Sie sind auf Seite 1von 18

Sports Biomechanics

ISSN: 1476-3141 (Print) 1752-6116 (Online) Journal homepage: http://www.tandfonline.com/loi/rspb20

Kinetics of the upper limb during table tennis


topspin forehands in advanced and intermediate
players

Yoichi Iino & Takeji Kojima

To cite this article: Yoichi Iino & Takeji Kojima (2011) Kinetics of the upper limb during table
tennis topspin forehands in advanced and intermediate players, Sports Biomechanics, 10:4,
361-377, DOI: 10.1080/14763141.2011.629304

To link to this article: http://dx.doi.org/10.1080/14763141.2011.629304

Published online: 29 Nov 2011.

Submit your article to this journal

Article views: 390

View related articles

Citing articles: 2 View citing articles

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=rspb20

Download by: [University Of South Australia Library] Date: 12 October 2016, At: 03:37
Sports Biomechanics
November 2011; 10(4): 361–377

Kinetics of the upper limb during table tennis topspin


forehands in advanced and intermediate players

YOICHI IINO & TAKEJI KOJIMA

Laboratory of Sports Sciences, Department of Life Sciences, The University of Tokyo, Tokyo, Japan

(Received 22 March 2011; accepted 6 September 2011)

Abstract
The purpose of this study was to determine the significance of mechanical energy generation and
transfer in the upper limb in generating the racket speed during table tennis topspin forehands. Nine
advanced and eight intermediate table tennis players performed the forehand stroke at maximum effort
against light and heavy backspin balls. Five high-speed video cameras operating at 200 fps were used to
record the motions of the upper body of the players. The joint forces and torques of the racket arm were
determined with inverse dynamics, and the amount of mechanical energy generated and transferred in
the arm was determined. The shoulder internal rotation torque exerted by advanced players was
significantly larger than that exerted by the intermediate players. Owing to a larger shoulder internal
rotation torque, the advanced players transferred mechanical energy from the trunk of the body to the
upper arm at a higher rate than the intermediate players could. Regression of the racket speed at ball
impact on the energy transfer to the upper arm suggests that increase in the energy transfer may be an
important factor for enabling intermediate players to generate a higher racket speed at impact in topspin
forehands.

Keywords: Energy generation, energy transfer, joint torque

Introduction
The topspin forehand known as ‘the forehand loop’ is the most effective attacking shot in
table tennis (Seemiller & Holowchak, 1997). The ability to generate high racket speed in the
forehand is an important factor for offensive players to win a match. This is because a higher
racket speed in the forehand results in a higher post-impact ball speed, a greater post-impact
ball spin, or both (Neal, 1991) and gives the opponent fewer chances to hit an attacking shot
(Geske & Mueller, 2010). Thus, understanding the ways by which high racket speed can be
generated in the forehand is of considerable interest for table tennis players and their coaches.
Several studies have investigated the kinematics of table tennis forehands (Neal, 1991;
Kasai & Mori, 1998; Lee & Xie, 2004; Iino & Kojima, 2009). Neal (1991) investigated upper-
limb kinematics of the forehand loop and smash shots in four international-level Chinese
players and four Australian junior players. He found that the hand velocities of Chinese
players were significantly higher than those of the Australian players. Kasai and Mori (1998)
qualitatively compared the kinematics of the whole body in forehand smashes in skilled and

Correspondence: Yoichi Iino, Laboratory of Sports Sciences, Department of Life Sciences, The University of Tokyo, 3-8-1 Komaba,
Meguro-ku, Tokyo 153-8902, Japan, E-mail: iino@idaten.c.u-tokyo.ac.jp

