Wisden on Murali's Action

[cnerd123]

OMG Pratters first Bradman now this.

plz stahp.

 

[Top_Cat]

Just to chime in with something about the video analysis done during the Champion's trophy to arrive at McGrath's flexion, it was not human beings looking at footage. The setup was that 2 cameras were used to view the bowling arm from two different angles.. Software exists that can convert two dimensional video to a 3d model of the movement.. So am not saying it is wholly accurate or anything just stating that it is not going to be as innacuarte as us looking at videos.

Almost certainly isn't accurate, converting from video to 3D is a hard problem. Lab testing is still king in terms of distances and measurements.

Elliott, et al, 2007, Sports Biomechanics, January 2007; 6(1): 99–108

That research is a bit old but more recent work (2012) hasn't solved the problem. Results of the full study are due next year.

The big issue isn't one of sensor technology but one of how to make sure the sensors interact with the players' bodies (correct positioning, skin movement, etc.). Also big errors (> 2.5 degrees) occur with inertial sensors and these compounded by natural anatomical variability. So if you're a bowler with a slightly longer forearm, there will be model bias and I don't think it's outrageous to say that many of us would enjoy being hung on evidence that error-prone. As identified by Elliott and co, there are also still problems with the definition, at the anatomical level, of a bowl vs a throw which are yet to be fully resolved. One example, an elbow doesn't just flex through 2 dimensions, there's some through the perpendicular axis with everyone so merely calculating 2D distances and angles between humerus and forearm isn't and should never be accepted as a valid approach. And, even if that problem is solved, natural variation is a huge influence on the final angle. As Murali himself has shown, he has a very unusual amount of flexibility in his shoulder at the point of delivery which most other people don't. This affects how you properly define what is a legal vs illegal delivery and the conclusions you reach for him as an individual will not generalise at all. And then there's how you deal with hyper-extension; where do you take measurements from in the case of someone with a highly flexible elbow like, say, Shoaib? These are only a few of the problems that come with this sort of caper.

The throwing law is an attempt to legislate in a general sense but it's often turning out to be naturally adversarial to those who are edge cases anatomically. The above issues (amongst many) make it a tough problem to solve if you want to keep these guys in the game. You might take the position that being physically unable to bowl correctly should disqualify you from playing and that's fine but then you have to weigh up whether the health of the game would have been better served to knock guys like Murali out early vs letting him entertain the **** out of us for 20 years and 800 Test wickets.

 

[indiaholic]

Almost certainly isn't accurate, converting from video to 3D is a hard problem. Lab testing is still king in terms of distances and measurements.

Elliott, et al, 2007, Sports Biomechanics, January 2007; 6(1): 99–108

That research is a bit old but more recent work (2012) hasn't solved the problem. Results of the full study are due next year.

The big issue isn't one of sensor technology but one of how to make sure the sensors interact with the players' bodies (correct positioning, skin movement, etc.). Also big errors (> 2.5 degrees) occur with inertial sensors and these compounded by natural anatomical variability. So if you're a bowler with a slightly longer forearm, there will be model bias and I don't think it's outrageous to say that many of us would enjoy being hung on evidence that error-prone. As identified by Elliott and co, there are also still problems with the definition, at the anatomical level, of a bowl vs a throw which are yet to be fully resolved. One example, an elbow doesn't just flex through 2 dimensions, there's some through the perpendicular axis with everyone so merely calculating 2D distances and angles between humerus and forearm isn't and should never be accepted as a valid approach. And, even if that problem is solved, natural variation is a huge influence on the final angle. As Murali himself has shown, he has a very unusual amount of flexibility in his shoulder at the point of delivery which most other people don't. This affects how you properly define what is a legal vs illegal delivery and the conclusions you reach for him as an individual will not generalise at all. And then there's how you deal with hyper-extension; where do you take measurements from in the case of someone with a highly flexible elbow like, say, Shoaib? These are only a few of the problems that come with this sort of caper.

The throwing law is an attempt to legislate in a general sense but it's often turning out to be naturally adversarial to those who are edge cases anatomically. The above issues (amongst many) make it a tough problem to solve if you want to keep these guys in the game. You might take the position that being physically unable to bowl correctly should disqualify you from playing and that's fine but then you have to weigh up whether the health of the game would have been better served to knock guys like Murali out early vs letting him entertain the **** out of us for 20 years and 800 Test wickets.

