“Aussies may face virtual Murali” – So they said in 2007 but what has transpired?

John Coomber, 3 August 2007 in http://www.brisbanetimes.com.au/news/cricket/aussies-may-face-virtual-murali/2007/08/03/1185648106244.html

Australia’s cricketers may soon be able to go into a virtual reality studio during a Test match and “shadow bat” against the bowlers they are about to face in the middle. Cameras will be set up to capture as nearly as possible a batsman’s-eye view of the opposition bowlers, and relay the feed to a studio near the Australian dressing room. Players padded up and waiting to bat will be able to rehearse their innings using images gathered from the middle, and projected life-size back into the pavilion.

If the system works as hoped, it will mean that a player like Mike Hussey can go out to face Muttiah Muralitharan having already got his eye in against him in real time, with the ability to replay deliveries he found difficult. He could, for instance, use the cameras to polish his technique against Murali’s deadly “doosra”, the apparent off-spinner that turns the other way.                                                                    

 Were they considering this Murali?     OR 

    this Murali?  …… this photo is courtesy of Dr Ravi Goonetilleke via hilal Suhaib’s good offices

“A batsman could pick up the cues in Murali’s bowling action on that particular day so they know when it’s coming,” said Marc Portus, manager of the sports science unit at Cricket Australia’s Centre of Excellence, which is this weekend marking its 20th anniversary. “We’re not quite at that stage yet, but it’s where we’re aiming to go — possibly as early as next summer. “We’ve done it with footage in a training environment and simulated game scenarios, but we’d really love to start doing it in matches. You can’t beat the real thing.” The virtual batting studio is unlikely to be functional in time for Australia’s two Tests against Sri Lanka this summer, in which Murali requires nine wickets to beat Shane Warne’s world record of 708 Test wickets.

Virtual batting is one of the startling technological developments aimed at maintaining Australia’s position as the No.1 cricket nation in the world. No longer do players rely solely on traditional net sessions and fielding practice to reach peak performance. Work is underway on a number of projects including GPS satellite tracking of players through matches (to help tailor specific training programs), computer analysis of career patterns to aid selectors, data mining of opposition teams, special goggles to help pinpoint how top batsmen use cues from the bowler’s arm to assess line and length before the ball is delivered, and so-called neural network software programs to predict performance and injury.

A GPS tracking device worn by fast bowler Nathan Bracken in a one-day international in Sydney last season showed he covered 18km during the course of the game. According to Portus, information like that can help physical conditioners devise specific fitness programs for individual players.

Research is now showing that the numerous sudden accelerations required in cricket are part of what makes it such a physically demanding game, even if to the casual observer, players appear to spend much of a day’s play doing not very much.

Data mining can give insights into the strengths and weaknesses of opposition players, and identify patterns which can be exploited. Cricket officials are reluctant to give details, but it is known that the Australian team used information gained from data mining to help wrest the Ashes back off England last summer.

Neural networks are another branch of data crunching about to have a big impact on professional sport. Adapted from manufacturing programs, they use performance-predicting software which has the ability to learn from its own mistakes.

“You feed information into it and get it to predict outcomes like when players might be injured, or averages, or results,” Portus said. “If it predicts the right outcomes – fantastic. But if it doesn’t, the sophistication in the software allows it to start to learn why it didn’t. It will make the necessary adjustment, so next time the information goes in it will learn from its past mistakes. “It’s quite sophisticated. We haven’t dived in to neural networks yet, but it’s only a matter of time.”