The LBW Umpire Referral Flaw

[Lillian Thomson]

The authorities are doing their best to minimise the effect that Hawkeye has had on LBW's - hence the alleged over use of the margin of error. A lot of older cricketers (and viewers) have been surprised (and alarmed) by the number of balls that are shown to be hitting the stumps that in the past would have been assumed not to have been. A finger spinner would never get an LBW if the batsman got a big stride in because the umpire would have to be guessing. Graham Swann for example got a lot of LBW's that previous generations of spinners would not have been given. Richie Benaud (RIP) once joked that if you gave out everything that Hawkeye is showing to be hitting the stumps the match would be over in a day. It's particularly those that are shown to be clipping leg stump when the naked eye based on passed assumptions believes it's missing.

 

[TheJediBrah]

The authorities are doing their best to minimise the effect that Hawkeye has had on LBW's - hence the alleged over use of the margin of error. A lot of older cricketers (and viewers) have been surprised (and alarmed) by the number of balls that are shown to be hitting the stumps that in the past would have been assumed not to have been. A finger spinner would never get an LBW if the batsman got a big stride in because the umpire would have to be guessing. Graham Swann for example got a lot of LBW's that previous generations of spinners would not have been given. Richie Benaud (RIP) once joked that if you gave out everything that Hawkeye is showing to be hitting the stumps the match would be over in a day. It's particularly those that are shown to be clipping leg stump when the naked eye based on passed assumptions believes it's missing.

I knew there was a reason he got wickets.

 

[Howe_zat]

I don't think it's reasonable to assume that the size of the 'umpire's call' region is set by how accurate Hawkeye is accepted to be by the manufacturer.

Even if it did by sheer fluke happen to be the exact same width as half a ball plus half a stump, you're still assuming that:

1. Hawkeye has the same absolute margin of error when the ball stops a variety of different distances from the stumps
2. Hawkeye has the same absolute margin of error regardless of where the cameras are set up or how far away from the pitch they are at different grounds
3. Hawkeye has the same absolute margin of error for balls travelling at completely different speeds
4. Hawkeye has always been incapable of improving its margin of error over several years

which make between little and no sense. It makes far more sense to assume the relative error may be the same, in which case all these variables will change the maximum absolute error. But because stumps and balls are always the same width, we use a standard size of 'umpire's call' region.

The 'umpires' call' region is the size it is so as commentators and TV viewers can understand it, give power to the umpire and carry on the tradition of making lbw decisions weighted in the batsman's favour. It has little to do with the accuracy of the technology, which remains frustratingly unclear.

 

[Lillian Thomson]

The 'umpires' call' region is the size it is so as commentators and TV viewers can understand it

Best post yet. If we could only get the players to understand it this thread would be redundant.

 

[Dan]

I've always seen this as an aid to on field umpires who make the final decision then and there, when technology gets involved I see no reason for such traditions to be observed,

All I am saying is that, for the benefit of people watching (i.e. umpires, as well as viewers), having two different varieties of Umpire's Call (ball fractionally missing stumps, ball fractionally clipping stumps) would be an absolute pain in the arse. Nobody seems to understand the system as it stands now, with only one variety of Umpire's Call.

I am suggesting that the Hawkeye people **** around with the size of the stumps displayed on-screen, so all Umpire's Call decisions, be they fractionally missing or fractionally clipping, appear to be the same. If they do that, I'm suggesting they ought to make them all look like they're clipping, because that gels with traditional decision making patterns a lot more than making them all appear like they're missing the stumps by a little bit.


Honestly, I've missed the obvious option for this -- colour-coded overlays:

Problem solved let's all go home.

Rule 1: Any bit of the ball hits the red zone, you're out (assuming all other conditions of LBW are met)
Rule 2: Any bit of the ball hits the yellow zone, it's the Umpire's Call (assuming all other conditions of LBW are met and Rule 1 is not met)
Rule 3: The entire ball is contained within the green zone, you're not out LBW.

 

[Dan]

The 'umpires' call' region is the size it is so as commentators and TV viewers can understand it, give power to the umpire and carry on the tradition of making lbw decisions weighted in the batsman's favour. It has little to do with the accuracy of the technology, which remains frustratingly unclear.

Yeah, all I've been suggesting is to fix the Umpire's Call region size to better match the accuracy of the technology, while still portraying it in exactly the same way to viewers.

