Technique Variables Associated with Fast Bowling Performance: A Systematic-Narrative Review

Abstract

The performance characteristics of fast bowling include high ball release speed, that reduces the shot execution time of the batter, accuracy, ensuring the ball follows the desired trajectory, and deception. This can lead the batter to misjudge the ball speed or path. Previous research has utilised a variety of biomechanical methodologies in order to further understand fast bowling techniques. The aim of this study was to systematically review biomechanical literature related to cricket fast bowling performance and narratively synthesise findings to provide a comprehensive summary of key performance characteristics. The articles were finalised according to Preferred Reporting Items for Systematic Reviews (PRISMA) guidelines. PubMed/MEDLINE, Scopus, CINAHL, SPORT Discus and Web of Science databases were searched and assessed by title, abstract, full-text and a bibliography assessment of included full-text articles. The search strategy included keywords of cricket fast bowling, biomechanics and performance analysis. A validated scale was developed to evaluate the quality of studies and risk of bias. Across the 15 studies finalised for the review (from 2000--2022), there was consensus in some of the characteristics reported to be associated with ball release speed; however, there was disagreement or limited evidence for others. A faster run-up, efficient and sequential transfer of linear to angular momentum, quick deceleration during the delivery stride with an extended front knee from front foot contact to ball release and delaying the bowling arm leading to ball release were the major techniques variables consensually associated with ball release speed. The lack of standardisation of cohort ability and protocols may have contributed to contradictory findings across studies.

1. Introduction

Fast bowling is a highly individualistic and specialised skill within the game of cricket. Bowling strategy is devised to restrict runs by either reducing scoring opportunities or achieving a dismissal. Fast bowlers aim to achieve these outcomes by using strategies such as generating high ball release speed, accuracy and deception [1]. The most widely researched strategy is maximising ball release speed (BRS), where bowlers have been reported to bowl in excess of 40 m/s (90 mph) [2]. High BRS reduces the batter’s reaction time and may hamper the ability to interpret the delivery, resulting in poor shot selection and execution [2]. To achieve these high speeds, the bowler completes a run-up to the crease before a leap into the delivery stride [2,3,4,5]. Mechanically, the technique then involves efficiently converting some of the whole-body linear momentum generated during the run-up into angular momentum which is sequentially transferred in a proximal to distal sequence from the torso, down the arm and into the bowling hand, eventually resulting in a high tangential release velocity of the ball (BRS) [2,6].

While the ability to generate high BRS is a major characteristic in defining a successful fast bowler, bowling accuracy [1,7] and consistency [1] are also important. Bowling accurately and consistency in combination with a well-placed field can restrict the opportunity for scoring runs and place a batter under stress [1]. Additionally, lateral variation of ball flight trajectory through movement in the air (swing) or movement off the pitch (seam), along with changes in pitch length, pitch line and ball speed, are strategies fast bowlers may use to create uncertainty for the batter. This, when coupled with speed and accuracy, often challenges the batter’s shot selection [5]. However, the majority of research in fast bowling has focused on BRS, with limited research exploring the biomechanics associated with these other bowling strategies.

Several studies have researched various combinations of biomechanical variables with the aim of developing a mechanical understanding of how fast bowling techniques are associated with BRS [2,8,9,10,11,12]. These variables may vary between individual bowlers due to differences in morphology and physical characteristics [13]. The majority of these studies have explored time-discrete kinematic and kinetic variables at specific time points, such as back foot contact (BFC), front foot contact (FFC) and ball release (BR), with inferences that are a generic representation of the cohort studied [2,3,10,12].

There have been contrasting findings in studies linking bowling technique to BRS. For instance, while Ferdinands et al. [3] found controlled rear leg knee flexion for just post half the delivery stride phase to be associated with BRS, Middleton et al. [10] found no correlation between these two variables. Moreover, the proposed role of the rear knee during the delivery stride is different across studies [3], and even within cohorts in the same study [10]. Consequently, the ambiguity of findings in the current literature provides a challenge in translating fast bowling performance research into applied practice.

