The Primetime Under Is Real — and It's Getting Stronger

You've heard the cliché: primetime games stay under the total. Most football clichés die the moment you check the data. This one doesn't — but the real story is more interesting than the cliché, because the effect isn't where you think it is, and it's growing rather than shrinking.

We pulled every NFL over/under result in the SpreadTrends database — every game since 2002 — and split it by broadcast window. Here's what 24 seasons actually say.

The all-time numbers: a lean, not a law

Across all games since 2002, closing totals are almost perfectly calibrated: 3,162 games finished over, 3,219 under (49.6% over, 50.4% under, excluding exact pushes). The market's expectation splits the league dead in half, exactly as it's designed to.

Primetime tilts that balance — modestly:

Window (2002–2025) Over–Under Under %
All games 3,162–3,219 50.4%
Monday Night Football 195–218 52.8%
Thursday Night Football 125–136 52.1%
Sunday Night Football 198–208 51.2%
All primetime 642–676 51.3%

Over the full 24-year sample, those are small deviations from the 50% norm — a percentage point or two. If the story ended there, "primetime unders" would be a rounding error, not a pattern.

It doesn't end there.

The last decade: the lean became a trend

Cut the same data to the last ten seasons (2015–2025) and the picture changes sharply:

Window (2015–2025) Over–Under Under %
All games 1,335–1,387 51.0%
Sunday Night Football 77–104 57.5%
Monday Night Football 78–101 56.4%
Thursday Night Football 74–81 52.3%
All primetime 292–340 53.8%

And in the most recent window (2020–2025) it's stronger still: 59.1% of SNF games finished under (55 of 93) and 58.8% of MNF games did (57 of 97). Even Thursday night crept up to 54.9%.

Put together: since 2015, the under side of Sunday and Monday night games combined is 205–155 — a 56.9% rate, against a league-wide baseline of 51.0%. Statistical trends this visible usually fade as the market adapts. This one has done the opposite.

2025: the acid test

The 2025 season leaned over league-wide — 146 overs to 135 unders (52.0%). If primetime unders were just riding a low-scoring era, the pattern should have washed out. Instead, Monday night games went 12–9 to the under (57.1%) in a season where the rest of the league leaned the other way. Sunday night was 10–9 to the under. The window effect survived a scoring environment that was actively working against it.

Why does this happen?

Two explanations fit the data:

1. Showcase games carry inflated scoring expectations. A primetime game is the only game on the screen, featuring marquee teams and stars, and expectations of fireworks run hottest for exactly these matchups. The posted totals reflect that collective optimism — and the games, on average, don't live up to it. Scoring expectations for the Sunday 1 PM slate are set soberly; scoring expectations for the Sunday Night Football stage are set with the highlight reel in mind.

2. It is not about short rest. The popular version of this theory — "Thursday teams are tired and sloppy, so games stay under" — is the one part that fails the data. Thursday is consistently the weakest primetime under window (52.1% all-time), and when we tested rest directly in our prediction-model research, totals in short-rest games came back essentially at the 50% norm (~50.5% under). The pattern lives on Sunday and Monday night — full-rest games — which points at inflated expectations, not tired legs.

The honest caveats

We publish these because most trend articles don't:

  • Recent-window samples are small. ~95 games per slot in the 2020–2025 window means the true rate could be several points off the observed one. The reason to take the pattern seriously is consistency: the under lean shows up all-time, strengthens in the 10-year cut, strengthens again in the 5-year cut, and held in an over-leaning 2025.
  • Markets adapt. A deviation this visible tends to get corrected eventually. If primetime totals start being set a point or two lower, the pattern will disappear into the number itself.
  • This is a lean, not a law. 57% means the over still hits more than 4 times in 10. Treat it as a documented statistical tendency, nothing more.

Check it yourself

Every number in this article comes straight from the SpreadTrends League Trends page — set the game filter to MNF, SNF, or TNF and pick your season window. The database covers every NFL game since 2002, and the filters run live against it.

Data: 2002–2025 seasons; games landing exactly on the total (pushes) are excluded from percentages.