New Zealand Rugby has partnered with Wellington data science and AI firms to build a new analytics platform, known as “Play in Grey”. The platform, which runs on Amazon Web Services’ public cloud, uses computer vision and machine learning to provide coaches with “near real time” analytics generated from game footage.
Each minute of game footage generates over one million data points on player actions and match events through machine vision algorithms created by Intela AI. The raw data is then converted into usable information via data science firm Dot Loves Data’s statistical modelling tools.
The companies say it gives coaches information on things like player locations, speeds, proximity, and running paths in “near real time” on the Play in Grey Platform.
Jason Healy, New Zealand Rugby’s performance analysis manager for high performance – the NZR division responsible for the country’s top teams including the All Blacks – told Which-50 the new platform holds a lot of potential and the final use cases are still being explored.
“There’s opportunities for us to see how it can enhance performance on the field [and] off the field,” Healy said at the AWS re:Invent event in Las Vegas earlier this month.
“[It could] provide us with better workflow design so we can buy more time, so we give coaches more time to do better planning or better thinking.”
Those time savings are particularly useful when teams are touring or playing major tournaments like this year’s Rugby World Cup, where New Zealand finished third.
Is technology changing rugby?
New Zealand Rugby already has an established analytics program with SAS as its official analytics provider. For a tournament like the World Cup analytics on both the opposition and the All Blacks performance typically come from a data warehouse. Analysts provide coaches and management with a wide range of information, Healy says, but how it is used will vary by coach, opposition, and certain scenarios.
The new Play in Grey platform should provide even more data points to add to the arsenal but ultimately it is only one tool, according to Healy.
“The bottom line is, no matter what we provide the coaches and what we provide the wider management teams with, in teams of information, players still make decisions on the field in the moment,” Healy told Which-50.
“Data plays its role but it’s just a tool.”
Asked whether the explosion in data collection and analytics has changed the game in the last 10 years, Healy said it still varies by teams, players and locations. Northern hemisphere teams, for example, have typically played in a tighter, less expansive style than their southern hemisphere counterparts.
“[Rugby] is still a sport that is really bound by some traditions … When the technologies come in I think that fulfils a different need for different people,” Healy said.
“So it could be observational recall [or] it could be about building their own knowledge set around what’s going on because they don’t know. For some it does form an important part of the decision making.”
How the technology is being used by coaches also varies, Healy explained. Some will add data insights after their own analysis of games, then combining it all to create messages for players. While others start with data, using the insights to identify particular moments in a game.
“I think it provides the toolbox for coaching and a toolbox for planning previewing and reviewing [games]. That’s changed a lot because there’s a lot more resources available to them then there has been in the past.”
Healy also notes that, unlike many American sports where analytics have led to fundamental changes in how the sport is played, the unpredictable nature of rugby means fans don’t yet need to worry about analytics creating a homogeneous play style.
“[In rugby] you could have 20 to 30 different variations of off a set piece. Even pattern recognition inside rugby is something that’s very difficult to do. I think the bottom line is [analytics] informs a player and they can better assess or evaluate their performances or better evaluate their opposition’s performances.”
The author traveled to AWS re:Invent as a guest of Amazon.