Article Text

Download PDFPDF

Why it is difficult to detect an illegally bowled cricket delivery with either the naked eye or usual two-dimensional video analysis
  1. K D Aginsky1,
  2. T D Noakes2
  1. 1Centre for Exercise Science and Sports Medicine, University of the Witwatersrand, Johannesburg, South Africa
  2. 2International Cricket Council (ICC) Accredited Testing Centre for Suspected Illegal Bowling Actions in Cricket, MRC/UCT Research Unit for Exercise Science and Sports Medicine, Department of Human Biology, Faculty of Health Sciences, The University of Cape Town and the Sports Science Institute of South Africa, Cape Town, South Africa
  1. Correspondence to Dr Kerith D Aginsky, PO Box 29332, Sandringham 2131, South Africa; kerith.aginsky{at}


Objective In this study, the influence of various anatomical and biomechanical factors on the ability to detect the magnitude of elbow extension during the cricket bowling action, with either the naked eye or from conventional two-dimensional video footage was evaluated.

Design Descriptive study

Setting One male bowler who had a large ℌcarrying angleℍ was studied by means of a three-dimensional biomechanical analysis using the Vicon MX13 motion analysis system when bowling at match pace. Photographs were taken from various angles during the bowling delivery action to illustrate factors influencing the visual appearance of an illegal delivery.

Main outcome measures The influence of extension of the elbow joint and the elbow's anatomically fixed position, known as the “carrying angle”, on the visual perception of an illegal (“thrown”) delivery were the main outcome measures.

Results The elbow joint rotates about 90° as a result of humeral rotation during the movement of shoulder circumduction produced by the bowler's delivery action. This causes the plane in which the elbow joint moves to change throughout the delivery action. This movement will also vary between bowlers. The movement of elbow planar flexion/extension can be viewed only when the viewer's eyes are at exactly 90u to the plane of elbow joint movement. Thus, a cricket umpire wishing to detect elbow extension during the bowler's delivery action would need continuously to change his viewing position as the bowler's arm moved from the horizontal to the vertical position at ball release. We show that by viewing the action from only one position, the umpire will be confused by the illusion of a “throw” produced by a large “carrying angle” in some bowlers.

Conclusion It is impossible for an umpire conclusively to cite a bowler for an illegal action based only on naked eye observation. To do this the umpire would need to be in at least three different positions throughout a single delivery. Similarly, video footage shot from at least three different angles during the bowler's delivery action would need to be viewed by the match referee. By viewing the delivery from only one position, the visual illusion of a “throw” will be created by the presence of a large “carrying angle”.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

On 26 December 1995 at the Melbourne Cricket Ground in Australia, Australian umpire Darrell Hair “no balled” Sri Lankan spin bowler Muttiah Muralitharan seven times on the grounds that his bowling action was illegal. During Sri Lanka's following tour to Australia in 1999, umpire Ross Emerson “no balled” Muralitharan from the bowler's end in a One-Day International.1 The International Cricket Council (ICC) subsequently ordered Muralitharan to undergo a three-dimensional biomechanical analysis at the University of Western Australia. Studies concluded that his bowling action was legitimate2 but that his action created the “optical illusion of throwing”3 as “from certain angles he does look suspect but from other angles there was not a problem“.1

This analysis was subject to original ICC regulations stating a bowler may not extend their elbow from the moment the arm is horizontal until ball release4 by more than 5° for spin bowlers, 7.5° for medium-paced bowlers (65–80 mph) and 10° for fast bowlers (>80 mph).5 These ICC regulations are in accordance with law 24.3, which defines what constitutes a legal bowling delivery.6 In March 2005, the ICC revised these regulations allowing a 15° elbow extension threshold for all bowling types.4 Since 1995, 36 out of the 39 times that bowlers have been cited was on the basis of video footage.

