KP's somewhat dubious LBW at the Oval yesterday prompted a debate about Hawkeye on TMS. Jonathan Agnew said he'd attended a recent demo of its powers at Lord's that had convinced him of its accuracy, despite previously holding the view of most people who watch a lot of cricket that some of its projections are about as accurate as an old umpire's eyesight in the summer dusk.
He cited a leg before decision on AB De Villiers from the winter, which all who saw it felt was plum, but was overturned by Hawkeye on height. It was replayed in the demo, and proven correct.
However, and here's the question: Hawkeye works on absolutes. It predicts the path of the ball based upon what has happened to it until it's been intercepted by the pad. Which would be fine if the flight of a cricket ball was always predictable. Hawkeye wouldn't, for example, be able to forecast one that goes past the stumps and then suddenly swings just before it reaches the keeper, as happens often at Lord's. The greater the distance it's asked to predict, the more fallible that prediction must be.
If the flight of the ball was always even, the batsman wouldn't miss it that often, would he? Wasim Akram's toecrusher, veering in at the absolute death, would defy Hawkeye. The problem with Pietersen's decision yesterday was that the device claimed the ball would have struck leg stump full face, when all of the visual and instinctive evidence suggested it might have brushed the outside of the stump at best.
Hawkeye is a good thing. It has improved the game. But it really can't claim to be infallible... after all, what is?