Add Category
Updated
The captain's challenge. The new 20/40. Outlawing tackles for players in mid-air. Midfield scrums.
Key points:
- The NRL hosted a competition for data scientists using player GPS data over the off-season
- The team that finished second was mainly filled with people without deep rugby league knowledge
- Their analysis of lateral movements in rugby league was praised by judges
All eyes will be on 2020's rule changes when the NRL season kicks off this weekend.
But despite all the tweaks, rugby league's essential character remains.
Rugby league is about taking the ball forward.
The fierce attack meeting a breaking wave of defence, surging their 10 metres back and forth. Field position means distance from the tryline. Run metres only count if they're forward. Forwards crave hit ups, not hit downs.
But to the rugby league community, the obsession about forward progress may have hidden the importance of something else.
The side-to-side.
At least that's what a team of mostly rugby league novices told a collection of the smartest NRL minds over the offseason.
As part of the NRL DataJam competition, a group of data scientists from Sydney firm Edge Red quantified just how important side-to-side movement is in rugby league success.
And it's not just the cut out pass, the crossfield kick or Benji Marshall's (early career) step … it's across all facets of play in attack and defence.
"Teams that run more laterally also happen to be the teams with better win records," explains Edge Red co-founder Howard Gu.
"Especially in the situations where the attacking team has a great opportunity to score a try … where the offensive team moved more side-to-side (than the defensive team), they almost had a 300 per cent greater probability of scoring."
Monica Ly, the firm's other co-founder, says most of her team did not have much of an appreciation of rugby league, including herself.
"I'm a huge data nerd and so when I heard about this, I just wanted to put my hand up to get my hands on GPS data to do some analysis on location."
As part of the competition, the NRL provided detailed location data of players from last season — information which is otherwise closely held and protected by clubs.
It tracked not only the movement of the ball side-to-side, but also recorded the location of every player, every second.
That allowed teams to measure how much ground each player covered during each play, including lateral movements — the distance players shift left and right if facing the tryline.
"We thought that we would just look at how fast people run towards the goal line," Ly says.
"But then when we're looking at the data, we realise that people actually make intricate movements from side to side."
First lateral measurements
Their effort was praised by judges such as Wests Tigers coach Michael Maguire and former NRL player Michael Ennis, and was awarded second place on the day, behind a play simulator developed by Suncorp.
Wests Tigers high performance manager Andrew Gray says Edge Red's work was fantastic.
"Lateral movement and the ability to maintain the integrity of a defensive line is the hallmark of a good rugby league team, and until this point, probably hasn't been measured," he says.
"It's something that coaches like myself work tirelessly on trying to improve within a team.
"It's something that tactical coaches speak about ... being shoulder to shoulder with your teammate, and keeping a line together.
"But to put some objectivity to that — and to show that better lateral movement leads to success — was fantastic."
Fresh perspective
Other teams participating in the DataJam were full of passionate rugby league fans.
Far from that being a disadvantage for Edge Red, Ly says she felt like that gave them a competitive edge.
"Our one goal was to communicate one thing about NRL to people who don't know the sport," she says.
Gray agrees a fresh perspective is often valuable.
"A lot of times when you look at a data problem with a preconceived conception of what you're going to find you have a bias, which changes the way you approach the question.
"You're probably looking for what you think you're already going to find.
"Having somebody look at the problem without any bias to what they think the answer is going to be, I think probably uncovers some insights."
Side-to-side smarts
Gu says this is only the first stage in helping improve rugby league using lateral thinking.
"Now that we can quantify it, we can now get into more detail around, you know, how can we optimise what is good versus what is bad?
"Our approach was to start that conversation, but then now continue pushing to try and optimise what we can now quantify."
Ly says her enjoyment working with the data has almost made her a rugby league fan.
"What this experience has given me is the awareness and the appreciation of some of the complexity around sport, which may look boring, but actually, it's quite interesting when you look into it."
She's also adamant some of the game's charms are beyond the reach of data science ... for now.
"Even if the probability says that you should be scoring, when you watch the video, and you watch what happens, there's actually a lot of things that happen in the background which contribute to whether or not you score a try.
"No matter how many mathematical analyses you throw at a problem, you can't 100 per cent pinpoint what's going to happen."
The Parramatta Eels play the Canterbury-Bankstown Bulldogs in the NRL season opener at 8:00pm today (Thursday).
Topics: sport, rugby-league, australia
First posted