What Did Empty Stadiums Teach Us About Home Advantage? What the Behind-Closed-Doors Seasons of 2020 and 2020-21 Showed

Empty stadiums showed that a crowd explains part of home advantage, not all of it. Studies of football behind closed doors in 2020 and 2020-21 found weaker home advantage in most leagues, with the clearest change in refereeing. Size varied by league and study. RubiScore match pages show the date, competition and venue needed to read such games.

This data study poses one question, whether a crowd creates home advantage, then reviews what researchers analysed, where the studies agree and differ, and which confounders limit them. It ends with a method for reading fixtures played with limited or no attendance. Match, competition and referee pages are listed at https://rubiscore.com.

What Question Did Empty Stadiums Answer?

Before 2020, almost every top-flight match had a crowd, so its effect was tangled up with team quality, travel and pitch familiarity. Tilp and Thaller name four candidate sources of home advantage: crowd support, territoriality, familiarity and travel fatigue.

The restart removed the crowd from otherwise familiar fixtures. The Bundesliga resumed on 16 May 2020 with nine rounds left and finished on 27 June. The Premier League restarted on 17 June and finished on 26 July, also behind closed doors. The same teams played in the same stadiums without spectators.

A crowd can act through officials and through players. The question therefore splits in two: did referee decisions change, and did home performance change?

Which Studies Analysed the Matches?

The findings below are reported as the authors state them.

• Bryson, Dolton, Reade, Schreyer and Singleton (Economics Letters, 2021): closed-door matches in several countries' 2019/20 league seasons. They found large, statistically significant effects on yellow cards: fewer for away teams, which reduced home advantage.
• Reade, Schreyer and Singleton (Economic Inquiry, 2022): visiting players were cautioned less often, by over a third of a yellow card per match, or once for every twenty-two fouls. Home advantage in results appears reduced, but the authors "cannot statistically reject no effect".
• Fischer and Haucap (Journal of Sports Economics, 2021): Germany's top three divisions in 2019/20. Home advantage fell in the first division and did not change in the second and third. The drop tracked lower stadium occupancy and faded over time.
• Sors, Grassi, Agostini and Murgia (European Journal of Sport Science, 2021): 841 closed-door matches in the first and second divisions of the four top-ranked UEFA countries. They found reduced home advantage and no referee bias.
• Scoppa (Journal of Economic Psychology, 2021): two divisions in each of Germany, Spain, England, Italy and Portugal, against the previous ten championships. Home advantage in points, goals and shots almost halved, and decisions on fouls, cards and penalties became more balanced.
• McCarrick, Bilalić, Neave and Wolfson (Psychology of Sport and Exercise, 2021): 4,844 games from 15 leagues in 11 countries. Home advantage in goals and points fell, reflecting weaker home performance, and home teams created fewer attacking opportunities. Referee bias was diluted but still significant for fouls and yellow cards, with no effect on red cards.
• Bilalić, Gula and Vaci (Scientific Reports, 2021): over 4,000 matches from 12 leagues. Absent fans cut home advantage by a third, through weaker home performance and disappearing officials' bias.
• Wunderlich, Weigelt, Rein and Memmert (PLOS ONE, 2021): over 40,000 matches, including more than 1,000 without spectators. Extra sanctions on away teams for fouls and yellow and red cards disappeared, and home shots fell, but the decrease in home advantage itself was non-significant.

The 2020-21 season gives a second look. Schank, Voigt and Orthey (2024 preprint) followed the Bundesliga from 2016/17 to 2023/24 and found 2020/21 home advantage very close to 2018/19, which they read as teams becoming accustomed to empty stands. Van Ours (Economics Letters, 2024) found that in five Eredivisie seasons, empty stadiums made home advantage disappear completely because home teams scored fewer goals. Alan Jackson's analysis in Significance, the Royal Statistical Society's magazine, put the Premier League's home-to-away win ratio at 0.94 in 2020/21 against an earlier average of 1.68.

Where Do the Studies Agree?

• Referees: Bryson, Reade, Wunderlich and Scoppa found fewer sanctions against away teams, and McCarrick found the bias diluted.
• Home attacking: Wunderlich and colleagues found fewer home shots, McCarrick and colleagues fewer attacking opportunities, and Schank and colleagues fewer home goals in the last nine Bundesliga rounds of 2019/20.
• Direction: Leitner, Daumann, Follert and Richlan's systematic review (Management Review Quarterly, 2023) covered 26 primary studies and concluded that ghost games had a considerable impact on home advantage, through reduced referee bias and the loss of emotional support from the stands.

Where Do the Studies Differ?

