The (mis)information about DRS
ShaneWarne: “I’m going to keep banging on about this. If a captain reviews a decision-then the on field umpires decision should be removed-as you can’t have the same ball being out or not outF ! Once this happens, it’s simple and clear-whether it should be out or not! @BCCI @ICC @HomeOfCricket”
This is one of the many tweet-rants by Shane Warne on the DRS. He frequently raises this objection which has resonated with many around the world, including, unfortunately, with God himself. Today, I will try and explain it as best as I can.
First, let us understand the need of DRS and how it was envisioned to help decision making in cricket. At the core of it, the DRS was introduced to remove umpiring howlers from the game. This effectively meant that teams should be able to overturn visibly wrong decisions by umpires. Think Steve Bucknor vs Sachin Tendulkar. But with the introduction of neutral umpires and excellent umpiring standards, the DRS quickly became a tool to punt on marginal calls and that is when the problems arose.
DRS components
The DRS consists of two separate technologies, Ultraedge and Hawkeye, as they are popularly known. In this, Ultraedge, mainly used for close catches, is the non-controversial sibling which only occasionally makes news but most of the times it keeps its head down, detects level of noise on a graph which the on field umpire may have missed and goes home to sleep peacefully every night.
On the other hand, we have the Hawkeye, which has gained so much notoriety that Marvel decided to cash in on the publicity and released an entire show about it. The Hawkeye’s job is comparatively bigger and more difficult. It first has to track the path of the delivery and then, post the point of obstruction, is required to predict the path of the ball. This throws up a fundamental difference between the two; for one, Ultraedge only reports changes in noise level, leaving the conclusion to the umpire, while Hawkeye also concludes, giving major FOMO to umpires. Secondly and more importantly, Ultraedge lives in the present (secret to its happiness), while Hawkeye is always called upon to predict the future, which is a task, degrees more complicated.
Objection
Fans and ex-cricketers (maybe current ones too if the ICC allowed), frequently call out to the “Umpire’s call” feature. The argument goes exactly like Warne’s tweet above: If it pitched in line, it did and if it didn’t, it did not. This gets even more vociferous when a batsman stays not out despite the ball clipping the stumps in the Hawkeye projected path. Because, even if the ball touches the stumps, the bails would come off and the batsman should be out.
Explanation
As an engineer by education, this confidence in technology by the critics is touching but misplaced. I think the technology here has shown great humility in admitting that it is not accurate.
When I used to work on a lathe machine, we learned of a concept called margin of error. The idea was that irrespective of how good or precisely you craft a metal shape on a lathe, it is highly unlikely that you will match the dimension specs precisely and hence, the specs always came with a range and not an exact number. This was typically x(+/-) 2%. So, if you had to make a rod with a radius of 100 mm, one with 102 or 98 mm was equally good and acceptable. As tools got automated, the margin of error kept reducing but never disappeared.
Now, apply this to the Hawkeye; a delivery that’s pitched/made impact in line to the naked eye, is sent upstairs. Now, if programmed for binary answers, the software will either rule for or against it. But, if it is so close to the line that it falls with in the margin of error, the technology politely bows, states its inability to decide and asks the human to take a call.
It is even more difficult to predict the likely trajectory. In my limited understanding, the software predicts multiple paths post impact and if we were to be shown that graphic, it would be seen as circle with a red centre, fading to a pinkish circumference, encompassing all the paths the ball can take, with the opacity of the colour depicting probability of a path. So, the ball is most likely to be at the centre red dot BUT this still has<100% certainty. This means that the ball seen as clipping the bails, is LIKELY to clip the bails but even a cm of away movement would be enough to save the batter.
It is very difficult to have visually seen half a ball hitting the stumps and then dissociate oneself from the animation but it is still only an animation and not actual turn of events. To elucidate this point, we have all seen deliveries, which is traveling in a straight line, on its way to the keeper but suddenly swings viciously to beat him and roll to the fence. The same delivery, DRS would predict going to the keeper’s gloves uneventfully.
So effectively, what the DRS is saying is this, “The ball is likely to have pitched/impacted/clipped BUT, it is so marginal that it falls within my margin of error and I am not certain about it.” This, I believe is a very honest and fair response from the software and I am sure it is checked rigorously to minimise the margin of error. As it gets better and consumes more data points, I think the margin of error will keep getting smaller, without ever disappearing.
Conclusion
The DRS was introduced to remove howlers, which it has effectively achieved, unless you count THE Headingley Test where Paine could not review a howler against Stokes because he had burned them all. What it now suffers from is not a gap in technology but in communication. It points to the ICC’s unwillingness to take it to the masses when it does not even consider educating commentators about the nuances who then spread their mistaken opinions to the viewers. I’d also like it to show all probable predicted paths to give viewers a sense of mechanism but I guess a ball clipping stumps is more box office.
PS 1: A friend recently pointed out that the DRS does not consider cases where its conclusion differs from the on field umpire in 2 or 3 separate dimensions. Consider this: The umpire thinks the ball pitched in line but missed the stumps and rules not out. When called upon, the DRS says that while the pitching was umpire’s call, the ball would have definitely hit the stumps. Ideally, I think this should be out but since it was an overall not out when it was sent upstairs, an umpire’s call means that the decision stays, irrespective of the umpire’s opinion on that dimension.
PS 2: And now for my own personal issue with the DRS. I think it’s high time that the DRS considers the probability of the delivery hitting middle of middle but the bails staying intact. Connoisseurs of the game know that this happens more frequently than the DRS will have you believe. Check this
Concluding with this lovely tweet to Shane (about DRS and not Spin bowling).