Principally, there is a need to distil and synthesise the literature to establish variables that are consistently and accurately associated with enhancing fast bowling performance. Hence, the objective of the review is to present a comprehensive, critical and objective analysis of the biomechanical characteristics that have been associated with fast bowling performance in the literature. It aims to provide coaches with evidence-based insights into the technique variables that are critical for optimising fast bowling performance, thereby informing training strategies.

2. Materials and Methods

The articles were systematically searched and finalised based on the PRISMA statement for conducting and reporting systematic reviews [14]. A narrative review methodology thereafter was adopted for comprehensive, critical and objective analysis of the technique parameters studied to be associated with fast bowling performance.

2.1. Search Strategy

PubMed/MEDLINE, Scopus, CINAHL, SPORT Discus and Web of Science databases were searched online. Only English-language, peer-reviewed journal articles were considered for the study. The review considered journal articles from the year 2001 until 2022, as data collection methods used before 2001 are mostly obsolete. The search used keywords with Boolean phrases (AND/OR/AND NOT; see

Table 1. Search string for initial filtering of the articles.

Search String

(“cricket bowling” OR “cricket bowlers” OR “cricket fast bowlers” OR “cricket fast bowling” OR “fast bowling” OR “fast bowlers” OR “pace bowlers” OR “pace bowling” OR “cricket pace bowlers” OR “cricket pace bowling”) AND (“Biomechanics” OR “performance analysis” OR “performance mechanics” OR “kinematics” OR “kinetics” OR “technique”) AND NOT (“injury” OR “injury prevention” OR “rehabilitation” OR “injury management”) AND NOT (“spin bowling” OR “finger spin bowling” OR “spin bowlers” OR “swing bowling” OR “swing bowlers”)

).

The articles were screened by title, abstract and full text for final inclusion. A second reviewer performed an independent assessment of the studies identified for the inclusion in the review.

The inclusion criteria applied were as follows:

(1)

A well stated research question aimed at understanding the relationships between biomechanical variables and fast bowling performance.

(2)

A well-stated, repeatable experimental methodological approach to study fast bowling performance biomechanics.

Both primary and secondary reviewers examined full texts of the articles according to the inclusion criteria. Further additional articles were identified by means of citations within the examined papers. The PRISMA flow diagram [14] was used for reporting the selection process, as shown in Figure 1.

2.2. Quality Assessment Scale

The quality of individual studies and risk of bias (Appendix A) were rated by two authors using the 20-question Appraisal Tool for Cross Sectional Studies [15] developed to assess standard of quality of reporting by cross-sectional studies. AXIS evaluates the following: study quality design (clarity of aims, design, rationale for sample size and selection); methodology (validity and reliability of data collection, analysis and statistics); reporting of results, discussion and conclusions; and declaration of funding sources, conflicts of interest, and statement of informed consent and ethical attainment. Three of the questions in this tool referred to non-responders, or response rates (Questions 7, 13 and 14), that were not applicable to biomechanical studies of fast bowling performance and were therefore excluded, so studies were graded out of 17. Studies with a total score of 14–17 were classified as high quality, 10–13 classified as medium quality and ≤9 as low quality [15].

3. Results

The initial search resulted in 262 articles from which 156 duplicates were excluded post exporting to Endnote (X11, Clarivate, Philadelphia, PA, USA). Additionally, nine articles were identified from other sources. A total of 15 studies met the inclusion criteria and were finalised for the review. The studies were scaled individually for risk of bias by the two reviewers. The reviewers were in mutual agreement regarding the quality of the studies graded. Most studies were classified as high quality (n = 10), a few as medium quality (n = 4) and one study as low quality. The low-quality study was included in the review to ensure a comprehensive evaluation of the literature on fast bowling performance technique. Despite its lower quality, the article contributes to a complete overview of current research, crucial for identifying gaps or inconsistencies in the current knowledge base.

