Does the presence of an opponent affect object projection accuracy in elite athletes? A study of the landing location of the short serve in elite badminton players

Abstract

The ability to accurately project (e.g. throw, kick, hit) an object at high speed is a uniquely human skill, and this ability has become a critical feature of many competitive sports. Nonetheless, in some sports, the target or end-point for a projected object is often not reached because an opponent intercepts or returns the object; thus, a player cannot use object landing location information to inform accuracy outcome. By comparing the landing location of serves performed without an opponent by elite badminton players to predicted landing points of serves delivered with an opponent, we aimed to determine whether object projection accuracy is affected by the presence of an opponent. Landing locations of serves to an opponent were predicted using a model developed through analysis of serves without an opponent present. The model predicted that 69% of serves to an opponent would have landed on or short (i.e. outside the permitted area) of the service line. Thus, serve trajectory in elite badminton players was considerably altered by the presence of an opponent, despite their aim to serve to a specific point on the court.

Keywords

Kinematics motion analysis perceptual-motor skill racket sport shuttlecock trajectory

Introduction

The ability to repeatedly throw, kick or hit an object with both high speed and accuracy is an important skill in a variety of sports.¹ Projecting an object rapidly towards a target location has been an essential skill in sports such as squash, tennis and badminton.² The skill of projecting an object to a specific point in space (e.g. on a court or field) has subsequently been a distinguishing factor for performance.^3,4 For instance, when fielding in cricket, the fielder throws the ball at the wickets in an attempt to run out the batsmen^5–7 and in baseball the fielder throws to a baseman with the same aim. Alternatively, in sports such as archery and darts, the projectile must hit a specific area within the target to achieve a higher score.^8–10

When aiming for a particular target, the physical position of the target influences the trajectory chosen by the individual projecting the object¹¹ so informational constraints can strongly influence movement patterns. For example, elite squash players’ shot responses were found to be dependent upon the opponent’s location on the court.¹² This could be due to visual information relating to the position of a moving target (i.e. the opponent), which can influence the specific movement patterns used by a performer.¹³

In elite sport, it has been reported that not only can accuracy be influenced by the presence of an opponent but also the desired landing location (e.g. short serve in competitive badminton).¹⁴ Previous research examining serve locations in squash have typically considered patterns of serves to different areas of the court as being suggestive of tactics. Whilst this does explain a general serve pattern, it does not discriminate between opponents but instead focuses solely on the server.¹⁵ In some sports, an opponent can position themselves between the individual projecting the object and the target, which causes an individual to move from their preferred trajectory pattern. For example, in basketball, the aim of the shooter remains the same regardless if an opponent is present or not. The question is then, does the decision to change the trajectory influence the landing accuracy despite the location of the desired landing location not changing? In some situations, the landing position of the object cannot always be determined (i.e. due to the opponent intercepting the object). For example, in badminton, the shuttlecock is generally returned by the opponent when the short serve is used, so the landing accuracy of the serve cannot be observed.³ Unlike tennis, in badminton a second serve is not permitted if a fault occurs.

The short serve in badminton (the most frequently used service stroke in the doubles discipline) requires a high degree of accuracy so that the shuttlecock has a downward trajectory as it passes the apex of the net, which forces the opponent to hit the shuttlecock back over the net with a high angle of trajectory,^3,8 yet would land just past the service square line if the shuttlecock was not returned. This makes it easier for the server to hit an offensive shot and score a point. In tennis, the shot direction has been shown to be partly based on the opponent’s positioning along the baseline,¹⁶ whilst in squash, the evidence suggests that the posture of the receiving player influences the shot choice of the server.¹⁷ The findings from these studies suggest that the actual on-court positioning of the opponent may not only influence shot selection but also the trajectory of the projectile.

In training, it is common for badminton players to practice short serve execution by using floor targets placed in the opponent’s service square without an opponent present, just as it is common in many sportspersons to throw, kick or hit and object to a target in practice. Performance changes with training are then evaluated based on the landing location of the shuttlecock within the target area.^3,18 However, in a match, the typical measure (i.e. landing accuracy) used to determine short serve accuracy is not available to the server because the shuttlecock is intercepted by an opponent before it can reach the end-point (i.e. the opponent’s service square).

Using landing location during practice removes the presence of an opponent, yet both elite coaches and players admit that their aim may be influenced by the presence of an opponent when performing the short serve.¹⁴ In sports such as soccer, it has been shown that a penalty taker may be inclined to direct the shot to a larger space to one side of the goalkeeper if the goalkeeper stands slightly off-centre.¹⁹ In basketball, the defending player uses their positioning on the court, outstretched arms, and jumping to not only intercept the shot but also in an attempt to put the shooter off.²⁰ However, in badminton, the receiver can also use their position on the court to influence the trajectory chosen by the server. But, in this situation, the actual trajectory of the shuttlecock, as well as the landing location, might change. Thus, the question arises as to whether elite exponents of these sports still use object trajectories that would be classified as ‘accurate’ based on likely landing location when an opponent is present, or whether they change their strategy despite retaining the aim of hitting a particular point or target regardless of the presence of an opponent.

