Patterson, C.W. (2015)
EXTENT AND FUNCTIONALITY OF MOVEMENT PATTERN VARIABILITY IN CRICKET BOWLING
An Investigation into Between-Bowler and Within-Bowler Approaches
The prevailing theory in coaching was that movement variability has a detrimental effect on performance while presenting an injury risk to the performer, while believing that elite performers show little, if any, variation. With the improvements of data collection equipment and protocols, a contradictory theory that these variances can help maintain a consistent performance by offering performers the opportunity to vary their technique to adapt to the changing task demands presented to them. In addition, some studies into movement pattern variation have indicated that the use of grouped data, also known as between-participant approaches, could cause significant results for one individual which could be prominent in a large population to be masked by other members in the study’s sample. Using a convenience sample of 7 amateur cricket bowlers, the study aimed to investigate the functionality of technique variations in maintaining performance consistency. In addition, the study aims to compare results from the individual within-participant approaches to a variety of between-participant approaches. The results highlighted both positive and negative influences on performance by technical variability and how these results alter between participants in the within-bowler approaches and the four between-bowler methodological approaches.
Movement pattern variability has become a major consideration for biomechanists and skill acquisition scientists over the recent years thanks to the development of technology and new theories. The prevailing belief to movement pattern variability is one that any variation to the movement pattern is detrimental to the performance of the task (Newell & Corcos, 1993; Atkinson & Spiers, 1998; Slifkin & Newell, 1999; Davids et al., 2003; Charlton et al., 2006; Glazier et al., 2006; Bartlett et al., 2006) and as such was defined as “background noise” in early researchers (Fitts, 1954; Schmidt et al., 1979). However, as technology has advanced, more and more evidence has been unearthed that even the elite athlete’s performance varies in their movement patterns (Arutyunyan et al., 1968; Arutyunyan et al., 1969; Morriss et al., 1997; Newell & Vaillancourt, 2001; Bartlett et al., 2007; Bartlett & Robins, 2008) thus contradicting the ideas that elite athlete’s replicate their movement pattern identically with the replication marking their mastering of the skill (Arutyunyan et al., 1968; Manoel & Connolly, 1995; Schoellhorn, 1997; Schoellhorn & Bauer, 1998; Scholz et al., 2000; Mullineaux, 2008). As such a contradicting theory has become more evident in that these variations in movement pattern help maintain the athlete’s level of performance due to where the variations are within the movement pattern (Arutyunyan et al., 1968; Arutyunyan et al., 1969; Clarke, 1995; Glass & Mackay, 1998; Scholz et al., 2000; Miller, 2002; Thompson & Stewart, 2002) or the elite’s performers ability to make slight alterations to adapt to changing circumstances (Vereijken et al., 1992; Newell & Vaillancourt, 2011; Davids et al., 2004; Newell & James, 2008). In addition to the change of impression of how the variability may impact performance, a theory has been proposed for the injury potential that such variability causes. Variability was frowned on as it loaded muscles unaccustomed to being loaded during the movement pattern thereby placing them at risk of being injured (Hamill et al., 1999; Fleisig et al., 2009). Occasional studies countered the theory by hypothesising that the variability could prevent injuries by relieving loads on the taxed muscles which can aid in the prevention of overuse injuries (Hamill et al., 1999; Bartlett et al., 2007).
To the author’s knowledge, only the study of Phillips et al. (2010) investigates the cricket bowling motion from a movement variability perspective. Following work investigating the playing levels impact on technical variability in fastball pitching by baseball pitchers (Fleisig et al., 2009), Phillips et al. (2010) identified one difference between junior and senior bowlers in the technical variabilities, that of the maximal braking ground reaction force. There are several cricket bowling studies identifying significant correlations between a range of technical components and either, or both, ball velocity and accuracy (Portus et al., 2000; Portus et al., 2001; Salter et al., 2007; Ferdinands et al., 2010), including those used in the Phillips et al. (2010) study. Despite these significant correlations, only the variations in lead knee flexion angles at ball release are significantly correlated with bowling velocity, albeit only in the junior player population group (Phillips et al., 2010). Salter et al. (2007) offers a suggestion on why this situation has occurred, the use of a between-bowler methodology which may cause critical individual findings to be masked by the remainder of the population sample.
