The Effects of an Eight over Cricket Bowling Spell upon Pace Bowling Biomechanics and Performance within Different Delivery Lengths

Pace bowlers must often perform extended bowling spells with maximal ball release speed (BRS) while targeting different delivery lengths when playing a multi-day match. This study investigated the effect of an eight over spell upon pace bowling biomechanics and performance at different delivery lengths. Nine male bowlers (age = 18.8 ± 1.7 years) completed an eight over spell, while targeting different lengths (short: 7–10 m, good: 4–7 m, full: 0–4 m from the batter’s stumps, respectively) in a randomized order. Trunk, knee and shoulder kinematics and ground reaction forces at front foot contact (FFC), as well as run-up velocity and BRS were measured. Paired sample t-tests (p ≤ 0.01), Hedges’ g effect sizes, and statistical parametrical mapping were used to assess differences between mean variables from the first and last three overs. No significant differences (p = 0.05–0.98) were found in any discrete or continuous variables, with the magnitude of difference being trivial-to-medium (g = 0.00–0.73) across all variables. Results suggest pace bowlers sustain BRS through a single eight over spell while tolerating the repeatedly high whole-body biomechanical loads as suggested by maintaining the kinematics or technique at the assessed joints during FFC. Practically, the findings are advantageous for bowling performance and support current bowling load monitoring practices.


Introduction
Cricket has a number of different match formats, which range in length from a few hours (e.g., twenty20 and one-day cricket) to several days (e.g., multi-day matches). Due to the length of the match, the multi-day format dictates a greater total workload (i.e., the combination of internal and external sport stressors) for all players [1]. Specifically, for pace bowlers, the increased total workload, is largely attributed to the greater number of deliveries bowled, often within an extended bowling spell. Despite the use of extended bowling spells within the longer match formats, pace bowlers are generally expected to be able to maintain a high ball release speed (BRS) each delivery. An increased BRS may disrupt the timing and stroke execution of an opposing batter, increasing the likelihood of dismissal. Pace bowlers will also employ various delivery lengths (i.e., short, good and full), defined as the distance the ball lands from the batter [2], to further assist with dismissing batters. kinematics, with respect to different delivery lengths. It was further hypothesized that BRS will remain constant throughout the eight over spell for each delivery length.

Participants
A convenience sample of nine healthy male cricketers was recruited for this study (age = 18.8 ± 1.7 years, height = 1.9 ± 0.1 m, weight = 80.7 ± 9.5 kg). Inclusion criteria required current or previous involvement in an Australian state cricket development pathway or state premier grade cricket, greater than 17 years of age, and free from any existing medical conditions that would be contradictory to participation. The sample included six right-arm and three left-arm pace bowlers. All participants, and where appropriate, guardians of participants under 18 years of age, received a clear explanation of the study, including the risks and benefits of participation, and provided written informed consent prior to participation. The research was approved by the Edith Cowan University Ethical Review Board (Approval 11948) and conformed to the policy statement with respect to the Declaration of Helsinki.

Procedures
A cross-sectional study design involving a single testing session was utilized to assess the biomechanical and performance differences within short, good, and full delivery lengths during a simulated eight over bowling spell among pace bowlers. The pace bowling assessment incorporated three-dimensional (3D) motion capture and in-ground tri-axial force plates to determine kinematic and kinetic discrete and continuous variables during FFC of the delivery stride. BRS was also recorded as a measure of performance. Mean comparisons between variables from the first and last three overs were used for analysis. Prior to the commencement of data collection, the participant's age, height, body mass, and anthropometric data were recorded. Anthropometric measurements were necessary for the motion capture analysis, which will be detailed later. A standardized warm-up, consisting of jogging, dynamic stretching of the upper and lower limbs, and progressive speed runs, was used for all participants.

