Major League Baseball during the COVID-19 pandemic: does a lack of spectators affect home advantage?

The study examines whether the absence of spectators during the COVID-19 pandemic affected home advantage in Major League Baseball (MLB). Home advantage—the tendency for home teams to win more than 50% of games—has been widely documented across professional sports, though MLB historically shows a smaller effect than other leagues. While crowd presence is often cited as a key driver of home advantage, evidence has been mixed and isolating the crowd effect is difficult because many contributing factors co-occur. The 2020 MLB season, played without spectators, provided a quasi-natural experiment to test the crowd’s role. The authors hypothesized that, given MLB’s relatively low home advantage compared to other sports, the absence of spectators would have minimal impact on MLB home advantage.

Prior research consistently documents home advantage across sports, with soccer showing particularly high levels and MLB among the lowest. Spectator effects are theorized to influence player psychology, opponent performance, and officiating, yet empirical findings are mixed. During COVID-19, some analyses of European football reported unchanged home advantage without crowds, while others found decreases and even reversals (e.g., Bundesliga). Natural experiments in same-stadium derbies suggest crowd support explains a large share of home advantage in NBA and Italian Serie A. Psychological mechanisms include increased confidence for home players and performance timing advantages, but also potential choking effects under pressure. Referee bias favoring home teams has been demonstrated in football and diminished in no-crowd settings during the pandemic; such bias is not evident in NBA and remains unstudied in MLB. Other contributors include facility familiarity and rarity (e.g., domes, turf), and travel-related fatigue and circadian factors, which may be attenuated in MLB due to series-based scheduling that increases visiting team familiarity and reduces travel strain. These mixed findings motivated examining MLB’s 2020 no-spectator season to isolate spectator effects on home advantage.

Design: Quasi-experimental comparison of MLB regular-season games from 2015–2019 (pre-pandemic, with spectators) versus 2020 (pandemic, no spectators). Data source: Retrosheet game logs for results and attendance across the 2015–2020 seasons. Total sample: 13,044 games (12,146 from 2015–2019; 898 from 2020). Variables: (1) Home win (game-level dummy: 1 if home team won, 0 otherwise). (2) Home advantage (team-season level: percentage of home games won). (3) Relative home advantage (team-season level: home wins/total wins, in percent) to account for team quality. Team quality: Sum of Wins Above Replacement (WAR) for all hitters and pitchers per team-season, categorized into season-specific tertiles. Attendance exposure: Attendance ratio per team-season (average home attendance divided by stadium capacity), categorized into season-specific quartiles. Stadium capacities were compiled from public sources. Statistical analyses: - Logistic regression at the game level: Dependent variable = Home win; Independent variable = Year (0 = 2015–2019; 1 = 2020) to estimate the effect of no spectators on the odds of a home win. - One-way ANOVA at the team-season level: Dependent variable = Relative home advantage; Independent variables tested separately: Year, WAR tertile, Attendance ratio quartile. - Factorial ANOVA models on Relative home advantage including combinations of Year, WAR, Attendance ratio, and their interactions (Model 1: Year + WAR; Model 2: Year + Attendance ratio; Model 3: Year + WAR + Attendance ratio + all pairwise and three-way interactions). Software: Data assembly in SAS v9.4; statistical analysis in SPSS 21.0. Significance threshold: p < 0.05.

• Descriptives: Average home advantage (team share of home games won) across 2015–2020 ranged approximately 52.77%–54.98%, consistent with prior MLB norms (~54%). Mean relative home advantage was significantly above 50% in each season (p < 0.01). Between 3 and 7 teams per season (2015–2019) and 4 teams in 2020 had relative home advantage below 50%. - Game-level logistic regression: The Year effect (2020 vs. 2015–2019) had an odds ratio of 1.068 (95% CI: 0.932–1.224; p = 0.344), indicating a nonsignificant 6.8% increase in the odds of a home win in 2020. - Team-level one-way ANOVAs on relative home advantage: Year (F = 1.328, p = 0.254), WAR tertile (F = 0.996, p = 0.371), and Attendance ratio quartile (F = 1.455, p = 0.228) were all nonsignificant. - Factorial ANOVAs on relative home advantage: No significant main effects or interactions in models including Year with WAR (Model 1), Year with Attendance ratio (Model 2), or Year, WAR, and Attendance ratio together (Model 3). All interaction terms (Year×WAR, Year×Attendance ratio, WAR×Attendance ratio, and the three-way interaction) were nonsignificant. - Overall: Neither the absence of spectators in 2020 nor team quality (WAR) nor typical attendance exposure (attendance ratio) had a significant effect on MLB relative home advantage.

The findings indicate that MLB home advantage did not change significantly during the spectator-free 2020 season compared to 2015–2019, suggesting that live spectators are not a primary driver of home advantage in MLB. This conclusion holds after accounting for team quality using WAR and examining attendance exposure proxies. The result contrasts with evidence from other sports—particularly European football and some NBA contexts—where crowd effects appear to play a larger role via player psychology and referee bias. MLB’s relatively modest and stable home advantage may be shaped more by non-crowd factors such as familiarity with unique stadium features (e.g., domes, turf, idiosyncratic park dimensions) and the league’s travel structure, which often reduces visiting team fatigue and increases field familiarity through multi-game series in a single location. The study adds to the literature by leveraging the unique COVID-19 natural experiment and by using relative home advantage to control for team strength, reinforcing that MLB differs from other leagues in the mechanisms underpinning home advantage.

A spectator-free 2020 MLB season did not significantly alter home advantage relative to 2015–2019, implying that spectators are not a critical determinant of MLB home advantage. The study’s contribution lies in exploiting a natural experiment to isolate crowd effects and in adjusting for team quality via WAR. The results point toward alternative mechanisms—facility familiarity and travel-related factors—as more plausible drivers. Future research should: (1) examine umpire decision patterns for potential bias in MLB; (2) disentangle effects of 2020-specific rule changes (e.g., seven-inning doubleheaders, universal DH in NL, extra-innings tiebreaker) on home advantage; (3) incorporate more granular scheduling and travel metrics to quantify fatigue and circadian factors; and (4) assess the impact of simulated crowd noise and in-stadium presentations used in 2020.

The design is quasi-experimental; multiple pandemic-era changes co-occurred with the absence of spectators in 2020, complicating causal inference. Differences included a shortened season, divisional-only opponents, more make-up games and doubleheaders, and player absences due to infection/quarantine—factors that may influence fatigue and performance for both home and visiting teams. Rule changes (seven-inning doubleheaders, universal DH in the NL, and the extra-innings runner/tiebreaker) could affect home advantage, with uncertain net effects. Teams also used artificial crowd noise and visual fan substitutes (cardboards/screens), which were not controlled. Consequently, results should be interpreted with caution regarding causality.