Australia have had a bittersweet day on Day 1 of the 3rd Test of the 2013 Ashes at Old Trafford. On the one hand, it was their best batting day of the series. Their captain made a century, and one of their up and coming players is 70 not out at the end of the day’s play. On the other hand, one of the three wickets they lost was an umpiring mistake. If the ICC plays to form, we will not learn any specifics of the communication between the TV umpire Kumar Dharmasena and umpire Tony Hill. Based on the available evidence, the decision to uphold the on field decision against Usman Khawaja can only be seen as an umpiring mistake. If Umpires are involved, there will be mistakes.
Yes, mistakes. Not errors. An Umpire may misread a video, or make a mistake in putting together the evidence from sound and the evidence from video. If he gets the chance to see it again later in the day, he may reach a different, correct, conclusion. He may correct his mistake. An error is a more complicated matter. Suppose an Umpire correctly reads the heat signature and conveys this evidence. Later, suppose he sees the ‘snickometer’, over an iced tea after the days play, which shows convincingly that hotspot had given a false positive. Here, the umpire has made an error. An error is a wrong decision which is reached despite the umpire making no mistake. Errors arise out of the very real possibility that evidence available to umpires can be contradictory, or worse, incomplete. Errors are, as I hope you will deduce from this definition, confounding than mistakes.
The decision against Usman Khawaja was not the worst instance of DRS use on the first day. The worst misuse of DRS came from England when they unsuccessfully reviewed an LBW appeal against Steven Smith.
It was the first ball of the 41st over. Smith played back to a ball from Graeme Swann which pitched outside off stump, gripped and turned back towards the stumps. It hit Smith above the knee roll, on the flap of his left pad. Smith is a right hander, so in cricket parlance, this would be his front pad. Tony Hill gave it not out. After prolonged deliberation with wicketkeeper Matthew Prior and bowler Swann, Alistair Cook requested a review “in the nick of time” as commentator David Gower observed live on air.
There was nothing inherently wrong about Hill’s decision. Umpires have long gone by the thumb rule that if the batsman is hit above the knee roll, the ball will tend to miss the stumps. On commentary, the height worried Sir Ian Botham as well.
Why did England review it? Let us assume here that they knew it hit above the knee roll. Given that Cook was standing at silly point, this is a fair assumption. Cook, Prior and Swann all appealed spontaneously (as they say) for the LBW decision in real time. So let us also assume that they genuinely thought it should have been given. The length of the deliberation between Swann, Prior and Cook is telling. It was not, evidently, an open and shut case where they were sure the Umpire had made a mistake. England reviewed a decision because it was close, not because it was clearly out.
As it happened, Smith was saved by a fraction of the width of the seam. It was that close.
Would Tony Hill’s decision have suddenly been a clear mistake if that ball had been a fraction of a millimeter closer to leg stump? Would Smith have been fairly declared Out had that ball been a fraction of a millimeter closer to leg stump? You can be sure that when Tony Hill made his decision, he did not say not out because it was that small fraction of a millimeter off the center of leg stump. Unless umpire Hill is a mutant cyborg, I am fairly sure that he doesn’t have a metric scale embedded in his brain. Umpire Hill’s mode of judgment (like Ian Botham’s) was probably more akin to a binary decision tree. Hit in line? Yes. Impact? Above the knee roll of the pad. Doubtful. Not out.
England reviewed the decision because (1) it was close, (2) it was still early in Smith’s innings, and (3) Australia was 135/3 at the time. Even if you don’t agree with the 2nd and 3rd reasons I outlined there, the fact that they appealed because it was close is a huge systematic problem for DRS. Here was a team requesting a review for a decision because it was close not because they thought it was not even remotely reasonably not out.
At least in Khawaja’s case, it is reasonably clear that what the umpires decided is at odds with the evidence available to us on TV. Smith escaped being dismissed in a Test Match by a fraction of a millimeter because the opposition requested a review of a perfectly reasonable decision by umpire Tony Hill.
When an umpire’s decision can be reversed by a fielding team even though it is perfectly reasonable using DRS, then the design of DRS as a system to overcome howlers is in serious jeopardy. As long as we have human beings umpiring Test matches (either on the field, or, in the not so distant future, possibly as technicians operating digital machines), we will see mistakes. That is unavoidable. Even the heat signature technology can be compromised by human beings making mistakes. But when a system permits the kind of review Cook requested against umpire Hill and Steven Smith, then it we have a more fundamental, and systematic problem.
This in a nutshell is the systematic problem of DRS. The system was devised to correct clear decisions, but has been repeatedly used to reverse close decisions. Cook’s use of DRS was the worst of the system on the 1st day at Manchester precisely because it could have been easily predicted that Cook would use DRS for a close decision. Close decisions are those which are not clearly out or not out, but could reasonably be either. When commentators say “that’s close”, they usually mean “that’s close to being out”. It does not mean that its clearly out. This applies especially in the case of LBWs where a judgment about things which don’t actually happen is required. Caught at the wicket, or run out decisions are basically different because the only things that need to be judged are things which actually take place.
Systematic problems are worse than the mistakes people make because they are inherently fallible as humans. But they can be solved. Cricket needs to decide if it wants LBWs to be judged by human umpires, or by ball trackers.
There is an argument to be made for letting the ball tracker determine LBW decisions. This would cause appeals like Swann’s against Smith to be given out. Perhaps this would be a good thing. If Cricket decides to use ball trackers to determine LBW appeals, I would be all for it. I think the technology is robust, and while it will, from time to time, produce decisions which appear at odds with the current convention in the cricket community about which appeals are out and which are not, these conventions will gradually change. It won’t be long before we accept ball tracking decisions as the norm.
There is also an argument to be made for letting umpires determine LBWs and allowing only limited reviews of things which actually happen – the pitching point, the point of impact, the presence (or absence) of an inside edge – and leaving the predictive path to the umpire’s judgment on the field. The first option may result in a large shift in the convention of what a good LBW appeal is. The latter would be the more conservative solution, leaving the conventions of the LBW decision more or less intact.
But that is a choice which Cricket will have to make affirmatively. It cannot hope to fall ass backwards into that kind of clarity.
Given that players are invited to play a central role in umpiring decisions, their use of the system must be scrutinized. A desperate, hopeless (and perhaps hopeful) review by the last remaining batsman for a plumb, clear LBW decision against him is perhaps more understandable and forgivable, and is a lesser sin, than a calculated gamble about a close decision. Similarly, an honest mistake by an umpire is a significantly smaller problem than a calculated gamble about a close decision like Cook, Prior and Swann attempted.
