Hypothesis

Is home advantage real in the A-League, or is it a myth? Home and away records are paired by team (13 home games, 13 away games each), so this uses paired tests rather than independent-sample tests.

Data: Soccerway, 2025–26 A-League Men season, home/away splits.

df <- read.csv("aleague 202526 home away.csv")

Points per game

diff <- df$home_ppg - df$away_ppg

shapiro.test(diff)
## 
##  Shapiro-Wilk normality test
## 
## data:  diff
## W = 0.91414, p-value = 0.241

Shapiro-Wilk is not significant (p = 0.241), so the differences are normally distributed and a paired t-test is appropriate.

t.test(df$home_ppg, df$away_ppg, paired = TRUE)
## 
##  Paired t-test
## 
## data:  df$home_ppg and df$away_ppg
## t = 1.3689, df = 11, p-value = 0.1983
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  -0.05844986  0.25078319
## sample estimates:
## mean difference 
##      0.09616667
mean(diff)
## [1] 0.09616667
d_cohen_ppg <- mean(diff) / sd(diff)
d_cohen_ppg
## [1] 0.3951802

Home teams average 0.096 more points per game than away teams, but this is not statistically significant (p = 0.198). Cohen’s d = 0.395 is a small-to-medium effect, with only 12 teams, the test is underpowered to detect it.

Goal difference

gd_diff <- df$home_gd - df$away_gd

shapiro.test(gd_diff)
## 
##  Shapiro-Wilk normality test
## 
## data:  gd_diff
## W = 0.92718, p-value = 0.3511

Normal again (p = 0.351), so paired t-test applies.

t.test(df$home_gd, df$away_gd, paired = TRUE)
## 
##  Paired t-test
## 
## data:  df$home_gd and df$away_gd
## t = 3.3568, df = 11, p-value = 0.0064
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
##  1.664196 8.002471
## sample estimates:
## mean difference 
##        4.833333
mean(gd_diff)
## [1] 4.833333
d_cohen_gd <- mean(gd_diff) / sd(gd_diff)
d_cohen_gd
## [1] 0.9690187

Home teams average +2.42 goal difference, away teams -2.42 which is a swing of 4.83 per team, and this is statistically significant (p = 0.0064), with a large effect size (d = 0.969).

Conclusion

Points say home advantage isn’t real in the A-League. Goal difference says otherwise. Points are a blunt instrument a 1-0 win and a 4-0 win count the same, so home advantage shows up in how convincingly teams win or lose, even when it doesn’t reliably convert into extra points.

list.files() file.exists(“aleague 202526 home away.csv”)