Hayden Crozier of the Western Bulldogs evades Richmond player Jason Castagna during their round 23 match in 2018. Image: Michael Dodge/Getty Images

Flying the Flag

In the endless hunt for premiership success, AFL clubs are leaving no stone unturned, with science and technology now seen as a key element in that quest.

A picture of fitness in his Western Bulldogs polo shirt, Sam Robertson looks ready to take on the AFL’s best as he greets me at Whitten Oval, the club’s home base in West Footscray. But Robertson isn’t a player. As the Bulldogs’ head of research and innovation, sport scientist Robertson makes his contributions to the team from the other side of the fence.

Robertson leads a team of four working to give the Dogs an edge over other clubs by using the latest technology. That word – technology – can mean different things to different people, even within the AFL. For fans it might conjure images of score reviews and the debatable benefits of goal-post cameras. For Robertson, it means something else altogether. He and his team use a discipline known as ‘machine learning’ in their aim to gain a performance edge.

“Machine learning,” says Robertson, “is a method of analysis that helps to identify patterns in data. As humans we’re good at identifying patterns. We use data analysis to do so with a non-human bias, picking up ‘non-linearities’ in that data.” In layman’s terms, that means uncovering variations in playing patterns that might reveal an advantage or disadvantage, for example in congested areas on a field. Indeed many believe the Western Bulldogs’ 2016 premiership success was largely attributable to the chains of manic handballs players used to extricate themselves from congestion. The unassuming Robertson won’t claim credit for that, but many believe his work played a significant role.

The GPS device being inserted into a player’s jumper. Image: Michael Dodge/Getty Images


Robertson works closely with Mat Inness, the Bulldogs’ physical performance manager, who lauds the part technology plays in monitoring player health and fitness: “We use the NordBord and GroinBar to help manage our players’ physical health.” These exotically named pieces of equipment aid in reducing the risk of soft tissue injuries, with Inness adjusting player workloads and training drills in response to the data they provide.

Robertson cites “hormonal profiles” as another way of identifying potential issues with a player’s health and wellbeing, sometimes even before players themselves are aware.

But the key to data analysis remains the GPS unit. Robertson and Inness have seen GPS technology transform footy since it was introduced. “When I arrived in 2012, the club had 15 GPS units for players to share,” says Inness. “Now every player on our list has their own.” That use of technology has also been extended to the Bulldogs’ AFLW players, as they seek to repeat their own 2018 premiership success.

The Bulldogs’ partnership with Victoria University also gives Robertson access to 31 students, all an active part of his team. To the best of his knowledge, that figure is unmatched by any other sports team.

From a tactical viewpoint, Robertson cites fixed cameras mounted above playing arenas as footy’s potential ‘next big thing’. “They’re already being used in major soccer leagues overseas,” he says. The technology could be easily adapted to larger AFL grounds and, combined with machine learning, identify tactical on-field advantages that might be gained.

Is the gathering of an ever more granular level of player data a privacy issue? “That’s a good question,” says Robertson. “It’s something we’d have to tick off before adopting the technology.”

Data sharing is another issue with which Robertson has to contend: “In the US, player data is openly shared between teams, whereas we have access only to Bulldogs player data, not the other 17 clubs.”

GPS devices lined up before a training session. Image: Michael Dodge/Getty Images


Technology is great for processing vast amounts of data almost instantaneously, thanks to high-speed data transfer that wasn’t available 15 years ago. Darren O’Shaughnessy, St Kilda’s recently appointed senior data analyst, recalls early days at Champion Data when he would send stats via a 28k modem. “Now Fox Footy receives an entire game’s statistics every few seconds on match day.”

O’Shaughnessy’s full-time appointment comes after seven part-time years with Hawthorn, reflecting AFL clubs’ ongoing drive to gain a competitive edge over rivals. The Saints’ general manager of football, Simon Lethlean, sees the appointment as the “way for us to go to the next level in how we analyse our recruiting, our game plan, and how we utilise all our data”.

Collecting data is one thing, but the key is to be able to use it intelligently. O'Shaughnessy cites one example: “We identified a player who produced outstanding numbers whenever he played in the midfield, and it seemed we’d uncovered an elite talent. But closer inspection of the data revealed coaches were only giving him a midfield role in latter stages of games, when the match was already won or lost. It provided no measurement of how he might handle the role when the heat is on.” This demonstrates the importance of separating the useful data from what O’Shaughnessy calls ‘noise’.

Therein lies the key to successfully using technology. It can help in many ways. Ultimately, though, says Robertson, it should be part of a suite of tools at a football club’s disposal, not necessarily a ‘magic bullet’. But if there is a piece of technology out there that might just give clubs a premiership edge, Robertson and his team will be only too willing to embrace it.