3.1. Study Characteristics

All the studies investigated fast bowling biomechanics associated with BRS. A total of 11 studies included in the review adopted a cross-sectional study design and the other four studies adopted a longitudinal comparative [16], retrospective comparative [17] or cross-sectional comparative [10,18] approach. All studies were conducted in a laboratory setting. While all studies tested male fast bowlers, the cohort varied across age (16–27 years), playing experience (recreational to elite) and bowling speed (25–40 m/s). Various data collection methodologies were used, namely (1) a combination of motion capture camera system with video cameras (n = 6), (2) only motion capture camera system (n = 4) and (3) only video cameras (n = 5). The bowling protocol required bowlers to deliver anywhere between 1 and 48 balls, and the selection of deliveries to be analysed varied greatly across studies (

Table 2. Study quality, input source, sample size, joints/segments studied, protocol, delivery analysed and statistical approaches adopted across studies.

Study	Study Quality	Study Design	Source	Sample Size	Mean Age (y)	Mean BRS (m/s)	Cohort	Protocol	Delivery Analysed	Statistical Test
Portus et al., 2004a [16]	Medium (13/17)	Longitudinal and comparative	2D video and force plate	N = 42	22.4 ± 3.5	Not mentioned	Australian Institute of Sport (AIS) high performance bowlers	Data during 1996–1999	Not mentioned	Pearson product moment correlations
Loram et al., 2005 [19]	Medium (13/17)	Cross-sectional	2D video	N = 12	16.6 ± 0.7	30.8 ± 1.8	School and college bowlers (South Africa)	3 deliveries	3 accurate deliveries	Multiple regression
Salter et al., 2007 [17]	Medium (13/17)—Single bowler (12/17)—All bowlers	Retrospective study. Comparative study	2D video and 3D motion capture	N = 1 & N = 20	22 ± 1	Between-bowler: 37.5 ± 1; within-bowler: 34.2 ± 1.6	English institute of sport and Australian county standard	20 deliveries	20 deliveries	Multiple stepwise regression between bowlers
Wormgoor et al., 2010 [11]	High (14/17)	Cross-sectional	2D video	N = 28	22 ± 3	34 ± 1.3	Premier club grade (South Africa)	6 deliveries	Single delivery	Two-tailed Pearson’s product-moment correlation coefficients
Ferdinands et al., 2010 [8]	High (14/17)	Cross-sectional	3D motion capture and force plate	N = 34	22.3 ± 3.7	32.1 ± 2.6	Premier club, district, first-class and international (country not mentioned)	10 deliveries	Single fastest good length delivery	Stepwise multiple regression
Zhang et al., 2011 [12]	High (14/17)	Cross-sectional	3D motion capture	N = 8	22.9 ± 2.9	29.76 ± 1.68	Senior club level (New Zealand)	8 deliveries under different conditions: sub-max, max, max with lower trunk flexion	Avg. of each condition	One way ANOVA
Ferdinands et al., 2013 [6]	Medium (13/17)	Cross-sectional	3D motion capture and force plate	N = 34	22.3 ± 3.7	31.9 ± 2.8	Premier New Zealand club grade	6 deliveries	Fastest delivery	Multiple linear stepwise regression model
Worthington et al., 2013a [2]	High (15/17)	Cross-sectional	3D motion capture and force plate	N = 20	20.1 ± 2.6	34.9 ± 1.7	England cricket board’s (ECB) elite bowling squad	6 max. velocity deliveries	3 fastest trials	Linear regression
Ferdinands et al., 2014 [3]	Low (9/17)	Cross-sectional	3D motion capture and force plate	N = 18	17.2 ± 1.7	Not mentioned	New South Wales (Australia) development squad	20 trials for each bowler	5 fastest trials of good length	Bivariate Pearson’s product-movement correlation coefficients
Glazier & Worthington, 2014 [9]	High (15/17)	Cross-sectional	3D motion capture and force plate	N = 20	20.1 ± 2.6	34.9 ± 1.7	England cricket board’s (ECB) elite bowling squad	6 max. velocity deliveries	3 fastest trials	Two-tailed Pearson’s product-moment correlation coefficients
King et al., 2016 [20]	High (15/17)	Cross-sectional	3D motion capture and force plate	N = 20	20.1 ± 2.6	34.9 ± 1.7	England cricket board’s (ECB) elite bowling squad	6 max. velocity deliveries.	3 fastest trials	Pearson product moment correlations
Spratford et al., 2016 [5]	High (14/17)	Observational and cross-sectional	3D motion capture	N = 69	21.28 ± 4.48	33.9 ± 1.5	AIS National under 19, senior state and international level bowlers	24 randomised deliveries varying in short, good and full length.	Mean of 24 deliveries	MANOVA
Middleton et al., 2016 [10]	High (16/17)	Cross-sectional Comparative	3D motion capture and force plate	N = 30	20.25 ± 2.75	29.4 ± 3.2	Australian state level and amateur competitors	5 overs at match intensity	Avg. of 4 deliveries	Pearson product moment correlations
Feros et al., 2019 [21]	High (15/17)	Observational and cross-sectional	2D video camera	N = 31	21.7 ± 4.7	Not mentioned	Australian club standard	8 overs (42 at match intensity)	Avg. of 4 max. effort deliveries	Spearman’s rank order correlations
Kiely et al., 2021 [22]	High (17/17)	Cross-sectional	3D motion capture and force plate	N = 20	22.1 ± 4.4	34.3 ± 1.9	U-17 and U-19 Australian state fast bowlers	12 max. intensity deliveries	Avg. of all	Stepwise regression

