Could high-tech help the All Blacks regain the Rugby World Cup?
New Zealand Rugby is trialing “Play the Grey” - a new video-analytics solution where AI (artificial intelligence) software “watches” a live video feed of the game, then feeds data about player and team performance to an iPad-totting coach on the sideline in real-time.
The idea is that by analysing this data, which includes details such as player locations, speeds, proximity, and running paths, coaches gain new insights that help teams gain competitive advantages. For example, if a team liked to shift the ball to the edges from set piece, push it infield, then kick to the corners, the coaches could devise a counter-strategy.
Its makers say it also has the potential for AI-powered talent scouting.
It would also depend on how useful others further down the food chain ultimately found the technology.
“There’s always an interplay between technology and people,” Healy said. “You can have a technology mindset, and technology is incredibly important, but people make the difference.”
For now, Healy stressed the union was still in an "exploratory" phase with the Play the Grey.
NZR wouldn’t enter a commercial contract, and start using the technology with tests and other matches until it had proved itself.
"It's a three-part process," he said. "We have to see if there's an appetite for it; we have to understand it and see if it produces useful learning - and then whether that learning can be transferred to the field in near real-time."
The All Blacks already use a lot of technology and a raft of historical stats and current GPS data was available to Steve Hansen as he scowled at his laptop in Yokohama.
But “Play the Grey” could take it to a whole new level. The solution was put together by two Wellington start-ups, Dot Loves Data and Intela AI and is powered by two services run by Amazon Web Services (AWS - the cloud computing division of Amazon): Amazon Elastic Container Service (Amazon ECS) and Amazon Simple Storage Service (Amazon S3).
Intela AI founder Asa Cox told the Herald that the trio decided to start with rugby, “because that’s the hardest sport.”
Football and netball are relatively easy for an AI to follow. The more compact rugby, where the location of the ball - or identities of players - are often hard to discern under a pile of bodies in a ruck or mall, posed a much bigger challenge.
Healy thought his product can overcome it, and get better over time. “AI needs to learn just like humans. It needs to be trained,” he said.
If everything works out, then Cox and co hope to market Play the Grey to sporting codes around the world, including NFL in the US and various football leagues around the globe.
POSTSCRIPT: Man vs Machine
Machine-learning software helping the All Blacks is one thing, but what about robo-ABs?
A couple of years ago, I attended an event that saw legendary Andrew Mehrtens take on three robots created by Massey and Canterbury universities and pneumatics company SMC in a kickoff.
Only one of the three robots could get the ball over the bar.
The carbon-based life-form known as Merts managed a perfect record.