FOR CLARITY: This post was entirely fabricated for April 1. The joke is apparently indiscernably subtle unless you follow AFL football quite closely.

Geelong Football Club’s league-leading use of big data has led to findings that it hopes will give the club the winning edge in 2016.

In concert with Geelong’s Deakin University and a grant from the Victorian Department of Innovation and Industry, the Club has spent the summer feeding data on team and player performance over the last 100 seasons into a supercomputer capable of running at 100 gigaflops.

The results have been as impressive as they have been influential, according to Geelong General Manager of Football, Steve Hocking.

“This data has really opened our eyes to things you’d never consider pursuing when you’re in the cut and thrust of football operations,” Hocking said.

One of the most important findings relates to factors that have historically led to a win. Hocking was eager not to give too much away. But he was able to describe a few factors the calculations turned up.

“What was really eye-opening from the big data study was the influence of previous games on the next game,” he said. “That came through loud and clear, and in many ways. We’re talking injuries, motivation, six day breaks, suspensions, the impact of travel, etc. While there’s always been anecdotal evidence they mattered, clubs have never really acted on it.”

This has led to a revolution at Geelong against the time-honoured “taking it one week at a time” preparation strategy. The age-old mantra is going the way of a steak and eggs on the morning of the game.

“It was hard to accept at first because it goes against so many years of football thinking, but we had to accept that inter-temporal linkages were a compelling factor in determining on-field success.”

According to Hocking, football season will be rearranged from a football operations perspective.  No longer will the club prepare for each round as it comes, but it will prepare for “blocs” of games.

“The season has been broken down into blocs of games,” he said. Some of those blocs are two games long, some three games, and the longest is a five-game bloc. All as suggested by the outputs of the study.

This is leading to substantial changes in practice for coaching staff, catering staff, and of course the men who wear the club jerseys onto the field on game day.

Preparing in the new big data-approved way has been something of a mental hurdle for the team, explained Geelong Captain Joel Selwood.

“We’ve been operating for so long on the basis of ‘taking it one game at a time’,” the Cats star midfielder said. “It’s actually been really hard to break that habit, and the coaches have had to institute systems to make sure we integrate the, um, intertemporal linkages into all our training.”

That includes media training.  On the side of training, while the rest of the team did skills work, Defender Harry Taylor was practising the new system with a media department intern in a mock post-match interview.

“Yeah, nah, the boys done good tonight and we’re just taking it … three weeks at a time?” said  the key position player and Cats vice-captain.

“Perfect,” said the young intern.

Published by


Thomas the Think Engine is the blog of a trained economist. It comes to you from Melbourne Australia.


  1. Thousands of variable data points on hundreds of millions of consumers obtained through Facebook is “big data”.
    Thousands of data points on each of the players in the Geelong Football Club is just “data”.
    The power of the computer you are analysing it with doesn’t make it “big”.
    Moneyball didn’t reflect a use of big data, neither does this.

    Liked by 1 person

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s