then you would have scored 45? i am just kidding may be 54 +45 =99 right. jokes apart i am a big fan of this guy, right vocabulary , right examples, right expressions and very fair and neutral and very consistent.
Did not play a Ranji trophy game did not play International Cricket just played A Div cricket but still he talks like a knowledgeable commentator with a 100 tests experience ...Big Fan of his commentry from my childhood ❤️❤️❤️
Well.. at least the technology won't be biased. Admit it or not but umpires do give biased decisions some times. The discussion revolving around removing the umpire's call is largely driven by the objective to remove biased decisions than the accuracy of the decisions.
@@DurgeshYadav-bc5nm accuracy is much needed then no point in technology Even in field catches runouts it's the same if you don't have enough evidence to overturn umpires call stays
@@DurgeshYadav-bc5nm just during these pandemic times there are only local umpires who could give biased decisions but neutral umpires do not do that and ICC panel local umpires are competent enough and if they err continuously and commit very basic errors then those umpires can be removed from the elite panel and put in the next lower level until they pass the test again and move up.
@@ishthiaqshaik1083 In case of catches, the distance should be a parameter. if the the catch has been taken outside the 30 yard circle or near the boundary ropes, the umpire's soft signal doesn't make sense, because the umpire can not be positively instinctive in those cases
@@DurgeshYadav-bc5nm that's the rule same is applied to lbw By cricket rules you should make you full use of umpires on feild and then refer to off feild
Interesting comment by Harsha Bhogle sir "ball tracker is constructed much like snicko." Please do a video on snicko as well, my thought the snicko was sound of the edge.
@silverrahul not so sure bout that.. for ball tracker, the point after impact has to be extrapolated, here Harsha uses the term "constructed" instead. However with Snicko it's just real time playback audio synced with the video. Can simply sync it using two systems timestamps.. ther is no extrapolation. ...right!
I really don't know why is it so difficult to understand Umpire's Call. The technology is still not 100% and so we should give the benefit of doubt in favour of Umpire who has the clear view.
@@abrahamjohn9457 The first step in the ball tracking part is to identify the 2D coordinate of the ball in the image from each camera. That requires ML based image processing algorithms. This output is then fed to a geometric algorithm (which is where the triangulation happens) to calculate 3D coordinates of the ball. This is done at a very high sample rate to track the ball accurately. Secondly, once the ball hits the batsman, the prediction of the ball path from there to the wickets is ML based.
@Rajesh Hamal Point is , although I don't know the exact details of ball tracking, but an educated guess is that it can't predict (accurately) very late swing, i.e. what if the ball started swinging after the impact point of the ball and the pad. Thus the compensation for side-wise movement should be made (but 50% seems too much of an error compensation). But the ball cant go upwards or downwards in mid-air, i.e. vertical movement can be predicted accurately. So, I don't agree with the error compensation for the bails.
I just feel that it's time to decrease the percentage. Make it 10 % now. If the ball hits less than 10 percent of the stumps it's umpires call. ( They say drs is 99% accurate then why does the ball needs to hit more than 50 percentage)
@silverrahul What if umpire gives out and 10/20% of ball is hitting stump? Do you change out to not out? We need to remember that it is a human making initial decision, and I am certain we have all seen scenarios where ball is clipping a bit of the stump, but umpire has given not out and umpire is actually applauded for making a good decision.
A positive part of Umpire's Call is that the review is retained, which is nice because it doesn't punish the reviewing team for a very close call that could have gone the other way.
Dear Harsha Sir, DRS is an exciting addition to the game. Umpire's call is also exciting as it still keeps the old school style of umpiring in which there is so much importance to what umpire first thought the ball was doing. This is then taken up by ball tracking and its either confirmed or refuted. The harsh decisions of umpire's call are when say the ball touching the stumps is in the range of 35-49%- that's when you think its very close and still its not given as the rule says 50% of the ball should touch. So number calculated by the software which are closer to 50% in the lower range like may be 30-49% and still not given out makes the blood boil sometimes and such decisions could be critical. To mitigate this the bandwidth can be slightly modified ,say more than 35% of ball instead of 50% if touches can be out. Otherwise all LBW decisions will be referred and anything clipping/touching the stumps will be straighway out.
I disagree. It can never be accurately predicted if the ball is hitting the stumps or not. That's why if the ball hits less than 50% the decision is given on basis of umpire's intuition.
@@omkardeokar it's not a prediction, mainly done to support the batsmen 9 out of 10 times you would see that if a ball less than 50% hit the stumps the bails would fall.
One correction here: there is a difference between resolution and FPS of a camera. Resolution is how many pixels you can fit in the camera sensor. Eg. 1080p, 2k, 4k and so on. FPS stands for frames per second. Normal match cameras have or shot at 24fps. High-speed footage are shot at 300 or 1000 fps. More the fps, smoother is the slow motion footage. In my opinion, to obtain a very good ball tracker or hawk eye, the cameras need to have a good resolution (to stuff as much data as possible onto the sensor ) and also have a higher fps (to have more frames per second, through which a smoother tracking can be obtained ). Please correct me if I am wrong.
@1:27 300 frames/second is not the resolution of the camera, it is the frame rate. The resolution of a camera is in pixels (or megapixels), for example a resolution of 4K (UHD) is 3840 pixels by 2160 pixels (a total of 8.29 megapixels).
What is said in the video is correct because 300 frames per second is translated as the time resolution of the camera. In DRS you have to resolve both time and space.
@@manish0013 ahahaha... Mate resolution has no meaning unless you specify what is it resolving. You could resolve anything , space, time, frequency, wavelength etc etc. Normally for the layman the camera companies or the mobile phone companies to make their life easier, term it as 'resolution'. But in this video they are referring to an entirely different scenario where they have to extrapolate something using simulation. The confidence of any simulation depends on its input data, in the case of DRS the inputs mainly depend on the spatial and time resolution which Bhogle clearly explain it as the number of points...
Agree with you sir. Umpire's fall's days are numbered. And it doesn't have to be 50% of the ball hitting. They can fix a certain percentage based on the experimental data and the margin of error of the ball tracking system.