The current study aimed to investigate the movement pattern variability in amateur cricket bowlers with a focus on the potential functionality that the variance may have in performance consistency. Additionally, the secondary objective of the study is to investigate the effect of differing methodological approaches has on the presented variances and their functionality. The study’s hypothesis predicts that significant differences will be present in comparisons between the within-bowler and between-bowler methodological approaches. Furthermore, there will be reported differences in the functionality of observed variances, differing between the different methodological approaches.
Participant Information & Familiarisation Sessions
A convenience sample of seven amateur cricket bowlers (age = 28.43 ± 11.09 years old; height = 1.82m ± 0.06m; mass = 94.43kg ± 8.51kg) volunteered for the study. Each participant was required to meet a series of inclusionary criteria, including a minimum of two seasons regular bowling experience and being free of musculoskeletal injuries for a period of 6 months prior to the start of the study, before participating in the study. Additionally, a participant’s inclusion in the study’s population sample required the completion of a familiarisation session, lasting between 1 and 2 hours in length, in accordance to ethical guidelines set by an independent ethics committee at Edinburgh Napier University. The familiarisation session consisted of a full briefing between the participants and the researchers which offered the participant the chance to have any queries answered before completing and signing the informed consent form and a detailed health questionnaire. During the session, the participant was able to experience not only the equipment, but the procedures used in data collection, albeit through a contracted version of the data collection procedures. The familiarisation protocols allowed for the bowler’s average bowling velocity to be calculated by the researchers to ensure that they met the 30-50mph range outline in the inclusion criteria as well as the bowlers to identify a starting position for their run-up approach to the bowling wicket.
Table 1. Participant anthropometric data and bowling classifications
Equipment and Laboratory Set-Up
Kinematic data was collected through a Qualysis motion capture system consisting of 12 infra-red cameras (Qualysis ProReflex MCU-500; Qualysis AB; Gothenburg, Sweden), calibrated by the manufacturers guidelines to a capture error <0.01mm and capturing at a frequency of 240Hz (Figure 1). Each of the cameras tracked the three-dimensional location within the capture volume of a series of 0.1cm diameter retro-reflective markers attached to specified anatomical locations either through individual markers or uniquely manufactured “clusters” (Figure 2). Within the capture volume of the Qualysis system was a regulation set of cricket stumps (Slazenger Spring Loaded Stumps; Dunlop Slazenger Ltd.; Shirebrook, UK) with a bowling crease marked out on the floor, 20.12m away from another set of cricket stumps to simulate bowling on a full length cricket wicket (Figure 1). Behind the batsman’s wickets, a zoned target (adapted from Portus et al., 2004) was hung with a 30cm gap between the base of the target and the floor (Figure 1) while a SpeedTrac radar gun (SpeedTrac X; SpeedTrac; Minnesota, USA) was positioned on the batting crease, the emitted infra-red beam angled across the line of the stumps to enhance to capture capabilities (Figure 1). All deliveries were bowled using a 5½ oz Slazenger cork-centred cricket ball (Slazenger Crown County Match Ball; Dunlop Slazenger Ltd.; Shirebrook, UK) .
