Bowling Report: Testing Muralitharan’s Doosra


By Staff Writers | July 12, 2010

Muttiah MuralitharanMuttiah Muralitharan’s ‘Doosra’ delivery was reported in 2004 by ICC match referee Chris Broad. As a result, Muralitharan underwent a round of testing at the University of Western Australia (UWA).

The following is the unedited bowling report released by the School of Human Movement & Exercise Science at the UWA.

At the time this report was released, the ICC playing regulations allowed bowlers to flex their elbows by 10 degrees for fast bowlers, 7.5 degrees for medium pacers, and 5 degrees for spin bowlers. Therefore, the report recommended remedial work on Muralitharan’s bowling action when bowling the Doosra in order to conform to the old regulations.

However, another ICC sanctioned study later revealed that all bowlers exceeded the limits allowed and that the human eye could only detect a flex above the 15 degree mark. As a result, the ICC changed the regulations to allow 15 degrees for all bowlers nullifying the need for any remedial work on Muralitharan’s bowling action.


Name: Muttiah Muralitharan
Biomechanics Team: Professor B.Elliott, Ms J.Alderson & Ms S. Reid
Session Date 1: April 1st, 2004
Technique Remediation

Session Date 2: April 7th, 2004


In response to a request from the Sri Lankan Cricket Board, directed through Mr Daryl Foster and the ICC (contact from Mr David Richardson) Mr Muttiah Muralitharan’s spin bowling action was assessed in the Biomechanics Laboratory of the School of Human Movement and Exercise Science at the University of Western Australia. This request followed the match referee (Mr Chris Broad) lodging a “suspect bowling action report” on his “doosra” delivery during the recent Australian tour of Sri Lanka. Testing was therefore restricted to analysis of his “doosra” delivery.

Mr Muttiah Muralitharan arrived in Perth on 31st March 2004, and his initial testing took place on Thursday the 1st of April. This testing comprised:

  • an anthropometric assessment of his bowling arm,
  • a three-dimensional (3D) analysis of his bowling arm during the complete bowling action, although elbow angles are only reported from a position where the upper arm is horizontal to the ground until ball release (the area covered by the laws of the game). This involved filming Mr Muralitharan using a 12-camera opto-reflective Vicon system operating at 250Hz (fields per second).

The testing occurred in a laboratory environment, which permitted a full bowling run up, such that a portion of the pitch was housed outside the laboratory. A laterally(side) placed video camera recorded his images during delivery to assist in the identification of ball release. A front-on video camera was used to assist in the selection of the best ”doosra deliveries”. Mr Bruce Yardley a former Australian spin bowler, who commented on the quality of each delivery, also assisted this task. He had been present in Sri Lanka during the recent Test series and could therefore comment on the quality of the laboratory compared with on-field bowling.

  • His best six deliveries were then analysed. Those selected satisfied the criteria of breaking the appropriate way off the pitch and being bowled with appropriate intensity by the spin bowling expert.

The results from this testing session were verbally provided to Mr Muttiah Muralitharan and Mr Daryl Foster. In an attempt to ensure that all efforts were made to comply with ICC Rules, Mr Muttiah Muralitharan, following consultation with the UWA biomechanics testing team then underwent a period of ‘technique remediation’ with Mr Daryl Foster, a renowned cricket coach. A final 3D analysis, following ICC guidelines was carried out on the 7th April. This report includes the results from both testing sessions. A preamble, prior to the presentation of these data is included, to assist with the interpretation of the results.


It is important when reading the following report that consideration is given to a number of issues. These include the accuracy of the measurement system used in bowling assessment (repeatability and validity of measures), range of acceptability of elbow angles in the critical region (from when the upper arm is parallel to the ground until ball release) and finally differences or similarities between fast and spin bowling actions.

Accuracy of measurement system

The opto-reflective 12-camera Vicon System that recorded at 250 pictures/sec has an error margin of approximately 1° in data collection. On-field recording systems, using a minimum of 3 high-speed video cameras for spin bowling, have accuracy levels of approximately 4°, although these error margins were recorded in a laboratory environment (Richards, 1999).