Marvelous post. Thank you. Had never thought about flex along the perpendicular axis.

 

[Top_Cat]

On that, from Elliott's paper.

Joint centres digitized from at least two camera views allowed the creation of a three-dimensional joint angle formed between two vectors. For instance, the line drawn from the wrist joint centre to the elbow joint centre, representing the forearm, formed the first vector, while the upper arm was similarly described as the vector from the elbow to the shoulder joint centre (Figure 1a). The resulting three-dimensional joint vector angle represents a combination of elbow flexion–extension, abduction–adduction, and pronation–supination.

Initial efforts to separate out the component parts of joint motion saw vector-based angles projected onto a two-dimensional plane (Figure 1b). However, the accuracy of such an approach may be compromised by “planar crosstalk”, whereby movement in one plane is incorrectly measured and interpreted as movement in another. For instance, upper arm (shoulder) internal–external rotation may dramatically affect the elbow flexion–extension angle, which has been inferred from a vector angle projected onto the sagittal plane. Consequently, the accuracy of this projected approach relies heavily on motion that remains predominantly planar during execution......

.....which, as found, elbow rotation most certainly is not. And, again, they don't even go into cases where the joints hyper-extend. Hard enough to fully understand the 'normal' case.

 

[indiaholic]

Thanks for the diagrams.. Makes a whole lot of sense and if I am interpreting this correctly, this will make actions like Murali's very difficult to judge considering the exaggerated elbow rotation due to the bent arm?

 

[indiaholic]

Plus the supination of the arm during the doosra would probably lead to even more elbow rotation and thus make it even more difficult to map it out in a 2d plane?

 

[Top_Cat]

Yeah, definiing his 'start point', since he has a bent elbow that bends not just in the transverse but more than most people in the saggital plane too, is hard enough. But as the papers also discuss, there is movement in the wrist and shoulder that influence where you draw vectors from in order to calculate the 3D angle. Get those a little bit wrong and you get the angle wildly wrong, depending on the method used.

This is exactly why the ICC taking the research on throwing out of the public domain is so problematic. Elliott has complained about it publically and how they're (mis)using his work. Definitions matter and they aren't sharing theirs, just handing out bans from up on high. Their research may be highly problematic or maybe they've made some breakthroughs, who knows. Either way, they're eminently challengeable and I'm guessing that, upon a key bowler from one of the big countries being cited and/or banned, this is exactly what will happen and they'll win.

 

[Top_Cat]

I should point out that this really isn't my area so if anyone has any other biomech or phys perspectives to add, would be real keen to know more.

 

[indiaholic]

Agreed that it is a huge problem. Any scientific study that is not accessible to people so that they can replicate it themselves is not very scientific at all.

This leads me to one of my pet peeves about so much research by public universities being behind paywalls but that is probably a topic for a different forum.

 

[TheJediBrah]

The only issue I have with the recent bans is that the ICC has let these bowlers do it their whole careers then all of a sudden cutting them off, it kind of leaves them in a pile of ****. It's hardly feasible for a Saeed Ajmal to go back and completely change his action. You're basically banning them from the game entirely in some cases.

And personally I like seeing off-spinners bowl the doosra, and these laws/testing is effectively making it an illegal delivery.

 

[smash84]

The only issue I have with the recent bans is that the ICC has let these bowlers do it their whole careers then all of a sudden cutting them off, it kind of leaves them in a pile of ****. It's hardly feasible for a Saeed Ajmal to go back and completely change his action. You're basically banning them from the game entirely in some cases.

And personally I like seeing off-spinners bowl the doosra, and these laws/testing is effectively making it an illegal delivery.

Ajmal wasn;t always a chucker though. His action markedly deteriorated.


This is his footage from the 2010 tour of England

https://www.youtube.com/watch?v=oIfQtE5dF4Q

and this is his footage from 2012

https://www.youtube.com/watch?v=MGHyYvgLu_A

 

[Migara]

Marvelous post. Thank you. Had never thought about flex along the perpendicular axis.

That's why I said I don't like the term hyper extension. Adduction and abduction are the correct terms.