Figure 1. Zoned target design and laboratory schematic used for data collection protocols
Figure 2. Anatomical locations for retro-reflective markers (green for calibration and data collection trials; red for calibration only) and tri-marker cluster attachments
Variable Definitions
Ball velocity (VEL), denoted in mph, was identified as the velocity of the ball as it reaches the batting crease’s popping line, breaking the infra-red beam of the SpeedTrac radar gun. Ball accuracy (ACC) was noted as the score given for each delivery based on which of the scoring zones the ball impacted in on the hung target behind the batting wicket with a direct strike on the wicket scoring the maximum score (Figure 1). Lead knee flexion (LKF) was determined by the angle of knee flexion created by the retro-reflective markers attached to the greater trochanter, the knee’s lateral epicondyle, and the ankle’s lateral malleolus (Figure 3A). Figure 2. Anatomical locations for retro-reflective markers (green for calibration and data collection trials; red for calibration only) and tri-marker cluster attachments The bowler’s pelvic alignment is expressed in regards to a line running from middle stump of the bowler’s wicket to the middle line of the batsman’s wicket (Figure 3B) and was recorded at the moment of back foot contact (BFC), the moment of front foot contact (FFC) and the moment of ball release (BR). Similarly, thoracic alignment (TA) was calculated in the same manner and again was recorded at BFC, at FFC, and at BR. The velocity of the changes in pelvic alignment and thoracic alignment from PABFC to PABR and from TABFC to TABR was used as the definitions for pelvic angular velocity and thoracic angular velocity, respectively. For the current study, the separation angle (SA) is regarded as the difference between the alignment of the thorax and the alignment of the pelvis at BFC, at FFC, at BR.
Figure 3. The definition of the lead knee flexion angle (A) and the pelvic and thoracic alignment at the crease at BFC, FFC and BR (B)
Data Collection & Analysis
The 31 retro-reflective markers and 7 specially manufactured “clusters” were attached to the body at the specified anatomical locations for the collection of a static calibration file, consisting of the participant standing in an anatomically neutral position for 30 seconds (Figure 2). Following the collection of the calibration file, many of the retroreflective markers were removed so only the tracking marker set was utilised for the remainder of the data collection session (Figure 2). Anthropometric data was collected and inputted into the model-building software as the participant completed a self-directed, individualised 5-minute warm-up. Following the completion of the warm-up, each bowler bowled 12 deliveries towards the batting stumps and zoned target, 12 deliveries split into 2 sets of 6 reflecting bowling overs as stated by the laws of cricket with any deliveries deemed to have been a “wide” or “no ball” being eliminated from the data set and a replacement delivery being bowled in their stead.
The raw kinematic data was collected through the Qualisys Track Manager software package (Qualisys Track Manager Version 2.0.381; Qualisys AB; Gothenburg, Sweden) where the retro-reflective marker locations were tagged for the creation of a three-dimensional (3D) model of the participant’s body throughout their bowling motion in Visual 3D (Visual 3D Stduent Edition Version 4.00.20; C-Motion Inc.; Germantown, Maryland, USA) . The 3D coordinate data was smoothed with a fourth order low-pass Butterworth filter with a 10Hz cut-off frequency before the smoothed data was imported into RStudio (RStudio Version 3.1.0; RStudio Inc.; Boston, MA, USA) for statistical analysis.
Where the within-bowler analysis presented the individual bowler’s results in isolation, the between-bowler analysis was done using four different methods of grouping data together. The first method is the grouping of all trials into one pool regardless of trial and/or participant number (“All”). The second approach is the participant average approach where the seven participant’s average for the 12 trials is employed in the groups vector matrix (“Participant”). The third grouping method employed in the study is the trial average approach where the seven participant’s performance on each trial are averaged for the group’s data set (“Trial”). Finally, the set average approach will be employed where the average from each set of six trials is imported into the group’s matrix resulting in each participant providing two values to the data set (“Over”).
For each of the variables, the absolute deviation from the mean was used as a measure of variance magnitude with larger deviations reflecting inconsistent performances. As functionality was defined as maintaining a consistent outcome, the association between the variance of the technical variables and the outcome variables was assessed through Pearson bivariate analysis. The bivariate analysis set a 95% confidence level thus significant correlations were identified as those with a p-value of less than 0.05. Using a scaled outlined in Fallowfield et al. (2005), the relationships were defined as “weak” (r-value between 0.0 and 0.49), “mild” (r-value between 0.5 and 0.7), “strong” (r-value between 0.71 and 0.9), and “very strong” (r-value above 0.91). To highlight the differences between the methodological approaches, a series of paired t-tests were run comparing each individual bowler’s variance values and each of the four between-bowler conditions, gain the p-value of less than 0.05 was used to identify significant differences.
