Meet the Australian cricket team's secret selector
- 11 January, 2017 15:24
There’s a new coach on the Australian cricket team's staff roster. This trainer is already proving hugely influential in deciding which players get picked and is one of high-performance chief Pat Howard’s most trusted advisers.
This star selector you’ve never heard of? A player performance platform powered by the Microsoft Cortana analytics suite.
The system, the first of its kind to be used by a cricket body, applies machine learning and predictive analytics to the vast body of data Cricket Australia collects on its elite player to assist coaches in making selections.
“It’s about giving our teams every possible edge,” explains Cricket Australia’s head of technology Mike Osborne. “The results will play out on the field.”
Over the past few years Cricket Australia’s sports scientists and coaching teams have become very good at collecting data. Every athlete’s training programme progress, including peak loads and chronic loads, their injury recovery progress and countless fitness and wellness indicators are continuously recorded. On match-days, a player’s every movement is tracked by GPS and every ball precisely measured and noted.
“Our high performance guys, they’ve spent the last number of years collecting data and getting very diligent and very good at collecting a lot of data around our elite players from around the country,” says Osborne.
“So we’re in this very lucky position of having just a wealth of data and the real trick now is what we take from that data and how do we help people consume it in a way that assists them in their jobs day to day.”
Those numbers are now being crunched by the platform which runs in Microsoft Azure and then presents the findings to coaching staff via a Power BI backed, visually rich dashboard.
“That will let you map the overall match readiness of the team and look at the players and understand who is in the right zone in terms of peaking in their training for the match and how injury programmes are progressing – will players be ready in time for a match?” says Osborne.
“It will look at how players perform in a given stadium, how they perform against different types of competition and give tips and insight for the coaching staff to help them train the players and help them decide on their tactics during an upcoming match.”
The project is being closely followed by Microsoft CEO and cricket tragic Satya Nadella who gushed about the possibilities during a recent visit to Sydney.
“Cricket Australia is beginning that journey using some of the same technology that is being used by industrial companies or healthcare companies or robotics companies,” he told developers in November last year. “And so the idea that developers can have or bend the curve even of cricket performance is perhaps one of the most empowering things I've seen in my life.”
Osborne, who worked for many years at Deloitte, says the use of data by Cricket Australia is comparable to any other business.
“Frankly it’s not dissimilar to the challenges that any enterprise faces,” he said. “Where you traditionally have lots of data but that data’s often siloed and so you really don’t get a holistic picture of whatever it is you’re looking at.
“You’re rarely going to make decisions on purely quantitative data, but what it’s trying to do, like any enterprise, is pull those data sources together to give you a much more holistic picture of what’s going on and provide an aid to whatever the process is you’re performing.”
Soon, the platform will offer coaches recommendations and suggestions, based on machine learning. Currently in a proof of concept stage, the platform is expected to be ready for next summer. It won’t, Cricket Australia is keen to emphasise, pick the team on its own and will remain simply a tool for human staff.
“The intent here is not to have a computer picking the Australian cricket team,” says Cricket Australia head of digital Finn Bradshaw.
“When you’re hiring a coach or selector – you don’t necessarily want to hire someone who is the best data analyst. Those guys have a lot on their plate – how can you present that information to them in a way that’s as digestible as possible and enables them to inform their decisions and the discussions they have with the players.”
“It’s not computer selection,” adds Osbourne. “It’s giving the selectors and coaches and staff more information so that when they’re talking they have better information than their competitors to help them put the right team in the right place on the field for the match.”