). Statistical approaches to determine the relationships between biomechanical variables and fast bowling performance (ball speed, accuracy) also varied across studies (Table 2).

3.2. Biomechanical Variables Studied

The common variables studied were as follows: front knee kinematics (n = 10) at front foot contact (FFC) and ball release (BR); delivery stride length (n = 7); run-up speed (n = 3); shoulder kinematics (n = 7) at FFC and BR; centre-of-mass (COM) kinematics (n = 6); shoulder and hip/pelvis separation angles (n = 4) at back foot contact (BFC) and BR; trunk kinematics (n = 5) at BFC, BR and between FFC and BR. Among these, the variables that were consistently associated with high BRS were knee angle at BR, run-up speed, shoulder angle and trunk flexion between FFC to BR (

Table 3. Summary of major biomechanical variables linked to fast bowling performance.

Technique Component	No. of Studies	Study Quality	Phase/Event	Relationship
Run-up	positive relationship (n = 3)	3—High	Run-up	Strong evidence that high run-up speed is linked to BRS
Stride length	positive relationship (n = 1) no relationship to BRS (n = 6)	6—High 1—Medium	Delivery stride	Most studies found no significant relationship to BRS
Bowling Shoulder	positive relationship (n = 3) no relationship (n = 4)	4—High 3—Medium	Delivery stride	Moderate evidence that a delayed bowling arm linked to BRS
Front knee	positive relationship (n = 8)	6—High 3—Medium	BR	Strong evidence that an extended front knee is linked to BRS
Back knee	positive relationship (n = 1) no relationship (n = 1)	1—High 1—Low	Delivery stride	Weak evidence that the back knee movement is linked to BRS
Trunk	positive relationship (n = 2) no relationship (n = 2)	3—High 1—Medium	FFC to BR	Moderate evidence that trunk flexion between FFC and BR is linked to BRS
COM kinematics	positive relationship (n = 6)	5—High 1—Low	Delivery stride	Strong evidence that deceleration of COM linked to BRS

). Stride length, COM kinematics and shoulder/hip separation angles were not consistently associated with BRS across studies.

Two studies investigated the relationship between technique variables and accuracy. Spratford et al. [5] found a functional difference between short and long stride length group bowlers; as stride length increased, accuracy decreased across the bowlers. Zhang et al. [12] found that accuracy decreased when bowlers bowled at their maximum speed.

4. Discussion

The review highlights the breadth of recent research exploring biomechanical technique parameters associated with fast bowling performance. It further provides a comprehensive, critical and objective analysis of the nature and methodology of the variables studied, and their association with fast bowling performance metrics. The majority of fast bowling performance research considered BRS to be the primary indicator of performance and hence technique parameters associated with BRS were typically studied. Some variables were consistently found to be linked to fast bowling performance across all studies, while a few variables were reported to have an inconsistent relationship to BRS.