Performance Correlations
Within-bowler analyses reported 14 significant correlations, 9 in the relationship between technical variances and velocity variances and 5 in the relationship between those technical variances and variances in the accuracy results (Table 2 and Table 3). Of the 9 correlations reported in the velocity analysis, 4 were mildly significant whilst the other 5 were strongly correlated with ball velocity (Table 2). In addition, there was a single mildly significant inverse correlation for both the velocity and accuracy correlation assessments, participant 7’s thoracic alignment and the moment of front foot contact and participant 2’s pelvic alignment at the moment of back foot contact, respectively (Table 2 and Table 3). The between-bowler approach reported 7 significant correlations between technical variances and variances in ball velocity, one of which was a mildly inverse correlation while the others were 4 weakly significant correlations, a mildly significant correlation, and a single strongly significant correlation (Table 2). Whereas, the analysis of relationships between technical variances and variances in the accuracy findings in the between-bowler approach only recorded a single weak inversely significant correlation, that of every trial’s pelvic alignment at the moment of front foot contact (Table 3).
Table 2. Summary of correlations between technical variables and bowling velocity for within- and between-bowler approaches
(* signifies a significant difference where p-value <0.05)
Table 3. Summary of correlations between technical variables and bowling accuracy for within- and between-bowler approaches
(* signifies a significant correlation where the p-value <0.05)
Differences Between the Two Approaches
Each assessment between the within-bowler approach and each of the different between-bowler groupings consisted of 98 individual comparisons, totalling 392 comparisons being assessed. All bar 5 variables in the 392 comparisons between the within-bowler and the 4 between-bowler approaches stated at least one significant difference (Table 4). The insignificant results were spread between the lead knee flexion angle, for participant 6 only, thoracic alignment at the moment of ball release, for both participants 2 and 7, and the thoracic-pelvic separation angle at the moment of front foot contact, for participants 1 and 3 (Table 4). 79 of the 98 (or 80.61%) of comparisons between the “All” grouping condition and the seven participants withinbowler results reported significant differences in variable variances. The other three between-bowler approaches found 77 (78.57%), 64 (65.13%), and 52 (53.07%) of the comparisons were significantly different for the “Trial”, “Participant”, and “Over” groupings respectively.
Table 4. Summary of significant differences between the results recorded for the four between-bowler approaches and the 7 within-bowler approaches
The current study aimed to investigate the potential functionality of movement variability in amateur cricket bowling and how differing methodological approaches can influence the findings. The majority (98.72%) of comparisons between the within-bowler approach and the four between-bowler methodological approach options resulted in a significant difference between the approaches using a 95% confidence level. Only 5 of the variables within the variance comparisons failed to register any significant differences between the within-bowler and any of the between-bowler approaches. Across both the within-bowler and between-bowler approaches there were multiple significant correlations reported between the technical and performance variables, four of which were inversely significant correlations. More correlations were identified in the analysis of technical components with ball velocity than ball accuracy which can be accounted by the large zoned target design and the large decrease in scoring between the zones which may have misled the accuracy correlations reports.