ISSN 1476-3141 print/ISSN 1752-6116 online q 2011 Taylor & Francis


http://dx.doi.org/10.1080/14763141.2011.629304
362 Y. Iino & T. Kojima

non-skilled players. They reported that the skilled players twisted the trunk and bent the
knees to a greater degree than the non-skilled players did. Lee and Xie (2004) investigated the
kinematics of topspin forehands in four elite male players during international competitions.
These studies, however, did not determine the values of racket speed and determined limited
kinematic variables (e.g. only the velocities of the centers of mass of the upper-limb
segments). Thus, how table tennis players can generate a high racket speed during forehand
strokes remains unclear.
Iino and Kojima (2009) investigated the kinematics of the upper limb during table tennis
topspin forehands in collegiate advanced and intermediate players. They found that lower
trunk axial rotation, upper trunk axial rotation relative to the lower trunk, shoulder flexion,
and shoulder internal rotation were the major contributors to the racket speed at ball impact
for both the groups of players. They also found that while the racket speed at impact did not
significantly differ between the two groups, racket speed in the case of the advanced players
increased at a significantly higher rate than it did in the case of the intermediate players.
However, the forces that caused the joint motions in the forehands were not determined in
this study. Thus, how the players produced the racket speed at impact in the forehand in
terms of kinetics and how the advanced players increased racket speed at a higher rate than
the intermediate players did remain unclear.
Determination of joint torques and forces has been a fundamental approach for
understanding the causes of human motions (Winter, 1990). However, to our knowledge, no
study has investigated the kinetics of the upper limb in table tennis forehands.
The table tennis forehand is an open kinematic chain motion that aims to accelerate the
most distal segment, i.e. the racket. When two adjacent body segments rotate in the same
direction, mechanical energy can be transferred from one segment to the other by the
tensions of the muscles that cross the joint formed by the segments (Robertson & Winter,
1980). In table tennis forehands, energy may be transferred from the trunk to the upper limb
segments during forward swing owing to trunk and upper-limb rotations. The translational
kinetic energy of a racket is proportional to the square of the racket speed. Thus,
quantification of the mechanical energy generation and transfer during the forehands may
help table tennis players and their coaches to improve the understanding of what is
biomechanically essential for generating a high racket speed at impact.
The purpose of this study was to determine the significance of the mechanical energy
generation and transfer in the racket arm in the generation of the racket speed during the
topspin forehands in advanced and intermediate table tennis players. The first hypothesis of
this study was that a large part of the mechanical energy of the racket arm at ball impact is
due to energy transfer through the shoulder joint. The second hypothesis was that owing to
higher rates of energy transfer by the limb joint torques and forces, the advanced players
increase racket speed at a higher rate than the intermediate players do.

Methods
Participants and experimental procedures
Data reported in this study were obtained from a previous study (Iino & Kojima, 2009) and
the details of the participants and experimental procedures are provided in that study. Nine
advanced (Division I) and eight intermediate (Division III) male collegiate table tennis
players participated in the study. Three of the advanced players were international players.
Mean ^ SD age, height, and body mass were 20.6 ^ 1.2 years, 1.71 ^ 0.06 m, and
66.2 ^ 9.2 kg, respectively, for the advanced players and 20.6 ^ 1.5 years, 1.70 ^ 0.08 m,
Kinetics of the upper limb 363

and 59.0 ^ 5.7 kg for the intermediate players. Each player used his own shake-hands
grip racket with an inverted rubber sheet on the forehand side. The participants provided
written informed consent before participating in the study. The experimental procedures
were approved by a local ethics committee.
Reflective markers of 20 mm in diameter were attached to the racket arm, shoulders, and
pelvis of each participant (Iino & Kojima, 2009). Three reflective markers of 30 mm in
diameter were also attached to the lateral aspects of the racket. The participants were asked
to hit crosscourt topspin forehand strokes at maximum effort against light and heavy
backspin balls that were projected by a ball machine. Five synchronized high-speed video
cameras (HAS-200, Ditect Co., Ltd., Tokyo, Japan) were used to record the forehands
against both spins at 200 fps. For each participant, at least three successful strokes for each
type of forehand stroke were recorded.
The light and heavy backspin balls with some ink marks were projected by the ball machine
to determine the spin rates immediately after the first bounce on the table. The ball motions
were recorded at 500 fps from a lateral side using a high-speed camera (VFC-2000, FOR.A
Co. Ltd., Tokyo, Japan). The spin rates of the former and latter backspin balls were
11.4 ^ 1.5 rps (n ¼ 7) and 36.8 ^ 2.5 rps (n ¼ 8), respectively.