Wow such a lovely insight. Before clicking the video I thought there would be nothing new to know as I was aware of the 50% rule of the ball hitting the stumps and umpire call on the same. But there was so much more to the video. Also I feel if even a fraction of ball is hitting the stumps then it should be out as in cricket out is based on the Funda of bails being dislodged. And if the ball is clipping the stumps at pace it will dislodge the bails even if it's a miniscule contact.
What's the reasoning behind umpire's call when the fundamental reason why batsmen/bowlers go for a DRS review is that they are opposing the umpire's call. That's the essence of DRS, right? In existing setup, a similar incident can have two different outcomes based on what umpire called. That seems fundamentally flawed.
My question is just because one asked for the review do we have to favor him. What about the opposition player. We should give benefit of the doubt to the umpire instead of favouring a team, just because a team asked for the review we should not favour that team. I like umpire's call, I think it is needed to be there. Instead we should focus on increasing the accuracy and precision of the drs. And I think they should consider the top of bail instead of the base just like they did with the edges of the stumps.
My question is ( and is genuine so listen carefully) Why is there a need to check umpire's call at 'pitching' and 'impact' It should be either in line(>50% ball) or not (
respect from Pak! the two major limitations with accurate ball tracking are no of frames and learning pattern of the simulating computer. It can be made more accurate if the back tracking adapts itself during the whole game while keeping track of every ball. The behavior of the pitch, bowler's specific skills to lift/swing the ball, etc can then be included
I was thinking before i started watching what he's say that i already don't know, but i knew one thing today that i have don't know i.e., 50 percent of the ball has to hit the bottom of the bail. Thank you Harsha for the insight.
It is very easy to explain but few do it properly. The idea of DRS was to back the decision of umpires, which had been used for over a century, and only override their decision in the event of a glaring mistake - the so called “howler”. As Harsha explains, the technology is based on simulated predictions which can never be 100% guaranteed accurate. So if the technology shows there is less than 50% chance the umpire is wrong then the decision will be to support the umpire. Only when it is clear the umpire has made a mistake, when more than 50 per cent is hitting the wicket, then should they be overruled. Not sure why this is so hard for the public, and indeed many commentators, to understand. Thoughts?
The technology has a margin of error, yet the way the information is conveyed currently this is not easily apparent to those that don’t understand mathematical simulations - ie the average cricket fan. When the ball hits the pad, the simulation needs to change from being an image the size of the ball to a smaller red ball about half the size. This should make it clear it is a simulation , and then only if half or more of the ball is hitting will It show an impact. I’m not sure if I have explained this well, but this simple solution could revolutionise the system and remove a lot of the controversy and misunderstanding. If Harsha or anyone involved in the game could pass it on, I believe it would solve a lot of problems.
While ball tracking uses known points to extrapolate and predict the target point, using the data collected during the test match would certainly improve the accuracy of this prediction. Having said that the elements of ball movement off the ground .. dip, drift, seam, swing plus the nature of pitch ..variable bounce etc ... will invariably affect the accuracy of the prediction. Add to that the images themselves that are not as clear as those generated by the visual eye .. which has got a sense of depth etc. On some pitches like the one in Motera with variable bounce and prodigious spin, the accuracy of the ball tracking predictions could be called in to question.
I know all this before watching the video because I have been watching cricket since 2010. But to express our knowledge in words specially if things are complicated as umpires call.. well only Harsha can do that.
It’s a mathematical model that is generated. It has confidence interval and a point estimate. Where we see the ball is a point estimate. 50% of confidence interval is below the ball and 50% above it. Considering, what proportion of confidence interval is within wicket zone, decision is made.
fun fact : following was a UPSC question aked in GS 3 paper of 2016 Mains Examination. Examine the technology involved in the Decision Review System. Why its use in the cricket was debated? Do you think its use would benefit cricket? :)
Umpire call in DRS should be stopped immediately. If the ball is touching the stump then the batsman is out and if it is not touching then not out. Simple!!
Very good explanation of DRS. I am great fan of your commentary. I like your flow of language. Now there are only few commentators who can give lively commentary like you. Tks for sharing.
Great explanation, but It surely can’t be satisfactory that whether a batsman is given out or not out depends on which side of a very fine line the umpire originally came down on. I am in favour of using technology to arrive at correct decisions, but am very much against it being used to predict something unpredictable. A good example was when Stokes was bowled by Abbas last summer. Had it hit his pad, whether he would have been given lbw would depend on where his pad was at the time. The ball initially seemed to be going straight (out); it then swung as if it was going to miss off stump (not out); and then moved back in (probably umpire’s call). Fine to use it for checking whether it pitched in line, for no balls etc. but not for lbw predictions.
Explain the intent first. Then the rules. The intent of this technology is to correct the umpire in LBW decisions only when he's really off the mark. So, what is the benchmark for being off the mark? It is set at 50% of the ball size. If more than 50% of the ball is within the wicket zone, then the decision is taken by technology. If less than 50% of the ball is within the wicket zone, the decision is taken by the umpire. So, upon being referred for a technology based review of the umpire's LBW decision, there are two possibilities - when the technology takes the decision (more than 50% of the ball within the wicket zone), it is always OUT and when the technology does not take the decision (less than 50% of the ball in the wicket zone), it's the umpire's call - If he's given it OUT, then it's OUT and if he's given it NOT OUT, then it's NOT OUT. Simple. 😊 PS : Phew! This is as long as the video itself... 😄
Yes the technology can get better and margin of error reduced but can it be eliminated entirely ? I presume when the equipment is installed at the stadium it gets calibrated and again maybe at the start of each innings. So at the start of the game the equipment is perfectly calibrated or as close as possible. But as the game evolves conditions change, pitch, ball and weather, the margin of error will increase because not possible to re-calibrate during a session. Some of these like temperature can be measured and adjusted for during the game based on past experience, but never perfectly. It would help if the on-screen graphics showed the margin of error for each decision. For example best/worst case or % chance is out. To the TV viewer the tracker is always presented as if perfect. Easy for the viewer to forget the margin of error. So always need some kind of allowance for error, might not be called umpires call but must be there somewhere. This is especially true for first class games, less so for one day matches, then again day/night games ?
