/ 20 January 2010

Inventor of Hawk-Eye defends technology

South Africa cricket coach Mickey Arthur has become the latest prominent figure in the sport to express reservations about the Hawk-Eye technology which is now being used to check leg-before-wicket decisions.

“I’m not 100% convinced about the predictive element of Hawk-Eye and I don’t think many players are either,” Arthur wrote in the February issue of Wisden Cricketer which goes on sale this Friday.

Hawk-Eye uses data produced by video cameras to predict whether a ball that has hit the batsman on the pads would have gone on to hit the stumps, the so-called lbw decision which can be among the most contentious in cricket. The technology is used in tennis to decide disputed line calls.

Under the Umpire Decision Review System (UDRS) implemented in the recent four-match series between South Africa and England and in six Tests in Australia, a third umpire uses the Hawk-Eye to help decide whether the on-field official has made a correct decision in the event of an appeal from either side.

The case against Hawk-Eye was put by former Australia fast bowler Dennis Lillee who said: “There’s no way Hawk-Eye can tell if a delivery is going to skid a bit more than normal or hit a crack, or a damp or worn patch, or a bit of grass on the wicket.”

It is a case that is robustly rejected by Paul Hawkins, the inventor of Hawk-Eye and a former player with the English minor county Buckinghamshire who graduated from Durham University with a doctorate in artificial intelligence.

“Dennis Lillee’s accuracy as a bowler is far greater than the accuracy of his journalism,” Hawkins said in a telephone interview. “He has no understanding of the technology. He has got no idea what he is talking about.”

“It still takes people time to buy into things they can’t understand, that’s probably the hardest thing.

Correct decisions
“Hot spot [images produced by infrared cameras which can show whether the ball hit the bat or not] scored very high on believability but actually its accuracy and reliability aren’t very good.

“We score very high on our accuracy but our believability is what is harder because we use a virtual world.”

Hawkins said Hawk-Eye calculated the trajectory of a ball after it had bounced and hit the batsman’s pad, based purely on the behaviour of that particular ball.

“You only need about a foot [30cm] to travel and you get an accurate prediction of the line, you need a little bit more to predict the height.

“If there is very little distance between pitching point and hitting the batsman, that is the hardest time to make an accurate prediction. That’s exactly the same for an umpire.

“We don’t measure before the ball is bowled, we measure what has happened on that ball.

“We don’t know if it’s kept low because it has hit a crack or skidded or not hit the seam or whatever. But we have seen what has happened to that ball by measuring it post-bounce just as an umpire does.”

Hawkins said the umpires in each of the 10 Tests in which Hawk-Eye had been used to help determine lbw decisions believed each decision had been correct.

“We have now 10 Test matches and most of the referrals have been lbws. We have certainly got every single one right,” he said. – Reuters