Reconstruction of 3D coordinates and smoothing


Each participant assessed his own strokes on the basis of the smoothness of swing and the
quality of the ball flight and bounce after impact. The most successful stroke was selected for
each type of forehand stroke for a subsequent analysis. The relevant markers were automatically
digitized using the software Frame-DIAS II (DKH Co., Ltd, Tokyo, Japan). Then, the 3D
coordinates of the markers were determined using a direct linear transformation method
(Abdel-Aziz & Karara, 1971). The coordinate data were smoothed using a fourth-order
Butterworth zero-phase lag digital low-pass filter. The cut-off frequencies of the filter were
determined using a residual analysis (Winter, 1990); the frequencies ranged from 5.8 to 14 Hz.

The segmental and joint coordinate systems


The coordinate systems fixed to the upper trunk, the upper arm, forearm, and the hand of
the racket arm were defined according to a previous study (Iino & Kojima, 2009). The joint
coordinate systems of the shoulder, elbow, and wrist were also defined to decompose the
joint angular velocities and joint torques into anatomically interpretable components such as
shoulder internal rotation velocity and shoulder internal rotation torque.
For the shoulder joint coordinate system, the internal/external rotation axis was a vector
from the elbow joint center toward the shoulder joint center. The flexion/extension axis was
the cross-product of the shoulder internal/external rotation axis and a vector from the wrist
joint center toward the elbow joint center. The adduction/abduction axis was the cross-
product of the shoulder flexion/extension axis and the internal/external rotation axis. For the
elbow/forearm joint coordinate system, the pronation/supination axis was a vector from the
wrist joint center toward the elbow joint center. The flexion/extension axis was the cross-
product of the shoulder internal/external rotation axis and the forearm pronation/supination
axis. The varus/valgus axis was the cross-product of the elbow flexion/extension axis and the
pronation/supination axis. For the wrist joint coordinate system, the internal/external
rotation axis was a vector from the head of the third metacarpal toward the wrist joint center.
The radial/ulnar deviation axis was the cross-product of the wrist internal/external rotation
axis and a vector from the radial styloid toward the ulnar styloid. The palmar/dorsi flexion
364 Y. Iino & T. Kojima

axis was the cross-product of the wrist internal/external rotation axis and the radial/ulnar
deviation axis. Joint angular velocity and joint torque components mean the anatomically
interpretable components in this study unless otherwise stated.

Determination of the joint forces and torques of the racket arm


The joint forces and torques of the racket arm were determined using inverse dynamics. The
hand and the racket were assumed to move as a single rigid segment during the swing; the
angular velocity of the segment was defined as that of the racket. This assumption was made
because the angular velocity of the hand was similar to that of the racket, but the former
velocity seemed less accurate than the latter velocity owing to the small dimension of the
hand. The mass, moments of inertia, and the position of the center of mass of each segment
were determined using the regression equations of Ae et al. (1992). The mass of the racket of
each participant was measured using scales.
The moments of inertia of the rackets were estimated under the assumption that they had
the same radii of gyration as those of a simple table tennis racket model. A racket with a mass
of 180 g was modeled as an object consisting of a circular column and three elliptic columns
(Figure 1). The three elliptic columns piled up so that the contacting ellipsoid surfaces
completely overlapped one another. The middle elliptic column and the circular column
represented the blade of the racket and had an equal uniform density. The two outer elliptic
columns represented the rubber sheets and had a uniform density. The long axis (Y-axis) of
the circular column was aligned with the major axis of the middle elliptic column. The
parameters of the columns and the moments of inertia and radii of gyration of the racket
model about the principal axes of inertia are shown in Table I.
The joint torque of each joint was decomposed into anatomically interpretable
components in its joint coordinate system.

