'DRS has affected the game more than we thought it would'
David Richardson, the ICC's general manager of cricket, talks about the impact of the review system, and the tests currently being conducted to analyse the effectiveness of ball-tracking technology
Interview by George Dobell
14-Feb-2012
The use of the DRS may have amplified the trend of umpires giving more lbw shouts out these days • Getty Images
Dave Richardson, the ICC's general manager of cricket, speaks:
Tell us about the testing of the DRS system currently going on at Cambridge University.
Well, it's a company called Computer Vision Consulting Limited, staffed by post-grads from Cambridge University with expertise in the supply of research and performance assessment services, particularly within the computer vision and broadcasting technology industries. We've asked them to review the level of accuracy and reliability of the two ball-tracking companies - Hawk-Eye and Virtual Eye - that have been accredited for international cricket. In terms of accuracy we want to know whether their virtual depictions of where the ball has pitched and where the ball has impacted the batsman accord with the reality, and whether their predictions as to where the ball would have hit the stumps are correct.
Well, it's a company called Computer Vision Consulting Limited, staffed by post-grads from Cambridge University with expertise in the supply of research and performance assessment services, particularly within the computer vision and broadcasting technology industries. We've asked them to review the level of accuracy and reliability of the two ball-tracking companies - Hawk-Eye and Virtual Eye - that have been accredited for international cricket. In terms of accuracy we want to know whether their virtual depictions of where the ball has pitched and where the ball has impacted the batsman accord with the reality, and whether their predictions as to where the ball would have hit the stumps are correct.
The reliability issue is slightly different. We want to know the percentage of times they can deliver an accurate tracking. If, in a Test, there are 60 lbw appeals and the ball-tracking technology is only able to deliver an accurate tracking on, for example, 50 of those occasions, then they would not be regarded as very reliable. On the other hand, if they were getting it right on 97 out of 100 occasions, we would probably regard that as being acceptable.
Presumably there was testing of the DRS before implementation?
Yes, we did some manual testing, but it was pretty basic. We asked the MCC to assist us with these tests to provide some level of independence. We were satisfied that the systems were accurate enough for us to have enough confidence to implement them for DRS. Certainly the initial testing led us to believe that the systems would be far more accurate than the human eye could ever be.
Yes, we did some manual testing, but it was pretty basic. We asked the MCC to assist us with these tests to provide some level of independence. We were satisfied that the systems were accurate enough for us to have enough confidence to implement them for DRS. Certainly the initial testing led us to believe that the systems would be far more accurate than the human eye could ever be.
This is a far more detailed review by a completely independent party. Hawk-Eye and Virtual Eye both tell us how accurate they are, but this will verify those claims..
So is it also a test to see which of Hawk-Eye and Virtual Eye works better?
It may allow us to draw some comparisons, yes.
It may allow us to draw some comparisons, yes.
Who is paying for it and how much is it costing?
The ICC is paying. I do not have the exact figure, but it's not too expensive. Certainly it is worth the cost.
The ICC is paying. I do not have the exact figure, but it's not too expensive. Certainly it is worth the cost.
The testers have to develop appropriate software first. Their expertise is analysing video footage. Their skills and technology are incredibly sophisticated. For example, if you can tell them the time, the date and the place at which the delivery has taken place, they can even work out from the shadow of the ball how high it is off the ground at a particular point. This helps them verify the accuracy of the ball tracking. At the risk of over-simplifying the method, they will be using actual video footage of a particular delivery from different angles and, by synchronising the footage, creating a 3D picture which can be used to verify the accuracy of the ball tracking.
Presumably part of the role of the tests is to convince everyone - including India - that the DRS works and that it should be accepted as a positive development?
Exactly. Whether it will help to change their minds on DRS, even assuming the report comes back positive, or whether they have other reasons for not wanting DRS, we will have to see.
Exactly. Whether it will help to change their minds on DRS, even assuming the report comes back positive, or whether they have other reasons for not wanting DRS, we will have to see.
"You see a replay on TV and say, "That looks as if it was hitting leg stump." But then Hawk-Eye shows it just missing. What people don't realise is that the camera for the slow-motion replay might not have been behind the bowler's arm"
What is the time scale for the testing?
They are hoping to have something ready for us by our next ICC Cricket Committee meeting, in May 2012.
They are hoping to have something ready for us by our next ICC Cricket Committee meeting, in May 2012.
How do you respond to the accusation that you should have arranged independent testing before implementation?
We were satisfied with the testing we did. We tested the accuracy as far as we could, and to an extent it showed that the technology was at least accurate and reliable enough for it to be used in the manner that we have used it.
We were satisfied with the testing we did. We tested the accuracy as far as we could, and to an extent it showed that the technology was at least accurate and reliable enough for it to be used in the manner that we have used it.
