PDA

View Full Version : Hawkeye gets it wrong


Nuke
06-12-2008, 07:10 AM
This probably won't be much of a surprise if you've watched a lot of matches on TV, but Hawkeye is being called on the carpet:

http://www.guardian.co.uk/sport/2008/jun/12/wimbledon.tennis

The accuracy of Hawk-Eye and similar line-calling technologies has been called into question by new research from Cardiff University's social sciences department.

In a paper called You Cannot Be Serious! Public understanding of technology with special reference to Hawk-Eye, the researchers claim that errors made by the machine can be greater than 3.6mm - the average error stated by its manufacturers.

Hawk-Eye has been shown to call a ball 'in' by 1mm when its true position was 'out' by 5mm. This result would pass the International Tennis Federation's testing system, despite being inaccurate.

"Technologies such as Hawk-Eye are meant to relegate line-call controversies to the past. However, our analysis has shown that Hawk-Eye does not always get it right and should not be relied on as the definitive decision-maker," said professor Harry Collins.

Hawk-Eye can still cause problems even when it makes the correct call because it can contradict what the human eye 'sees', as balls moving at certain speeds and angles will fool the human brain into thinking they are out when technically they are in. A correct call by Hawk-Eye in these situations will contradict the evidence of the majority of observers, causing controversy, the study claims.

The research will be welcomed by defending Wimbledon champion Roger Federer, a persistent critic of Hawk-Eye. During his victory over Rafael Nadal in last year's final, he protested that the machine was "killing him". Hawk-Eye will be implemented for the second time at this year's championships.

Automatix
06-12-2008, 07:17 AM
It's still more accurate than a human eye and that's the whole point...

flyer
06-12-2008, 07:26 AM
It's still more accurate than a human eye and that's the whole point...

That is not enough though because some balls are obviously out, as in leave a mark on the court, etc and hawkeye calls them in....thats whe it becomes a problem

perhaps the call should on ly overturned in the ball is in (or out) by more than the possible margin of error, that way only the calls that can be 100% certain to be accurate will be overturned

Duzza
06-12-2008, 07:28 AM
Watch out, it may backfire and start killing people>
http://blog.wired.com/gadgets/ed209.jpg

Rickson
06-12-2008, 07:32 AM
It's still more accurate than a human eye and that's the whole point...

That's for sure, but for some reason, they must have positioned the hawkeye cameras incorrectly at last year's Wimbledon final because Nadal's challenges were definitely wrong. Some of those balls had to be at least an inch out, but hawkeye had them right on the line. Federer should have done the same thing and challenged any ball he struck that was within 2 inches out.

sureshs
06-12-2008, 07:35 AM
During one of the clay tournaments this year, the mark showed the ball in but HawkEye/Shotspot called it out by a couple of mm, within the margin of error. I forget the details but it was on TTC.

Janne
06-12-2008, 07:38 AM
I remember seeing Hawk Eye making a ridiculous call when Mauresmo was playing last year. The whole stadium together with the umpire and the players laughed a little when they saw it.

sureshs
06-12-2008, 07:42 AM
That is not enough though because some balls are obviously out, as in leave a mark on the court, etc and hawkeye calls them in....thats whe it becomes a problem

perhaps the call should on ly overturned in the ball is in (or out) by more than the possible margin of error, that way only the calls that can be 100% certain to be accurate will be overturned

I think the "margin of error" is already factored in and cannot be added this way. If you are a vendor and your scale has a +/- 2% error, can a customer ask you to add 4% more material to anything that you sell to compensate for it? It doesn't work that way. The tolerance has been calculated already in a statistical fashion, and it has to be accepted as such. The 3.6 mm or whatever is also not a 100% figure - it is probably from some statistical analysis also.

Also, bringing the human eye into the picture will only complicate the matter. The two systems cannot be combined in some simple way.

dropserve
06-12-2008, 07:43 AM
When a call is challenged there is almost always NO REPLAY. Isn't that a little bit strange?

Why don't we get a replay just to compare?

That's very suspicious...

bsandy
06-12-2008, 07:43 AM
Everybody knows it's +/- 6mm, but we won't have to worry about the worst call in history -- Serena at the US Open.

. . . Bud

AAAA
06-12-2008, 07:55 AM
Imagine the scene

It's matchpoint to Basher Bill against Fast Eddie. Bill hit's a screaming shoot to the baseline. Eddie tries desparately to reach it but only manages to see the ball just clip the outside edge of the baseline, Eddie sees the ball was just in by a hair's width. The crowd see in, the line judges see in, Eddie sees in, the umpire sees in and Bill sees in. Now Eddie and every player knows about the imperfections of Hawk-eye so Eddie tries his luck and challenges hoping on this call that Hawk-eye will make an error and show the ball out, lucky for him Hawk-eye makes a mistake. Since Hawk-Eye is the absolute law, Eddie saves the matchpoint and then goes on to win.
It goes both ways but the other player may never get as good an opportunity(ball just in) in the match to exploit the flaws of Hawk-eye.

There are other situations Federer has mentioned which I can't think of right now where a player can use Hawk-eye to gain an advantage they couldn't gain before Hawk-eye was introduced.

In the old days Mac could intimidate line-judges to reverse decisions but the umpire had the power to over-rule.

Now substitute 'intimidate' for 'exploit technological flaw' the Umpire has no power to over-rule because Hawk-eye is the law.

