SEC Dominance, Big Ten Defense, and Pac-12 After Dark: Conference Myths Tested

Analysis · by Rick's Picks Analytics

Every college football conference has a reputation. The SEC wins everything. The Big Ten plays smashmouth defense. The Big 12 is a track meet. But how many of these stereotypes hold up under statistical scrutiny? We tested six conference hypotheses with a decade of data.

Correcting for Conference Realignment

Before any analysis, we had to solve a major data problem: conference realignment. Texas played in the Big 12 through 2023, then joined the SEC in 2024. If we naively used current conference affiliations for historical games, we would attribute 2020 Texas games to the SEC -- a look-ahead bias that corrupts every finding.

Our solution: a season-aware conference mapping system that returns the correct conference for any team in any season. Every game in our dataset is tagged with the team's actual conference at the time the game was played.

Hypothesis 1: SEC Cross-Conference Dominance

The question: Does the SEC win more than 50% of its games against other Power conferences?

The test: Binomial test against 50%. We filter for games where one team is SEC and the opponent is from another Power conference (using season-correct conference assignments).

What the data shows: The SEC does win cross-conference Power matchups at a rate above 50%. This is the most well-supported conference stereotype in our data. The SEC's recruiting advantages, depth, and coaching quality translate into measurable on-field superiority against peer conferences.

How we use it: In cross-conference predictions, SEC teams receive a small quality bonus above what their ELO rating alone suggests. This accounts for the conference's developmental advantage that pure ELO may underweight.

Hypothesis 2: Big Ten Defensive Scoring

The question: Do Big Ten conference games produce lower total scoring than other Power conference games?

The test: Welch's t-test comparing mean total scoring in Big Ten conference games versus conference games in other Power leagues. Cohen's d measures the practical size of any difference.

What the data shows: Big Ten conference games do tend to score slightly lower, but the effect size matters. If Big Ten games average 48 total points versus 51 for other conferences, that 3-point difference might be statistically significant with thousands of games, but it may not be practically meaningful for prediction purposes.

The evolving picture: With recent conference realignment bringing Oregon, USC, and other high-scoring programs into the Big Ten, this defensive stereotype may weaken in coming seasons.

Hypothesis 3: Big 12 Shootout Style

The question: Does the Big 12 produce the highest-scoring conference games among Power leagues?

The test: We compute mean total scoring for all Power conference games, rank them, and then use Welch's t-test to compare Big 12 scoring against the combined average of other Power conferences. Cohen's d quantifies the difference.

What the data shows: The Big 12 has historically been the highest-scoring Power conference, consistent with its reputation for spread offenses and relatively less emphasis on defense. However, the gap has narrowed as spread concepts have proliferated across all conferences.

Hypothesis 4: Group-of-5 Upset Rate

The question: How often do G5 teams upset ranked Power opponents?

The test: Binomial test against a 17.5% baseline upset rate. We define upsets as G5 straight-up wins over ranked Power opponents.

Caveat: Rankings come from the teams table (current season), not game-time polls, introducing look-ahead bias. Results are directional.

What the data shows: G5 upset rates are low but not negligible. When they happen, they tend to be massive ATS events (since G5 teams are typically large underdogs against ranked Power opponents). This informs our "heavy underdog" adjustments.

Hypothesis 5: Pac-12 After Dark

The question: Did Pac-12 games have higher scoring variance than other conferences? The "Pac-12 After Dark" phenomenon suggested that late-night West Coast games were uniquely chaotic.

The test: Levene's test for variance equality, Welch's t-test for mean scoring differences, Cohen's d. Restricted to seasons through 2023 (pre-dissolution).

What the data shows: Pac-12 games did show higher scoring variance compared to some conferences. Whether this was "chaos" or simply the result of more extreme talent gaps within the conference (elite programs like Oregon alongside struggling programs like Colorado at the time) is debatable. Regardless, the Pac-12 as we knew it no longer exists after 2024 realignment.

Hypothesis 6: ACC Coastal Chaos

The question: Did the pre-realignment ACC Coastal Division (through 2021) produce more parity, measured by the percentage of games decided by 7 points or fewer?

The test: Two-proportion z-test comparing the rate of close games in ACC Coastal matchups versus other Power conference games. Cohen's h measures the effect size for proportions.

What the data shows: The ACC Coastal was indeed chaotic -- a different team seemed to win the division every year, and games were disproportionately close. This validated the "Coastal Chaos" narrative with actual data.

Conference Signals in Our Algorithm

Conference identity enters our model in two ways: (1) a conference strength modifier based on cross-conference results (SEC gets a small boost, historically weaker conferences get a small reduction), and (2) conference-specific scoring tendencies that inform over/under projections (Big 12 games trend higher, Big Ten games trend lower). Both are recalibrated each season as conferences evolve.

All tests use Bonferroni correction across six hypotheses (adjusted alpha = 0.00833). Conference assignments use season-aware mappings. Results validated on 2023-2024 holdout data.

Rick's Picks publishes statistical analysis of college football for informational and entertainment purposes. Nothing here is betting advice. 21+. If gambling is affecting you or someone you know, call 1-800-GAMBLER.