 

[Top_Cat]

Tbh, the lack of published info on Hawkeye's accuracy is exactly why some of us don't like that throwing assessments have retreated from public view and scrutiny. We're all completely in the dark about how they've assessed the accuracy of it, how they've improved it, by how much, etc. Even the public statements of the founder don't really tell you a lot. For example, aside from what exactly is meant by 5mm, they claim 5mm accuracy but 10mm in some circumstances; which circumstances and how often do they tend to occur in a match? If it's 5mm accuracy for anything stump high and gets worse the shorter the ball, that's a very different thing to the converse.

 

[Flametree]

My gripes with the system :

Do the laws state what they mean by "pitching outside leg stump" or "hitting in line with stumps"? I doubt the pre-DRS law did - was it "if any part of the ball pitches outside leg it can't be lbw", or "if all of the ball pitches outside leg...", ie if any part pitches in line with the stumps it can be lbw. Same with the impact being in line - some of the ball or all of the ball? As far as I'm concerned, like in tennis, where it's measuring something which has actually happened, it should be 100% yes / no, and there shouldn't be an umpire's call on where it pitched or where it impacted the pad.

Second, the margin of error for the prediction should vary based on how far from the wickets the point of impact is. If you play a forward defensive shot, get hit on the pad and the ball is shown to be clipping leg stump, say 40% of the ball hitting some part of the stump, then I'm fine an umpire giving that not out on a "can't be sure" basis and having technology uphold the decision. If you go right back to an offspinner and miss it, and again technology shows 40% of the ball hitting the leg stump, then to me that should be out whatever the umpire initially says. Yes, less than half the ball is predicted to hit the stumps exactly like the previous ball, but the likelihood that none of the ball is about to hit the stumps surely differs depending on the distance left for it to travel.

Visually you could present it like the strike zone earlier - if any part of the ball is hitting within a strike zone it's out, but the strike zone varies in size depending on the point of impact. Big stride forward, strike zone shrinks to say middle half of the off and leg stumps and 2-3 cm from the top of the bails; hit on the back foot, strike zone is basically the size of the stumps minus a coat of varnish or two...

 

[Maximas]

@Dan, looks good to me

 

[TheJediBrah]

My gripes with the system :

Do the laws state what they mean by "pitching outside leg stump" or "hitting in line with stumps"? I doubt the pre-DRS law did - was it "if any part of the ball pitches outside leg it can't be lbw", or "if all of the ball pitches outside leg...", ie if any part pitches in line with the stumps it can be lbw. Same with the impact being in line - some of the ball or all of the ball? As far as I'm concerned, like in tennis, where it's measuring something which has actually happened, it should be 100% yes / no, and there shouldn't be an umpire's call on where it pitched or where it impacted the pad.

Second, the margin of error for the prediction should vary based on how far from the wickets the point of impact is. If you play a forward defensive shot, get hit on the pad and the ball is shown to be clipping leg stump, say 40% of the ball hitting some part of the stump, then I'm fine an umpire giving that not out on a "can't be sure" basis and having technology uphold the decision. If you go right back to an offspinner and miss it, and again technology shows 40% of the ball hitting the leg stump, then to me that should be out whatever the umpire initially says. Yes, less than half the ball is predicted to hit the stumps exactly like the previous ball, but the likelihood that none of the ball is about to hit the stumps surely differs depending on the distance left for it to travel.

Visually you could present it like the strike zone earlier - if any part of the ball is hitting within a strike zone it's out, but the strike zone varies in size depending on the point of impact. Big stride forward, strike zone shrinks to say middle half of the off and leg stumps and 2-3 cm from the top of the bails; hit on the back foot, strike zone is basically the size of the stumps minus a coat of varnish or two...

A lot of your suggestions, while sensible, would simply be too complicated for commentators and general audience to get their heads around IMO

 

[wellAlbidarned]

The only changes we need are

1. Margin of error is changed to any part of the ball hitting the middle of leg or off stump

2. Top 3 batsmen are banned from reviewing LBWs because they suck at using them

 

[Bahnz]

In my opinion, the change that really needs to be made is shrinking the margin of error to what the Hawkeye's evidence actually says it is (I remember reading somewhere that it was 2mm, though Dan's link to a 5-10mm moe seems reasonable as well). A much smaller and more realistic margin of error would greatly reduce the current imbalance between "out" DRS reviews and "not out" DRS reviews. Ideally I'd like the moe extended in both directions, but it would be a less significant issue if there wasn't such a massive and arbitrary moe in the first place.