A lot of people have misgivings about DRS because they are misled by what they see on TV or from their particular angle of sight. People see a replay on TV and say, "That looks as if it was hitting leg stump." But then Hawk-Eye shows it just missing and you say to yourself, "That just cannot be." But what people don't realise is that the camera for the slow-motion replay might not have been behind the bowler's arm. There are three cameras in a row and the one used for slow-motion replays is one of the ones on the side. So, often the picture you see on your TV screen is slightly misleading. The Hawk-Eye cameras are set-up and calibrated perfectly to ensure they provide accurate results. We need to provide the evidence to prove that [people] should trust the evidence provided by the ball-tracking technology, not what they might see on television.
There are concerns that the third umpire is not always using the best technology or seeing the best pictures from the best angles. Is that all part of the learning process of using DRS?
Yes, it has been part of the learning process. But nowadays we have a technical expert who sits with the third umpire in every match where DRS is used. His job is to ensure everything is set up correctly. He monitors that the third umpire has the best available television, with the best screen definition and picture clarity. All that is checked. But it's not always possible to ensure that high definition is available - that is dependent on the cabling at the venue, I believe.
Yes, it has been part of the learning process. But nowadays we have a technical expert who sits with the third umpire in every match where DRS is used. His job is to ensure everything is set up correctly. He monitors that the third umpire has the best available television, with the best screen definition and picture clarity. All that is checked. But it's not always possible to ensure that high definition is available - that is dependent on the cabling at the venue, I believe.
Presumably the broadcasters have a responsibility too?
There are minimum specification requirements for the third umpire TV room, some of which are applicable to the broadcasters. Some broadcasters exceed what is required but all have no problem in meeting the minimum requirement. But we're taking a pragmatic approach. We believe that having a TV that is not high definition is better than not having one at all.
There are minimum specification requirements for the third umpire TV room, some of which are applicable to the broadcasters. Some broadcasters exceed what is required but all have no problem in meeting the minimum requirement. But we're taking a pragmatic approach. We believe that having a TV that is not high definition is better than not having one at all.
There was a dismissal in the third Test in the UAE when Simon Taufel, the on-field umpire, gave Mohammad Hafeez not out but the third umpire, Shavir Tarapore, overruled him despite the fact that Hot Spot seemed to show a thin edge. Are these the inevitable teething problems of implementing such a system? Will we have more knowledge in five years?
We'll have more experience. The umpires will know to look out for a tiny fleck of white. The umpires will learn by experience. So, yes, that is the case. Technology itself will also improve with time.
We'll have more experience. The umpires will know to look out for a tiny fleck of white. The umpires will learn by experience. So, yes, that is the case. Technology itself will also improve with time.
The DRS seemed to have a huge impact on the series in the UAE. Was that a surprise or a concern?
It's too early to say. Prior to the use of DRS there was a trend that spin bowlers were [beginning to get] more lbws. Even though we weren't using DRS, the umpires were being assessed by it. They might be giving a batsman not out if the ball hit them on the front foot, as they could not be sure that it would have gone on to hit the stumps. Then they would go and watch the replay, see what Hawk-Eye said and realise the ball was hitting halfway up middle. The TV commentators would be saying, "Why hasn't he given it out?"
It's too early to say. Prior to the use of DRS there was a trend that spin bowlers were [beginning to get] more lbws. Even though we weren't using DRS, the umpires were being assessed by it. They might be giving a batsman not out if the ball hit them on the front foot, as they could not be sure that it would have gone on to hit the stumps. Then they would go and watch the replay, see what Hawk-Eye said and realise the ball was hitting halfway up middle. The TV commentators would be saying, "Why hasn't he given it out?"
That changed umpiring. Umpires realised they could give more decisions out if the ball was heading towards leg stump. They also realised that in subcontinent conditions the ball was rarely going to bounce over the top of the stumps, so they started giving more front foot lbws, too.
So the trend had started before DRS came in. It may be, though, the use of DRS has amplified things. Umpires may have realised that if they give someone out and DRS shows it was not out, then their decision can be rectified. So they might, I suppose, have the courage of their convictions a bit more and take a less conservative approach to giving the batsman out.
I think if we're totally honest, DRS has affected the game slightly more than we thought it would. In the Pakistan-England series in particular. The pitches in the UAE have been relatively low, especially for the spinners. Because of the lower bounce there has been an increase in the number of lbws.
It seems to be creating a better balance between bat and ball, but is there a concern that Tests might not last five days?
I wouldn't be too hasty in saying that. I agree that the Pakistan-England series may have suggested that, but maybe a few batsmen were out of form and the bowlers of both sides were very good. There was a bit more in the pitches than most expected too. It's too early to draw too many conclusions.
I wouldn't be too hasty in saying that. I agree that the Pakistan-England series may have suggested that, but maybe a few batsmen were out of form and the bowlers of both sides were very good. There was a bit more in the pitches than most expected too. It's too early to draw too many conclusions.
One positive aspect of the DRS should be that it encourages batsmen to use their bats more.
Yes, it is going to bring about a change of technique. And that will improve things for everybody.
Yes, it is going to bring about a change of technique. And that will improve things for everybody.
That point you made - about the balance between bat and ball - that is key. A year ago, every Tom, Dick and Harry was averaging more than 50 in Test cricket. The balance between bat and ball had got out of kilter and experts were complaining. I think that using DRS may help redress that balance.
