Hawkeye: How exactly was this technology validated for cricket?

[cricketrulez]

I posted the same query on RSC. Didn't get a response with the correct information. Hawkeye is trying to predict the result of "turbulence event" which didn't actually happen. This is a bit like weather forecasting w/o forecast models. Has anyone seen any Independent experimental validation of of the predictions by Hawkeye. Not some Dog and Pony show from Hawkeye innovations. I'm talking about feeding the software incomplete information from camera's and check it prediction against actual path. All this w/o any location specific in put such as pitch condition, wind etc. Has anyone ever been shown this info?

[gaurav92]

All thanks to this man : :hail::hail::hail:

[Sidhoni]

All thanks to this man : :hail::hail::hail:

:haha::haha::haha:

[NareshK]

I posted the same query on RSC. Didn't get a response with the correct information. Hawkeye is trying to predict the result of "turbulence event" which didn't actually happen. This is a bit like weather forecasting w/o forecast models. Has anyone seen any Independent experimental validation of of the predictions by Hawkeye. Not some Dog and Pony show from Hawkeye innovations. I'm talking about feeding the software incomplete information from camera's and check it prediction against actual path. All this w/o any location specific in put such as pitch condition, wind etc. Has anyone ever been shown this info?

I dont really understand how this Hawk eye works. Anyhow, why dont they just shoot some footage of a bowler, a spinner preferably bowling some deliveries at the stumps without a batsman standing in front of the wickets and compare it with Hawk eye and see how accurately it predicts the line ? Let it be examined by an independent panel from all countries and if the accuracy is good, then nobody should have a problem adopting it and if there isnt, it should be shelved. Its simple, isnt it ?

[cricketrulez]

I dont really understand how this Hawk eye works. Anyhow, why dont they just shoot some footage of a bowler, a spinner preferably bowling some deliveries at the stumps without a batsman standing in front of the wickets and compare it with Hawk eye and see how accurately it predicts the line ? Let it be examined by an independent panel from all countries and if the accuracy is good, then nobody should have a problem adopting it and if there isnt, it should be shelved. Its simple, isnt it ?

Yup, thats what I'm talking about. It shouldn't be restricted to spinners and it should in realistic pitches which age with time during a test.

[Raghav_12]

Yup' date=' thats what I'm talking about. It shouldn't be restricted to spinners and it should in realistic pitches which age with time during a test.[/quote'] I believe they would have done that analysis about predictions of Hawkeye and how much of these predictions come true if you allow normal course of things to take place. Without that, it wouldn't have been put in practise. No reason for this to be restricted to spinners only.

[mishra]

@cricketrulez - I believe you are correct. predicting the future path of the ball or bounce of the ball is just very difficult. Say in England on a day it may happen that wind speed/direction may change every second. Thats why we see those late swings. Frequently we see those near impossible trajectory of the ball. Now unless hawkeye can measure wind density and speed in realtime and calculates possible effect it cant predict the correct path. On the contrary the onfield umpire can feel the gush of the wind and factor in the bouncy condition of track. There could be a reason why Swanny and English get so many decision correct on hawkeye. They probably know the actau logic of hawkeye or may have been the users during requirement phase

[cricketrulez]

I believe they would have done that analysis about predictions of Hawkeye and how much of these predictions come true if you allow normal course of things to take place. Without that, it wouldn't have been put in practise. No reason for this to be restricted to spinners only.

Then how come it is not available to the public? In all the marketing material for H/E, this would be key piece of information which would convince the naysayers and shut them up for good. If you claim that you can predict the path with incomplete data, as in when it hits the pads, you should be able to do that with incomplete data, by stopping the feed to the software from a given distance and demonstrate its accuracy by comparing it with the actual event. they have put in a 2.5m rule. Why? how did they come up with this 2.5m number? by theory or practice?

[punjabi_khota]

1). There is no such thing as "turbulence event". 2). The pitch doesn't matter in prediction as the (last) measurement is taken after the ball has bounced. I know Nirvanam had hard time grasping this concept, but hopefully others won't.

[wobblydoggy]

I dont really understand how this Hawk eye works. Anyhow, why dont they just shoot some footage of a bowler, a spinner preferably bowling some deliveries at the stumps without a batsman standing in front of the wickets and compare it with Hawk eye and see how accurately it predicts the line ? Let it be examined by an independent panel from all countries and if the accuracy is good, then nobody should have a problem adopting it and if there isnt, it should be shelved. Its simple, isnt it ?

What a sensible answer.