The ICC has established a panel of experts to oversee the issue of illegal bowling actions and accredited four laboratories to perform biomechanical testing once the bowlers have been reported either by the on-field umpire or the match referee viewing video footage during a cricket match.4

At one such accredited laboratory, several bowlers with suspect actions have been evaluated over the past 3 years. During these evaluations, the researchers noted that the majority of bowlers tested had a large “carrying angle” (fig 1C), above 15°, as does Muralitharan (18°).2 In evaluating the possible contribution of this anatomical defect to the visual illusion of an illegal delivery, the researchers noticed that between arm horizontal and ball release in the bowler's delivery action, the humerus rotates either externally or internally through about 90° depending on the nature of the delivery. The presence of this rotation has been mentioned by Portus et al (2003), where a group of fast bowlers all presented with some degree of humeral rotation in the longitudinal axis.7 As a result of this experience, it was realised that even if the naked eye was able to distinguish between elbow extensions of less or more than 15° during the bowling action, the complexity of the action prevents the optical detection of an illegal delivery during play. For the umpire to achieve this detection, he would need to be in at least three different locations during the last hundred or so milliseconds of the bowler's delivery action. Similarly, detection by camera would require at least three high-speed cameras positioned in different areas. Therefore, the aim of this study was to assess whether a large “carrying angle” created the illusion that a bowler appeared to have an illegal bowling action and the subsequent complexity in determining this using the naked eye or two-dimensional video footage.

Figure 1

Note that A, C and E were taken from three different positions while the subject stood in anatomical position. Similarly, B, D and F were taken with the elbow flexed to 65°. (A) “Carrying angle” (17°) viewed facing 90° to the plane in which the elbow joint flexes and extends. (B) Elbow flexion (65°) viewed from the same position in (A). (C) “Carrying angle” of 17° viewed with the elbow crease facing the camera. (D) Elbow flexion of 65° and “carrying angle” of 17° viewed with the elbow crease facing the camera. (E) “Carrying angle” viewed from behind with the olecranon process facing the viewer. (F) Elbow flexion of 65° and the “carrying angle” of 17° viewed from behind with the olecranon process facing the viewer.


A male, off-spin bowler aged 24 playing cricket at international level was assessed. The subject previously bowled with an illegal action; however, following remediation of his bowling action, it still appeared to the naked eye and using video footage that his action was illegal.

An anthropometric assessment of the subjects' bowling arm as per ICC guidelines was performed.4 The “carrying angle” (fig 1C) was measured with a goniometer at the centre of the cubital crease midway between the humeral epicondyles with the arms in the static anatomical position.8,,11 The ICC defined the “carrying angle” as “the angle between the longitudinal axis of the upper arm and forearm, in the frontal plane, with the average range of carry angle for the human population being between 5°–15°.”1 The average “carrying angle” for men has previously been reported to range between 6.8° and 12.5°.8

The three-dimensional analysis of the subjects' bowling arm was assessed throughout the subjects' bowling action using an eight-camera opto-reflective Vicon MX13 motion analysis system operating at 200 frames per second. Reflective markers were placed on specific anatomical sites in accordance with the units' cricket bowling model to calculate elbow angles. The cameras were calibrated within the delivery space provided before data collection, with an accuracy within 2°. Only elbow angle through the position from back foot contact (including arm horizontal, maximum flexion) to ball release and 10 frames after ball release, as per ICC guidelines, were reported.4 The bowler was instructed to bowl 10 off-break deliveries at full match pace. Ball speed was measured using a radar gun and was recorded as 77.1 (1.8) kph. To further investigate and explain the visual phenomena, photographs filmed at 200 Hz were taken from front-on, side-on and behind during his bowling action.

Ethics approval for this study was obtained from the University of Cape Town Human Research Ethics Committee.

Results and discussion

Part 1: Positioning of the elbow joint during the delivery action

The elbow joint being a hinge joint theoretically only moves through flexion/extension. Elbow flexion (fig 1A,B) is the forward and upward movement of the forearm from the anatomical position, with elbow extension the return movement. To detect elbow flexion/extension, the viewer must be positioned at exactly 90° to where the elbow joint's crease anteriorly or olecranon process posteriorly is pointing (fig 1A,B). If the viewer can see either the elbow crease or the olecranon process (fig 1C-F), they will be unable to detect the extent of elbow flexion/extension.