Rates of work and the work done by the joint torque components, and the rates of energy transfer and
the amount of energy transfer by the joint forces and torque components
The rate of work done by a joint torque component, which is hereafter referred to as the joint
torque power, was defined as the product of the torque component and its matching joint
angular velocity component. The rate of work done on a segment by a joint force was defined
as the dot product of the joint force vector and the velocity vector of the joint center; this rate
of work done was the same as the rate of energy transfer by the joint force according to

Figure 1. A table tennis racket model and the coordinate axes fixed to the model.
Kinetics of the upper limb 365

Table I. Parameters of the circular and elliptic columns that constitute a racket model and the inertial properties of
the model.

Moment of Radius of
inertia (kg m2) gyration (m)

Radius (m) Height (m) Mass (g) X Y Z X Y Z

Circular column 0.0125 0.090 24.2 – – – – – –


Middle elliptic column 0.085 £ 0.075 0.006 65.8 – – – – – –
Outer elliptic column 0.085 £ 0.075 0.0035 45 – – – – – –
Whole – – 180 0.00065 0.00022 0.00087 0.060 0.035 0.070

Newton’s third law. The rate of work done on a segment by a joint torque component was
defined as the product of the torque component and its matching angular velocity
component of the segment. The rate of energy transfer by a joint torque component was
determined using the formulation provided by Robertson and Winter (1980).
The work done by a joint torque component was determined through numerical integration
of its joint torque power. Likewise, the amount of energy transferred by each joint torque
component and each joint force was determined through numerical integration of the rates of
energy transfer by the torque component and the joint force, respectively. Energy generation
and transfer in the racket arm segments were minimal until the beginning of pelvis forward
rotation. Thus, the period from the beginning of pelvis forward rotation to ball impact was
adopted as the duration of these numerical integrations. The amount of positive and negative
work done by the shoulder internal rotation torque was determined through numerical
integration of the positive and negative rates, respectively, of work done by the torque over the
same duration.

Energy transfer ratio of the racket arm


The amount of energy transferred through the shoulder joint to the upper arm from the
beginning of pelvis forward rotation to ball impact was determined through summation of the
amounts of energy transferred by the shoulder joint force and joint torque components to the
upper arm. The increase in mechanical energy of the racket arm over the same duration was
determined through summation of the amount of net work done on the racket arm segments
by the shoulder, elbow/forearm, and wrist joint torques. The ratio of the amount of energy
transferred through the shoulder joint to the racket arm to the increase in mechanical energy of
the racket arm was also calculated and is hereafter referred to as the energy transfer ratio of the
racket arm.

Normalization
Considering the differences in body mass among the participants, the joint torques, powers,
work, and energy were normalized by the body mass of each participant.

Statistical analysis
Two-way repeated measure ANOVAs were performed to test the effects of performance level
and ball spin on the dependent variables; the maximum joint torques, maximum rates of
work done by the joint torque components, maximum rates of energy transfer by the joint
366 Y. Iino & T. Kojima

forces and joint torque components, amount of work done by the joint torque components,
amount of energy transferred by the joint forces and joint torque components, and energy
transfer ratios of the racket arm. Statistical significance should be adjusted to reduce type I
error rates in light of the number of ANOVA tests performed, whereas common adjustment
methods such as the Bonferroni correction may lead to higher type II error rates because
of the number of the tests. With this in mind, statistical significance was set at p , 0.01.
Two-tailed paired t-tests were performed for pair-wise comparisons within each performance
level and ball spin separately, at 0.01 level of significance, when significant interactions were
found.
Simple linear regression analyses were carried out to examine the relationships between
the square of the racket speed at impact and each of the following independent variables for
each performance level group: the amount of normalized energy transferred by the shoulder
joint force and joint torque; the amount of normalized energy transferred by the shoulder
joint torque; the amount of normalized energy transferred by the shoulder joint force; and
the amount of normalized energy generated by the shoulder, elbow/forearm, and wrist joint
torques. The square of the speed was taken as a dependent variable in consideration of the
dimension of the normalized energy. Statistical significance was set at p , 0.01.