The identification of elbow and shoulder joint centres in on-field data collection, where a shirt is worn also involves large errors. In a match the ability to differentiate anatomical movements such as “elbow extension” by digitising segment end-points, particularly if you have segment rotations, is extremely difficult and prone to error. This is certainly the case with spin bowlers. It is therefore not surprising that laboratory testing is preferred, particularly for spin bowlers, where an appropriate pitch length and run up can be structured. This is clearly the only way to test players, where data would be able to withstand scientific and therefore legal scrutiny.

Range of acceptability of elbow angles

The International Cricket Council (ICC) guidelines have been structured around fast bowling, so ranges of acceptability (10° – fast bowling; 5° – spin bowling) may in fact need to be modified for spin bowling. Portus et al. (2003), the only published work in the area of changes of elbow angle during fast bowling, suggested the ICC range of acceptability should be increased to 15° if a large number of current fast bowlers are not to be subject to scrutiny and then remediation (none have been called for “throwing”). The logic in reducing the margin for fast bowlers compared with spin bowlers is based on the lower speed delivery of this classification of bowler. However, while run-up speed and length of arm are generally higher for fast bowlers, spinners such as Muttiah Muralitharan actually have a similar rotational speed of the arm system. Mr Muralitharan recorded a similar time (»0.08s), from arm horizontal to release”, to that recorded by Shabbir Ahmed Khan the Pakistan fast bowler recently tested by this team. Therefore a case can certainly be made for some spin bowlers such as Mr Muralitharan to have the same range of acceptability in elbow angle to that of fast bowlers.


  Mr Muttiah Muralitharan Normal
Wrist flexion-extension  78° flexion, 50° extension Not applicable
Wrist abduction- adduction 26° abduction, 26° abduction Not applicable
Forearm abduction angle (“carry angle”) 18°
Elbow flexion – extension * Static 35° (flex) Dynamic 24° (flex) 0° (full extension)
Shoulder internal rotation 68° 40°
Shoulder external rotation 102° 90°

* The dynamic value is the smallest flexion angle recorded while bowling (i.e. under load).

The anthropometry assessment clearly shows that Mr Muralitharan has a natural 35° of elbow flexion during standing, which during the delivery action (under load) reduces to a value of approximately 24°. Therefore any biomechanical assessment of his bowling action must take this 24° angle into account. In practical terms this means that his elbow joint, depending on the load, will always display at least some flexion. His elbow abduction angle is also such that it displays a relatively large “carry angle”.

Mr Muralitharan’s shoulder external rotation range is higher than normal, which allows him a greater range of motion during delivery. While this is an advantage in the development of speed, it also is a natural occurrence and does not therefore fall outside the bounds of human normality nor the rules of cricket. While this may be an advantage in bowling, it does not directly impact on the extension of the elbow.

However, the external rotation at the shoulder, combined with the 18° “carry angle” and 24° of permanent elbow flexion (see dynamic flexion above) will give the impression of “preparation for a throw”. This is particularly true when the action is viewed in two-dimensions (e.g. television, or when observed by an umpire from a fixed position).


Mr Muralitharan attended the biomechanics laboratory at the School of Human Movement & Exercise Science on April 1st 2004. The results from this initial testing session are presented below.

Session 1 Results

Following a warm-up, markers were attached to Mr Muralitharan as shown in Figure 1. The mean velocity of six deliveries selected for analysis was 64 km/hr. A mean elbow extension range of 14° was recorded for these six “doosra” deliveries (Table 1, Figures 2). The curves graphed in Figure 2 clearly show that each delivery was bowled with a similar action. One can then be confident that Mr Muralitharan bowls with a similar action in his “doosra” delivery. While one could argue that this extension is acceptable it is outside the current extension threshold of 5° set by the ICC. Hence a period of remediation followed aimed at reducing the level of elbow extension from upper arm horizontal to release.

Figure 1

Figure 1: Defining elbow flexion-extension axis.

  • To view a 3D reconstruction of Mr Muralitharan’s elbow angle from testing session one [and two] click here.

Table 1: Mean Changes in elbow angle from upper arm horizontal to ball release (6 deliveries)

Delivery Type Range of Extension (°) Speed (km/hr)
Doosra 14° (± 2°) extension 65 (± 3.0)

Figure 2: Elbow angle changes over the period from upper arm horizontal to ball release.

Figure 2


See attached Remediation Report (Mr Daryl Foster).


Following the period of remediation a second biomechanical analysis of Mr Muralitharan’s “doosra” delivery was conducted on the 7th April, 2004. The results from this testing session are presented below.