The elbow is also in a fixed state of abduction, which is the movement of the arm away from the midline of the body when standing in a static anatomical position. This elbow abduction is referred to as the “carrying angle” (see part 2) and is best viewed when either the elbow crease (fig 1C,D) or olecranon process is pointing directly at the viewer (fig 1E,F), thus the viewer is not in a position to view elbow flexion/extension. Lloyd et al (2000), defined the abduction/adduction axis as being perpendicular to the flexion/extension axis.2 Importantly, both of these movements, flexion/extension and abduction/adduction, occur in different planes and can be assessed only when viewed from a position exactly 90° to the plane in which the specific movement occurs.

The practical implication is that a stationary viewer unaware of this phenomenon might assume that they were viewing a single movement in one plane, for example, elbow flexion/extension, when, in fact, a more complex combined movement of elbow flexion/extension and abduction/adduction was taking place. This occurs during the bowling action as a result of circumduction of the shoulder as the arm moves from arm horizontal with the elbow pointing directly downwards (fig 2A) to arm vertical at which moment the elbow points towards mid-on (fig 3). As a result, the elbow must rotate through 90° during the last hundred milliseconds of the delivery action, which importantly is likely different for seam, finger spin and wrist spin bowlers and depends on one's specific bowling action. Thus, the optimum position to view elbow flexion/extension during the delivery action will possibly vary for each bowler.

Figure 2

(A) Arm horizontal with the elbow facing downwards with camera positioned at 90° to the elbow joint. Note that the marker on the elbow is placed on the medial epicondyle. (B) The front-on view and (C) the view from behind the bowler.

Figure 3

Viewing positions on the field to detect elbow flexion/extension during the delivery action of the specific bowler evaluated in this study. (A) Directions from which elbow flexion/extension can be detected when the arm is horizontal during the delivery action. (B) Directions from which elbow flexion/extension can be detected when the arm is vertical before ball release. (C) Directions from which elbow flexion/extension can be detected at the moment of ball release. RHB, right hand batsman, WK, wicketkeeper.

In the case of the bowler tested, when his elbow faced downwards at arm horizontal, the only position from which the extent of elbow flexion/extension could be viewed was from side-on with the medial elbow epicondyle pointing towards the side on camera (fig 2A; A in fig 3). If one viewed the bowler from either in front (fig 2B) or behind (fig 2C; C in fig 3), the extent of elbow flexion/extension was not visible as the movement occurred in the same plane as the direction of the camera image.

When the bowler's shoulder rotated to arm overhead, the elbow rotated to face diagonally away from the bowler with the olecranon process pointing in the direction of mid-on (fig 3; fig 4A). To view elbow flexion/extension at this point, thus being 90° to the elbow joint, the viewer must be positioned at either mid-off (B in fig 3) or at square-leg umpire (B in fig 3). Therefore, no assessment of the extent of elbow flexion/extension could be made from any of the views in fig 4. However, the square-leg umpire would be in a favourable position to detect the extent of elbow flexion/extension (B in fig 3) at this point in the delivery action.

Figure 4

Arm vertical with the elbow facing diagonally away from the body towards mid-on for a right-hand batsman. (A) Side-on view with the marker positioned on the medial epicondyle with the olecranon process pointing towards mid-on; (B) Front-on view with the anterior elbow crease pointing towards point; (C) View from behind with the olecranon process pointing to mid-on and the lateral epicondyle pointing towards mid-off; (D) Elbow at ball release side-on view; (E) Front-on view and (F) view from behind with lateral epicondyle (marked) pointing towards the camera.

As this bowler's humerus rotated internally immediately before ball release, the olecranon process came to a point parallel to the popping crease (fig 4D). Thus, elbow flexion/extension could be viewed either directly in front (fig 4E) or directly behind (fig 4F) the bowler (C in fig 3).