Given the above, the purpose of the present study was to determine whether there was a difference between the targeted landing location of the shuttlecock in training (i.e. without an opponent) and the predicted landing location in a match-like environment (i.e. with an opponent). We hypothesised that when an opponent was present, the shuttlecock would sometimes fall short of the opponent’s service line, which may be due to the presence of an opponent, and also to the trajectory being a more important factor for short serve accuracy than the landing location.

Methods

Two separate data collection sessions were completed (Sessions 1 and 2) with 13 players tested in total (age: 23.4 ± 5.1 years, body mass: 73.2 ± 11.1 kg, height: 175 ± 8.6 cm). In Session 1, five elite badminton players performed 30 short serves each without an opponent so that the trajectory of the shuttlecock from racquet contact to landing could be recorded. The full shuttlecock trajectory from Session 1 was used to fit a model in order to predict the landing location when an opponent was present (i.e. Session 2 – where the full trajectory is not available since an opponent returns the shuttlecock (see Figure 1)). To determine the accuracy of the developed model, the full trajectory data were used to compare predicted landing position to the actual landing position. In Session 2, eight elite badminton players performed 30 short serves with an opponent present (i.e. opponent returned the shuttlecock). The prediction model was applied to these serves to estimate landing location with an opponent present. The study was approved by Edith Cowan University Human Ethics Committee, and players gave informed consent prior to testing.

Figure 1.

Top – schematic of Session 1, where the full shuttlecock trajectory (i.e. racquet contact to ground) was recorded in order to create a landing prediction model. Bottom – schematic of Session 2, where only a partial trajectory of the shuttlecock was captured because the opponent returned the serve.

Testing protocol

All players were free from injury at the time of testing, wore their normal match attire and used their own racquets. A badminton court was marked out, and a net was placed in accordance with the International Badminton Federation standards in the biomechanics laboratory at Edith Cowan University. The motion-capture system consisted of 8 cameras in Session 1 and 22 cameras in Session 2, set to record at a 250-Hz sampling rate (VICON Oxford Metrics Ltd, Oxford, UK). Prior to data collection, the capture volume was calibrated using standard VICON procedures. Reflective tape was placed around the base of the head of the shuttlecock to allow visualisation by the system. Players performed badminton-specific play for warm-up and to familiarise themselves with the testing environment. The players then performed 30 serves to the most common area during a competitive match – the edge of the opponents’ service square near the line that bisects the service squares. Shuttlecock trajectory data were filtered using a 6-Hz low-pass fourth order Butterworth filter, which was determined using residual analysis.²¹

Prediction model and error distribution of shuttlecock landing

In both data collection sessions, a nonlinear least-squares regression model was used to characterise the horizontal position ( $y$ ) of the shuttlecock as a function of time ( $t$ ), for the data beginning shortly after when the shuttlecock left the racquet and ending with the apex (highest point) of the shuttlecock trajectory. With simple ballistics, horizontal velocity remains constant as there are no horizontal forces acting on the mass. However, the horizontal velocity of the shuttlecock decreases over time due to viscous damping effects (i.e. drag), which remove kinetic energy from the system. As such, a power fit with offset ( $y = a t^{b} + c$ ) was used to allow for a decreasing slope ( $b < 1$ ) over time, where $c$ is the initial horizontal position of the shuttlecock and $a$ and $b$ are regression parameters fit the data.

A second nonlinear least-squares regression model was used to characterise the vertical position ( $z$ ) of the shuttlecock as a function of horizontal position ( $y$ ). Specifically, a third-order polynomial ( $z = {c_{3} y}^{3} + c_{2} y^{2} + c_{1} y + c_{0}$ ) was used instead of a second-order polynomial (appropriate for simple ballistics) as the angle of the shuttlecock varies during flight, and the altered air flow causes a change in trajectory. A shuttlecock has an open conical shape formed by feathers which are embedded into a cork or rubber base. This unique shape causes the shuttlecock to ‘tumble’ immediately after the racquet has made contact since most of the mass is held within the base of the shuttlecock. Aerodynamic damping eventually stabilises the tumbling until the shuttlecock travels with curvilinear (but not rotational) motion. This results in a trajectory more complex than that of simple ballistics.²² Although the third-order polynomial fit does not specifically describe the effects of aerodynamic damping in a physical sense, this model was chosen to add a higher degree of freedom allowing for trajectory slopes not fully explained by simple ballistics.

