02 / Home advantage Complete
Which players benefit most from playing at home?
Finding
Everyone does better at home by about the same amount: +27% xG per minute for the same player, in the same team and season. Not one of 4,665 players has a reliable personal home edge.
Home and away splits are one of the first things people reach for: this striker is a home player, that winger travels badly.
If those splits were real, part of a player's value would depend on where he plays — and a transfer would change it.
Is home advantage a property of the player, or of the match?
Comparing a player only with himself removes ability, team quality and season form. What's left is the effect of playing at home.
Within-player on per-minute output with ; errors clustered by match.
Individual effects get calibrated standard errors from pseudo-splits and shuffles within each player's own matches, a DerSimonian–Laird estimate of the real between-player spread, and .
Persistence is measured against shuffled data. , and for every headline effect.
Per minute, for the same player: +27%, +25%, shots +22%, +21%, goals +24%.
Home edge per minute, same player
Home ÷ away − 1, with fans. Lines are 95% intervals.
Every metric rises by roughly a quarter at home; the intervals are narrow and none overlaps zero.
Study 02 · within-player PPML · 2015/16–2024/25
Every attacking number moves together, by roughly a quarter.
Behind closed doors the xG edge fell from +28% to +12%, and came back to +25% with fans. The betting market priced the same loss: the home/away win-probability ratio for evenly matched teams fell from 1.58 to 1.31.
Player xG home edge, by period
Same player, home ÷ away − 1.
The edge fell from +28% to +12% without fans and came back to +25%. The market's home/away win-probability ratio for evenly matched teams fell from 1.58 to 1.31.
Study 02 · player level · betting market from football-data.co.uk
La Liga has the largest home edge (xG +35%), Serie A the smallest (+23%).
xG home edge by league
Same player, with fans. Lines are 95% intervals.
La Liga has the largest edge (+34.8%), Serie A the smallest (+22.6%).
Study 02 · per-league within-player models
Maybe the average is shared, but a few players are genuine home specialists.
0of 4,665players with a reliable personal home edge
Each player's estimate is only 5–8% . Within a team-season, individual differences vanish for xG, shots, key passes and goals.
Each player's personal home edge
4,606 players in this export (the reliability test reports 4,665), xG home vs away. Each row is scaled to its own peak; outer bars collect everything beyond ±200%.
As measuredspread ±77 pts
After removing noisespread ±2 pts
Measured naively, players' home edges look wildly different. Once match-to-match noise is removed, 4,390 of 4,606 land between +20% and +30%.
Study 02 · empirical-Bayes shrinkage
Against their own league's average, clubs differ by about ±1%: 0 of 121 clubs are reliably different, and a club's past home edge does not predict its future one.
The six “strongest home clubs”, before and after removing noise
Club xG home edge relative to its league's average.
Raw, these clubs look 20–30 points better at home than their league. After shrinkage, each is within about 1%.
Study 02 · 121 clubs, empirical-Bayes shrinkage
What the raw data suggested
The 2015–19 “top 10 home players” looked like genuine home specialists.
What happened after we checked
In 2022–25 they were back near the all-player average: +43%, against +33% for everyone. The correlation between the two periods was 0.018.
Lesson
A home/away split is mostly noise. Ranking players by it is in action.
SourcesUnderstat · football-data.co.uk · Open-Meteo ERA5 · Wikipedia/Wikidata · OpenStreetMap
Analysis period2015/16 — 2024/25 · frozen dataset · Methodology
08The next question