Liked the explanation but resolution is the wrong term to use. What he meant is frame rate (aka fps). Resolution is the number of pixels in each frame which is not relevant here.
The drs LBW system is a mathematical equation based on lots of variables. The ball tracing is a expected value of the prediction.There is a certain degree of uncertainty involved in the prediction. That’s why it is quite understandable to weigh more on umpires decision than rely totally on technology.
As much as Harsha wanted to make this prediction, I'd say the days aren't numbered for umpires call because.... its umpires we are talking about.. they are an essential part of the game. They govern the game. They have the power to kick out abusive spectators too. We need them. And for that, it is very essential that they hold power...even in the most miniscule kind of things they can. And this is why I believe even wise men wouldn't want to have the days numbered for umpire's call.
A great explanation, but lost me when he suggested that in the case where the umpire gives it out, HawkEye agrees but thinks less than 50% is hitting, that it should be overturned. I think they should just set up some automatic error calculation (surely not beyond the wit of these companies?) and give some defined margins at which it becomes umpire's call. They've tried to simplify what is essentially a pretty well-defined scientific process (error calcs) and in doing so have defeated the purpose.
But the solution for now is, they should include the top of the bales like they included outside of the stumps. "Are numbered" but can stay for now. Till resolution improves.
Whenever we achieves accuracy with hitting part till that date, now onwards we should at least change the percentage of ball hitting the stumps. So instead of giving out at 50% clipping, give it even if it's hitting 30%. 70% - 30% system will be fair!!
He told bowl has to hit the wicket zone(which is bottom of the bail to below) for batsman to get out in case of lbw, but when it is not a lbw and bowl hits top of the bail and it flies away they give it out. What is this now??
Snicko is not constructed harsha, that is real time audio capture through a highly sensitive microphone. The noise dampening and filtration is done, to show the waveform, but it is real audio and not constructed.
Thank you bhogle for explaining it very clearly if we reject umpire call and only stick on to the Drs what is the use of keeping umpires on the field we should respect the decision of umpires
Umpires call is a farce! If any portion of the cricket ball is predicted to touch the stumps in its path of trajectory it should be assumed that the velocity of the ball can and will dislodge the bails, thereby proclaiming the batsmen out and not out if the ball is predicted to be missing the stumps. It should be kept simple and can be a win-win situation with everyone keeping their jobs including the umps. 🤷🏽♂️
The whole point is that it can't be predicted accurately whether or not the ball will hit the stumps until 50% or more of the ball hits. That's the margin of error.
1. if it's a umpire call. Bowlers sacrifice a stump space approx out of 3.. that's hard on the bowler. 2. If it's given not out you're including half ball on both sides of corner stumps which means 3+ stumps area. that's hard on batsmen. Not sure what's the advantage of having umpire call
When you take outside of the stump as a reference why cant you take outside or top outline of the bails. Its not a straight line, but reference could be a stepped line covering top of the bails.
I don't like that when DRS video image does that 360 degree from front on to rear view, it never stops & shows you exactly side on where a batsman was struck on the pads in relation to tracking.. don't believe me go watch any number of replays and it always whizzes past that and goes from in front to behind very quickly and not allowing anyone to see the side on where ball strikes pad. Also I'm not sure tracking takes into account ball speed because I see so many spinners deliveries bouncing over the wicket that look from front on like they would not do so at .. but why isn't an image/vision of the batsman mixed in with drs image of the side on impact where ball strikes his pad..? just go look at as many replays as you want, the DRS tracking replay does this 360 view spin from in front to back of stumps but never ever stops side on to show you exactly in relation to the crease where the impact is. WHY..?
Gabba test me pujara ko 2nd innings me umpire call pe out diya tha..jbki ball tracking me dekhne pe ball wa not hitting the stumps at all..please explain sir
@@Ksramana that’s what they say they are doing now... they should just put it in plain view. As soon as the ball hits the wicket, a number should pop right up. Will save me a precious 30 minutes explaining the decision to myself and then my novice parents or friends. Too many controversial calls or should I say uncertain calls at least to the naked eye watching!
How precise are these ball tracking systems. Since it has these margin of errors for the same ball, will it generate the same trajectory multiple times?
Yes I think the level of error has got smaller and this should be reflected, but I think there is a bigger problem with the current rules, Umpire's call is also applied to weather the impact was inline, this is makes no sense because this is not a projection so has no margin for error. We know for a fact if the ball pitched inline with the stumps, why are we still using umpires call for that.
Hawk eye and Virtual eye are bogus.. It just completes the degree of inclination after the impact.. It does not take into account the speed of the ball, bowler's height, wicket conditions, reverse swing, slower ball, knuckle ball, cross seam ball and so many factors that can cause the bounce or angle to vary... ICC should come up with a different system to these lbw saga
All of the above-mentioned factors will have an impact either before or immediately after the ball pitches. Once the ball pitches and takes an inclination the ball will follow that very inclination irrespective of any factors that you mentioned above. And in 99% of cases DRS is used in cases where the ball has already pitched and has already started its inclination
@@SwapnilShanbhag I beg to differ.. If you have been watching cricket closely, ball tends to behave differently even while it's reaching wicket keeper.. My point is a new ball bowled by a pace bowler at 90mph must behave differently compared to an 80 over old ball bowled by a wrist spinner on different conditions.. unfortunately these systems doesnt take into account these factors and only completes the trajectory after the impact.. There are many flaws in these so called technologies and its high time to come up with fresh ideas
@@shivaaakumar3 Except the speed of the ball..rest all is taken care of in DRS... whether you bowl cross seam or seam up.. reverse swing or seam or spin..after hitting the pad thats all gone. After the impact its all irrelevant. And we can see what its doing before hitting the pad so that is very much plottable.
@@shivaaakumar3 Even the late swings are captured in the ball tracking. If that's what you are talking about. And it should not take into account the factors such as the shine of the ball because it really does not need to know all those. The behaviour of the ball will be apparent by the swing or reverse swing and also the movement off the pitch if any.. and those things do get tracked by the current tech.