[punjabi_khota]

@cricketrulez - I believe you are correct. predicting the future path of the ball or bounce of the ball is just very difficult. Say in England on a day it may happen that wind speed/direction may change every second. Thats why we see those late swings. Frequently we see those near impossible trajectory of the ball. Now unless hawkeye can measure wind density and speed in realtime and calculates possible effect it cant predict the correct path. On the contrary the onfield umpire can feel the gush of the wind and factor in the bouncy condition of track. There could be a reason why Swanny and English get so many decision correct on hawkeye. They probably know the actau logic of hawkeye or may have been the users during requirement phase

:omg: did i enter the wrong forum :wall:

[cricketrulez]

1). There is no such thing as "turbulence event". OK, turbulent event 2). The pitch doesn't matter in prediction as the (last) measurement is taken after the ball has bounced. I know Nirvanam had hard time grasping this concept, but hopefully others won't. So do you (or hawkeye) have any independent experimental data to support your conclusion? Please explain the concept one more time if you don't mind. If you are suggesting that prior measurements can accurately predict the future path of a projectile governed by turbulence, I'd say you would need to read up a bit.

[punjabi_khota]

1). There is no such thing as "turbulence event". OK, turbulent event

:haha: I am not trying to correct your English here. I don't know where you are getting your terminology, but turbulence is not even remotely related to what we are talking about. I am assuming you are referring that there are uncertainties in the system such as wind speed, air drag etc.

2). The pitch doesn't matter in prediction as the (last) measurement is taken after the ball has bounced. I know Nirvanam had hard time grasping this concept, but hopefully others won't. So do you (or hawkeye) have any independent experimental data to support your conclusion?

Here's a thought experiment. Remeber we are only talking about the effect of the pitch here, so assume no disturbances. Throw a ball down on the floor, and say you are able to record the velocity and position of the ball, say 10 millisecs after it bounces. Now do you need to know what kind of ground is it to predict the path of the ball ?

[cricketrulez]

I am assuming you are referring that there are uncertainties in the system such as wind speed, air drag etc. Yup Here's a thought experiment. Remeber we are only talking about the effect of the pitch here, so assume no disturbances. Throw a ball down on the floor, and say you are able to record the velocity and position of the ball, say 10 millisecs after it bounces. Now do you need to know what kind of ground is it to predict the path of the ball Yes, you would. The path of the ball is governed by its rotation/spin which in turn is governed by all the factors you mentioned earlier If you are going about though experiments, you are talking about predicting the location of the ball ~65 milliseconds after it hits the pads (assuming a bowling speed of 140kph, we will leave the deceleration aspect alone. 38 m/sec 2.5m rule 50milli sec to travel 2.5m it would take ~65 milliseconds ) The question that is being asked is, has anyone "measured" this accuracy under realistic conditions.

[punjabi_khota]

Man, I asked you to ignore all the other factors and focus on the pitch. JUST THE PITCH. I am claiming that once you take all the measurements AFTER the bounce, the pitch doesn't matter :wall: Think about this. At 10 milliseconds AFTER THE BOUNCE, there is a miracle and the pitch is changed. Will the path of the ball change ?

[ganeshran]

Pokay is right. The pitch comes into play only if the batsman is struck on the full. If the ball has already bounced then the pitch has no meaning because the pitch's affect on the ball is already concluded and all that remains is calculation of its trajectory using the path that its already taking

[cricketrulez]

OK, we will go with Pitch doesn't mater bit for now. Let me repeat my original question. Please focus your answer on this question Has there been any experimental validation of H/E predictions for cricket purposes?

[mishra]

Never Heard of any. May be BCCI/ICC can actaully make a independent body for the validation. QA is definitely missing or questionable for the DRS. I still dont think that they have factored anything which we have been discussing. Each time i see that hawkeye it reminds me of the movement of ping pong ball on ping pong table without those spin shots. And I am certain that hawkeye prediction is just simple as that

[cricketrulez]

According to this paper (only independent peer reviewed paper) http://pus.sagepub.com/content/17/3/283 All testing by the vendor was done using a bowling machine! At this point, I'm fully behind BCCI! F#$k H/E Bowling machine? are they serious?

[Blue-no1]

Pokay is right. The pitch comes into play only if the batsman is struck on the full. If the ball has already bounced then the pitch has no meaning because the pitch's affect on the ball is already concluded and all that remains is calculation of its trajectory using the path that its already taking

Today during SL-Eng game they were showing highlights of SL innings due to rain, did any one see this: The ball started out as an outswinger as it left the bowlers hand, pitched moved in and as as it left the batsman ended as an outswinger. I could not believe it and Matt prior collected the ball awkwardly above his head after moving in the wrong direction. Does Hawkeye account for post pitching variations due to dew/air factors?