For the umpire to be in a position to detect the extent of elbow extension as the arm travels from horizontal to the instant of ball release, he must be positioned square to the bowler when the bowler's arm is horizontal (fig 2A; A in fig 3), at the square-leg umpire's position (B in fig 3) when the bowler's arm is overhead (fig 5B) and directly in front (fig 4E) or behind (fig 4F; C in fig 3) the bowler a few milliseconds later at the instant of ball release. Therefore, the square-leg umpire, who is required to cite a bowler for “throwing”, is only in the optimum position for an instant, as the medial epicondyle of the elbow pointed directly at the umpire, when the arm of this bowler was directly overhead. This would not provide him with all the information required to detect an illegal delivery.

Figure 5

View from umpire's position at square-leg (A) arm at horizontal, (B) arm vertical with the medial epicondyle pointing towards the square-leg umpire and (C) arm at ball release.

Importantly, the umpire would need to be positioned in three places simultaneously (fig 3) to view the full range through which the bowler's elbow may have extended during a single delivery action from arm horizontal to ball release, and thus to be in the correct positions to cite a bowler. Similarly, to properly assess the legality of a delivery using two-dimensional video footage, the umpires or match referee would have to view images of each delivery taken at high speed from cameras positioned in the same arc (fig 3) relative to the bowler.

Part 2: Influence of the “carrying angle” on the visual perception of “throwing”: a visual illusion

The second problem facing umpires and match referees is produced by the “carrying angle” or cubitus valgus. It is created by the slight valgus angulation of the trochlear notch of the ulna with respect to the shaft, coupled with the valgus angulation of the distal humerus.11 The “carrying angle” is a natural feature of the elbow joint and when in static anatomical position, it cannot be altered voluntarily. The “carrying angle” allows the arms freely to clear the side of the body when the arms are swinging during walking, or when carrying objects. This “carrying angle” is reported to remain constant after the age of 15 years.9 10

The “carrying angle” is best viewed when the elbow crease is viewed directly from the front (fig 1C) or directly facing the olecranon process from behind (fig 1E). The “carrying angle” cannot be detected from the side-on position (fig 1A). One must, however, take into account the laxity of the bowler's elbowas this will have some influence on the movement while bowling.

As described, the shoulder undergoes circumduction during the delivery action producing external humeral rotation at arm horizontal causing the olecranon process to face downwards and the elbow crease upwards (fig 2A), resulting in the “carrying angle” being obscured in the lateral view and only being viewed from either directly above or below. As the arm moved overhead (fig 4A), the humerus rotated internally so that the olecranon process pointed towards mid-on (fig 3) and the elbow crease pointed towards point (fig 3), allowing the “carrying angle” to be viewed from either position. At ball release due to continued humeral internal rotation, the elbow crease and the olecranon process faced side-on, thus the “carrying angle” was best viewed when positioned side-on to the bowler (fig 4D).

Figure 6A shows that when viewing the arm overhead from either in front of or behind the bowler, one is viewing a combination of elbow flexion and the “carrying angle” because as detailed in fig 4, the elbow joint pointed diagonally towards mid-on (fig 3). However, at ball release (fig 6B) the olecranon process pointed parallel to the popping crease (fig 4E,F) so that exclusively at that point, elbow flexion/extension was viewed from in front of or behind as is the video footage provided. As a result, a bowler with a large “carrying angle” will appear to “straighten” his elbow when viewed from either in front of or behind as his arm moves from arm vertical (fig 6A) to ball release (fig 6B). The important point is that this illusion will occur whether or not the elbow actually extends during this part of the delivery action. An umpire or match referee, unaware of this illusion caused by the bowler's large “carrying angle” and humeral rotation during the bowling action, will be in no doubt from this position that the bowler shown in fig 6 has an illegal action. However, an elbow extension angle of 11.1° was recorded from arm horizontal to ball release for this bowler and thus a legal action according to the ICC regulations.

Figure 6

(A) Arm overhead, (B) at ball release, (C) view from fine leg of Muralitharan's delivery action from arm vertical to ball release (“carrying angle” is 18°) (fig 6C, copyright of Touchline photo).