Why did Harsha say "the ball tracker much like snicko, is constructed"? How is Snicko "constructed"? I thought it was realtime sound capture and playback synced with the cameras.. what part is constructed? 🤔
Well with due respect to the technology- we should not forget that after all the umpires are human and they are in the best position to make a decision but there could be some judgement error but they do that in real time where it is done in nano seconds but where as technology and the 3rd umpire does it post factor so i would not say removing that aspect is good for cricket. It adds excitement and also shows how good a player or a captain or a wicket keeper's judgement is and there is always DRS review which can correct those errors but removing it totally is absurd and pointless then otherwise they should bring in robots to do the umpiring. Yes only thing which i would say should be corrected is if there is umpires call and the ball is hitting the stumps or even brushing in that zone then that decision can be over ruled but other than that there is no need to tinker with umpire's call rules. The decision to take DRS review should be reduced to 10 seconds and that should be precise we have seen some errors in that when a delayed decision has been allowed to one team and a similar situation other team is denied which is not fair. And also in case of many run outs it is sheer wastage of time when umpires can easily give out but the umpires always ask for a umpires review, yes i can understand if it is very tight , a touch and go decision but sometimes it looks wierd and silly when umpires are asking for reviews in cases where it is so clear and there is no need for playing safe with their decisions which is not right.
The batsman has always been entitled to the benefit of the doubt. The DRS can (so long as captains use it correctly) eliminate the howler. I like umpires call it applies equally to both teams
I don't understand where the 50% margin comes from. Hawkeye affirms its error in projection to be about 3 mm on average. In an espncricinfo debate it was shown that empirical data of balls reaching wicketkeeper show that over 95% of balls had errors within 1 mm. Ball radius is roughly 7.1 cm. Which means if around 4% of ball is hitting stumps, hawkeye says ball WILL hit stumps. 4%, not 50%. Even if u take max error using gaussian distribution, 1 in a billion balls will show error above 10%. Of course this is real life and errors may be higher due to other random factors, but I don't recall any technology that shows a 12 times jump in error due to real life chaos. At max twice. Umpires call can at max be set to 10% of ball if u want to be ultra safe. Does ICC not trust hawkeye accuracy? Idk. If I am missing something and thus have made a mistake, kindly clarify.
Can you please explain that to commentators who seem to think that ball tracking is 100% accurate and perfect all the time! So fed up of hearing complaints from people on TV and radio when umpires call is the result. Even in Tennis (which uses Hawkeye) the technology is wrong and that's dealing with judging where a ball lands, not predicting where it was going to go...
Harsha, this needed a side-by-side comparison of what is out and what is not. The concept of wicket zone could also have been better qualified, especially the bail part.
Commentary panel consists of former crickters and harsha bhogle. You just can't exclude him from the panel due to his sheer knowledge about the game. Commentary team needs you!!!!
If 2 runs are required to win of the last ball, the umpire gives a batsman out, the batsman knows he is not out so he completes two runs and calls for DRS and he is not out. What will be the result?
India Playing XI for Chennai 1st Test Match 1. R.Sharma 2. S.Gill 3. C.Pujara 4. V.Kohli(C) 5. A.Rahane(VC) 6. R.Pant(WK) 7. W.Sundar/A.Patel 8. R.Ashwin 9. K.Yadhav 10. J.Bumrah 11. I.Sharma/M.Siraj
If Harsha Bhogle explains me Christopher Nolan movies, I'll quietly listen to him all day.
Maybe he should explain david dhavan movies.
He is Amitabh of cricket commentary
haha I know right... amazing way Harsha explains love it...
Tenet movie
@@yesh1nth he is far superior than those Bollywood shit
Now time to explain Duckworth-Lewis-Stern method 🙄
Its actually simplier than you think
Yeah because it's a simple formula. There is no probability there.
It's not Drs man
It means
Aao sab cheating kare 😂😂
Ikrrrrrrrrrr
No seriously, this man can read the phonebook itself and it will still sound like the best fairy tale ever!!!
This mans eyes smile along with his face. Very effective talker and most liked commentator in cricket.
🌈
Harsha is becoming the David Attenborough of cricket
He already is... *Harsha of Cricket*
Sir David Attenborough you mean 😅
@@bhargavdesai7984 true but not David Attenborough he is special
Wo kon h
The voice of cricket in India
If Harsha was my maths teacher in school, I wouldn't have scored 54.
😂😂😂
then you would have scored 45? i am just kidding may be 54 +45 =99 right. jokes apart i am a big fan of this guy, right vocabulary , right examples, right expressions and very fair and neutral and very consistent.
Because he would teach you only those things, you already know.
If ravi Shastri taught you mathematics, you'd have scored 36
@@your_dad_18 that's alot if ravi shastri was the teacher 😂😂
Did not play a Ranji trophy game did not play International Cricket just played A Div cricket but still he talks like a knowledgeable commentator with a 100 tests experience ...Big Fan of his commentry from my childhood ❤️❤️❤️
I personally find it okay to blend technology and visible human instincts as neither are 100 percent correct
Well.. at least the technology won't be biased. Admit it or not but umpires do give biased decisions some times. The discussion revolving around removing the umpire's call is largely driven by the objective to remove biased decisions than the accuracy of the decisions.
@@DurgeshYadav-bc5nm accuracy is much needed then no point in technology
Even in field catches runouts it's the same if you don't have enough evidence to overturn umpires call stays
@@DurgeshYadav-bc5nm just during these pandemic times there are only local umpires who could give biased decisions but neutral umpires do not do that and ICC panel local umpires are competent enough and if they err continuously and commit very basic errors then those umpires can be removed from the elite panel and put in the next lower level until they pass the test again and move up.
@@ishthiaqshaik1083 In case of catches, the distance should be a parameter. if the the catch has been taken outside the 30 yard circle or near the boundary ropes, the umpire's soft signal doesn't make sense, because the umpire can not be positively instinctive in those cases
@@DurgeshYadav-bc5nm that's the rule same is applied to lbw
By cricket rules you should make you full use of umpires on feild and then refer to off feild
Expectations from the video: Explanation of what is umpire's call.
Reality: Hasha explains how the camera works.😂
Interesting comment by Harsha Bhogle sir "ball tracker is constructed much like snicko." Please do a video on snicko as well, my thought the snicko was sound of the edge.
I had the same qn.