A commentator once said on air he asked a tournament organiser why they didn't use Hawk-eye and Cyclops to be doubly sure. Organiser said they don't incase they contradict each other.

sureshs
06-12-2008, 08:01 AM
The problem in factoring in other decisions from the players and linespeople is that if the underlying assumption is that HawkEye is superior, the combined system will render its use invalid. If the assumption is that HawkEye is not superior, it need not be used. Let us say you measure something with a low precision instrument to get an approximate idea, or because it is cheaper to use most of the time. Sometimes, you bring in a more precise instrument from time to time to take a reading. Then, if you combine the low and high precision readings by some averaging, then the entire measurement system becomes invalid.

sureshs
06-12-2008, 08:03 AM
A commentator once said on air he asked a tournament organiser why they didn't use Hawk-eye and Cyclops to be doubly sure. Organiser said they don't incase they contradict each other.


Yes, that is the statistically correct way to do it. You should not use both independently and then combine.

Thud and blunder
06-12-2008, 08:20 AM
This is a social science paper about the perception of technology, not about the technology itself. So no idea why the papers are reporting it as the latter.

cknobman
06-12-2008, 09:21 AM
Watch out, it may backfire and start killing people>
http://blog.wired.com/gadgets/ed209.jpg

OMG LOL I just watched Robocop the other day and cracked the F up when this thing went psycho in the board room and totally wasted this guy pretending to hold a gun. LOL!!!

cknobman
06-12-2008, 09:24 AM
During one of the clay tournaments this year, the mark showed the ball in but HawkEye/Shotspot called it out by a couple of mm, within the margin of error. I forget the details but it was on TTC.


Also stated by the commentator on TTC was that since hawkeye was not official and there for TV only that there were fewer cameras positioned around the court which increased the innaccuracy of the system. So what you saw on clay is not indicitive of how Hawkeye is used when it is official.

Djokovicfan4life
06-12-2008, 09:28 AM
That is not enough though because some balls are obviously out, as in leave a mark on the court, etc and hawkeye calls them in....thats whe it becomes a problem

perhaps the call should on ly overturned in the ball is in (or out) by more than the possible margin of error, that way only the calls that can be 100% certain to be accurate will be overturned

On clay I agree that Hawkeye is pretty substandard. But unfortunately it's still our best bet on other surfaces, IMO. Those "marks" seem pretty questionable sometimes on hard courts, and obviously grass is grass, so it's just something we're gonna have to live with for a while.

Duzza
06-12-2008, 06:27 PM
OMG LOL I just watched Robocop the other day and cracked the F up when this thing went psycho in the board room and totally wasted this guy pretending to hold a gun. LOL!!!

LOL so did I. We couldn't stop laughing at the pointlessness of the violence. *10 billions shots to the legs and head* = "oww.....I'll come back as half man half robot"

SempreSami
06-12-2008, 06:33 PM
Fed's little rant at last year's Wimbledon final was pathetic.

"How in the world was that ball in?"
<Carlos Ramos says something>
"****!"

I hope the BBC commentators apologised out of how cringeworthy Fed sounds when he's annoyed.

krprunitennis2
06-12-2008, 06:42 PM
I think tennis is just getting too fancy and riled up with the hawkeye.

I'm not entirely against it though. IMO, it adds some intensity in matches when people really have no clue if the ball was in or out "Oooooooooo. AAHHH!" I don't think people should overthink it though xO

Tennis was fine a few years ago w/o the Hawkeye.

Max G.
06-12-2008, 06:48 PM
I'll say the same thing I always do. Hawkeye should be available to the chair umpire to help him make decisions. The chair should be able to show himself replays of any balls he feels like, and then make the best decision he can, taking into account both Hawkeye (and its margin of error), the call of the linesman, and what he saw. I presume that in most cases, this would mean just going with Hawkeye, since (on average) it's better than the linesmen or the chair, but he'd have the discretion to overrule Hawkeye in the case of a 'clear error', as with linesmen.

Of course, that's in the ideal world. In our current situation, despite Hawkeye's flaws, it's better at calling the lines than the linesmen are, and thus should always be given priority; I can agree with that much.

Max G.
06-12-2008, 06:51 PM
Imagine the scene

It's matchpoint to Basher Bill against Fast Eddie. Bill hit's a screaming shoot to the baseline. Eddie tries desparately to reach it but only manages to see the ball just clip the outside edge of the baseline, Eddie sees the ball was just in by a hair's width. The crowd see in, the line judges see in, Eddie sees in, the umpire sees in and Bill sees in. Now Eddie and every player knows about the imperfections of Hawk-eye so Eddie tries his luck and challenges hoping on this call that Hawk-eye will make an error and show the ball out, lucky for him Hawk-eye makes a mistake. Since Hawk-Eye is the absolute law, Eddie saves the matchpoint and then goes on to win.
It goes both ways but the other player may never get as good an opportunity(ball just in) in the match to exploit the flaws of Hawk-eye.


However, there's the opposite situation. Linesman sees the ball in, chair doesn't see it well enough to overrule it, Eddie challenges and finds out the ball was actually out. If Hawkeye is more accurate than the linesmen, this situation would be more common than the exploit-hawkeye-to-get-a-bad-call, thus resulting in, overall, more correct calls and fewer bad calls.