Hawkeye: More fallible than they'd like us to think.

[kwigibo]

I'd love a reliable technological alternative to the on field umpire, but Hawkeye is looking more and more flawed. Case in point, Ntini's dismissal last night, which confirmed what I've suspected for a while, Hawkeye sucks at accurately displaying changes in trajectory off the pitch. Looking at the video, the ball spun enough that it was not missing off, but was likely a bit high. The commentators felt this way too until hawkeye told them what to think, which was the opposite, not necessarily high, but missing off.

It does make sense that these errors are made, and I've noticed it before numerous times, particularly when the ball is spinning generously. The technology is not as magical as the proprieters wish us to believe, to be able to discern otherwise relatively small but crucial changes in trajectory is probably beyond their tracking system. From what I know of it there is an amount of extrapolation involved, and in a dismissal such as this, I doubt there is enough time between ball pitching and hitting the pad to recieve sufficient data to make an accurate inference.

Anyone else dissillusioned?

Anyone still unhappy they haven't adopted it for official umpiring use?

[Autobahn]

Originally Posted by kwigibo

I'd love a reliable technological alternative to the on field umpire, but Hawkeye is looking more and more flawed. Case in point, Ntini's dismissal last night, which confirmed what I've suspected for a while, Hawkeye sucks at accurately displaying changes in trajectory off the pitch. Looking at the video, the ball spun enough that it was not missing off, but was likely a bit high. The commentators felt this way too until hawkeye told them what to think, which was the opposite, not necessarily high, but missing off.

It does make sense that these errors are made, and I've noticed it before numerous times, particularly when the ball is spinning generously. The technology is not as magical as the proprieters wish us to believe, to be able to discern otherwise relatively small but crucial changes in trajectory is probably beyond their tracking system. From what I know of it there is an amount of extrapolation involved, and in a dismissal such as this, I doubt there is enough time between ball pitching and hitting the pad to recieve sufficient data to make an accurate inference.

Anyone else dissillusioned?

Anyone still unhappy they haven't adopted it for official umpiring use?

Well they use the same technology for guided missiles so i doubt the accuracy is a problem and again hawkeye only follows the trajectory, it doesn't predict them as you seem to be suggesting.

The only thing that's widely accepted to make hawkeye go wrong is if the cameras aren't algined properly.

[SJS]

Originally Posted by Autobahn

Well they use the same technology for guided missiles so i doubt the accuracy is a problem and again hawkeye only follows the trajectory, it doesn't predict them as you seem to be suggesting.

The only thing that's widely accepted to make hawkeye go wrong is if the cameras aren't algined properly.

In one of the test matches in the recent series, Kumble clean bowled a batsman. Hawkeye showed that the ball would have missed the stumps by over an inch !!

This is a fact.

I think Hawkeye should be trusted only for where the ball pitched and for the rest it should be shown with a caveat 'Hawkeye is approximate"

[Smudge]

Originally Posted by SJS

In one of the test matches in the recent series, Kumble clean bowled a batsman. Hawkeye showed that the ball would have missed the stumps by over an inch !!

This is a fact.

The cheating Indian keeper must have moved the stumps!!!

[Francis]

I was actually going to start a similar thread because I had the exact same thoughts. I personally felt that the ball may have gone over the top as it barely hit Ntini's leg. However, it wasn't missing off. In the Ashes the C4 technitians had to re-alanyse a Warne ball that confused and confounded the Hawkeye machine. So they had to re-analyse it.

I like having the Hawkeye when watching cricket. But I am by no means convinced by it.

[SJS]

Originally Posted by Voltman

The cheating Indian keeper must have moved the stumps!!!

Actually Dravid did it from first slip but please dont tell anyone

[16 tins of Spam]

Even if Hawkeye could somehow be proven to be 100% accurate, I still don't want it.

With all the advocation of technology going on these days, the natural course of things will eventually render the umpire nothing more than a strangely dressed fellow who tells the players when it's time for a break.

It may be an old-fashioned attitude, but part of cricket's appeal lies in its connection with a more gentile past, and I'd like to keep the umps firmly planted at the non-strikers end, making the big calls.

Anyway, sport would be a lot less interesting without controversy.

[kwigibo]

Originally Posted by Autobahn

Well they use the same technology for guided missiles so i doubt the accuracy is a problem and again hawkeye only follows the trajectory, it doesn't predict them as you seem to be suggesting.

The only thing that's widely accepted to make hawkeye go wrong is if the cameras aren't algined properly.