@silverrahul not so sure bout that.. for ball tracker, the point after impact has to be extrapolated, here Harsha uses the term "constructed" instead. However with Snicko it's just real time playback audio synced with the video. Can simply sync it using two systems timestamps.. ther is no extrapolation. ...right!
I really don't know why is it so difficult to understand Umpire's Call. The technology is still not 100% and so we should give the benefit of doubt in favour of Umpire who has the clear view.
The technology is far, far better than the umpire's view
@little god of mischief Why should cricket rely on the umpires' judgement when the technology is superior? I think there's nothing sad about it
Apply Machine Learning. That will increase the accuracy.
Yeah 👍
The prediction is already ML based.
It's already plotting prediction based paths. We don't want ML models to create it's own simulated frames
@@AlaapLove Ball tracking system are based on the principles of triangulation using visual images. No ML is applied.
@@abrahamjohn9457 The first step in the ball tracking part is to identify the 2D coordinate of the ball in the image from each camera. That requires ML based image processing algorithms. This output is then fed to a geometric algorithm (which is where the triangulation happens) to calculate 3D coordinates of the ball. This is done at a very high sample rate to track the ball accurately.
Secondly, once the ball hits the batsman, the prediction of the ball path from there to the wickets is ML based.
This makes me to feel like Virat should have been given notout against NZ in semis of WC19
😢😢
@Rajesh Hamal ..it was clipping the bails.. not even 50%
@Rajesh Hamal yes! But if umpire gives not out.. What you said( bullshit ) would also be given not out!
@Rajesh Hamal Point is , although I don't know the exact details of ball tracking, but an educated guess is that it can't predict (accurately) very late swing, i.e. what if the ball started swinging after the impact point of the ball and the pad. Thus the compensation for side-wise movement should be made (but 50% seems too much of an error compensation). But the ball cant go upwards or downwards in mid-air, i.e. vertical movement can be predicted accurately. So, I don't agree with the error compensation for the bails.
@@sandeepatmakuru1176 but umpires calk was out ..so any part of ball hitting stumps is fine
I just feel that it's time to decrease the percentage. Make it 10 % now. If the ball hits less than 10 percent of the stumps it's umpires call.
( They say drs is 99% accurate then why does the ball needs to hit more than 50 percentage)
That's based on the margin of error of the ball tracker. Those numbers cannot be played with unless the error reported in ball tracking is reduced
@silverrahul What if umpire gives out and 10/20% of ball is hitting stump? Do you change out to not out? We need to remember that it is a human making initial decision, and I am certain we have all seen scenarios where ball is clipping a bit of the stump, but umpire has given not out and umpire is actually applauded for making a good decision.
@silverrahul but basis of LBW law remains that benefit of doubt goes to batsman.
I'm pretty sure a controversy would still pop up if the ball hits 9% of the stumps lol.
The engineering side of Harsha came on forefront this time :-D
This man is wow.....I love his language. I will listen to him all day.
Why are u not a professor at my college☹️☹️
Because it doesn't pay as much :)
So insightful....Harsha Bhogle..as usual impeccable...
I wish Harsha Bhogle was my science teacher. Everything he explains seems to be simple 😍👌
Any video with Harsha Bhogle can be watched just for his SMILE.👍👍
A positive part of Umpire's Call is that the review is retained, which is nice because it doesn't punish the reviewing team for a very close call that could have gone the other way.
The only positive point
Dear Harsha Sir, DRS is an exciting addition to the game. Umpire's call is also exciting as it still keeps the old school style of umpiring in which there is so much importance to what umpire first thought the ball was doing. This is then taken up by ball tracking and its either confirmed or refuted. The harsh decisions of umpire's call are when say the ball touching the stumps is in the range of 35-49%- that's when you think its very close and still its not given as the rule says 50% of the ball should touch. So number calculated by the software which are closer to 50% in the lower range like may be 30-49% and still not given out makes the blood boil sometimes and such decisions could be critical. To mitigate this the bandwidth can be slightly modified ,say more than 35% of ball instead of 50% if touches can be out. Otherwise all LBW decisions will be referred and anything clipping/touching the stumps will be straighway out.
If ball reaches stumps it should be out, forget umpire's call
Watch bumrah getting smith bowled at mcg, barely 25% was hitting the stumps.
That's how they counter human error and machine error
Totally agree
I disagree. It can never be accurately predicted if the ball is hitting the stumps or not. That's why if the ball hits less than 50% the decision is given on basis of umpire's intuition.
Its not about popular opinion here, its studied by experts who believe in scientific margin of error calculations, since this is a prediction.
@@omkardeokar it's not a prediction, mainly done to support the batsmen
9 out of 10 times you would see that if a ball less than 50% hit the stumps the bails would fall.
One correction here: there is a difference between resolution and FPS of a camera. Resolution is how many pixels you can fit in the camera sensor. Eg. 1080p, 2k, 4k and so on.
FPS stands for frames per second. Normal match cameras have or shot at 24fps. High-speed footage are shot at 300 or 1000 fps. More the fps, smoother is the slow motion footage.
In my opinion, to obtain a very good ball tracker or hawk eye, the cameras need to have a good resolution (to stuff as much data as possible onto the sensor ) and also have a higher fps (to have more frames per second, through which a smoother tracking can be obtained ).
Please correct me if I am wrong.
@1:27 300 frames/second is not the resolution of the camera, it is the frame rate. The resolution of a camera is in pixels (or megapixels), for example a resolution of 4K (UHD) is 3840 pixels by 2160 pixels (a total of 8.29 megapixels).
true
What is said in the video is correct because 300 frames per second is translated as the time resolution of the camera. In DRS you have to resolve both time and space.
@@toneytf resolution and time resolution are two different things you can't take them as same way.
@@manish0013 ahahaha
@@manish0013 ahahaha... Mate resolution has no meaning unless you specify what is it resolving. You could resolve anything , space, time, frequency, wavelength etc etc. Normally for the layman the camera companies or the mobile phone companies to make their life easier, term it as 'resolution'. But in this video they are referring to an entirely different scenario where they have to extrapolate something using simulation. The confidence of any simulation depends on its input data, in the case of DRS the inputs mainly depend on the spatial and time resolution which Bhogle clearly explain it as the number of points...