 

[honestbharani]

I agree with Migara and others here who have said this: all bowlers do need to be tested and then, AND ONLY THEN, can we actually understand what the exact parameters can and should be. For instance, I find that it is almost impossible to actually bowl a cricket ball as we do in cricket without flexing the elbow to some extent (straightening a bent arm or bending a straight arm). So given that it is a question of degrees allowed, a better way is required to set these limits.


Just on this topic, TC, I believe the issue with the 2004 tests were not the accuracy. If I am not wrong, they did the 2D analysis of guys' who had already been tested live and figured that the deviation was not so bad and then they tested every bowler in the world assuming the same margin of error. So while the actual degree of flexion might vary upon real testing for every bowler, I dont think a normalized curve of those flexion numbers would be different from what they obtained using the 2D analysis. Is my understanding of this correct?

 

[honestbharani]

Almost certainly isn't accurate, converting from video to 3D is a hard problem. Lab testing is still king in terms of distances and measurements.

Elliott, et al, 2007, Sports Biomechanics, January 2007; 6(1): 99–108

That research is a bit old but more recent work (2012) hasn't solved the problem. Results of the full study are due next year.

The big issue isn't one of sensor technology but one of how to make sure the sensors interact with the players' bodies (correct positioning, skin movement, etc.). Also big errors (> 2.5 degrees) occur with inertial sensors and these compounded by natural anatomical variability. So if you're a bowler with a slightly longer forearm, there will be model bias and I don't think it's outrageous to say that many of us would enjoy being hung on evidence that error-prone. As identified by Elliott and co, there are also still problems with the definition, at the anatomical level, of a bowl vs a throw which are yet to be fully resolved. One example, an elbow doesn't just flex through 2 dimensions, there's some through the perpendicular axis with everyone so merely calculating 2D distances and angles between humerus and forearm isn't and should never be accepted as a valid approach. And, even if that problem is solved, natural variation is a huge influence on the final angle. As Murali himself has shown, he has a very unusual amount of flexibility in his shoulder at the point of delivery which most other people don't. This affects how you properly define what is a legal vs illegal delivery and the conclusions you reach for him as an individual will not generalise at all. And then there's how you deal with hyper-extension; where do you take measurements from in the case of someone with a highly flexible elbow like, say, Shoaib? These are only a few of the problems that come with this sort of caper.

The throwing law is an attempt to legislate in a general sense but it's often turning out to be naturally adversarial to those who are edge cases anatomically. The above issues (amongst many) make it a tough problem to solve if you want to keep these guys in the game. You might take the position that being physically unable to bowl correctly should disqualify you from playing and that's fine but then you have to weigh up whether the health of the game would have been better served to knock guys like Murali out early vs letting him entertain the **** out of us for 20 years and 800 Test wickets.

I personally think it would be stupid if we start disqualifying people from the game because of superior physical or anatomical abilities... I mean, come on, it is like punishing a batsman for being able to judge the length of a ball earlier than most others or a batsman having superior reflexes.

 

[Son Of Coco]

Yeah the process was a joke. The alternative of using our eyes and a protractor that jedibrah our resident scientist has suggested is much more scientific in it's approach.

The part of the whole chucking drama that I found intriguing was the "99% of bowlers throw" line that the ICC came out with, I've always wondered what interpretation of the rule they used to come up with this, what they measured as 'flex', and at what point in the delivery did they measure it. I assume the report is available to read somewhere, but my limited googled skills haven't revealed anything yet.

 

[Son Of Coco]

And FWIW, it is just about impossible to "bowl" a cricket ball at the speeds they do in FC or international cricket with the revs they put on the ball and the flight and the dip (for spinners) as well as the swing and drift without flexing the elbow (either straightening the bent arm or bending the straight arm).

I'd argue that neither of these would be classified as chucking anyway if you're talking about hyperextension due to the forces exerted on the arm during delivery, and then the arm returning to its normal position after the ball has been released. I'd certainly hope that wasn't what was measured in the ICC's wide-ranging video study. Chucking is not this.

 

[the big bambino]

"Chucking is not this". Agree with that. One of the worst aspects of this whole affair has been the impugning of bowlers like McGrath and Pollock (and now Swann and Warne in this thread) whose actions have never been questioned. It is a poor justification for the advent of the changed law, which can probably claim good reasons necessitating it without reference to other bowling actions.

