LBW is a unique law in cricket, possibly in all of sport. The law first appeared in 1774. At the time, it said the striker would be out if he puts his leg before the wicket with a design to stop the ball and actually prevent the ball from hitting his wicket. The law hasn't changed much for 237 years now.

In its current form, an lbw decision is judged at two distinct levels. The first judgement is about what actually happens - where the ball pitched, where it hit the pad, whether there was bat involved and whether the batsman offered a shot. The second judgement is about what doesn't happen. Would the ball have gone on to hit the stumps had the pad not been in the way? These two judgements cannot be made with equal certainty. The first event can be observed, but the second requires conjecture. Ball-tracking provides an alternative method of making this conjecture. If we are to assess this new alternative, it has to be compared reasonably to the traditional method used by umpires.

This comparison has already been made. Through internal tests, ball-tracking providers claim the technology is much more accurate than human umpires. The method used for this test has not been published. Nevertheless there are people who don’t need tests to be assured that ball-tracking is accurate and gives a definitive prediction about the path of the ball. Then there are those who don't trust ball-tracking because it has contradicted the expert judgement of umpires and players on some occasions. Still others are uncomfortable with it because precious little is publicly known about the method used in the prediction, and whatever little is known suggests that the method is removed from the manner of cricketing judgment.

What everybody should agree is that as a disembodied data-processing system, ball-tracking will always be far more consistent than even the best human umpire. When it comes to accuracy, if accuracy is measured in millimetres, a fair comparative study between ball-tracking and the umpire should take into account the capacity of the umpire to make mistakes. This is to say that in some lbw appeals an umpire may rule one way in real time, but when shown the replay might conclude that he made a mistake. This difference between an error and a mistake is crucial if one has to make a reasonable comparison between the relative merits of ball-tracking technology. It is not clear if this distinction is made in ball-tracking providers claims.

There is a difference between an error and a mistake as they are applied in this debate. A margin of error provides a range of answers within which the actual answer is claimed to be. A mistake is by definition outside the margin of error. So how might the margin of error be measured for human umpires?

Take the 50 best umpires in the world. Show each umpire the same set of 100 lbw appeals in random order, and record how the umpire rules for each appeal. Allow the umpire to see each appeal as many times as he wants, in order to minimise mistakes - not errors - in judging where the ball pitched, where the point of impact was, and whether or not there was an inside edge. If you then put together the range of rulings for each lbw appeal, you will find that some appeals will have been ruled out by a majority - or even all - of umpires, others will have been ruled not-out by a similar majority. A third set of appeals is likely to divide expert opinion. A few may even have half the umpires rule them out, and the other half not-out. So you will get some appeals in which the decision is out with a high degree of certainty, while in others the decision is out but with marginal certainty. For example, if 45 of the 50 umpires rule that a given appeal is out, then one can conclude with 90% certainty that the correct decision in the case of this appeal was out.

Here, the margin of error is not measured in terms of the distance from the stumps in millimetres. Instead, it is a measure of the consensus of expert judgement, with some effort to control for the human capacity to make a mistake.

If an umpire rules not-out in real time in a case similar to one that the vast majority of umpires in our experiment above ruled to be out, then the umpire has made a mistake. In other, more marginal cases, both decisions are acceptable, as is often the case in some lbw decisions. Two recent examples involving Sachin Tendulkar come to mind. In the World Cup semi-final, he was given out lbw by Ian Gould, only for the decision to be reversed. It was a perfectly reasonable, albeit marginal, lbw decision. On day five of the Lord's Test, Billy Bowden ruled Tendulkar not-out to a Stuart Broad appeal. It looked close, but Tendulkar had made a forward stride, and the not-out decision was reasonable. In both instances, other good umpires might have reasonably disagreed with the decisions on the field.

In the current DRS method, ball-tracking is used to verify an umpire's judgment. This can only be considered fair and reasonable if it is established that umpires reach their conclusions about lbw appeals using the same method as ball-tracking, but with lesser rigour. This is not the case. If Gould had a chance to see a replay of the lbw decision he gave against Shivnarine Chanderpaul in Barbados, when an offbreak hit him outside off and was turning further away, he would surely reverse his original decision. However, if the ball-tracking technology was asked to re-compute the decision against Tendulkar in the World Cup semi-final, it would come back with the same answer.

That is the difference that ball-tracking brings to lbw decisions. Under DRS Tendulkar was definitely not-out, albeit by a wafer-thin margin. With umpires ruling, marginal decisions remain truly marginal which is to say, that if the exact same appeal happened multiple times, there is no guarantee that the same decision would be reached on each occasion. Ball-tracking manufactures certainty. It is worth remembering that this certainty is currently situated beyond the actual ball-track, in all the arbitrary boundaries that have been set through the use of the umpire's-call zone and the 2.5m rule.

Ball-tracking gives cricket a choice. The lbw law could reasonably be given over completely to ball-tracking technology. Even existing technology would be good enough for this. The umpire on the field need not get involved at all. The ICC can introduce an elaborate set of arbitrary limits on what should be given out and what shouldn't, and we would get consistent lbw decisions. The umpire's-call zone could be far more sophisticated than it currently is. It could be redrawn to take in account whether the bowler is bowling over or round the stumps. It could be much stricter on leg stump than on off stump (cricket is a side-on game, and the leg side is considered to be the blind side, which is why you can’t be lbw to balls pitching outside leg stump). The umpire could be spared the humiliation of being put to the test every single time.

Or the umpire's hand can be strengthened, by allowing the third umpire to pro-actively correct errors about pitching point, point of impact and inside edges in lbw appeals, and, most crucially, by allowing the third umpire to use his expert judgment to advise the umpire if his on-field decision is obviously wrong with respect to the predicted path. This will not require-ball tracking. A simple replay would do.

Where ball-tracking will come in handy is to record each lbw appeal, and compare it to other appeals in other games. We could learn a great deal about lbw decisions this way, with ball-tracking not deciding an appeal but being used to build a database of lbw appeals and decisions.

Bowlers will probably prefer the first option, and will vigorously want rules that require lbws to be given even if the ball is shown to be grazing the leg bail. Batsmen will probably also prefer the first option, but will want the arbitrary limits to be very strict, and the benefit of doubt to be interpreted generously in their favor.

Whatever choice is made, the game will do well to keep in mind this basic fact in mind. The problem is not that umpires lack the expertise. It is that umpires are human.