Varied methods were adopted to study the major contributors known in the literature to be associated with fast bowling performance. With a sample size of 18 and above across most studies, the studies provide a foundational understanding of fast bowling techniques associated with BRS. Although studies with limited sample size may provide valuable initial insights, the generalizability of the findings to a broader population should be carefully considered [12,19]. Relationships between run-up speed [2,20,21], front knee angle at BR [2,10,11,17,19,20,21], rear knee angle at BFC [3,10], trunk kinematics from FFC to BR [2,6,12,20], position of the bowling arm at BR [2,8,11] and BRS have been widely studied.

4.1. Performance Parameters Associated with Fast Bowling Performance

The bowling action is usually broken down into run-up, pre-delivery stride, delivery stride and follow through. The run-up phase is initiated at the starting point of the approach run until BFC. The pre-delivery stride is defined as the stride prior to the delivery stride, which may also be referred to as the jump and gather phase. The delivery phase is defined from BFC to BR. Within the delivery phase, the events of BFC, FFC and BR are the major time points studied. Lastly, follow-through involves the continuation of the bowling action post BR.

4.1.1. Run-Up and Pre-Delivery Stride

The run-up phase is important for the fast bowler as it allows for the development of the required linear momentum for the delivery phase [21]. The run-up velocity has been positively correlated to BRS [2,20,21], with run-up velocities ranging from 4.7 m/s to 6.7 m/s [2,21]. Faster run-up velocity results in greater linear momentum that can be transferred into angular momentum of the torso and arm, potentially resulting in greater BRS [2]. Similarly, Middleton et al. [10] reported that COM velocity at BFC (estimated as run-up speed) was correlated to BRS for a high performance group, and Glazier and Worthington [9] found that horizontal velocity of the COM at BFC and FFC were significantly correlated with BRS.

4.1.2. Delivery Phase

The delivery stride length has been popularly studied for its relationship with BRS by several researchers (n = 7; six high quality and one medium quality). It is most likely that there is little effect of delivery stride length on BRS, where six studies found no association between delivery stride length and BRS [5,8,10,11,17,22], with just one finding a positive association between delivery stride length and BRS [21]. Greater stride length, however, may be detrimental to bowling accuracy. Spratford et al. [5] reported that the increase in stride length could be associated with the speed accuracy trade-off, where an increase in stride length positions the body COM further behind the front foot, possibly resulting in decreased accuracy.

4.1.3. Rear Leg Kinematics

Two studies, one low [3] and one high [10] quality, investigated rear leg kinematics during the delivery phase. Ferdinands et al. [3] demonstrated that the rear leg undergoes controlled knee flexion over more than half of the delivery phase before it initiates extension. Bowlers who produced rapid extension of the knee as part of this effective rear leg drive mechanism were inclined to produce higher BRS. Meanwhile, Middleton et al. [10] found no significant correlation between back knee flexion and BRS.

4.1.4. Front Leg Kinematics

Front knee kinematics have been widely researched (quality of studies; low n = 1, medium n = 3, high n = 6). Studies have consistently demonstrated that a more extended front knee at FFC and BR has been associated with greater BRS. An extended front knee provides a rigid lever, slows down the linear pelvis velocity and transfers momentum generated in the run-up into more distal segments (such as the trunk), permitting them to rapidly rotate over the front leg [2,11,19,20].

Middleton et al. [10] reiterated a long-held [16,23] belief that a flexor-extender knee action is ideal. Portus et al., 2004 [16] postulated that flexing the front knee post FFC could reduce the impact forces, which may be beneficial for bowlers with a propensity for back injuries. Subsequently, extending the knee from FFC to BR may cause COM deceleration and result in transfer of energy to the trunk and bowling arm, as shown by Ferdinands et al. [8].