Pace Bowling Performance Test
An eight over spell, comprising 48 maximum effort deliveries of either a short, good, or full length ( Figure 1) were performed by each participant [2,3,16,17]. This length of bowling spell is not uncommon within high level cricket and has also been used in previous research for the assessment of kinematic changes in pace bowling technique [3]. The dimensions of the laboratory allowed each participant to use their normal, full length run-up and follow-through, while bowling deliveries on the equivalent of a standard-sized cricket pitch. A four-minute active rest period was provided between each over to comply with match regulations [3,6,17], however this is atypical of match-play as participants were not required to undertake fielding actions. The justification for the use of an active recovery, exclusive of fielding movement patterns is based upon the suggestions of Stronach, Cronin, and Portus [14]. Stronach, Cronin, and Portus [14] outlined that it is the repeated bouts of eccentric contractions associated with the delivery stride of pace bowling during an extended bowling spell which will lead to high levels of fatigue, not low-level intermittent bouts of running. The classification of the various delivery lengths was based upon match and training analysis of pace bowling effectiveness from a state squad competing within Australian national competitions. Assessment protocols routinely used by Cricket Australia also require bowlers to target delivery lengths of short, good, and full [8,18]. Each delivery length was targeted six times during the first and last three overs of the eight over spell in a randomized order. Delivery length was visually identified by two researchers [8]. If a delivery landed on the line between two target lengths, the delivery was classified as of hitting the intended length [3]. Although infrequent, if a participant failed to land the delivery in the intended length, the order of deliveries was revised to ensure the required number of different deliveries was attained.
All bowlers used a red Kookaburra four-piece (156 grams) cricket ball (A.G. Thompson Pty. Ltd., Melbourne, Australia) and wore their own bowling spikes during testing. A Stalker Pro II speed radar gun (Stalker Radar, Richardson, TX, USA) was located behind the batting stumps net and aimed at the ball release point to measure BRS.

Kinematic Data Collection and Analysis
All trials were recorded using the XSENS motion analysis system (MVN system, XSENS Technology, Enschede, Netherlands) to determine pace bowling kinematics. Following the warm-up and prior to testing, each participant was fitted within a lycra suit (XSENS Technology, Enschede, Netherlands), which housed 17 inertial sensors (0.038 × 0.053 × 0.021 m, 0.03 kg) with a sample frequency of 120 Hz. The inertial sensors were located on the participant's pelvis, sternum, head and left and right shoulders, upper arms, forearms, hands, upper legs, lower legs, and feet. The inertial sensors (XSENS Technology, Enschede, Netherlands)) consisted of a 3D accelerometer, gyroscope, and magnetometer. Each body segment's position and orientation can be estimated by integrating the gyroscope and double integrating the accelerometer data, in combination with a biomechanical model [19]. The XSENS motion analysis system has previously been found to be a reliable and valid means of assessing joint angles in dynamic movements [19][20][21]. Additional lengths were also determined between sensor locations and landmarks once the participant was fitted within the lycra suit. A static capture was undertaken prior to pace bowling performance testing and required participants to adopt a stationary N-pose, as per the manufacturer's recommendations. Following the fitting of the lycra suit and calibration procedures, participants then bowled as many deliveries as required to become familiar with the test environment.
Data collected from the XSENS motion capture system was exported to Visual3D (Version 6.00.18; C-Motion, Inc., Germantown, MD, USA) for filtering and kinematic analysis. A full-body model was recreated in Visual3D based upon previous cricket pace bowling models [11,22] using the exported 64 virtual marker set that is calculated by the manufacturer software (MVN Studio Version 3.5.3, XSENS Technology, Enschede, The Netherlands) using the measured segments orientations and body dimensions. A second order low pass Butterworth filter with a cut-off frequency of 6 Hz, as determined by a fast Fourier transformation, was applied to all kinematic data [23,24]. Segment orientation was such that the z-axis pointed upward along the longitudinal axis, the x-axis to the participant's right, and the y-axis forward. Similarly, the global coordinate system was defined with