In order to extrapolate the shuttlecock’s trajectory utilising both regression models, horizontal position was first predicted for an arbitrary but sufficient length of time (since predicted time of contact with the ground was not yet known). Next, the vertical position was extrapolated as a function of predicted horizontal position. Finally, the time of predicted ground contact was determined as the point at which the predicted vertical position changed from positive to negative (i.e. the shuttlecock would fall below the ground), and the final horizontal position of the shuttlecock was recorded at this time. This value was then compared to the location of the serving line.

To assess the accuracy of extrapolating the shuttlecock’s trajectory, error was quantified by comparing predictions via the approach detailed above with the measured trajectories of serves allowed to fully reach the ground (in Session 1). The 150 short serves performed in total by the five subjects were characterised with the regressions, and the predicted landing position was compared to measured values. Bland–Altman plots were used to assess the error between the curve-fitting model and the recorded trajectory (i.e. between model predictions and measurement) against mean value (horizontal and vertical landing position).

Using the error distribution quantified by the Bland–Altman plots in Session 1, the predicted landing position in Session 2 was determined before or over the service line if the distance was greater than the 95% limit of agreement. However, if the distance was less than the limit of agreement, the serve was considered to fall on the serving line. The proportion of serves falling before, on or over the line was determined for each player as well as the average across all players.

Results

The results from Session 1 revealed no significant bias between the actual landing position and predicted landing position (i.e. no systematic error). The mean horizontal error (anterior–posterior) was 6.0 ± 14.5 cm, and the vertical error (up–down) was 9.1 ± 16.3 cm.

The prediction model was subsequently applied to the shuttlecock trajectories with an opponent present (Session 2). The predicted landing locations for all serves for each player are shown in Figure 2. Overall, the predicted landing location of the shuttlecock before, on or after the service line was normally distributed (p > 0.56). Table 1 shows the individual player landing predictions for serves landing before, on or over the line. Data relating to serves ‘on the line’ were included as a separate variable because a large proportion of serves were predicted to land on some part of the service line, which is 40 mm wide. It was apparent that for some players a higher proportion of serves landed in certain areas: for players 5 and 8, serves were predicted to land over the line; for players 1, 2, 3 and 6, the majority of serves were predicted to land on or before the line for players 4, 5, 7 and 8.

Figure 2.

Predicted landing location for all serves for all players in Session 2 (with an opponent). The service line is located at 0 m on the y-axis. Positive values indicate that the serve landed over the service line, whilst negative values indicate that it landed before the service line. Sixty-nine percent of serves were predicted to have landed on or before the service line.

Table 1.

Individual landing location predictions for serves landing before, on or over the service line for each player (Session 2).

Landing location (%)
Player	% before	% on the line	% over
1	29	55	16
2	40	50	10
3	2	65	33
4	0	20	80
5	70	30	0
6	50	35	25
7	0	10	90
8	69	31	0
Mean	33	37	32
SD	30	18	35

Discussion and implications

The aim of the present study was to determine whether players changed the service trajectory of the shuttlecock sufficiently to change its landing location when an opponent was present. Shuttlecock trajectory data were captured from servers without an opponent present in order to create a model to predict the shuttlecock landing location from the flight path. The prediction model was then applied to shuttlecock trajectory data captured with an opponent present in order to predict the landing location of the shuttlecock, despite it being hit by the opponent. Based on predictions from this model, 69% of all serves performed with an opponent present would have landed on or short of the service line (i.e. they were ‘inaccurate’ serves). The results suggest that the server may choose a different landing location when serving during a match because of the presence of an opponent. The prediction error could have been due to the tumble of the shuttlecock during flight.

The results from the present study indicate that the trajectory of the serve may be influenced by the presence of an opponent. Rojas et al.¹³ reported similar results when analysing the basketball jump shot; i.e. players shot with a different ball trajectory when a defender was present in order to reduce the chance of the defender intercepting the shot. In the badminton short serve, the opponent will often move towards the net to ‘attack’ the serve in an attempt to hit the shuttlecock from as high a point as possible.²³ Therefore, in order to reduce the receiver’s advantage, the server may choose to hit the shuttlecock with a different trajectory to cause it to land on or before the service line. Because a large proportion of the serves (69%) landed on or before the line, it may be inappropriate to practice serving to ground-based targets without an opponent present; it seems imperative that practice replicates the performance environment faced by the athlete as closely as possible. These findings may not only impact serving in badminton, but other racquets sports (i.e. squash and tennis) should consider replicating match-like environments when practicing the serve, as the presence of an opponent may not only influence the end-point of the serve but also the trajectory.

