It’s not a great time to be an Australian cricket fan. The recent home Test series against South Africa was a sign of just how far the team have fallen, with Australia only able to snap a five-Test losing streak with a seven-wicket win over the Proteas on home soil. So many pieces have been written about the decline of Steve Smith’s side that it’s difficult to envisage the captain finding any truly original solution to the problem. But Satya Nadella, Microsoft CEO and lifelong cricket fan, is stepping in to help.
Before Nadella visited Australia in November, the AFR revealed that Microsoft is to assist Cricket Australia with its data analytics, ‘deploying a new performance analytics system that will be trialled this summer,’ according to Business Insider. The platform - which aims to make it easier for coaches to interpret data and turn it into actionable insight - will be used by coaching and support staff from both the test and state teams to try and improve performance.
The Melbourne Cricket Ground was one of the first landmarks Nadella visited, and his interest in both cricket and analytics have clearly conflated in something of a passion project for the CEO. ‘It’s been one of the dreams all of my life,’ the Indian-born head of Microsoft said. ‘And I know it’s not the best day to talk cricket in Australia. I was glad to see how Cricket Australia plans to use machine learning. Cricket is one of the richest sports when it comes to using machine data. But how could you harness the power of data to even start having a more intelligent informed conversation about performance of teams, performance of players?’
Microsoft’s technology seeks to answer one of sports analytics’ key issues: often there is simply too much data for analytics teams to make sense of. So many metrics are recorded in all sports that an incredible amount of data is hoarded, leaving those involved in sports analytics all too often struggling to find the valuable insight among the noise. This is why those less willing to embrace technology in sport see analytics as a distraction, a swarm of data points disconnected from the reality of the pitch, the court, or the field.
Context is key. In 2009, Bill Belichick baffled Patriots fans in what’s become known as the ‘fourth-and-2 game.’ Forgoing the punt on his team’s own 28-yard-line, Belichick went for it, his team couldn’t convert and Payton Manning won the game for the Colts. Fans were baffled, but the context of both the game and the stats backed up Belichick’s unconventional call. If the fourth down had been converted, the Patriots would’ve held possession, giving them a 92% chance of seeing out the game. This drops to 79% had the team punted. 79% is good odds, you might think, but given that over half (55.7%) of fourth-and-2 attempts had been successful the previous two seasons, Belichick’s call makes analytical sense.
The decision makes even more sense when you consider the context of the game, though. Manning was having a trademark excellent game, and punting would’ve put the ball right back into his hands. Analytics isn’t yet on Belichick’s level - had the play come off, he would’ve been lauded for it - because it alone is unable to take into account the level of context a coach can. Microsoft’s intelligent platform hopes to bridge this gap with a ‘visual and intuitive dashboard,’ and the ‘latest in machine learning and predictive analytics technology.’
The company claims that the work it is doing with Cricket Australia, with a system far and above that which is used in dressing rooms and on pitch side at present, will ‘shape the use of data in other sports as well.’ If it can find a way of genuinely making data more intuitive and actionable, it just might.