Results
Joint torques, their power, and work done by the joint torques
General trends in the time-courses of the joint angular velocity, joint torque, and joint torque
power at the racket arm joints against light backspin were similar to those against heavy
backspin (Figures 2– 4). Thus, the figures for the forehands against the former spin were
omitted. The shoulder adduction and internal rotation torques and elbow varus torque
peaked just before the beginning of shoulder internal rotation irrespective of performance
level and ball spin, whereas the elbow flexion torque peaked around ball impact.
Significant effects of performance level on the maximum shoulder internal rotation, elbow
varus, and wrist radial deviation torques were observed (Table II). A significant effect of ball
spin on the maximum elbow flexion torque was also observed.
The shoulder internal rotation torque showed first negative and subsequently positive
power bursts irrespective of performance level and ball spin (Figure 2). The maximum
positive and negative powers of the torque were significantly larger for the advanced players
than for the intermediate players (Table III). The amount of work done by the shoulder
adduction torque was substantial, but the amount of work done by the shoulder internal
rotation torque was minimal. The amount of negative work done by the latter torque was
significantly larger for the advanced players than for the intermediate players.

Rates of energy transfer and the amounts of energy transferred by the joint torque components of the
racket arm
The maximum rates of energy transfer by the shoulder internal rotation, elbow varus, and
wrist radial deviation torques were significantly larger for the advanced players than for the
intermediate players (Table IV). No significant effects of ball spin on the maximum rates of
energy transfer by the racket arm joint torques were observed. The shoulder adduction and
internal rotation, elbow varus, and flexion torques transferred substantial amount of
mechanical energy to distal segments. No significant effects of performance level and ball
spin on the amounts of energy transferred by the joint torques were observed.
Kinetics of the upper limb 367

Figure 2. The time-courses of the joint angular velocity, joint torque, and joint torque power at the shoulder joint of
the racket arm against heavy backspin for an advanced and an intermediate player with body masses of 61.4 and
62.6 kg, respectively. Vertical dashed and dotted lines indicate the beginnings of pelvis forward rotation and shoulder
internal rotation, respectively. Time of 0 s corresponds to ball impact. Add, adduction; Abd, abduction, Flex,
flexion; Ext, extension; Int rot, internal rotation; Ext rot, external rotation.

Rates of energy transfer and the amount of energy transferred by the joint forces of the racket arm
The rates of energy transfer by the joint forces at the shoulder, elbow, and wrist joints were
substantial (Table V). The maximum rate of energy transfer by the shoulder joint force was
significantly larger against heavy backspin than against light backspin. A significant
interaction of the effects of performance level and ball spin on the amount of energy
transferred by the elbow joint force was observed. Pair-wise comparisons revealed that
the amount against heavy backspin was larger than that against light backspin for the
intermediate players. No significant effects of performance level and ball spin on the amount
of energy transferred by the shoulder and wrist joint forces were observed.
368 Y. Iino & T. Kojima

Figure 3. The time-courses of the joint angular velocity, joint torque, and joint torque power at the elbow/forearm
joint of the racket arm against heavy backspin for the same advanced and the same intermediate players as in
Figure 2. Vertical dashed and dotted lines indicate the beginnings of pelvis forward rotation and shoulder internal
rotation, respectively. Time of 0 s corresponds to ball impact. Flex, flexion; Ext, extension; Pro, pronation; Sup,
supination.