The findings from the present study suggest that landing accuracy alone may not be a good indicator of performance accuracy and that the trajectory of the shuttlecock should be considered when quantifying short serve accuracy. Furthermore, coaches and players competing at an elite level should consider using a training partner when training for short serve accuracy.

Conclusion

In conclusion, the present study revealed that a large proportion of badminton short serves would be ‘inaccurate’ (i.e. would fall short of the opponent’s service square) when an opponent was present. Serving the shuttlecock short might be a subconscious strategy to reduce the advantage of the opponent advancing towards the net in order to hit the shuttlecock from as high a position over the next as possible. Future research should examine the specific factors that might influence the chosen trajectory, including aspects of player (opponent) positioning within the court, stance and anthropometric profile (arm span, height, etc.). Additionally, it would be useful to identify whether elite-level players would benefit from training their ability to select which serves that should not be returned because they would fall short of the service line.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported with funding from Edith Cowan University and Australian Institute of Sport.

ORCID iDs

Shayne Vial

Jodie Cochrane Wilkie

References

Bartlett

Wheat

Robins

Is movement variability important for sports biomechanists?

Sport Biomech 2007; 6: 224–243.

Phomsoupha

Laffaye

The science of badminton: game characteristics, anthropometry, physiology, visual fitness and biomechanics.

Sport Med 2015; 45: 473–495.

Duncan

Chan

Clarke

, et al. The effect of badminton-specific exercise on badminton short-serve performance in competition and practice climates. Eur J Sport Sci 2017; 17: 119–126.

Jardine

Martin

NG.

Spatial ability and throwing accuracy.

Behav Genet 1983; 13: 331–340.

Duffield

Portus

Comparison of three types of full-body compression garments on throwing and repeat-sprint performance in cricket players.

Br J Sport Med 2007; 41: 409–414; discussion 414.

Freeston

Rooney

Throwing speed and accuracy in baseball and cricket players.

Percept Mot Ski 2014; 118: 637–650.

Portus

Farrow

Enhancing cricket batting skill: implications for biomechanics and skill acquisition research and practice.

Sport Biomech 2011; 10: 294–305.

Edwards

Waterhouse

Atkinson

, et al. Effects of time of day and distance upon accuracy and consistency of throwing darts. J Sport Sci 2007; 25: 1531–1538.

Etnyre

BR.

Accuracy characteristics of throwing as a result of maximum force effort.

Percept Mot Ski 1998; 86(3 Pt 2): 1211–1217.

10.

Smeets

Frens

Brenner

Throwing darts: timing is not the limiting factor.

Exp Brain Res 2002; 144: 268–274.

11.

Chow

Davids

Button

, et al. Dynamics of movement patterning in learning a discrete multiarticular action. Motor Control 2008; 12: 219–240.

12.

Vučković

James

Hughes

, et al. The effect of court location and available time on the tactical shot selection of elite squash players. J Sport Sci Med 2013; 12: 66–73.

13.

Rojas

Cepero

Ona

, et al. Kinematic adjustments in the basketball jump shot against an opponent. Ergonomics 2000; 43: 1651–1660.

14.

Vial

Cochrane

Blazevich

, et al. Using the trajectory of the shuttlecock as a measure of performance accuracy in the badminton short serve. Int J Sports Sci Coach 2019; 14: 91–96.

15.

Murray

James

Perš

, et al. Using a situation awareness approach to determine decision-making behaviour in squash. J Sports Sci 2018; 36: 1415–1422.

16.

Loffing

Hagemann

On-court position influences skilled tennis players’ anticipation of shot outcome. J Sport Exerc Psychol 2014; 36: 14–26.

17.

McGarry

Franks

IM.

Modeling competitive squash performance from quantitative analysis. Hum Perform 1995; 8: 113–129.

18.

Edwards

Lindsay

Waterhouse

Effect of time of day on the accuracy and consistency of the badminton serve.

Ergonomics 2005; 48: 1488–1498.

19.

Masters

van der Kamp

Jackson

RC.

Imperceptibly off-center goalkeepers influence penalty-kick direction in soccer.

Psychol Sci 2007; 18: 222–223.

20.

Csataljay

James

Hughes

, et al. Effects of defensive pressure on basketball shooting performance. Int J Perform Anal Sport 2013; 13: 594–601.

21.

Gabriel

Noble

, et al. Estimate of the optimum cutoff frequency for the Butterworth low-pass digital filter. J Appl Biomech 1999; 15: 318–329.

22.

Chen

Pan

Chen

YJ.

A study of shuttlecock’s trajectory in badminton.

J Sport Sci Med 2009; 8: 657–662.

23.

Pinder

Davids

Renshaw

, et al. Representative learning design and functionality of research and practice in sport. J Sport Exerc Psychol 2011; 33: 146–155.