I'd be really interested to see a statistical analysis of whether "not out" decisions have become more common for lbws in matches since the DRS came into play, as currently it provides a very strong incentive for umpires to err on the side of caution for those 70:30 umpiring shouts.

 

[RossTaylorsBox]

2. Top 3 batsmen are banned from reviewing LBWs because they suck at using them

Ross Taylor to first drop then.

 

[Dan]

In my opinion, the change that really needs to be made is shrinking the margin of error to what the Hawkeye's evidence actually says it is (I remember reading somewhere that it was 2mm, though Dan's link to a 5-10mm moe seems reasonable as well). A much smaller and more realistic margin of error would greatly reduce the current imbalance between "out" DRS reviews and "not out" DRS reviews. Ideally I'd like the moe extended in both directions, but it would be a less significant issue if there wasn't such a massive and arbitrary moe in the first place.

2mm is the company-quoted figure for tennis Hawkeye (i.e. measuring what has actually happened), whereas the predicted path has a slightly larger one.

I agree with everything T_C says, too. As per usual.

 

[cnerd123]

What about the 2.5 meter rule?

World Cup 2011: ICC issues revised guidelines for 2.5m rule | Cricket News | ICC Cricket World Cup 2011 | ESPN Cricinfo

FWIW I can't find anything official yet to explain the half-ball rule, except that the ICC feels that the margin for error is too significant. No data or accuracy figures.

 

[dfrinku]

My gripes with the system :

Do the laws state what they mean by "pitching outside leg stump" or "hitting in line with stumps"? I doubt the pre-DRS law did - was it "if any part of the ball pitches outside leg it can't be lbw", or "if all of the ball pitches outside leg...", ie if any part pitches in line with the stumps it can be lbw. Same with the impact being in line - some of the ball or all of the ball? As far as I'm concerned, like in tennis, where it's measuring something which has actually happened, it should be 100% yes / no, and there shouldn't be an umpire's call on where it pitched or where it impacted the pad.

Second, the margin of error for the prediction should vary based on how far from the wickets the point of impact is. If you play a forward defensive shot, get hit on the pad and the ball is shown to be clipping leg stump, say 40% of the ball hitting some part of the stump, then I'm fine an umpire giving that not out on a "can't be sure" basis and having technology uphold the decision. If you go right back to an offspinner and miss it, and again technology shows 40% of the ball hitting the leg stump, then to me that should be out whatever the umpire initially says. Yes, less than half the ball is predicted to hit the stumps exactly like the previous ball, but the likelihood that none of the ball is about to hit the stumps surely differs depending on the distance left for it to travel.

Visually you could present it like the strike zone earlier - if any part of the ball is hitting within a strike zone it's out, but the strike zone varies in size depending on the point of impact. Big stride forward, strike zone shrinks to say middle half of the off and leg stumps and 2-3 cm from the top of the bails; hit on the back foot, strike zone is basically the size of the stumps minus a coat of varnish or two...

I think these are two crucial points. Does anyone know the answer to the first point? I don't think there should be any umpires call on where the ball pitches or hits. And I completely agree that the margin of error should and would be smaller if the ball has less to travel which is why I asked that question a few pages ago. Good points.

 

[Top_Cat]

What about the 2.5 meter rule?

World Cup 2011: ICC issues revised guidelines for 2.5m rule | Cricket News | ICC Cricket World Cup 2011 | ESPN Cricinfo

FWIW I can't find anything official yet to explain the half-ball rule, except that the ICC feels that the margin for error is too significant. No data or accuracy figures.

I'll be shocked if you do. It's a proprietary product and, tbh, I wouldn't be surpsied if Hawkeye never provided any supporting data to the ICC on the basis they wouldn't understand it anyway.

 

[Lillian Thomson]

This will go around in circles until they move the goalposts to stop protecting the umpires. Even if the technology was proven to be 100% accurate they still wouldn't use it as such under the current remit of reversing "glaring errors". The current purpose of Hawkeye is show whether or not the umpire made a reasonable decision based on what he saw with his eyes in real time. If it shows that he did the decision will always stand even when everyone suspects it's wrong. The margin for error is based on human error, not technological error.

 

[weldone]

2. Top 3 batsmen are banned from reviewing LBWs because they suck at using them

They'll just drop to 4-6 and use it, then.

 

[Dan]

How about we ban reviewing LBWs because everyone sucks at using them and the umpire is usually right anyway and they're not 100% accurate? #BCCIThoughts