Agree with you sir. Umpire's fall's days are numbered. And it doesn't have to be 50% of the ball hitting. They can fix a certain percentage based on the experimental data and the margin of error of the ball tracking system.
Wow such a lovely insight. Before clicking the video I thought there would be nothing new to know as I was aware of the 50% rule of the ball hitting the stumps and umpire call on the same. But there was so much more to the video.
Also I feel if even a fraction of ball is hitting the stumps then it should be out as in cricket out is based on the Funda of bails being dislodged. And if the ball is clipping the stumps at pace it will dislodge the bails even if it's a miniscule contact.
It would be really great of you cricbuzz if you start an dedicated 'explained' series featuring harsha for such esoteric topics.
What's the reasoning behind umpire's call when the fundamental reason why batsmen/bowlers go for a DRS review is that they are opposing the umpire's call. That's the essence of DRS, right? In existing setup, a similar incident can have two different outcomes based on what umpire called. That seems fundamentally flawed.
My question is just because one asked for the review do we have to favor him. What about the opposition player.
We should give benefit of the doubt to the umpire instead of favouring a team, just because a team asked for the review we should not favour that team.
I like umpire's call, I think it is needed to be there. Instead we should focus on increasing the accuracy and precision of the drs.
And I think they should consider the top of bail instead of the base just like they did with the edges of the stumps.
My question is ( and is genuine so listen carefully) Why is there a need to check umpire's call at 'pitching' and 'impact'
It should be either in line(>50% ball) or not (
respect from Pak! the two major limitations with accurate ball tracking are no of frames and learning pattern of the simulating computer. It can be made more accurate if the back tracking adapts itself during the whole game while keeping track of every ball. The behavior of the pitch, bowler's specific skills to lift/swing the ball, etc can then be included
I was thinking before i started watching what he's say that i already don't know, but i knew one thing today that i have don't know i.e., 50 percent of the ball has to hit the bottom of the bail. Thank you Harsha for the insight.
Great analysis 🔥,
You r legends
It is very easy to explain but few do it properly. The idea of DRS was to back the decision of umpires, which had been used for over a century, and only override their decision in the event of a glaring mistake - the so called “howler”. As Harsha explains, the technology is based on simulated predictions which can never be 100% guaranteed accurate. So if the technology shows there is less than 50% chance the umpire is wrong then the decision will be to support the umpire. Only when it is clear the umpire has made a mistake, when more than 50 per cent is hitting the wicket, then should they be overruled. Not sure why this is so hard for the public, and indeed many commentators, to understand. Thoughts?
The technology has a margin of error, yet the way the information is conveyed currently this is not easily apparent to those that don’t understand mathematical simulations - ie the average cricket fan. When the ball hits the pad, the simulation needs to change from being an image the size of the ball to a smaller red ball about half the size. This should make it clear it is a simulation , and then only if half or more of the ball is hitting will It show an impact. I’m not sure if I have explained this well, but this simple solution could revolutionise the system and remove a lot of the controversy and misunderstanding. If Harsha or anyone involved in the game could pass it on, I believe it would solve a lot of problems.
While ball tracking uses known points to extrapolate and predict the target point, using the data collected during the test match would certainly improve the accuracy of this prediction. Having said that the elements of ball movement off the ground .. dip, drift, seam, swing plus the nature of pitch ..variable bounce etc ... will invariably affect the accuracy of the prediction. Add to that the images themselves that are not as clear as those generated by the visual eye .. which has got a sense of depth etc.
On some pitches like the one in Motera with variable bounce and prodigious spin, the accuracy of the ball tracking predictions could be called in to question.
Very well explained...especially the decision about ball hitting the 50% below the baseline of bails....that clears lot of doubts...thank you again.
I know all this before watching the video because I have been watching cricket since 2010. But to express our knowledge in words specially if things are complicated as umpires call.. well only Harsha can do that.
It’s a mathematical model that is generated. It has confidence interval and a point estimate. Where we see the ball is a point estimate. 50% of confidence interval is below the ball and 50% above it. Considering, what proportion of confidence interval is within wicket zone, decision is made.
fun fact : following was a UPSC question aked in GS 3 paper of 2016 Mains Examination.
Examine the technology involved in the Decision Review System. Why its use in the cricket was debated? Do you think its use would benefit cricket? :)
Very well explained..love it!!
Umpire call in DRS should be stopped immediately. If the ball is touching the stump then the batsman is out and if it is not touching then not out. Simple!!
That's not how it works.
It doesn't even work that way when the ball actually hits the stumps. It is not always out.
Very nicely explained by Mr. Harsha Bhogle
I think it would be really helpful if in future ball% hitting stumps would be shonw side by side to ball tracker.
Yeah that would be interesting.
Isn't it impossible.. how can you know the exact percentage when the trajectory itself is constructed on predictions.
Waiting for
"Come on Ash"
Very good explanation of DRS. I am great fan of your commentary. I like your flow of language. Now there are only few commentators who can give lively commentary like you. Tks for sharing.
Harsha .. had you been a college/univ Prof, you would have had 100% attendance all life in your class. Thank you for explaining this.
Great analyst of the game 👏🏻
Great explanation, but It surely can’t be satisfactory that whether a batsman is given out or not out depends on which side of a very fine line the umpire originally came down on. I am in favour of using technology to arrive at correct decisions, but am very much against it being used to predict something unpredictable.
A good example was when Stokes was bowled by Abbas last summer. Had it hit his pad, whether he would have been given lbw would depend on where his pad was at the time. The ball initially seemed to be going straight (out); it then swung as if it was going to miss off stump (not out); and then moved back in (probably umpire’s call).
Fine to use it for checking whether it pitched in line, for no balls etc. but not for lbw predictions.
This adds up to the already complicated laws of cricket. I found myself explaining this to my German friends, no one understood me.
Keep it simple. As a batsman never let the ball hit your pad, meet it with the middle of your bat! Worked alright for Don Bradman for the most part.
Explain the intent first. Then the rules. The intent of this technology is to correct the umpire in LBW decisions only when he's really off the mark. So, what is the benchmark for being off the mark? It is set at 50% of the ball size.