"A year ago, every Tom, Dick and Harry was averaging more than 50 in Test cricket. The balance between bat and ball had got out of kilter and experts were complaining. I think that using DRS may help redress that balance"
Are the umpires on side?
The majority are. None of them like making mistakes, and sometimes they might feel that the DRS makes their mistakes more embarrassing. But credit to them, the DRS - more often than not - shows what a good job they do. The stats are unbelievable. Umpires are proved right after reviews 75% of the time. Umpires do not like their decisions having to be overturned but the fact is that at the end of a series where DRS is used, the umpires are normally congratulated for a job well done and people are amazed as to how many decisions they actually get right. In non-DRS series, even where the umpires might have made fewer mistakes, they are inevitably criticised for those that they do make.
The majority are. None of them like making mistakes, and sometimes they might feel that the DRS makes their mistakes more embarrassing. But credit to them, the DRS - more often than not - shows what a good job they do. The stats are unbelievable. Umpires are proved right after reviews 75% of the time. Umpires do not like their decisions having to be overturned but the fact is that at the end of a series where DRS is used, the umpires are normally congratulated for a job well done and people are amazed as to how many decisions they actually get right. In non-DRS series, even where the umpires might have made fewer mistakes, they are inevitably criticised for those that they do make.
It wasn't too long ago that there were just two umpires - local umpires, at that - and none of this technology or manpower. Can the game afford the investment?
You're right, it is expensive. It costs as least $10,000 a day to have Hot Spot and ball-tracking. In my view, cricket will pick up the cost - of using technology for umpiring purposes, if not in the short term then in the long term - because eventually the broadcasters will include it in their budgeting and consequent negotiations with the member boards in determining broadcast rights fees. They'll pay X amount less in broadcast rights each time in order to pay for the technology that is required for DRS. We'd like to have Hot Spot and ball-tracking at every Test. And, if we did, we might be able to do a deal with the companies to provide it at a better price.
You're right, it is expensive. It costs as least $10,000 a day to have Hot Spot and ball-tracking. In my view, cricket will pick up the cost - of using technology for umpiring purposes, if not in the short term then in the long term - because eventually the broadcasters will include it in their budgeting and consequent negotiations with the member boards in determining broadcast rights fees. They'll pay X amount less in broadcast rights each time in order to pay for the technology that is required for DRS. We'd like to have Hot Spot and ball-tracking at every Test. And, if we did, we might be able to do a deal with the companies to provide it at a better price.
Is it worth the cost? Well, if we use DRS, the percentage of correct decisions increases by approximately four to five percentage points, from 93% to 98%. I think it is worth it. Technology is here to stay. If the broadcasters are going to continue to use it, we have to use it.
Are similar tests planned on Hot Spot?
No. Hot Spot is real. It's not a virtual picture. What they will be working upon is making their cameras even more sensitive, so that the smallest of touches will be visible. They're learning about the best way to position and set their cameras. Experience has shown that the priority is the sensitivity of the camera, i.e. its ability to generate a visible heat mark as opposed to clearly defined pictures which look nice but do not provide the required level of sensitivity to pick up the faint edges. Whilst Hot Spot cameras may not be able to pick up the very faint touches, they will never show a mark where there is no touch and they are also very useful in distinguishing between whether the ball has touched the bat or gloves, as opposed to, for example, the thigh guard, arm guard, shoulder or helmet.
No. Hot Spot is real. It's not a virtual picture. What they will be working upon is making their cameras even more sensitive, so that the smallest of touches will be visible. They're learning about the best way to position and set their cameras. Experience has shown that the priority is the sensitivity of the camera, i.e. its ability to generate a visible heat mark as opposed to clearly defined pictures which look nice but do not provide the required level of sensitivity to pick up the faint edges. Whilst Hot Spot cameras may not be able to pick up the very faint touches, they will never show a mark where there is no touch and they are also very useful in distinguishing between whether the ball has touched the bat or gloves, as opposed to, for example, the thigh guard, arm guard, shoulder or helmet.
And Snicko: will you test that?
Not yet, no. Snicko is just a visual depiction of the sound from the stump microphone. At this stage the third umpire does not use it. The third umpire gets the sound from the stump microphone.
Not yet, no. Snicko is just a visual depiction of the sound from the stump microphone. At this stage the third umpire does not use it. The third umpire gets the sound from the stump microphone.
They don't just use Hot Spot; they can use what the on-field umpire uses. They can use what they see and what they hear. Even if there is no Hot Spot but they hear a noise, assuming there is no evidence that the noise was caused by something other than the ball hitting bat or glove, they are quite entitled to advise the on-field umpire that the batsman must have got a faint edge. Viewers might not understand that this is the protocol, I know, but hopefully they will understand how it works in time. In time, we'd like the communication between the on-field and third umpire to be heard by the TV viewers. Hopefully, when the umpires are so confident in the system and so well versed in using it, we will be able to do that. That's the aim.
George Dobell is a senior correspondent at ESPNcricinfo