4.1.5. Trunk, Pelvis and Centre-of-Mass Kinematics

Trunk and pelvis kinematics have been reported by several studies (n = 10 high-quality and n = 5 medium-quality studies). Trunk and pelvis rotation contributed most to high BRS following upper arm rotation contribution [12]. Increased trunk flexion allows for large rotational energy to be transferred to distal segments and eventually to the ball [6]. Specifically, the magnitude of upper trunk flexion between FFC and BR was reported to be associated with higher BRS [2]; the mechanism leading to increased upper trunk flexion remains unexplained. Additionally, the hip–shoulder separation angle, defined as the angle between the vector joining the hip joint centres and the vector joining the shoulder joint centres, has also been considered an important variable associated with bowling speed (n = 4; two high quality and two medium quality). A relationship between maximum hip–shoulder separation angle during the delivery stride was observed, with bowlers with larger separation angles demonstrating higher BRS [16]. However, Ferdinands et al. [6] found no correlation between hip–shoulder separation angle and BRS.

The kinematics of the mass centre were considered in several studies across the review (n = 6; five high-quality and one medium-quality study). The kinetic link principle specifically indicates that, for effective transfer of full-body momentum gained during the run-up to the bowling arm, a sequential deceleration of proximal segments during the delivery stride is required [8]. Therefore, for efficient transfer of momentum, COM deceleration promotes rapid accelerations of the distal segments (via trunk flexion and bowling arm circumduction), generating high ball speed. Ferdinands et al. [8] found that COM deceleration in the delivery stride phase was the most important kinematic factor associated with ball speed. Middleton et al. [10] also stressed the importance of quick COM deceleration in addition to a quick run-up for generating high BRS. Supporting previous findings, Kiely et al. [22] observed that COM acceleration during the delivery stride and FFC to BR phase had a large negative correlation with ball speed. However, rather than during the entire delivery stride, Glazier and Worthington [9] found a significant negative relationship between BRS and average horizontal COM acceleration during FFC to BR only.

4.1.6. Bowling Arm and Shoulder Kinematics

The association between shoulder kinematics (at FFC and BR) and BRS was assessed by a total of seven studies included in the review (study quality; high, n = 4 and medium, n = 3). Shoulder flexion angle at FFC has been observed to be strongly correlated to BRS [2,11,16]. A larger shoulder flexion angle at FFC allows for a greater range of motion, during which the arm can rapidly extend and may result in greater torques to generate BRS [11,16].

Worthington et al. [2] found that the individual factor most strongly linked to high BRS was shoulder angle at BR. Mechanically, having the bowling arm further back relative to the upper trunk as the ball was released delayed the onset of arm circumduction and led to larger trunk flexion, resulting in more rotational energy being transmitted to the ball [2,6,12].

4.1.7. Segmental Sequencing

The kinetic link principle states that there is a relationship between the temporal order of segmental motions, referred to as segmental sequencing, and end effector velocity (in cricket the final segment to impart velocity to the ball) [6]. Specifically, a proximal-to-distal sequence optimises end-effector velocity in activities involving the coordination of multiple body segments. Two studies observed segmental sequencing in fast bowlers using different methods [6,12]. Zhang et al. [12] studied the segmental sequencing adopted by the bowler by calculating the angular velocities of the pelvis, torso, thorax and upper and lower bowling arm and found a proximal-to-distal sequence of body segment motion executed by all bowlers. Ferdinands et al. [6] also found similar results and, in addition, found a distinct temporal sequencing of peak linear and rotational kinetic energy and peak angular velocities, which followed a general proximal-to-distal order.

4.1.8. Bowling Delivery Length, Accuracy and Type

Another important aspect of bowling kinematics is delivery length. The environmental and pitch conditions, match format and opposing batter can influence the fast bowlers’ tactical decision to experiment with varying delivery length along with type of delivery and BRS [24]. Frequent variation in trajectory of the bowling delivery may increase the difficulty of batting. It is therefore advantageous for the fast bowler to possess the skill required to vary ball trajectories. The review found no study investigating the relationship between technique and bowling delivery length and type.