Energy transfer ratio and the relationship between the square of the racket speed and each of the
amount of energy transfer and generation
Irrespective of performance level and ball spin, more than 76% of the increase in mechanical
energy of the racket arm was due to energy transfer through the shoulder joint (Table VI).
Kinetics of the upper limb 369

Figure 4. The time-courses of the joint angular velocity, joint torque, and joint torque power at the wrist joint of the
racket arm against heavy backspin for the same advanced and the same intermediate players as in Figure 2. Vertical
dashed and dotted lines indicate the beginnings of pelvis forward rotation and shoulder internal rotation,
respectively. Time of 0 s corresponds to ball impact. Flex, flexion; Dev, deviation; Int rot, internal rotation.

The increase in mechanical energy was significantly larger against heavy backspin than
against light backspin.
The regressions of the square of the racket speed at impact for the advanced players were
not significant for all independent variables (Figure 5). The regressions for the intermediate
370 Y. Iino & T. Kojima

Table II. M ^ SD normalized maximum joint torques of the racket arm during the forehands against light and heavy
backspins (Nm/kg).

Advanced Intermediate ANOVA p-value

Light Heavy Light Heavy Performance Spin Interaction

Shoulder adduction 0.91 ^ 0.14 0.93 ^ 0.15 0.78 ^ 0.28 0.80 ^ 0.28 0.223 0.153 0.816
Shoulder flexion 0.29 ^ 0.10 0.31 ^ 0.12 0.30 ^ 0.13 0.31 ^ 0.15 0.922 0.548 0.697
Shoulder internal rotation 0.58 ^ 0.13 0.61 ^ 0.10 0.37 ^ 0.10 0.39 ^ 0.11 0.001* 0.095 0.710
Elbow varus 0.61 ^ 0.10 0.63 ^ 0.09 0.42 ^ 0.13 0.43 ^ 0.14 0.003* 0.109 0.764
Elbow flexion 0.78 ^ 0.16 0.85 ^ 0.14 0.65 ^ 0.13 0.72 ^ 0.13 0.072 0.001* 0.931
Wrist dorsi flexion 0.07 ^ 0.03 0.07 ^ 0.03 0.08 ^ 0.04 0.08 ^ 0.04 0.781 0.206 0.165
Wrist radial deviation 0.13 ^ 0.03 0.14 ^ 0.04 0.08 ^ 0.02 0.09 ^ 0.03 0.009* 0.013 0.412

*Significance ( p , 0.01).

players were significant for the normalized energy transferred by the shoulder joint force and
torque against both the spins and the normalized energy transferred by the shoulder joint
force against light backspin (Figure 6).

Discussion and implications


This study determined the significance of mechanical energy generation and transfer in the
racket arm in generating the racket speed for the table tennis topspin forehands used by
advanced and intermediate players.
The shoulder internal rotation torque observed around the beginning of shoulder internal
rotation motion appeared to cause the angular acceleration of the joint motion (Figure 2).
Shoulder internal rotation contributed substantially to the racket speed at impact (Iino &
Kojima, 2009). The elbow flexion torque observed at ball impact appeared to control the
elbow flexion angle of 0.91 –1.52 rad from full extension at ball impact (Iino & Kojima,
2009), because angular accelerations of elbow flexion were small around ball impact
(Figure 3). The flexed elbow angle increased the effectiveness of the shoulder internal
rotation in generating the racket speed and might have helped the rotation contribute
substantially to the racket speed at impact.
The maximum rates of energy transfer by the shoulder internal rotation torque, the elbow
varus torque, and the wrist radial deviation torque were higher for the advanced players than
for the intermediate players (Table IV). On the other hand, the maximum rates of the energy
transfer by the shoulder, elbow, and wrist joint forces were not significantly different between
the two groups of the players (Table V). Thus, the second hypothesis that owing to higher
energy transfer rates by the limb joint torques and forces, the advanced players increase
racket speed at a higher rate than the intermediate players was partly verified.
The rate of energy transfer by the shoulder internal rotation torque depended on both the
angular velocity component of the upper trunk parallel to the axis of shoulder internal
rotation and the internal rotation torque. The advanced players exerted a significantly larger
shoulder internal rotation torque than the intermediate players (Table II). Thus, a higher
rate of energy transfer by the shoulder internal rotation torque for the advanced players was
at least partly due to a larger shoulder internal rotation torque. This suggests that physical
training for the shoulder internal rotators may be important for intermediate players to
generate the racket speed at a higher rate. The means of the angular velocity component at
the time of the maximum rate of energy transfer by the torque for the advanced players were
Table III. Maximum joint torque powers of the racket arm joints and the amounts of net, positive, and negative work done by the shoulder joint torque components during
the forehands against light and heavy backspins (M ^ SD).