These threads will always be buzzing with anger from people with entrenched opinions. I'm not saying the opinions (for and against the changed law) are unjustified but the anger expressing them can be a real turn off. There seems to be a consideration now of just how hard this issue has been to define, regulate and enforce. Apart from the difficulties of measuring 3D events in 2D parameters (hopefully not a misunderstanding on my part of the exchange btwn TC and indiaholic) it seems throwing and bowling are related in as much both share the trait of flexion. Bcos of that it becomes difficult defining one from the other if you measure that commonality. Perhaps a rethink on some other aspect that differentiates them is required.

 

[honestbharani]

I'd argue that neither of these would be classified as chucking anyway if you're talking about hyperextension due to the forces exerted on the arm during delivery, and then the arm returning to its normal position after the ball has been released. I'd certainly hope that wasn't what was measured in the ICC's wide-ranging video study. Chucking is not this.

This is what has been considered to be chucking since the old days though. They just used to report and ban anyone who "looked" different to what they were used to seeing. You are right. Chucking is not this but neither the ICC nor the scientists involved have yet been able to define exactly what constitutes "deliberate" chucking. To me, that is the whole issue. There is one thing about a "legal" bowling action but the biggest thing ICC needs to figure out is how to stop a bowler gaining an unfair advantage intentionally. For all we know, some bowler with a squeaky clean action could still be changing his elbow or arm flexion intentionally to generate better speed/accuracy/turn right now in international cricket. To me, this is the bigger question cricket needs to address. Obviously, the definition of a "legal" bowling action would follow from this. And then there is the question of how to handle exceptional cases where people have congenital irregularities in their anatomy that only allows them to bowl in a certain way. It is a huge huge headache simply because of how vague and impractical the original idea of a "proper" bowling action was.

 

[Top_Cat]

The part of the whole chucking drama that I found intriguing was the "99% of bowlers throw" line that the ICC came out with, I've always wondered what interpretation of the rule they used to come up with this, what they measured as 'flex', and at what point in the delivery did they measure it. I assume the report is available to read somewhere, but my limited googled skills haven't revealed anything yet.

Fairly sure the report has never been made public, tbh. They didn't need to measure anything anyway, the measurements presumably informed the development of thresholds. Before that, though, the old throwing rule said any partial or complete straightening prior to letting go of the ball was a throw:

Law 24: No Ball

(a) Definition of a Throw
A ball shall be deemed to have been thrown if, in the opinion of either Umpire, the process of straightening the bowling arm, whether it be partial or complete, takes place during that part of the delivery swing which directly precedes the ball leaving the hand. This definition shall not debar a Bowler from the use of the wrist in the delivery swing.

Testing of 130 players confirmed that which should be obvious; if you move a lever with three flexible points fast through 3-dimensional space, there will be some flexing and bending at those points so literally everyone threw to a degree. Labelling everyone a filthy chukka, of course, is silly so there needed to be thesholds invented which reflected biomechnical reality. Testing initially supported the 5/7.5/10 degrees for spin/meds/quicks after using the 'visible throw' guideline for where the cut-offs should be but, as Elliott himself said, the threshold needed to be extended.

Bruce Elliott, one of three biomechanics who conducted the research, denied the issue was specifically about Muralitharan. Elliott was happy with the recommendation.

"I'm happy with that [15 degrees]. It's a step in the right direction," Elliott said. "They've seen that a lot of bowlers today sit about 10 degrees ... and it's been shown in studies that's there's no real benefit until you get beyond 15 degrees.

"People assumed everyone was zero and if you had asked me two years ago I'd would have guessed that five degrees was a reasonable number of most bowlers but research has shown otherwise.

I happen to agree with Elliott about there probably being no serious benefit before 15 degrees but I guess it is arguable, unless there's unpublished data he's provided to support that claim.

 

[Furball]

Ajmal wasn;t always a chucker though. His action markedly deteriorated.


This is his footage from the 2010 tour of England

https://www.youtube.com/watch?v=oIfQtE5dF4Q

and this is his footage from 2012

https://www.youtube.com/watch?v=MGHyYvgLu_A

His action in 2010 was diabolical so I'd hate to see what it had "deteriorated" to.