"Home field is worth three points" is a habit, not a law. Home field advantage modeling starts by dropping it. Instead, you separate a team's neutral-site strength from a venue effect, so home results don't get counted twice. A worked example moves a neutral margin of 3 to 4.8 at home and back to 3 at a neutral site. Then you estimate those venue effects by shrinking small samples toward the league norm, since one stadium's history is thin and noisy. Finally, you test the whole thing with backtests in time order. Done well, it tells you when a team's home edge is real and when the market is paying for a reputation.
Learning objectives
- Separate neutral team ratings from a home or venue effect.
- Work the example from a neutral margin of 3 to 4.8 at home and back to 3 at a neutral site.
- Estimate venue effects with regularization and chronological backtests.