If more than 50% of the ball is within the wicket zone, then the decision is taken by technology. If less than 50% of the ball is within the wicket zone, the decision is taken by the umpire.
So, upon being referred for a technology based review of the umpire's LBW decision, there are two possibilities - when the technology takes the decision (more than 50% of the ball within the wicket zone), it is always OUT and when the technology does not take the decision (less than 50% of the ball in the wicket zone), it's the umpire's call - If he's given it OUT, then it's OUT and if he's given it NOT OUT, then it's NOT OUT.
Simple. 😊
PS : Phew! This is as long as the video itself... 😄
Great video, but you are confusing resolution with capture speed aka fps!
" If ball hit the stumps...then the batsman was out...thats simple rule of LBW"- Sir Don Bradman
Yes the technology can get better and margin of error reduced but can it be eliminated entirely ? I presume when the equipment is installed at the stadium it gets calibrated and again maybe at the start of each innings. So at the start of the game the equipment is perfectly calibrated or as close as possible. But as the game evolves conditions change, pitch, ball and weather, the margin of error will increase because not possible to re-calibrate during a session. Some of these like temperature can be measured and adjusted for during the game based on past experience, but never perfectly. It would help if the on-screen graphics showed the margin of error for each decision. For example best/worst case or % chance is out. To the TV viewer the tracker is always presented as if perfect. Easy for the viewer to forget the margin of error. So always need some kind of allowance for error, might not be called umpires call but must be there somewhere. This is especially true for first class games, less so for one day matches, then again day/night games ?
Liked the explanation but resolution is the wrong term to use. What he meant is frame rate (aka fps). Resolution is the number of pixels in each frame which is not relevant here.
Damn, this guy should have been my physics teacher
Love You So Much This Voice And Akash Vani Voice
The drs LBW system is a mathematical equation based on lots of variables. The ball tracing is a expected value of the prediction.There is a certain degree of uncertainty involved in the prediction. That’s why it is quite understandable to weigh more on umpires decision than rely totally on technology.
As much as Harsha wanted to make this prediction, I'd say the days aren't numbered for umpires call because.... its umpires we are talking about.. they are an essential part of the game. They govern the game. They have the power to kick out abusive spectators too. We need them. And for that, it is very essential that they hold power...even in the most miniscule kind of things they can. And this is why I believe even wise men wouldn't want to have the days numbered for umpire's call.
A great explanation, but lost me when he suggested that in the case where the umpire gives it out, HawkEye agrees but thinks less than 50% is hitting, that it should be overturned.
I think they should just set up some automatic error calculation (surely not beyond the wit of these companies?) and give some defined margins at which it becomes umpire's call. They've tried to simplify what is essentially a pretty well-defined scientific process (error calcs) and in doing so have defeated the purpose.
But the solution for now is, they should include the top of the bales like they included outside of the stumps. "Are numbered" but can stay for now. Till resolution improves.
This was very much needed
Whenever we achieves accuracy with hitting part till that date, now onwards we should at least change the percentage of ball hitting the stumps. So instead of giving out at 50% clipping, give it even if it's hitting 30%.
70% - 30% system will be fair!!
This is actually one of few comments that make sense.
He told bowl has to hit the wicket zone(which is bottom of the bail to below) for batsman to get out in case of lbw, but when it is not a lbw and bowl hits top of the bail and it flies away they give it out. What is this now??
You are a cricket scientist Harsha.. enjoyed each word of your explanation in this video.. thank you for a clear explanation..
Make this man a professor in some iit
Snicko is not constructed harsha, that is real time audio capture through a highly sensitive microphone. The noise dampening and filtration is done, to show the waveform, but it is real audio and not constructed.
Excellent & easily explained.
FPS is not a unit of resolution. It's frame rate. I wonder how the editor of the video missed that!
please explain how the calibration of DRS is done before every match
Thank you bhogle for explaining it very clearly if we reject umpire call and only stick on to the Drs what is the use of keeping umpires on the field we should respect the decision of umpires
Umpires call is a farce! If any portion of the cricket ball is predicted to touch the stumps in its path of trajectory it should be assumed that the velocity of the ball can and will dislodge the bails, thereby proclaiming the batsmen out and not out if the ball is predicted to be missing the stumps. It should be kept simple and can be a win-win situation with everyone keeping their jobs including the umps. 🤷🏽♂️
The whole point is that it can't be predicted accurately whether or not the ball will hit the stumps until 50% or more of the ball hits. That's the margin of error.
@@Srikar-on9ib Well at least then it can not be predicted for either teams right..? At least this will get rid of the umpire's biased calls
1. if it's a umpire call. Bowlers sacrifice a stump space approx out of 3.. that's hard on the bowler.
2. If it's given not out you're including half ball on both sides of corner stumps which means 3+ stumps area. that's hard on batsmen.
Not sure what's the advantage of having umpire call
I think you're confusing frame rates with resolution
When you take outside of the stump as a reference why cant you take outside or top outline of the bails. Its not a straight line, but reference could be a stepped line covering top of the bails.
I don't like that when DRS video image does that 360 degree from front on to rear view, it never stops & shows you exactly side on where a batsman was struck on the pads in relation to tracking.. don't believe me go watch any number of replays and it always whizzes past that and goes from in front to behind very quickly and not allowing anyone to see the side on where ball strikes pad. Also I'm not sure tracking takes into account ball speed because I see so many spinners deliveries bouncing over the wicket that look from front on like they would not do so at .. but why isn't an image/vision of the batsman mixed in with drs image of the side on impact where ball strikes his pad..? just go look at as many replays as you want, the DRS tracking replay does this 360 view spin from in front to back of stumps but never ever stops side on to show you exactly in relation to the crease where the impact is. WHY..?
Greatest Cricket knowledge of you I Love And Respect from Pakistan
Gabba test me pujara ko 2nd innings me umpire call pe out diya tha..jbki ball tracking me dekhne pe ball wa not hitting the stumps at all..please explain sir
What if they calculate by seeing the percent of ball hitting , if its greater or equal to 50, than decision should remain out
That's what they are doing now!