Kinematic Data Collection and Analysis
All trials were recorded using the XSENS motion analysis system (MVN system, XSENS Technology, Enschede, Netherlands) to determine pace bowling kinematics. Following the warm-up and prior to testing, each participant was fitted within a lycra suit (XSENS Technology, Enschede, Netherlands), which housed 17 inertial sensors (0.038 × 0.053 × 0.021 m, 0.03 kg) with a sample frequency of 120 Hz. The inertial sensors were located on the participant's pelvis, sternum, head and left and right shoulders, upper arms, forearms, hands, upper legs, lower legs, and feet. The inertial sensors (XSENS Technology, Enschede, Netherlands)) consisted of a 3D accelerometer, gyroscope, and magnetometer. Each body segment's position and orientation can be estimated by integrating the gyroscope and double integrating the accelerometer data, in combination with a biomechanical model [19]. The XSENS motion analysis system has previously been found to be a reliable and valid means of assessing joint angles in dynamic movements [19][20][21]. Additional lengths were also determined between sensor locations and landmarks once the participant was fitted within the lycra suit. A static capture was undertaken prior to pace bowling performance testing and required participants to adopt a stationary N-pose, as per the manufacturer's recommendations. Following the fitting of the lycra suit and calibration procedures, participants then bowled as many deliveries as required to become familiar with the test environment.
Data collected from the XSENS motion capture system was exported to Visual3D (Version 6.00.18; C-Motion, Inc., Germantown, MD, USA) for filtering and kinematic analysis. A full-body model was recreated in Visual3D based upon previous cricket pace bowling models [11,22] using the exported 64 virtual marker set that is calculated by the manufacturer software (MVN Studio Version 3.5.3, XSENS Technology, Enschede, The Netherlands) using the measured segments orientations and body dimensions. A second order low pass Butterworth filter with a cut-off frequency of 6 Hz, as determined by a fast Fourier transformation, was applied to all kinematic data [23,24]. Segment orientation was such that the z-axis pointed upward along the longitudinal axis, the x-axis to the participant's right, and the y-axis forward. Similarly, the global coordinate system was defined with the y-axis pointing down the wicket (towards the batter), the x-axis to the right and the z-axis representing the upwards vertical. Joint angles were calculated as Cardan angles with an axis order rotation of XYZ [11]. Five discrete kinematic parameters were calculated for each trial: Horizontal run-up velocity, front knee angle at FFC and ball release, trunk flexion from FFC to ball release, and bowling arm shoulder angle at ball release. Horizontal run-up velocity in the global frame was determined as per Worthington, King, and Ranson [11]. Front knee angle was calculated as the relative angle between the thigh and shank (full extension = 0 • , flexed > 0 • ). Trunk angle was relative to the global coordinate system (neutral spine position = 0 • , flexed > 0 • ). Shoulder angle was defined as relative to the trunk (anatomical position = 0 • ) [9,11]. Due to the unique actions of the shoulder during pace bowling, which require full circumduction of the humerus with varying degrees of internal and external rotation, current methods of joint angle calculation are not well defined for the entire motion, leading to errors during calculations [25]. Therefore, while not customary, a separate axis order of rotation (ZYZ) was used to determine shoulder joint angle at the instance of ball release [25]. The use of a separate axis order rotation was undertaken such that an appropriate representation, which is comparative to previous research [9,11] could be provided for the shoulder joint angle variable. The continuous joint-time curve of the front knee from FFC to ball release was also assessed for each delivery. A customized MATLAB (R2015b, The MathWorks Inc, Natick, MA, USA) script was used to time normalize the data to 100% of the stance phase from FFC to ball release using a spline fill pattern prior to statistical analysis of continuous variables using statistical parametric mapping (SPM).

Kinetic Data Collection and Analysis
An in-ground tri-axial force plate (9287CA, Kistler Group, Winterthur, Switzerland) sampling at 960 Hz was used to collect GRF data during FFC of the delivery stride. FFC corresponded to the first instance at which the vertical GRF exceeded 20 N [26]. Both the XSENS motion capture system and the in-ground force plate were time synchronized through an analog board (Kistler Group, Winterthur, Switzerland) which allowed the XSENS recording software to trigger data capture within the force plate software (Bioware Version 5.3.0.7, Kistler Group, Winterthur, Switzerland) via a voltage rising edge configuration. All trials were filmed with a high-speed video camera (Apple Inc, Cupertino, CA, USA) at 240 Hz from a position perpendicular to the delivery stride to sync FFC on the force plate and ball release using video analysis software (Kinovea-0.8.15, Kinovea, Bordeaux, France) [3,27]. Flooring surface (Mondo S.p.A., Alba, Italy) of the laboratory and on-top of the force platform was consistent.
Discrete kinetic variables were all measured from FFC to ball release, and included peak vertical and braking forces, vertical and braking impulses, and average vertical and braking loading rates (peak force divided by time from FFC to peak force). The force platform software (Bioware Version 5.3.0.7, Kistler Group, Winterthur, Switzerland) was used for analysis of each delivery bowled. All kinetic variables were normalized to body weight (BW). The continuous force-time curve in the vertical and anterior/posterior axes of the entire FFC was also assessed for each delivery. As previously stated, a customized MATLAB script was used to time normalize the data to 100% of the stance phase with the only variation being that force-time curves in both the vertical and anterior/posterior axes were normalized to the entire FFC phase, and not just to ball release.

