« WORST SLIP | Main | THE LONG LEGBYE »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

David Barry

The original paper that the article is based on was pretty disappointing. It does talk about cricket, but it doesn't tell us anything that a diligent reader of Cricinfo wouldn't know already - indeed, the authors openly acknowledge that most of their material for the cricket section comes from a couple of Cricinfo articles.

All it says that there's a margin of error, and it would be more honest and correct to put a probability on the Hawkeye decisions, rather than an apparently certain out/not-out. In the case of balls pitching close to the pad, the accuracy is still within 20mm (but what is accuracy? 95% of balls within 20mm? 99%? Hawkeye haven't said.), and is typically better than 5mm if there's more than two feet between the pitch of the ball and the impact on the pad.

Adsy

Hawkeye in tennis just has to measure where it thinks the ball will bounce.. not whether or not the ball would have hit the linesman in the schnozz on the way through. Cricket has to do this (substitute schnozz for stumps) and thats where the uncertainty lies.

As Tony said, you can use SSMo for seeing where the ball pitches, hits on the pad, and even the height of the ball from side on to see if its likely to go over the top. The rest is educated speculation at best, and something that has to take into account all the things that go into making an LBW decision.

My problem with cricket decisions (and I'm no maths guru) is that isn't the program just extrapolating the current flight trajectory of the ball on the way through, with no real way of calculating the spin or swing once it hits the pad? I guess that what an umpire has to go on as well, but they have a little more intuition (you'd think) than a computer program.

Adsy

Hawkeye in tennis just has to measure where it thinks the ball will bounce.. not whether or not the ball would have hit the linesman in the schnozz on the way through. Cricket has to do this (substitute schnozz for stumps) and thats where the uncertainty lies.

As Tony said, you can use SSMo for seeing where the ball pitches, hits on the pad, and even the height of the ball from side on to see if its likely to go over the top. The rest is educated speculation at best, and something that has to take into account all the things that go into making an LBW decision.

My problem with cricket decisions (and I'm no maths guru) is that isn't the program just extrapolating the current flight trajectory of the ball on the way through, with no real way of calculating the spin or swing once it hits the pad? I guess that what an umpire has to go on as well, but they have a little more intuition (you'd think) than a computer program.

David Barry

I disagree with that Adsy. If you've got the trajectory of the ball, extrapolating's easy. The LBW law assumes that the ball would have continued on the path it was on when it hit the pad, so that any subsequent swing (or spin in the case of it hitting the pad on the full) is ignored.

Russ

David is right. The article raised a lot of questions worth answering, but in the absence of any answers didn't tell us much. I'm curious about the use of bowling machines for testing. If they aren't using cricket balls then they may not be seeing the particularly interesting cases of spin and swing that would cause large uncertainties.

Adsy, umpires aren't really supposed to predict swing or spin, except that which they have already seen. Though, obviously a ball that they see starting to swing viciously is going to introduce more doubt, which hawkeye will ignore. In both cases, the biggest issue is the limited amount of information you get between the ball bouncing and hitting the pad. The trajectory before bouncing is largely irrelevant for any sort of close call, and both umpires and hawkeye are going to struggle with balls seaming a little and hitting the batsman on the half volley. It also never helps that the ball effectively jumps when it hits and is absorbed by the pad, so the important frame that shows whether a batsman is being hit "in line" is never quite right.

Having said all that, 20mm is only 1/10th the width of the ball. Applying a benefit of the doubt principle would allow both an increasingly large cone of uncertainty - an area within which the complete ball is 99% likely to be contained - and a diminishing cone of certainty - an area within which some part of the ball is 99% likely to touch. As long as the cone of certainty touches the stumps the decision is probably okay; at some level though, some decisions are probably unknowable.

Which raises an interesting point about the "benefit of the doubt". Strictly speaking it doesn't exist, it is just accepted for general play. At almost every level of cricket I've watched though, umpires don't actually tend towards a "benefit of the doubt" so much as a "benefit of the shot". Players who play forward and get hit on the pad in front are generally reprieved regardless of the outness of the ball; players who get hit on the leg playing back, or who don't play a shot at all, or who get some vicious delivery that beats them all ends up, are invariably given out. I don't actually mind this either, there is a certain natural justice to making a decision based on who of the bowler or batsman "won", rather than strictly technical grounds.

Russ

ah... what am I talking about. That should be 1/3-1/4 the width of the ball, which is fairly significant when you start diminishing it from either side (and particularly when you need to add an initial measurement uncertainty of around 5mm). At any rate, you can probably be reasonably sure that if the centre of the hawkeye ball hits the stumps, then some part of the true ball would hit the stumps; if hawkeye says it is clipping the stumps then there is probably a better than 50% chance of it hitting, but still sufficient doubt to question it.