@@Ksramana that’s what they say they are doing now... they should just put it in plain view. As soon as the ball hits the wicket, a number should pop right up. Will save me a precious 30 minutes explaining the decision to myself and then my novice parents or friends. Too many controversial calls or should I say uncertain calls at least to the naked eye watching!
This video cleared all my doubts regarding DRS.👍
How precise are these ball tracking systems. Since it has these margin of errors for the same ball, will it generate the same trajectory multiple times?
Yes I think the level of error has got smaller and this should be reflected, but I think there is a bigger problem with the current rules, Umpire's call is also applied to weather the impact was inline, this is makes no sense because this is not a projection so has no margin for error. We know for a fact if the ball pitched inline with the stumps, why are we still using umpires call for that.
Hawk eye and Virtual eye are bogus.. It just completes the degree of inclination after the impact.. It does not take into account the speed of the ball, bowler's height, wicket conditions, reverse swing, slower ball, knuckle ball, cross seam ball and so many factors that can cause the bounce or angle to vary... ICC should come up with a different system to these lbw saga
All of the above-mentioned factors will have an impact either before or immediately after the ball pitches. Once the ball pitches and takes an inclination the ball will follow that very inclination irrespective of any factors that you mentioned above. And in 99% of cases DRS is used in cases where the ball has already pitched and has already started its inclination
@@SwapnilShanbhag True Bro! Hope others understand 🙏🏻
@@SwapnilShanbhag I beg to differ.. If you have been watching cricket closely, ball tends to behave differently even while it's reaching wicket keeper.. My point is a new ball bowled by a pace bowler at 90mph must behave differently compared to an 80 over old ball bowled by a wrist spinner on different conditions.. unfortunately these systems doesnt take into account these factors and only completes the trajectory after the impact.. There are many flaws in these so called technologies and its high time to come up with fresh ideas
@@shivaaakumar3 Except the speed of the ball..rest all is taken care of in DRS... whether you bowl cross seam or seam up.. reverse swing or seam or spin..after hitting the pad thats all gone. After the impact its all irrelevant. And we can see what its doing before hitting the pad so that is very much plottable.
@@shivaaakumar3 Even the late swings are captured in the ball tracking. If that's what you are talking about. And it should not take into account the factors such as the shine of the ball because it really does not need to know all those. The behaviour of the ball will be apparent by the swing or reverse swing and also the movement off the pitch if any.. and those things do get tracked by the current tech.
Why did Harsha say "the ball tracker much like snicko, is constructed"?
How is Snicko "constructed"? I thought it was realtime sound capture and playback synced with the cameras.. what part is constructed? 🤔
Well with due respect to the technology- we should not forget that after all the umpires are human and they are in the best position to make a decision but there could be some judgement error but they do that in real time where it is done in nano seconds but where as technology and the 3rd umpire does it post factor so i would not say removing that aspect is good for cricket. It adds excitement and also shows how good a player or a captain or a wicket keeper's judgement is and there is always DRS review which can correct those errors but removing it totally is absurd and pointless then otherwise they should bring in robots to do the umpiring.
Yes only thing which i would say should be corrected is if there is umpires call and the ball is hitting the stumps or even brushing in that zone then that decision can be over ruled but other than that there is no need to tinker with umpire's call rules. The decision to take DRS review should be reduced to 10 seconds and that should be precise we have seen some errors in that when a delayed decision has been allowed to one team and a similar situation other team is denied which is not fair.
And also in case of many run outs it is sheer wastage of time when umpires can easily give out but the umpires always ask for a umpires review, yes i can understand if it is very tight , a touch and go decision but sometimes it looks wierd and silly when umpires are asking for reviews in cases where it is so clear and there is no need for playing safe with their decisions which is not right.
While explaining, if graphic is shown then it should have been much clearer for those who are new to the game.
The batsman has always been entitled to the benefit of the doubt. The DRS can (so long as captains use it correctly) eliminate the howler. I like umpires call it applies equally to both teams
I don't understand where the 50% margin comes from. Hawkeye affirms its error in projection to be about 3 mm on average. In an espncricinfo debate it was shown that empirical data of balls reaching wicketkeeper show that over 95% of balls had errors within 1 mm. Ball radius is roughly 7.1 cm. Which means if around 4% of ball is hitting stumps, hawkeye says ball WILL hit stumps. 4%, not 50%. Even if u take max error using gaussian distribution, 1 in a billion balls will show error above 10%.
Of course this is real life and errors may be higher due to other random factors, but I don't recall any technology that shows a 12 times jump in error due to real life chaos. At max twice. Umpires call can at max be set to 10% of ball if u want to be ultra safe. Does ICC not trust hawkeye accuracy? Idk. If I am missing something and thus have made a mistake, kindly clarify.
Can you please explain that to commentators who seem to think that ball tracking is 100% accurate and perfect all the time! So fed up of hearing complaints from people on TV and radio when umpires call is the result.
Even in Tennis (which uses Hawkeye) the technology is wrong and that's dealing with judging where a ball lands, not predicting where it was going to go...
Thank you 🙏
Harsha, this needed a side-by-side comparison of what is out and what is not. The concept of wicket zone could also have been better qualified, especially the bail part.
The only issue is, the technology can have mixed results, specially in unfavourable conditions, eg high winds.
Amazingly explained and Harsha Bhogle is THEEEE BESTTTTTTTT!!!
Wonderful explanation HARSHA
Thank you Hasha, i fully agree. One verdict.
Commentary panel consists of former crickters and harsha bhogle. You just can't exclude him from the panel due to his sheer knowledge about the game. Commentary team needs you!!!!
Very well explained Harsha, Thank you
Wow concept clear :)
Explain DLS and beamer also
Very nicely explained
If 2 runs are required to win of the last ball, the umpire gives a batsman out, the batsman knows he is not out so he completes two runs and calls for DRS and he is not out. What will be the result?
India Playing XI for Chennai 1st Test Match
1. R.Sharma
2. S.Gill
3. C.Pujara
4. V.Kohli(C)
5. A.Rahane(VC)
6. R.Pant(WK)
7. W.Sundar/A.Patel
8. R.Ashwin
9. K.Yadhav
10. J.Bumrah
11. I.Sharma/M.Siraj