The basic rule for the analysing functionality in this manner is that inverse correlations between technical variances and performance variances show a functional role in the technical variances that aid in maintaining a consistent performance. The correlation between lead knee flexion angles and ball release velocity has been well documented (Foster et al., 1989; Portus et al., 2000; Portus, 2001; Portus et al., 2004; Loram et al., 2005; Worthington & King, 2013) and it could be assumed that there be similar relationships also seen in the analysis between variable variations. Three of the four between-bowler approaches, similarly to those used in the studies reporting the cruciality of the component in bowling performance, would agree with that assumption despite only one participant showing such a relationship in the within-bowler approach. The findings of the between-bowler correlation analysis echoed the findings of Phillips et al. (2010) whose nationally-contracted population group also reported a significant relationship between variances in lead knee angles and ball velocity. A potential reason could be due to the inconsistency in defining the role of the knee joint within the bowling action, with some research suggesting a fully extended, braced knee is best for performance (Davis & Blanksby, 1976; Elliott et al., 1986; Burden & Bartlett, 1990) while others suggest a knee that extends from landing to ball release (Mason et al., 1989; Burden & Bartlett, 1990; Elliott et al., 1992; Bartlett et al., 1996; Portus, 2001). The variance in role definition could indicate that how the joint contributes to the bowling action is depending on the individual and thus the joint could have little implication on the balls velocity. The exceedingly weak insignificant correlations between lead knee flexion and the accuracy component of performance are unsurprising due to the aforementioned target design and scoring method. In addition, the knee’s flexion angle would only influence the length of the delivery and thus the height of the impact on the target which offers little chance of the ball impacting in a differing zone due to the target’s design.
One participant recorded a significant correlation between the variances in pelvic alignment at both the moments of back foot and front foot contact with ball velocity, however none of the between-bowler groupings reported anything more than a weak insignificant correlation. The pelvic alignment at back foot contact has little influence on performance as the pelvis is free to move in open space thus changing its alignment until the lead foot is planted, only then does it provide a base for the thorax to rotate against. At ball release, the pelvic alignment shows thrice as many of the significant relationships with ball velocity although one relationship was identified as an inverse correlation. The pelvic alignment at ball release is indicative of the timing of the rear foot push-off with the more open pelvic reflecting the earlier push-off by the rear foot. The timing of this push-off ends the pelvic-thoracic counter-rotation which has been identified as a significant component contributing to ball velocity (Portus, 2001; Ferdinands et al., 2010). Little relationship was found between pelvic alignment variances and ball accuracy variances with none at the moment of ball release thus potentially highlighting that inconsistencies in the pelvic alignment can be remedied by the actions of the thorax and bowling arm later in the bowling motion.
The thoracic angular velocity reports more significant correlations between its variance and ball velocity than the pelvic angular velocity’s variances, reflecting the more important role that the thoracic rotation has on ball velocity. As well as being further along the kinetic chain, unlike the pelvis, the thorax is not in a permanently stationary position for a duration of the bowling action thus allowing for more variances to occur. Neither reported any significant correlations with bowling accuracy, indicating that variances in angular velocities may be overcome by the actions of the bowling arm. Similarly to the pelvic alignment throughout the delivery stride, variances in the thoracic alignment only reported significant correlations with ball velocity variance in a couple of individuals, despite the reported relationship between the thoracic alignment and ball velocity in other studies (Portus, 2001). Except for two bowlers, the separation angle variances had insignificant correlations with the performance variables despite previous studies indicting a relationship between the actual separation angles, or the counter-rotation movement they represent, and ball velocities (Portus, 2001; Salter et al., 2007; Ferdinands et al., 2010). The author proposes that similarly to the individual variances in pelvic and thoracic alignments, variations in these variables can be masked by the kinematics of the bowling arm, a proposed theory that could be confirmed or rejected through collecting data on the bowling arm’s motion.
It should be noted that across both the between-bowler and within-bowler approaches, various insignificantly mild and a few insignificantly strong correlations between technical and performance variables have been identified with these correlations split evenly between positive and negative correlations. The analysis of the significant correlations alone would indicate that, with a couple of exceptions, variances in technical components are detrimental to performance, however the inclusion of the insignificant correlations, insignificant for this population sample, could question the consistency of this generalisation.