Francis Xavier Holden

Trivial. Haven't you got any kids to teach?

Adsy

Isn't that rule only applicable for balls that hit the batsman on the full? Eg. low full tosses etc? If a ball bounces, it has to pitch in line (generally), hit in line AND go on to hit the stumps. There are many times where bowlers pitch in line, hit in line and batsmen get the "benefit" due to the umpires interpretation that it may have indeed gone on to miss the stumps. This "interpretation" part is what I'm worried about.

Russ

Adsy, the general principle is always to predict based on the observed trajectory of the ball prior to impact. The full toss rule is to prevent conjecture (that the ball will spin after it bounces, for instance) that can only lead to confusion.

Tony T

A Test batsman who misses a fully on his pads is asking for trouble.

Key phrase... apart from Cone of Certainty: "the biggest issue is the limited amount of information you get between the ball bouncing and hitting the pad."

If Hawk had up to now shown that it was capable of assessing late-and-sharp movement off the pitch and swing through the air, then I wouldn't have a problem with it. But I have my doubts about whether it can operate/update/process quick enough to account for late movement.

Take a sharp offie to a right-hander. Hawk often underestimates the turn down leg and has the ball hitting the stumps instead of clearly missing leg as is shown on the TV footage, both normal speed and slo-mo. This says to me trajectory is calculated earlier in flight and projected as is, without the turn. That it does so is even more apparent with big in-swingers which are shown via a large to-leg curves before the ball bounces, but with a straightening of the trajectory after the ball has pitched.

As always, though, it's not the fact Hawk-Eye misreads trajectory that gets under my skin; I'm an engineer, so I like technical advancements and am prepared to accept their introduction as long as their application is explained and justified. Nine never did that. They immediately deferred to Hawk's judgment. At some point they must have had a meeting in which they addressed how to handle Hawk, and in which they concluded they would back it in regardless, even when they knew it was wrong. This both hit at their credibility, such as it is, and played us for suckers, such as we are. At least they have recently pulled back their sell.

Mind you, tennis doesn't need all this. All it needs is the ability to spot the ball on the ground.

PS: It's about time the AGB got a Hawk-Eye category.

PPS: I'm on holidays. I mean, self-directed duties.

Big Ramifications

What are you going to do - go up to the net and start having a debate or asking fans to phone in and vote?"

*END OF EXPLANATION* Why does he have to get on his high horse after that?

Do they really have nothing more important to do?

What, like chase balls around for the entertainment of drunken fat yobbos? What a pretentious git.

Paul Hawkins said: "Anyone living in the sane world would say Hawk-Eye is sufficiently accurate."

"No system is ever going to have zero margin of error but at some point you have to make a decision.

Exactly. Personally, I’d rather have a computer error than a human error. I get the feeling that computer errors will at least be a lot more CONSISTENT. Rather than Rudi “Check me out ladies. Watch how quick I can raise my finger.” Koerzten, or Billy “Gee aren’t I a wacky guy? Look at me make an abortion of every cricket signal in the book. Wouldn’t you just love to have me around at your next dinner party?” Bowden, or Morgan "You shot who in the what, now?" Freeman, or Aleem “Nice little earner” Dar.
.
.
.
And what everyone said above about super slomo. I would trust my own judgment with super slomo over Hawkeye. I can’t understand why Hawkeye hasn’t improved in the last few years. Surely all it would take would be to increase the frequency of data points. Possibly improve the cameras. Maybe the tracking technology has a ways to go.

Give it time, it will get a lot better.

ps: Get yer finger out, Paul Hawkins. Time to spend a bit more coin on R&D.

Tony T

Those nicknames really roll off the tongue, Big Baby.

But you is right. Hawkins ought to spend less time on smart-arse put-downs, and more on improving Hawk "I often get it wrong when the ball pitches very close to the batsman" Eye.

Big Ramifactions

Those nicknames really roll off the tongue, Big Baby.

I have climbed the mountain. And I have SEEN the Promised Land.

Professor Rosseforp

When the commentators can start pronouncing the tournament as WIMBLEDON instead of WIMPLETON, we might start to look at headier matters such as hawkeye.
This issue came up at my previously-mentioned annual trivia night, where several participants challenged the host's Wimpleton pronunciation (it happens every year), thereby incurring a $1.00 fine, but the host was good enough to fine himself for mispronunciation.

Soulberry

A combination of tricks may work for cricket. Add Hot Spot to whichever combination of technology cricket has decided to accept.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment

Your Information

(Name and email address are required. Email address will not be displayed with the comment.)