DRS at Lord's: Kallis sawn off, AB, SA survive by a whisker

Jacques Kallis's final delivery in the first innings of the Lord's Test  in the picture below. The review against de Villiers follows immediately after. The first days play was a poor one for DRS, for reasons laid out in this post.

There are two things of interest in this episode, apart from the fact that Kallis was definitely given out incorrectly even if we agree that the ball hit his glove.

Law 32(1) says the following:

"The striker is out Caught if a ball delivered by the bowler, not being a No ball, touches his bat without having previously been in contact with any fielder, and is subsequently held by a fielder as a fair catch before it touches the ground."

As with many such instances where rules are defined, "touches his bat" in this case does not literally mean the ball touches his bat. "Bat" is a technical term in the Laws of Cricket, and is defined under Law 6(8) as follows:

8. Contact with the ball
In these Laws,
(a) reference to the bat shall imply that the bat is held in the batsman’s hand or a glove worn on his hand, unless stated otherwise.
(b) contact between the ball and
either (i) the bat itself
or (ii) the batsman’s hand holding the bat
or (iii) any part of a glove worn on the batsman’s hand holding the bat
or (iv) any additional materials permitted under 3, 5 or 6 above
shall be regarded as the ball striking or touching the bat or being struck by the bat.

In Kallis's case, the only part of his body the ball could have hit, is the bottom hand after it had already come off his bat. Almost 7 years to the day after a Test Match was decided because Billy Bowden failed to notice the exact same thing, South Africa's best batsman has been done in due to this oversight.

But what of the more basic question? Does the heat signature and/or the replays show conclusively that the ball hit the glove at all? In my opinion, they don't. There is no camera angle available from which the part of the glove where the ball apparently hit is visible before the ball hit. From the front on angle (behind the bowler's arm) the view is obscured because the part of the glove the ball hit is on the back of the hand when viewed from this standpoint. From the side on angle, the view is obscured because the part of the glove behind the bat handle and front hand. Hotspot is not even clear about whether or not there was any contact, let alone be conclusive about this. This is also awell known problem with Hotspot, and has been acknowledged by Warren Brennan of BBG Sports, the company that produces the device.

What of the deflection? The only camera from which a deflection can be reliably identified is one which is absolutely steady and in which the foreshortening (i.e. the zoom, in a simple sense) does not change at all from start to finish. In all the camera angles we saw, the foreshortening does change, and the camera pans along with the ball (see the camera angle from leg slip for this). In such an instance, it is very easy to see deflections where there are none. The other tell tale sign when there is a deflection - an abrupt change in the rotation of the ball is not visible either.

All the technology is inconclusive in Kallis's case. What I suspect happened is that as Kallis's bottom hand came off the bat, his glove brushed the shoulder of his bat. This was the source of the noise.

Umpire Tucker made a mistake, not an error in this case simply because he disregarded Law 6(8) which guides Law 32. For an explanation of the distinction, see this. But what of the original on-field decision? Even if the problem with Law 6(8) had not arisen, it would still have to be classed at worst as an error (due to the unaccounted for noise), since the replay did not provide any information to suggest that the on-field decision was obviously wrong.

In the case of Kallis's decision, Umpire Tucker's mistake had several consequences. A subsequent review against A B de Villiers, which should never have been available to England became available. They had already lost a review when Steven Finn caught Alviro Petersen above the knee roll on the front pad early in SA's innings.

DRS survived by a whisker, for that AB de Villiers LBW decision (which he survived by a whisker) demonstrates by DRS in its present form does not distinguish between obvious and marginal decisions. A miillimetre closer to off stump, and de Villiers would have been sawn off by a decidedly marginal and bold LBW decision going against him due to a review which should not have been available to England in the first place.

Michael Atherton's comment on live commentary was telling. He said "I think they'll review this one. It looked very very close!". Not only does it betray utter contempt for the Umpire, it also implicitly legitimizes the use of the review for "close" decisions, as opposed to obvious errors.

When seen live, and then when seen on the replay, it was always clear that the decision was going to be marginal on off-stump as far as point of impact was concerned. The ball-track confirmed this. The new rule change (which has not, to the best of my knowledge gone into effect yet), expanding the umpire's call zone for point of impact, would make this a far easier decision. At Lord's AB de Villiers was merely lucky that the point of impact was in the ball-track's ridiculously narrow zone of uncertainty. Even if that decision had gone against de Villiers and Umpire Dharmasena, it could not be reasonably said that Dharmasena made an obvious mistake. In fact, he made absolutely the right decision - given that he was not absolutely sure whether the impact was within the width of the stumps, he gave the benefit of doubt to the batsman.

In a sophisticated version of DRS, the third Umpire should be able to look at that replay of the appeal against de Villiers, and determine that this was not an obvious mistake by the Umpire.

This keeps happening in DRS. Appeals are reviewed because they are "close". By the ICC's own admission, DRS was not supposed to consider for such appeals, but the way it has been designed, it is forced to.

This is the deepest (and in my view, fatal) problem with DRS.