Salter et al. (2007) presented an initial look at the difference between using a within- or between-bowler methodology suggesting that the between-bowler methodology could present the “optimal” performance which fails to match the performance of any of the bowlers participating in the study. Salter et al. (2007) in conjunction with Bates (1996) and Dufek et al. (1995) have highlighted that case studies or within-participant methodologies provide evidence to counter, or eject, relationships stated in between-participant methodologies. The reasoning for such action, as highlighted in the varied reported correlations in the current study, is the “masking” of one individual’s significant findings by the insignificant findings from other individuals within the grouped between-bowler, especially when using sample sizes of lower numbers. From the current study, one such example of this “masking” effect is the thoracic alignment at front foot contact’s relationship with ball velocity where the between-bowler approach stated only an insignificantly weak correlation despite one participant’s significantly mild correlation which countered the results of the other participants. The other aspect of the “masking” is highlighted by the relationship of the lead knee’s flexion angles variance and the variance of ball velocity where the between-bowler results give three significant correlations and one strong insignificant correlation despite only one of the seven participant’s data indicating such a relationship. In these instances, the performance of the other participants has altered the correlations and their significance, causing a misreporting of results, further highlighted by the difference in correlations reported between the four grouped conditions were only the number of trials in the analysis alters, It is relatively unsurprising that such differences in the relationship between variables variability given the significant differences in reported variance between the within-bowler approach and each of the four between-bowler approaches for the vast majority of the variables.
The key findings of this study are three-fold. Firstly, there are, with few exceptions, significant relationships that indicate that increasing variance in certain technical variables will result in increases in performance variance. Secondly, the strengths and significances of such relationships vary between different bowlers and these individual variances maybe weakened or lost when using between-participant analysis. The third key finding is on how the choice of the methodological approach and the way that data is pooled for between-participant analysis can alter the presented results of the study thus the methodical approach can be manipulated to present desired results. It would be suggested that the choice of using a within- or between-participant methodology would be determined by the research question and the population sample being utilised. For elite performers or questions regarding movement variability, within-participant approaches would be recommended. As for sweeping generalisations that are custom in between-participant studies, Salter et al. (2007) predicts that the generalisation could be confirmed or countered once enough within-participant studies and trials have been conducted although this could require a excessive amount of trials and time to reach such a level to comment on the findings from between-participant studies..
-
Arutyunyan, G.H., Gurfinkel, V.S., and Mirskii, M.L. (1968). Investigation of aiming at a target. Biophysics, 13: 1162-1167.
-
Arutyunyan G.H., Gurfinkel, V.S., and Mirskii, M.L. (1969). Organisation of movements on execution by man of an exact postural task. Biophysics, 14: 1162-1167.
-
Atkinson, G., and Spiers, L. (1998). Diurnal variation in tennis serve. Perceptual and Motor Skills, 86: 1335-1338.
-
Bartlett, R., Stockill, N., Elliott, B., and Burnett, A. (1996). The biomechanics of fast bowling in men’s cricket: A review. Journal of Sports Sciences, 14: 403-424.
-
Bartlett, R.M., Wheat, J., and Robins, M. (2007). Is movement variability important for sports biomechanists? Sports Biomechanics, 6: 224-243.
-
Bartlett, R.M., and Robins, M. (2008). Biomechanics of throwing. In: Handbook of Biomechanics and Human Movement (edited by Y. Hong and R. Bartlett), pp.285-297. London, UK: Routledge.
-
Bates, B.T. (1996). Single-subject methodology: An alternative approach. Medicine and Science in Sports and Exercise, 28: 631-638.
-
Burden, A.M., and Bartlett, R.M. (1990). A kinematic investigation of elite fast and fast-medium cricket bowlers. In: Proceedings of the VIIIth International Symposium of the Society of Biomechanics in Sports (edited by M.Nosek, D.Sojka, W.E.Morrison, and P.Susanka), pp. 41-46. Prague: Conex.
-
Charlton, L.G., Chow, J.W., and Shim, J. (2006). Variability in motor output and Olympic performers. In: Movement System Variability (edited by K. Davids, S. Bennett, and K. Newell), pp.85-108. Champaign, IL: Human Kinetics.
-
Clark, J.E. (1995). On becoming skilful: Patterns and constraints. Research Quarterly for Exercise and Sport, 66: 173-183.