4.1.9. Inter and Intra-Individual Analyses

Experimental research generally involves averaging data derived from a range of athletes to develop a mechanical understanding of a skill [25]. However, while examining elite athlete performance, intra-individual analyses is crucial, as individual differences exist when athletes execute a skill [26]. Salter et al. [17] was the only study that compared between-bowler to within-bowler methodological approaches to identify associations between technique and BRS. The within-bowler method provided detailed information about how the individual bowler’s technique was associated with BRS. Specifically, they explained 87.5% of the variation in the bowler’s BRS using a stepwise multiple regression including run-up velocity, angular velocity of the bowling arm, vertical velocity of the non-bowling arm and stride length. Hence, they found the within-bowler methodology provided a mechanical understanding of the influence of technique on BRS in an individual bowler that was absent when a between-bowler methodology was adopted. However, the comparison between a within-bowler approach and a between-bowler approach was studied with just a single bowler.

5. Limitations and Future Directions

Time-discrete variables coinciding with key bowling events have been identified and linked to fast bowing performance; nevertheless, the underlying technique adopted by individuals to achieve these performance parameters has not been extensively studied. Hence, the translation of these technique factors to coaching practice and improved bowling performance has been limited.

Both consistent and inconsistent associations between fast bowling performance and technique parameters have been reported in the literature. The ambiguity of these relationships between technique variables and performance could be a result of grouping bowlers in ways which resulted in inter-individual differences being more or less pronounced. Inconsistencies in research design, cohort, testing protocol, variable definitions and data analysis techniques could also contribute to the observed ambiguity.

Given that the articles in the review studied exclusively male fast bowlers, there is a need to be cautious when applying the identified technique parameters to female fast bowling coaching practices. Moreover, these technique parameters have typically been associated with BRS. Although producing a high BRS is a major aim of fast bowlers, the ability to bowl accurately, and to alter delivery length and type, are also important. Hence, there is a greater need to study the variations in technique required to achieve these aims and ultimately improve fast bowling performance.

Traditionally, statistical methods usually require reduced, simplified datasets that have the potential to limit extensive understanding of the data. Advancements in data science and technology have created new opportunities in the field of human performance research that can overcome the limitations of traditional methods of notational analysis (for example, manual time-intensive tasks). Advances in statistical software, like statistical parametric mapping (SPM12), has demonstrated the potential to study spatio–temporal biomechanical data for statistical inferences. Additionally, machine learning has demonstrated the potential to handle complex datasets, gain new insights from data by identifying relationships between variables and discovering patterns in data. Although machine learning has seen recent applications in the estimation of kinetics and kinematics in sport, it is yet to be progressively explored in sport biomechanics performance research.

All studies were restricted to the lab environment and differed in measurement techniques, data collection protocol, cohort, deliveries analysed and statistical approaches. Inertial measurement units (IMUs) have demonstrated high potential as an alternative to the 3D motion capture technologies, allowing biomechanical testing in the field. The measurement of physical loads and the analysis of movement parameters have seen recent advancements with the introduction of these sensors. With all studies in the review, based in the laboratory environment, future research should potentially consider the evaluation of fast bowling performance biomechanics on-field in real-world environments.

6. Conclusions

The aim of this review was to identify and summarise the variables associated with fast bowling performance studied in the literature to date. Most studies used ball release speed as their sole performance variable. To summarise, bowling techniques linked to increased ball release speed involved the efficient transfer of linear momentum of the centre of mass generated during the run-up to angular momentum of the distal body segments. Bowlers achieve this by maintaining an extended front knee from front foot contact to ball release, generating angular momentum of the torso, leading to an increased range of thoracic motion between front foot contact and ball release. Delaying circumduction of the bowling arm eventually allows a greater transfer of momentum generated by the proximal segments to the ball. Additionally, several variables identified were not consistently associated with ball release speed across all studies. This ambiguity could be potentially due to differences within sampling population, data collection techniques, or statistical approaches. There is a need to rationally standardise data collection protocols, biomechanical modelling methods and data analyses techniques to progress towards accurate, reliable and precise results across studies. Lastly, although ball release speed is undoubtedly an important performance variable, there is also a need to study the ability of fast bowlers to vary bowling trajectories consistently and accurately to successfully confuse batters.