Big Stadiums, Big Advantage? How Stadium Size Affects Spread Coverage

Analysis · by Rick's Picks Analytics

Michigan Stadium holds 107,601 fans. Wake Forest's Allegacy Federal Credit Union Stadium holds 31,500. Does that 76,000-person difference actually show up in how often the home team covers the spread?

The Dataset

We collected capacity data for 65 major FBS programs and categorized them into four tiers:

• Small (under 50,000): Programs like Wake Forest, Duke, Vanderbilt
• Medium (50,000-70,000): Programs like Iowa State, Kentucky, Arizona
• Large (70,000-90,000): Programs like Florida, Michigan State, Auburn
• Massive (90,000+): Programs like Michigan, Ohio State, Penn State, Texas A&M, Tennessee

We then matched every home game from 2009 onward with the stadium's capacity and computed ATS outcomes using our canonical spread math.

Hypothesis 1: Correlation Between Size and Cover Rate

The question: Is there a positive correlation between home stadium capacity and home ATS cover rate?

The test: Point-biserial correlation between stadium capacity (continuous) and whether the home team covered (binary). This tells us whether bigger stadiums systematically produce more home covers.

What the data shows: There is a positive correlation, but it's weak. Stadium size alone is a poor predictor of ATS outcomes. This makes sense -- a 100,000-seat stadium doesn't help if the team is 3-9.

Hypothesis 2: Massive vs Small

The question: Do stadiums with 90,000+ capacity produce a different home cover rate than sub-50,000 venues?

The test: Chi-square test comparing cover rates, with Cohen's d on the ATS margin difference. This is the sharpest contrast in our data.

The finding: Massive stadiums do show a higher home cover rate than small stadiums. The crowd noise, intimidation factor, and general atmosphere of a 100,000-seat stadium create a measurable home-field advantage beyond what the spread already accounts for.

The caveat: Massive stadiums belong to elite programs (Michigan, Ohio State, LSU, Tennessee). These teams also have superior recruiting, coaching, and talent. Separating the stadium effect from the team-quality effect is challenging. Our analysis controls for this by looking at ATS performance (which already factors in team quality via the spread) rather than straight-up wins.

Hypothesis 3: Conference Games Amplify the Effect

The question: Is the stadium-size advantage larger in conference games, where visiting fans are outnumbered and hostility is maximized?

The theory: A non-conference cupcake game in a massive stadium might not produce the same atmosphere as a top-15 conference showdown. The crowd is more engaged, louder, and more hostile when the opponent matters.

The test: We split games into conference and non-conference, then compare the stadium-size effect within each group.

What we find: Conference games in massive stadiums do show a slightly amplified effect compared to non-conference games. However, this interaction effect is smaller than the main stadium-size effect and may not survive Bonferroni correction.

Hypothesis 4: Top-Performing Individual Stadiums

The question: Which specific stadiums consistently outperform ATS expectations?

The requirement: Minimum 50 home games in the dataset to ensure statistical stability.

The analysis: For each qualifying stadium, we compute the home cover rate and test it against 50% with a binomial test. We then rank stadiums by ATS outperformance.

Key insight: The stadiums that top this list aren't always the biggest. Some mid-sized venues with passionate, tight-packed fanbases (think Autzen Stadium, Kinnick Stadium) outperform larger venues with less engaged crowds. Stadium culture matters as much as raw capacity.

A Critical Bug We Fixed

Our original analysis had an inverted spread convention -- it was marking home covers as losses and vice versa. This was caught during our February 2026 refactor. The old code also applied an unjustified 6x multiplier when converting cover-rate differences into point adjustments. Both bugs have been corrected, and all results now use our canonical spread_utils module where negative spread equals home favored.

How Stadium Size Enters Our Algorithm

Stadium capacity provides a small ATS adjustment in our prediction model. Massive stadiums get a slight home-team boost, small stadiums get a slight reduction, and medium/large venues use the baseline. The adjustment is amplified slightly for conference games. This factor alone won't make or break a prediction, but stacked with weather, travel distance, and public bias, it contributes to our composite edge.

Analysis uses point-biserial correlation, chi-square tests, binomial tests, and Cohen's d, with Bonferroni correction across four hypotheses. Data from 2009-2024 with temporal validation on 2023-2024.

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.