-
Davids, K., Glazier, P., Araujo, D., and Bartlett, R.M. (2003). Movement systems as dynamic systems: The functional role of variability and its implications for sports medicine. Sports Medicine, 33: 245-260.
-
Davis, K., and Blanksby, B. (1976). A cinematographic analysis of fast bowling in cricket. Australian Journal for Health, Physical Education and Recreation, 71(suppl.): 9-15.
-
Dufek, J.S., Bates, B.T., Stergiou, N., and James, C.R. (1995). Interactive effects between group and single-subject response patterns. Human Movement Science, 14: 301-323.
-
Elliott, B., Foster, D., and Gray, S. (1986). Biomechanical and physical factors influencing fast bowling. Australian Journal of Science and Medicine in Sport, 18: 16-21.
-
Elliott, B., Hardcastle, P., Burnett, A., and Foster, D. (1992). The influence of fast bowling and physical factors on radiologic features in high performance young fast bowlers. Sports Medicine, Training and Rehabilitation, 3: 113-130.
-
Fallowfield, J.L., Hale, B.J., and Wilkinson, D.M. (2005). Using Statistics in Sport and Exercise Science Research. Chichester, UK: Lotus Publishing.
-
Ferdinands, R., Kersting, U.G., Marshall, R.N., and Stuelcken, M. (2010). Distribution of modern cricket bowling actions in New Zealand. European Journal of Sport Science, 10: 179-190.
-
Fitts, P.M. (1954). The information capacity of the human motor system in controlling the amplitude of movement. Journal of Experimental Psychology, 47: 381-391.
-
Fleisig, G., Chu, Y., Webber, A., and Andrews, J. (2009). Variability in baseball pitching biomechanics among various levels of competition. Sports Biomechanics, 8: 10-21.
-
Foster, D., John, D., Elliott, B.C., Ackland, T., and Fitch, K. (1989). Back injuries to fast-bowlers in cricket: A prospective study. British Journal of Sports Medicine, 23: 150-154.
-
Glass, L., and Mackey, M.C. (1998). From Clocks to Chaos: The Rhythms of Life. Princeton, NJ: Princeton University Press.
-
Glazier, P.S., Wheat, J.S., Pease, D.L., and Bartlett, R.M. (2006). The interface of biomechanics and motor control: Dynamic systems theory and the functional role of movement variability. In: Movement Systems Variability (edited by K. Davids, S. Bennet, and K. Newell), pp.49-72. Champaign, IL: Human Kinetics
-
Hamill, J., van Emmerik, R.E., Heiderscheit, B.C., and Li, L. (1999). A dynamical systems approach to lower extremity running injuries. Clinical Biomechanics, 14: 297-308.
-
Loram, L.C., McKinon, W., Wormgoor, S., and Rogers, G.G. (2005). Determinants of ball release speed in schoolboy fast-medium bowlers in cricket. Journal of Sports Medicine and Physical Fitness, 45(4): 483-490.
-
Manoel, E.J., and Connolly, K.J. (1995). Variability and the development of skilled actions. International Journal of Psychophysiology, 19: 129-147.
-
Mason, B., Weissensteiner, J., and Spence, P. (1989). Development of a model of fast bowling in cricket. Excel, 6(1): 2-12.
-
Miller, S. (2002). Variability in basketball shooting: Practical Implications. In: International Research in Sports Biomechanics (edited by Y. Hong), pp.27-34. London, UK: Routledge.
-
Morriss, C., Bartlett, R.M., and Fowler, N. (1997). Biomechanical analysis of the men’s javelin throw at the 1995 World Championships in athletics. New Studies in Athletics, 12: 31-34.
-
Mullineaux, D.R. (2008). Research Methods: Sample Size and Variability Effects on Statistical Power. In: Biomechanical Evaluation of Movement in Sports and Exercise: The British Association of Sport and Exercise Sciences Guidelines (edited by C.J. Payton and R.M. Bartlett), pp. 153-175. New York, USA: Routledge.