Statistical Analyses
Descriptive statistics (mean ± standard deviation) were used to profile each measured parameter. Normality of data was assessed by visual analysis of the Q-Q plots [28,29]. To assess the reliability of all discrete variables, intra-class correlation coefficient (ICC) and coefficient of variance (CV) were determined. An ICC ≥ 0.70 [30] and a CV ≤ 10% was deemed acceptable [31]. A two-tailed paired-samples t-test was used to determine significant changes in the mean of the first and last three overs for each discrete variable within delivery lengths. Due to this statistical approach and to decrease the chances of making a Type 1 error, the criterion for significance was set at p ≤ 0.01 [32]. Hedges' g effect sizes and 95% confidence intervals were also calculated for the first and last three over comparisons. The magnitude of effect was assessed on the following scale: Less than 0.2 was considered a trivial effect, 0.2 to 0.49 a small effect, 0.5 to 0.79 a medium effect, and greater than 0.8 a large effect [33]. These statistics were computed using the Statistics Package for Social Sciences Version 23.0 (IBM, Armonk, NY, USA).
Statistical parametric mapping (SPM) was used to evaluate if significant changes in kinematics and kinetics occurred at points other than the commonly assessed peaks and other discrete measures. Briefly, SPM uses random field theory to objectively identify field regions which co-vary significantly with the experimental design [34,35]. Two-tailed paired sample t-tests were performed on the normalized time series data during FFC to determine if a significant (p < 0.05) change occurred between the first and last three overs of the testing protocol, with respect to the front knee kinematics and vertical and anterior/posterior axes GRF curves. The SPM analysis required four steps as outlined in De Ridder, et al. [36]. All SPM analyses were implemented in MATLAB R2015b (The MathWorks Inc, Natick, MA, USA) using the open source package "rft1d" located at http://www.spm1d.org/ [37].

Results
All investigated variables were deemed to be normally distributed as determined by the Q-Q plot analysis. All discrete reliability measures (ICC = 0.72-0.99, CV = 1.46-9.70%) were deemed acceptable, except for the vertical loading rate measure for the first (CV = 19.39-20.97%) and last (CV = 10.33-20.08%) three overs for all delivery lengths. There were no significant differences between BRS between the first and last three overs with respect to the measured delivery lengths (Table 1). Tables 2 and 3 outline the kinematic and BW normalized kinetic results measured during the first and last three overs of the testing protocol, respectively. There were no significant differences between the assessed variables across the short, good and full delivery lengths. However, there was a 14% increase in horizontal run-up velocity for the short delivery length from the first to last three overs of the testing protocol, which had a medium effect. However, this was non-significant. Figures 2-4 depict the differences between the first and last three overs of an eight over bowling spell, as assessed using SPM upon front knee kinematics, and vertical and horizontal GRF during FFC, respectively. The SPM curve analysis revealed no significant differences.      describing the strength and slope of the relationship between the first three overs and last three overs testing measures. The dotted horizontal line indicates the random field theory thresholds for significance, and p values indicate the likelihood that a random process of the temporal smoothness would be expected to produce a suprathreshold cluster of the observed size.  (i), is the mean ground reaction force trajectory with standard deviation clouds (first three overs = blue, last three overs = red). (ii), displays the paired samples SPM{t}: The t statistic as a function of time, describing the strength and slope of the relationship between the first three overs and last three overs testing measures. The dotted horizontal line indicates the random field theory thresholds for significance, and p values indicate the likelihood that a random process of the temporal smoothness would be expected to produce a suprathreshold cluster of the observed size.