As previously argued, the square-leg umpire can only view the extent of elbow flexion/extension when the arm of the bowler that we studied was overhead (fig 5B). At arm horizontal (fig 5A) and ball release (fig 5C), the umpire is viewing a combination of elbow flexion and the “carrying angle”. Thus, the umpire's view changes from a combination of elbow flexion/extension and the “carrying angle”, to only elbow flexion/extension at arm vertical (fig 5B) and then back to a combination at ball release. It is also noted that the square-leg umpire is positioned far away from the bowler so that the complexity of the movements he must interpret and the speed at which they occur make it impossible for the umpire to conclusively determine if a bowler has an illegal bowling action.

Figure 6C shows still photographs of Muralitharan taken from fine leg during two One-Day Internationals (left: 17 January 2006 and right: 24 January 2006). Muralitharan's voluntary threedimensional biomechanical analysis was performed beginning February 2006. Since his action was deemed legal, one must assume that these images are of a legal bowling action.

Muralitharan's shoulder rotates the same as the bowler in our study, and he has a “carrying angle” of 18°, 2 thus the findings in fig 6 of our bowler would hold true for Muralitharan's bowling in fig 6C. This is supported by Lloyd et al, who previously reported that Muralitharan's abnormal fixed flexed and abducted elbow position gives the impression of a “throw” mainly when viewing various arm position changes in the delivery.2

Thus, the researchers argue that the greater the bowler's anatomically fixed “carrying angle”, the greater is the illusion that the elbow is extending as the arm moves from arm horizontal to ball release. When viewed from the unsuitable positions of either umpire or from two-dimensional video footage, it is impossible to conclusively determine whether the bowler has an illegal action.

The researchers acknowledge that the laboratory environment may not mirror the match environment, as the surface and surroundings are different. This study only reports a single bowler tested; however, taking into account the report of Portus et al (2003) that all the fast bowlers they tested rotated their humerus during the analysis phase of the delivery action,7 one can assume that the finding in our bowler would, in fact, be found in other bowlers.


The upper arm goes through a complex three-dimensional movement during the bowling action, but the ICC ruling focuses exclusively on the legality of movement in a single plane, elbow flexion/extension. This movement can only be observed by a viewer who is positioned at 90° to the elbow joint with the bowler's medial or lateral epicondyle pointing directly towards them. For the bowler evaluated in this study, the arcs to which his epicondyles point during his bowling action are shown in fig 3; therefore, to detect an illegal delivery in this bowler, an on-field umpire and video footage would need to view the delivery from at least three positions in that arc during the delivery action.

The researchers have also shown that bowlers with a large “carrying angle” will give the impression that their elbow has straightened during the final phase of the bowling action when viewed from either the umpire's or video footage positions.

Since no on-field umpire can be in three places at once nor can an umpire distinguish with certainty between elbow extension and an illusory straightening in a bowler with a large “carrying angle”, it is impossible for him to be certain whether a bowler has an illegal action. Nor is it possible for a match referee to distinguish between these possibilities on the basis of twodimensional video footage shot from unsuitable positions.

In view of the uncertainties highlighted in this study, it is proposed that caution be taken when citing and not to have this information released to the media until a detailed biomechanical analysis has shown that his delivery action is illegal according to the current ICC rulings. It is also recommended that umpires and match referees should be further educated in the anatomy and biomechanics of the bowling action, so they are in a position to make more informed decisions. It is also recommended that further research on different bowlers be performed in this area.


Mr Hilton Houghton from ReelSport and photographer Mr Carl Fourie are acknowledged for the permission to use the photographs of the bowler and Dr Michele van Rooyen and Mr Jacques Gay for providing technical assistance.Permission was also obtained from Touchline Photo to include images of Muttiah Muralitharan for publication.



  • Funding The research undertaken in the unit was funded by the Harry Crossley and Nellie Atkinson Staff Research Funds of the University of Cape Town, the Medical Research Council of South Africa, Discovery Health, Bromor Foods and the National Research Foundation of South Africa through the THRIP initiative.

  • Competing interests None.

  • Patient consent Obtained.