coin flip throughout.

Pooled over 2002-2025, home teams matched or beat the spread 48.8% of the time across 6,161 games (2,926-3,070-165). That is slightly short of 50%, and consistent in direction, but it is not a strong result: a gap that size would turn up by chance roughly one time in sixteen. We would describe it as a lean, not a finding.

Home favorites against road favorites

The usual next step in home-field analysis is a split by spread size. Each row is how the favorite did against the spread, with the home and road sides kept apart.

Twelve cells is a lot of chances for one to look unusual by accident. So before reading anything into this table, we split the data into two halves, 2002-2013 and 2014-2025, and asked which patterns showed up in both.

Most of the table is noise. Two rows are not.

The three-point home favorite

A home team favored by exactly three points has matched the spread 43.9% of the time: 240 games ahead of the number, 315 behind it, 56 landing on it exactly. A road team favored by the same three points sits at 50.3%.

It holds in both halves of the data (44.0% and 43.7%), and in every six-season block since 2002:

The road column swings back and forth around 50%, which is what noise looks like. The home column never gets there. And this one survives the multiple-comparison problem: even allowing for the twelve cells we looked at, a gap this size would turn up by chance less than one time in fifty.

The mechanism is not subtle. Home teams favored by three lost the game outright 45.8% of the time (280 of 611). Road teams favored by three lost outright 41.6% of the time (189 of 454). Straight up, the three-point home favorite won 54.2% of its games against 58.4% for the three-point road favorite. These are home teams the market rated as better than their opponent once home field was included, and they lost almost as often as they won.

One reading, and we would put it no more strongly than this: a three-point home favorite is often a game the market sees as close to even, with home field supplying most of the margin. If home field is worth less than the number assumes, those are exactly the teams that would come up short most often. The data is consistent with that. It does not prove it.

What it does show is that the pattern has not faded as home field shrank. The most recent block, 2020-2025, is the weakest of the four.

The big road favorite: worth watching, not established

The other row that stands out is the large road favorite. Road teams favored by 7.5 or more matched the spread 43.3% of the time (132-173-3), and the result is low in both halves of the data.

But it does not clear the same bar. It rests on 308 games, and in one of the four six-season blocks (2014-2019) it ran the other way at 54.3%. Grouping 7.5-9.5 and 10-plus together is also a choice we made after seeing the table, which flatters it. We are listing it because it is the second-largest gap in the table and it is consistent in direction, not because we think it is real.

And 2026?

Through Week 7, home teams are 51-49-1 straight up with an average margin of -0.07 points, almost exactly the 2020 numbers. Against the spread they are 42-54-5, or 44.1%.

It is tempting to read that as home field disappearing. It is 101 games, and a 101-game stretch at 44% is well within what an ordinary season can produce by chance. The market is crediting home teams about 1.75 points this season, the same as the four seasons before it. If this keeps up for another two months it becomes interesting. Seven weeks is not enough.

What this adds up to

1. Home field is real, and it is smaller than it was. Home teams win about 54% of their games straight up now, against 57% for most of the previous two decades, and the average margin is closer to two points than two and a half.

2. The point spread tracked it. The home credit built into the spread fell by about a point, and home teams have matched the spread between 48% and 50% of the time in every era. As a whole, home field is priced.

3. The exception is the three-point home favorite. At 43.9% against the spread over 611 games, low in every six-season block, it is the one corner of home-field analysis that has held up under every check we ran. The number most people still associate with home field is the one spot where home field seems to deliver less than it promises.

This is the same pattern as our look at the point spread itself: the market absorbs the obvious things. The interesting question is never whether home field matters. It is whether the spread already knows how much.

The honest caveats

• Regular season only, neutral sites excluded. Playoff hosting is decided by seeding, which mixes home field with team quality, and neutral sites have no real home team. 68 regular-season games were removed: the 62 international games and the Bills' six designated home games in Toronto, 2008-2013.
• "Home credit in the spread" is an average, not a published figure. The spread does not report its home-field component. Averaging the home team's spread across a full season is a standard approximation, because every team plays at home and on the road, but it is an estimate.
• Closing spreads only. Nothing here looks at opening numbers or how they moved.
• Pushes count as half in every percentage, the same rule the site uses. The raw records are in the tables and the backing file.
• The 2020 season is not a clean no-crowd sample. Several stadiums admitted limited crowds during the season. It is still the closest thing to a controlled test the sport has.
• We looked at twelve spread bands. The three-point home favorite is the only cell that survives a multiple-comparison correction and holds in every block; the big road favorite does not, and we say so.

See it yourself

• Spreads: results broken down by the size of the spread, season by season
• League Trends: home teams' record against the spread and the season's average home margin
• Standings: every team's 2026 record, straight up and against the spread
• Predictions: our graded record, published in full, wins and losses alike