Discussion
This is the first study to investigate the GRF experienced during FFC, in addition to bowling kinematics, for an eight over bowling spell with respect to different delivery lengths among pace bowlers. The results reinforced previous research [3][4][5][6]8], as the pace bowlers in this study were able to maintain their technique and BRS throughout a single extended bowling spell. The ability to maintain similar biomechanics was contrary to the study hypothesis and may suggest that pace bowlers are able to tolerate the repeatedly high whole-body biomechanical load (as determined by GRF) experienced within an eight over spell within the joints assessed. The results from this study provide valuable information about the loading experienced by a pace bowler during an eight over spell, as well as providing support for current bowling load monitoring practices of counting the total number of deliveries bowled.
The ability to maintain a high BRS throughout an extended bowling spell could reduce the decision and stroke execution time for an opposing batter. In accordance with the study hypothesis, there was no significant difference in BRS between the first three and last three overs of an eight over spell. The lack of significant change in BRS with trivial to small effect, is similar to previous extended bowling spell investigations which found no significant changes among elite (spell 1 = 34.92 ± 1.41 m·s −1 , spell 2 = 34.83 ± 1.25 m·s −1 ) [8] and state first grade and high-performance pace bowlers (2nd over = 31.73 ± 0.78 m·s −1 , 8th over = 31.93 ± 0.29 m·s −1 ) [3]. The ability to maintain BRS throughout an extended bowling spell is beneficial to pace bowling performance, as it should reduce the decision-making time for batters.
With respect to different delivery lengths among pace bowlers, there was no significant change in the assessed kinematics between the first and last three overs of an eight over spell. There was, however, a 14%, medium effect size increase in horizontal run-up velocity for the short delivery length in the final three overs when compared to the first three overs of the eight over spell. This result supported the findings of Burnett, Elliott, and Marshall [7], who documented a non-significant increase (0.2 m·s −1 ) in run-up velocity, although different delivery lengths were not assessed. Burnett, Elliott, and Marshall [7] suggested that increases in approach speed during the later stages of an extended bowling spell may be required to counter any changes in other areas (e.g., trunk and knee kinematics) of the pace bowling system. Regardless, the current results showed there were no significant changes in bowling kinematics within any of the delivery lengths, and it should be noted that the data from the current investigation can only speculate that an increase in run-up velocity was in an effort to use momentum to maintain the kinematics displayed at FFC. As such a hypothesis would require an extended analysis of the system energy flow.
The results from this study also indicated no significant differences in GRF during FFC throughout an eight over spell, regardless of delivery length. The pace bowlers within this study were able to maintain appropriate GRF during FFC to ensure the maintenance of BRS for each delivery length. Therefore, it appears that the bowlers within the current study were able to tolerate the repeated bouts of eccentric contractions associated with the high GRF experienced during FFC of the delivery stride [14]. However, it is important to note that while it is not uncommon for pace bowlers to perform an eight over spell [3], they may also be required to bowl up to 50 overs (300 deliveries or more) during a multi-day match depending on the match circumstances and game strategy [38]. Although GRF may not change during an eight over spell, there is a need for future research to assess the loading experienced by a pace bowler within an entire multi-day match, where multiple extended bowling spells will be completed.
The results from the SPM revealed no significant difference for the vector field analyzed GRF or front knee kinematics during FFC, regardless of the delivery length. These findings are in accordance with the discrete GRF and kinematic measures and further suggest that pace bowlers can appropriately maintain their technique in the joints assessed throughout a single eight over spell for each delivery length. It is also noteworthy that the use of SPM provided valuable new insight into the whole-body biomechanical load experienced in the vertical and braking/propulsive planes and sagittal knee kinematics by pace bowlers. The use of SPM visually presents the average GRF and knee kinematic traces of all pace bowlers, providing critical information about the loading and movement patterns present during FFC, not just the commonly assessed peaks or average values as suggested within previous research [39,40]. The use of SPM may allow future investigations to evaluate specific portions of FFC that may be most associated with specific alterations as a result of multiple extended bowling spells.
The consistent GRF reported throughout the eight over spell is of practical importance when considering current bowling load monitoring practices. Current practices stipulate that the total number of deliveries bowled, irrespective of the length of spell or delivery, be used as a tool to monitor bowling load [15,18,41]. The study findings provide support for current bowling load monitoring practices and suggest that strength and conditioning coaches do not need to track or prescribe bowling loads with respect to the length of spell or delivery for up to eight overs. However, it is important to note that while the results of the current study highlight consistency in whole-body external biomechanical load during FFC of the delivery stride throughout and extended eight-over spell and changes in delivery length, there is a need for greater research into the individual variations between bowlers. This is with respect to how joint-based load is distributed throughout the system when performing extended bowling spells which target different delivery lengths. As such, individuals with similar whole-body measures of load may be using different joint based strategies during performance that could have implications upon performance or injury which are not detailed by whole-body measures of load.
There are certain limitations for this study. The participant numbers utilized in this study were low (n = 9), although they were similar to previous research analyzing the effects of an extended bowling spell upon biomechanics and BRS in pace bowlers [7,8,42]. The full-body model utilized for kinematic data analysis involved an exported virtual marker set that is calculated instead of direct measured marker location. However, the estimated marker locations and subsequent model produced has previously been found to be a reliable and valid measure of joint kinematics during dynamic movements when in combination with a biomechanical model [19][20][21]. Furthermore, the model used within the current investigation was based upon previously published models for cricket pace bowlers [11,22], ultimately these factors indicate that the developed model was appropriate for the determination of joint kinematics. Therefore, as with interpretation of kinematic data, interpretation of the magnitude of effect and not just statistical significance is critical to interpretation of a meaningful or clinically relevant magnitude of change or difference. The use of laboratory-based testing may limit the ecological validity of the study. However, laboratory testing allowed for a more detailed analysis of the pace bowling action and has previously been used within the literature [3,7,8,42]. Future research should attempt to investigate the biomechanics of pace bowlers during an actual multi-day match or evaluate changes in timing or segmental coordination leading to FFC that could mediate in an effort to maintain the variables assessed in this study. Finally, the current investigation did not take an individual approach to understanding the variability and implications of an extended bowling spell within each pace bowler. However, the focus of the investigation was on the global characteristics which may alter as a result of an extended bowling spell, regardless of technique. Within the context of these limitations, this study still provides valuable insight into the loading experienced by a pace bowler throughout a single extended bowling spell, while also adding support to current bowling load monitoring practices.