Session 2 Results

The mean extension for the elbow from upper arm horizontal to ball release was 10.2° (Table 2, Figure 3). Variations in the elbow extension curves (Figure 3) and the small standard deviation for the 6 deliveries (Table 2), show that each of these deliveries is very close to a 10° level. He therefore bowls with a consistent action.

His mean delivery speed of 72 km/hr, which is at the higher end of his “test match range” of 65-75 km/hr, shows that he was bowling with intensity in this laboratory environment. The spin bowling expert also testified to the fact that the deliveries analysed deviated in the appropriate manner with “venom”. He rotated his upper arm from the horizontal to release in a mean time of 0.072s, which is quicker than in Test 1 and also quicker than the time taken to rotate through the same angle by Shabbir Ahmed Khan.

  • To view a 3D reconstruction of Mr Muralitharan’s elbow angle from testing session [one and] two – click here.

Table 2: Changes in elbow angle from upper arm horizontal to ball release.

Delivery Type Range of Extension (°) Speed (km/hr) Match Range (km/hr)
Doosra 10.2° (± 0.6°) extension 72 (± 0.3) 65-75

Figure 3: Elbow angle changes over the period from upper arm horizontal to ball release following remediation.

Figure 3

Following remediation Mr Muralitharan bowled with an increased flexion angle. However, the remediation had the effect of reducing elbow extension range from 14° to 10° (see figures 2 & 3; 3-D skeletal reconstructions found within report).


In making recommendations regarding Mr Muralitharan the following should be stated. While a full run up and standard pitch were used, data were collected in a laboratory environment. It is our considered opinion that this is the only way to record accurate and reliable 3D data of elbow movement, particularly for spin bowling. The key to the issue with reference to a spin bowler, is the quality of the delivery and the rate of rotation of the upper arm. In our case Mr Muralitharan produced high quality deliveries with an upper arm action that was similar in rotational speed to that of a fast bowler. The mean time for his upper-arm to move from the horizontal to release in Testing session 2 (» 0.072 s) was compared with the same movement recorded on video from the recent Sri Lanka v Australia Test series. This video was provided by Mr Muralitharan. While the positioning of cameras for data from the Test series was not ideal and video images were recorded at a slower rate (50 fps), it was evident that the time for the upper-arm to move from the horizontal to release was similar for the Test series and the laboratory testing. Mean ball velocity at testing session 2 of 72 km/hr was also at the upper end of the range commonly reported for Mr Muralitharan under Test conditions. We therefore contend that the bowling action recorded was similar to that used in a Test match.

A case may be made for Mr Muralitharan’s initial elbow extension to be acceptable at 14°. Particularly, when one considers the speed of his arm rotation is similar to that of a fast bowler and the only scientific data related to fast bowling suggested an increase in the acceptable extension threshold from 10° to 15° (Portus et al., 2003). However, the mean extension across 6 deliveries was outside current ICC guidelines for fast bowlers.

For this reason a period of technique modification was carried out to reduce the level of elbow movement during the delivery of his “doosra”. Following this remediation his level of elbow extension reduced to 10°, which is within fast bowling guidelines. We contend that because the speed of his upper arm rotation is as fast and in some cases quicker than fast bowlers, his level of acceptability for elbow extension should also be set at the 10° mark. With no spin bowling data base to make a comparison, this would seem both a wise and prudent recommendation. Following the findings from Portus et al. (2003) we would also recommend that the ICC consider increasing the fast bowling extension threshold to 15°.

Finally it is our considered opinion that Mr Muralitharan be permitted to continue bowling his “doosra” at least until a valid data base is collected on the various spin bowling disciplines. The relatively minor level of elbow extension following remediation over the period from arm horizontal to release is not believed to give Mr Muralitharan an unfair advantage over batsmen or other bowlers.



Portus, M., Mason, B., Rath, D. & Rosemond, C. (2003). Fast bowling arm actions and the illegal delivery law in men’s high performance cricket matches. Science and Medicine in Cricket. R. Stretch, T. Noakes & C. Vaughan (Eds), Com Press, Port Elizabeth, South Africa: 41-54.

Richards, J. (1999). The measurement of human motion: A comparison of commercially available systems. Human Movement Science, 18: 589-602.


Leave a comment

Leave a Reply