Home Ground Advantage: How Much Does It Matter?
Our initial analysis showed that home ground advantage is important, so I’ll need to build some aspect of it into my overall AFL predictive model. But to do so, I needed to answer a few key questions:
- Can we quantify home ground advantage?
- Does it exist when the away team is from the same state?
- Does the advantage increase with distance travelled?
- How many points is home ground advantage worth?
To answer these, I built out a dataframe with new features:
- Whether a game was home vs. interstate or intrastate opposition
- The distance travelled (using city location—so, e.g., Melbourne vs. Melbourne = 0)
- Note: I treated Geelong as interstate for this analysis, as they have a dedicated home ground outside Melbourne
- Teams playing frequent games at a neutral venue (e.g., Darwin, Tasmania) were considered to have "home" status for those matches
Key Results
The data lines up with what you’d expect:
| Category | Home Win Rate (%) | Avg Home Margin (pts) |
|---|---|---|
| All games | 57.0 | +6.7 |
| Interstate | 58.8 | +8.5 |
| Not Interstate | 52.5 | +2.0 |
Home teams (excluding neutral venues) win 57% of the time, but most of this advantage comes from interstate travel. When the away team travels interstate, the home team gains an average margin of 8.5 points. Against intrastate opponents, the advantage drops to just 2 points.
Does Distance Matter?
Regression analysis found that distance travelled and neutral venue both had large p-values—meaning they are not statistically significant predictors of home margin. For the full notebook and code, see GitHub.
Conclusion
Home ground advantage in the AFL is real, but almost entirely explained by interstate travel. Now that I’ve quantified this, I can move on to the next set of features for the predictive model.