Guiding missiles is different than tracking the exact trajectory of a small ball, on a scale of dimensions much smaller than that involved in any military application of the technology. And guided missiles aren't 100% accurate, far from it. And tracking guided missiles isn't a piece of cake either, otherwise the US missile defense shield wouldn't have been such a fiasco.

And Hawkeye is by definition a predictive system, it predicts where the ball would have gone had the batsman not got in the way. And the data collected could not possibly be so complete as to not involve some sort of inference. And of course that inference is going to be subject to error when accounting for the difficulty of establishing a consistent set of assumptions that apply to the fluid conditions involved in bowling a cricket ball.

If the cricket pitch were a mile long I'd say the margin of error would be negligible, but I don't think the technology is completely up to the task in the current circumstance.

[Dasa]

I'm fairly sure that the HawkEye animations shown on TV are just for the TV audience, and aren't as accurate as the real thing....it's just made into a slick looking animation to make it easier to understand for people watching at home.

Interesting points on the use of technology - http://content-aus.cricinfo.com/ci/c...ry/142121.html

[SJS]

Originally Posted by Francis

I like having the Hawkeye when watching cricket. But I am by no means convinced by it.

I like hawkeye too but some times commentators use it to butteress their own point of view. There are occasions when a ball appears to be going down the leg side. Umpire does not give the benefit of the doubt to the batsman and sends him packing, hawkeye shows that the ball would have got the faintest of kisses to the outside edge of the leg stump.

I have heard commentators admiring the 'great' decison by the umpire in these cases. I dont understand how the umpire in his regular, full speed, hawk-eye-deprived vision could have guessed that it would have not gone a millimeter on the other side? Let alone the fact that Hawkeye could be wrong.

The reverse has also happened at times.

[16 tins of Spam]

Originally Posted by Dasa

Interesting points on the use of technology - http://content-aus.cricinfo.com/ci/c...ry/142121.html

In places, that article is bizarre. Who has ever heard someone complain that microwaves remove the human element from cooking?

That has to be, hands down, the worst analogy in support of an argument I've come across. Ever.

[Dark Hunter]

It's always irritated me when commentators use it as evidence for their argument. "We'll take a look at hawkeye, yep, it was just going to hit, i'm right." Yet, if hawkeye shows it missing, and the commentator is saying it's out, they usually argue that hawkeye got it wrong.

IMO it's never going to be a hundred percent right. Hence, we shouldn't make it available to umpires. Yeah, 9/10 it'll probably help him make the right decision, but what if it get's one wrong, or doesn't compute properly. It's happened before, SJS gave us an example earlier. People seem to forget that it's just another bit of technology and it can have bugs and such, just like any other program.

One thing i would like to see done, is hawkeye following the ball after it's hit. Just imagine a batsman dancing down the pitch, then smashing the ball out of the park. Hawkeye shows the ball making it about halfway down the pitch, and all of a sudden it just changes direction and goes flying.

[16 tins of Spam]

Originally Posted by Dark Hunter

One thing i would like to see done, is hawkeye following the ball after it's hit. Just imagine a batsman dancing down the pitch, then smashing the ball out of the park. Hawkeye shows the ball making it about halfway down the pitch, and all of a sudden it just changes direction and goes flying

We kind of have that here in NZ with Virtual Spectator. It's like the old "wagon wheel", but in 3D - so those really big hits look spectacular in VS.

[FaaipDeOiad]

Originally Posted by mundaneyogi

In places, that article is bizarre. Who has ever heard someone complain that microwaves remove the human element from cooking?

That has to be, hands down, the worst analogy in support of an argument I've come across. Ever.

I don't know about that, but microwaves do make the food taste like crap. Is that the same complaint?

Seriously though, Hawkeye is quite obviously flawed quite a lot of the time. It'd be insane to attempt to use it in cricket right now instead of the umpire, as it isn't even close to reliable. How many times does the Kumble thing happen? It's only rarely that we see Hawkeye when it isn't on an LBW shout, and then we can't really test how accurate it is. If we're seeing clear examples of it being inaccurate on the rare occasions we as viewers can actually test it, it's obvious why it isn't being used yet.

Obviously Hawkeye is perfectly good at determining the trajectory of a straight ball, but it sucks awfully at picking up where the ball is going to go after it deviates, whether it be in the air or off the surface. The reasons for that are fairly obvious, if you think about it.

[andyc]

Originally Posted by Dark Hunter

It's always irritated me when commentators use it as evidence for their argument. "We'll take a look at hawkeye, yep, it was just going to hit, i'm right." Yet, if hawkeye shows it missing, and the commentator is saying it's out, they usually argue that hawkeye got it wrong.

Haha yeah that annoys me too.