• Size: Wunderlich and colleagues give home, draw and away shares of roughly 45:27:28 across ten seasons with spectators and 43:25:32 without. Scoppa reports an almost halved advantage, Bilalić and colleagues a third, and the same Wunderlich paper a non-significant overall decrease.
• League: the Bundesliga moved from 46:24:30 to 33:23:45, while the Premier League's matches without spectators moved from 46:25:30 to 47:22:32. Benz and Lopez (AStA Advances in Statistical Analysis, 2023) used bivariate Poisson models on 17 leagues and found mixed results: a decrease in some, a possible rise in others.
• Mechanism: Wunderlich and colleagues saw referee bias vanish while home advantage mostly stayed, so fans are not the main driver. McCarrick and colleagues and Bilalić and colleagues found weaker home performance too. Fischer and Haucap found little evidence of referee or tactical channels and favoured psychological reasons. Van Ours judged referees an improbable intermediary.
• Sample and measure: the Premier League barely moved in Wunderlich's sample, yet Jackson's simple ratio of home to away wins fell to 0.94 across 2020/21. His ratio has no controls for team strength, so the two are not like-for-like.

What Confounders and Limits Apply?

• Restart congestion: the Premier League's "Project Restart" plan envisaged the 92 remaining matches in about six weeks, and the season finished on 26 July. Fatigue and rotation can differ from the weeks before the break.
• Five substitutes and VAR: on 8 May 2020 IFAB allowed up to five substitutes per team, in three windows, at each competition organiser's discretion. The Premier League adopted it. IFAB also let competitions stop using VAR on restart. Matches before and after the break differ in more than attendance.
• Neutral-venue completions: Project Restart first proposed neutral stadiums, which would have removed pitch familiarity as well as the crowd. The 2019-20 UEFA Champions League was completed in Lisbon in August 2020 as single-leg ties without spectators, so those games are not ordinary home fixtures.
• Short samples: Tilp and Thaller (Frontiers in Sports and Active Living, 2020) analysed 83 Bundesliga games without an audience in 2019/20 and counted more home losses (36) than home wins (27) in the last nine rounds. In 83 matches, one result moves the home-win share by about 1.2 percentage points and five results by about six.
• Selection and stakes: restart matches were the season's final rounds, with their own title, European and relegation stakes, not a random sample.
• Adaptation and partial crowds: the effect decayed over time in Fischer and Haucap's data. Schank and colleagues found a U-shaped link between home advantage and stadium utilisation in 2021/22. Bilalić, Graf and Vaci (Psychology of Sport and Exercise, 2023) found home advantage nearly 50% lower at EURO 2020, with one-third capacity, than in the same teams' qualifying matches.
• Non-crowd factors: Wunderlich and colleagues also found home advantage in almost 6,000 amateur Kreisliga A games, where crowds are small and travel short. They cite territorial behaviour in familiar surroundings.

What Is the Verdict?

The evidence supports a partial answer. A crowd explains part of home advantage, but not all. The best-replicated finding is the referee channel, and several studies also saw weaker home performance. Yet home advantage survived in many empty-stadium samples, and Jackson found the lower three English divisions still averaged a home-to-away win ratio of 1.2 in 2020/21.

Effect sizes run from non-significant to a third, to almost half, to complete disappearance in the Dutch data. No study isolates the crowd cleanly, because 2020 changed schedules, rules and routines at once. Direction and channel are firmer conclusions than any single number.

How Do You Read a Match Page With Restricted or No Attendance?

Platforms such as RubiScore combine match, competition, referee and stadium pages, which lets each check below be lined up quickly.

• Step 1: Check the date. Matches between the 2020 suspensions and the return of crowds belong to a different regime. Attendance rules changed within seasons, so split a club's home record into before, restricted and normal periods.
• Step 2: Check competition and venue. Confirm the fixture was a league home match at the club's usual ground, not a neutral or relocated one.
• Step 3: Compare cards and shots, not only results. Fouls, cards and shots give several observations per match, while a result gives one, so they show crowd effects sooner than win-draw-loss shares.
• Step 4: Use whole seasons. Compare 2018/19, 2019/20, 2020/21 and 2021/22 separately, and treat short stretches, such as the 83 Bundesliga games above, as fragile.
• Step 5: Do not use 2020-21 home form as a baseline. It reflects a no-crowd or partial-crowd environment and, in two studies, a period of adaptation.

Empty stadiums did not erase home advantage, and they did not leave it untouched. They showed that the crowd matters through officials and home performance, while the rest of the edge comes from other sources, such as the familiarity and territorial behaviour Wunderlich and colleagues describe.