Establishing Metrics for Margin, Total Score and Team Score Predictions

These days, I reckon I know what a good margin forecaster looks like. Any person or algorithm - and I'm still at the point where I think there's a meaningful distinction to be made there - who (that?) can consistently predict margins within 30 points of the actual result is unarguably competent. That benchmark is based on the empirical performances I've seen from others and measured for my own forecasting models across the last decade of analysing football. 

Read More

Season-to-Season Rises and Falls Across VFL/AFL History

As the 2016 AFL season proper looms and the window for more leisurely analyses slowly closes, today we'll wander across the expanse of footy history this time using MoSSBODS Team Ratings to decide which of the 1,442 teams that have played VFL/AFL football have been the big improvers, and which the big decliners (well ... you find an antonym then).

Read More

Assessing Team Abilities: Scoring Shots or Final Score?

In the previous post here on the Statistical Analyses blog we revisited the topic of Scoring Shot conversion and found that it appears to be unpredictable at a team level across entire seasons. That result, coupled with an earlier one where we found conversion rates to be unpredictable for a given team in a given game (and with some conversations I've had on Twitter subsequent to the more-recent analysis) makes it hard to reject the null hypothesis that team conversion rates are generated in a manner that's indistinguishable from a random variable.

Read More

Reoptimisation and the Fear of Overfitting : ChiPS 2016

Richard McElreath, in one of the lectures from his Statistical Rethinking course on YouTube aptly and amusingly notes that (and I'm paraphrasing) models are prone to get excited by exposure to data and one of our jobs as statistical modellers is to ensure that this excitability doesn't lead to problems such as overfitting.

Read More