Conclusions
In conclusion, an eight over spell targeting short, good and full delivery lengths, does not result in significant changes to a pace bowler's biomechanics or BRS. The results suggest that pace bowlers can tolerate the repeated high whole-body external biomechanical loading (as represented by GRF) experienced during FFC without any one isolated variable significantly changing technique. As GRF was maintained throughout the bowling spell, this provides support for current bowling load monitoring practices used in the applied setting. Ultimately, this research indicates that within an eight over spell, pace bowlers are able to maintain BRS and technique while employing a range of delivery lengths in an attempt to dismiss opposing batters. Additional research is required to determine the effects of repeated bowling spells upon the biomechanics and BRS of pace bowlers within the multi-day match format.
Author Contributions: Study design, data collection, data interpretation, manuscript writing, S.J.C.; study design, data interpretation, manuscript writing, R.G.L. and S.N.; data collection, manuscript writing, W.Y.; data interpretation and manuscript writing, R.F.C and W.A.A.
Funding: This study was funded by the Australian Postgraduate Award Scholarship as well as the Edith Cowan University Merit Award Scholarship.

Acknowledgments:
We would like to acknowledge our participants for their contribution, time and effort to this study. Many thanks also to the Western Australian Cricket Association, with whom this research could not have been possible. This study was funded by the Australian Postgraduate Award Scholarship as well as the Edith Cowan University Merit Award Scholarship. SN is a member of the Australian Centre for Research into Injury in Sport and its Prevention (ACRISP) at Edith Cowan University. ACRISP is one of the International Research Centres for the Prevention of Injury and Protection of Athlete Health supported by the International Olympic Committee (IOC).

Conflicts of Interest:
The authors have no conflict of interest.

Abbreviations
The following abbreviations are used in this manuscript: