There’s a betting narrative that gets repeated every January, whispered in sportsbooks and Twitter threads: the 1-seed under curse. Championship games are defensive battles. The stakes are too high for offensive fireworks. Take the under and cash.
Except the data tells a completely different story. Over the past 26 years, conference championship games have gone over the total 61.2% of the time. Not sometimes. Not in a coin flip. Nearly two out of every three games.
Where the Myth Came From
The narrative isn’t random. It has a source: 2021 and 2022.
In those two seasons, all four conference championship games went under. Every single one. The Bengals-Chiefs slugfest that ended 27-24? Under a 54.5 total. The Rams’ 20-17 win over San Francisco? Way under 46. For two straight years, the under was automatic.
Then came 2024. Both conference championship games didn’t just go over. They destroyed the total, exceeding it by an average of 21.5 points. The narrative shattered in real time.
This is classic small-sample storytelling. Two years of extreme results created a “curse” that never existed. Zoom out to 49 championship games since 1999, and the curse flips into its opposite: conference championships are higher-scoring than bettors expect.
The Numbers Don’t Lie
Conference championship games average 47.6 total points against a betting line of 45.6. That’s a 2.1-point surplus over market expectations. While that might not sound dramatic, it represents consistent upward pressure on scoring.
Compare championship games to the rest of the playoffs: Wild Card and Divisional rounds split almost perfectly 50-50 between overs and unders (48.2% overs, 51.1% unders). Conference championships? They jump 13 percentage points toward overs.
Even when you isolate games with home 1-seeds—the supposed epicenter of the curse—the pattern holds. These games go over 62.1% of the time. Eighteen overs, eleven unders. The top seed doesn’t suppress scoring. If anything, it correlates with offensive firepower.
Statistical tests confirm there’s no significant under trend. The binomial test returns a p-value of 0.152, well above the 0.05 threshold needed to declare significance. Translation: there’s no evidence championship games behave differently from a coin flip, and what directional lean exists favors overs, not unders.
The Nuance: Not All Matchups Are Equal
Here’s where it gets interesting. The overall myth is false, but dig into specific seed matchups and a pattern emerges.
1-seed vs. 3-seed games go under 75% of the time. Six unders in eight games. Same story for 2-seed vs. 4-seed matchups: 75% under rate across four games.
Meanwhile, 1-seed vs. 2-seed matchups—the most common at 18 games—go over 72.2% of the time. Thirteen overs, five unders. This is the premier battle of conference titans, and it consistently produces shootouts.
Lower-seed upsets follow the same pattern. When a 1-seed faces a 5-seed, all three historical games went over. When a 2-seed faces a 6-seed, five of six went over.
The pattern becomes clear when comparing matchup types: mismatched games where a dominant 1-seed faces a weaker 3-seed turn into defensive slogs. The underdog can’t score, and the favorite grinds out a win. But when two elite offenses meet—1 vs. 2—both teams have the talent to move the ball. The result is points.
Decade-Level Trends Show Market Adaptation
The 2000s were an over bettor’s paradise. Championship games went over 75% of the time (15 overs, 5 unders), averaging 4.03 points above the total. Markets hadn’t caught up to the NFL’s offensive explosion.
By the 2010s, the gap narrowed. Games still went over at a 60% rate, but the average surplus dropped to 1.12 points. Oddsmakers adapted, raising totals to match the new scoring environment.
The 2020s show a 50-50 split overall, but that hides wild year-to-year volatility. From 2019 to 2020, all four games went over. Then 2021-2022 brought four straight unders. Then 2024 delivered two massive overs.
This isn’t a stable trend. It’s variance playing out in a small sample. Championship games only happen twice per year, which means a single outlier season can distort perception.
What This Means for Bettors
First, stop blindly betting unders in conference championships. That strategy would have lost money over the past 26 years. The data doesn’t support it.
Second, consider the matchup. When you see a 1-seed vs. 2-seed showdown—two elite offenses with championship-caliber quarterbacks—the historical edge is on the over. Seventy-two percent is a meaningful edge, even if the sample size (18 games) requires caution.
Third, there may be value in targeting unders for mismatched 1 vs. 3 games, where the under has hit 75% of the time. But proceed carefully: eight games is a tiny sample, and one outlier season could reverse the trend.
Fourth, accept that championship game totals are high-variance bets. These aren’t regular-season matchups with hundreds of data points. They’re unique environments where weather, injuries, and game script can overwhelm historical patterns.
The 2024 Reality Check
This year’s conference championships obliterated the under narrative. The AFC game (Chiefs vs. Bills) flew over 47.5 with 63 combined points. The NFC game (Eagles vs. Commanders) crushed a 47.5 total with 55 points.
Both games featured elite offenses, tight spreads, and playoff intensity. Both delivered scoring. It’s exactly what the historical data predicts for premier matchups.
If you walked into Championship Sunday believing in the under curse, you walked out poorer. The myth doesn’t just lack evidence—it actively contradicts 26 years of results.
The Bigger Picture
This analysis isn’t about one specific bet. It’s about how narratives form in sports betting and how small samples create false patterns.
Two years of extreme results (2021-2022) generated a story that felt true. Defensive battles. High stakes. Championship pressure. It made intuitive sense, which made it sticky.
But intuition isn’t evidence. When you zoom out to the full dataset, the curse vanishes. What remains is a market that slightly undervalues scoring in championship games overall, and significantly undervalues it in premier 1 vs. 2 matchups.
The lesson: test your assumptions. Narratives are cheap. Data is expensive. And in betting, expensive data is what separates profit from loss.
Data source: nflverse (1999-2024 NFL seasons) Sample size: 49 conference championship games with betting lines Methodology: Binomial testing, matchup-specific stratification, decade-level trend analysis