The Footycast Model

Here, p represents the probability of the home team winning, while h and a represent the home and away teams' ratings, respectively. The home advantage value ha is either 0 or approximately 7 rating points (equivalent to a about a goal).

Following each round, team ratings are updated based on their expected performance against their opponents and their actual performance. As seen above, the ERS method generates a probability based on team ratings. This probability is then converted into a predicted margin using the formula:

One of the more technical aspects of our model is determining the parameter(s) that dictate the extent to which teams' ratings are updated. Notably, a key feature of our model is that the amount each team's rating is updated decreases in importance as the season progresses. This is based on the notion that we become more confident in a team's ability as the season unfolds, and therefore, we pay less attention to upsets or blowout margins.

In sum, our ERS-based AFL prediction model is grounded in robust statistical methodology and a commitment to simplicity and parsimony. We believe that this approach provides fans and followers with accurate and comprehensive predictions that are both intellectually rigorous and entertaining.

*This is not entirely true... there are other factors and parameters this is by and large the most important, followed by home ground advantage.