-
Newell, K.M., and Corcos, D.M. (1993). Issues in variability and motor control. In: Variability and Motor Control (edited by K. Newell and D. Corcos), pp. 1-12. Champaign, IL: Human Kinetics
-
Newell, K.M., and Vaillancourt, D. (2001). Dimensional change in motor learning. Human Movement Science, 20: 695-715.
-
Newell, K.M., and James, E.G. (2008). The Amount and Structure of Human Movement Variation. In: Handbook of Biomechanics and Human Movement Science (edited by Y.Hong and R. Bartlett), pp. 93-104. London, UK: Routledge.
-
Phillips, E., Marc, P., Davids, K.W., Brown, N. and Renshaw, I., (2010). Quantifying variability within technique and performance in elite fast bowlers: is technical variability dysfunctional or functional? In: Proceedings of the 2010 Conference of Science, Medicine and Coaching in Cricket (pp. 121- 124). Cricket Australia.
-
Portus, M.R., Sinclair, P., Burke, S., Moore, D., and Farhart, P. (2000). Cricket fast bowling performance and technique and the influence of selected physical factors during an 8-over spell. Journal of Sports Sciences, 18: 999-1011.
-
Portus, M.R. (2001). Relationship between cricket fast bowling technique, trunk injuries and ball release speed. In: Proceedings of the XIXth International Symposium on Biomechanics in Sports (edited J. Blackwell), pp. 231-234. San Francisco, CA: University of San Francisco Press.
-
Portus, M.R., Mason, B.R., Elliott, B.C., Pfitzner, M.C., and Done, R.P. (2004). Technique factors related to ball release speed and trunk injuries in high performance fast bowlers. Sports Biomechanics, 3(2): 263-284.
-
Salter, C.W., Sinclair, P.J., and Portus, M.R. (2007). The association between fast bowling technique and ball release speed: A pilot study of the within-bowler and between-bowler approaches. Journal of Sports Sciences, 25: 1279-1285.
-
Schmidt, R.A., Zelaznik, H., Hawkins, B., Frank, J.S., and Quinn, J.T. (1979). Motor output variability: A theory for the accuracy of rapid motor acts. Psychological Review, 86: 415-451.
-
Schoellhorn, W.I. (1997). Self-organising maps for the analysis of complex movement patterns. Neural Processing Letters, 5: 193-199.
-
Schoellhorn, W.I., and Bauer, H.U. (1998). Identifying Individual Movement Styles in High Performance Sports by Self-Organising Kohonen Maps. In: Proceedings of the XVI Congress of the ISBS (edited by H.J. Reihle and M. Vieten), pp. 754-758. Konstanz: Konstanz University Press
-
Scholz, J.P., Schoner, G., and Latash, M.L. (2000). Identifying the correct structure of multi-joint coordination during pistol shooting. Experimental Brain Research, 135: 382-404.
-
Slifkin, A.B., and Newell, K.M. (1999). Noise, information transmission, and force variability. Journal of Experimental Psychology: Human Perception and Performance, 25: 837-851.
-
Thompson, J.M., and Stewart, H.B. (2002). Nonlinear Dynamics and Chaos Second Edition. New York, USA: Wiley.
-
Vereijken, B., Whiting, H.T., Newell, K.M., and van Emmerick, R.E. (1992). Freezing degrees of freedom in skill acquisition. Journal of Motor Behaviour, 24: 133-142.
-
Worthington, P.J., and King. M.A. (2013). Relationships between fast bowling technique and ball release speed in cricket. Journal of Applied Biomechanics, 29: 78-84.
The author would like to thank Dr Susan Brown and the technical support staff of Edinburgh Napier University for all their support and assistance throughout the duration of the study. A special note of gratitude is expressed to Miss Jemma Oswald for all her assistance during the data collection phase of the study. Lastly, the author wishes to thank the participants for the final study, and those in the early pilot studies, who volunteered their time, without which the study would not have occurred.