Advanced Intermediate ANOVA p-value

Light Heavy Light Heavy Performance Spin Interaction

Maximum joint torque power (W/kg)


Shoulder adduction 4.3 ^ 1.2 4.2 ^ 1.6 3.7 ^ 1.2 4.2 ^ 1.3 0.707 0.344 0.163
Shoulder flexion 1.0 ^ 0.8 1.0 ^ 1.0 1.5 ^ 1.1 1.3 ^ 1.1 0.405 0.757 0.695
Shoulder internal rotation, positive 3.1 ^ 0.8 3.3 ^ 0.8 1.7 ^ 0.8 1.7 ^ 1.1 0.003* 0.474 0.583
Shoulder internal rotation, negative -4.0 ^ 1.9 -4.4 ^ 1.8 -1.3 ^ 0.8 -1.7 ^ 0.9 0.002* 0.002* 0.606
Elbow flexion 1.4 ^ 0.6 1.5 ^ 0.6 1.1 ^ 0.3 1.1 ^ 0.3 0.179 0.681 0.495
Wrist dorsi flexion 0.2 ^ 0.2 0.2 ^ 0.1 0.1 ^ 0.1 0.2 ^ 0.2 0.680 0.432 0.146
Wrist radial deviation 0.3 ^ 0.1 0.3 ^ 0.2 0.2 ^ 0.1 0.2 ^ 0.1 0.093 0.269 0.964
Net work done by the joint torque (J/kg)
Shoulder adduction 0.23 ^ 0.14 0.24 ^ 0.16 0.22 ^ 0.11 0.25 ^ 0.08 0.983 0.212 0.542
Shoulder internal rotation -0.09 ^ 0.05 -0.09 ^ 0.07 -0.01 ^ 0.04 -0.04 ^ 0.03 0.020 0.279 0.069
Positive and negative work done by the shoulder internal rotation torque (J/kg)
Positive 0.11 ^ 0.04 0.12 ^ 0.03 0.07 ^ 0.04 0.06 ^ 0.04 0.018 0.659 0.138
Negative -0.20 ^ 0.06 -0.21 ^ 0.06 -0.09 ^ 0.05 -0.11 ^ 0.04 0.001* 0.018 0.234

*Significance ( p , 0.01).
Kinetics of the upper limb
371
372 Y. Iino & T. Kojima

Table IV. Maximum rates of energy transfer by the joint torque components and the amounts of energy transfer
during the forehands against light and heavy backspins (M ^ SD).

Advanced Intermediate ANOVA p-value

Light Heavy Light Heavy Performance Spin Interaction

Maximum rate of energy transfer (W/kg)


Shoulder adduction 5.2 ^ 1.9 5.1 ^ 2.2 3.6 ^ 3.2 4.0 ^ 3.5 0.326 0.653 0.490
Shoulder internal rotation 6.6 ^ 2.1 7.0 ^ 1.6 3.4 ^ 1.2 3.9 ^ 1.6 0.001* 0.065 0.979
Elbow varus 8.3 ^ 2.0 8.7 ^ 1.8 5.2 ^ 2.1 5.6 ^ 2.5 0.006* 0.179 0.863
Elbow flexion 9.5 ^ 2.0 10.9 ^ 2.7 9.9 ^ 3.0 10.5 ^ 3.2 0.999 0.027 0.345
Wrist radial deviation 1.9 ^ 0.3 2.1 ^ 0.5 1.3 ^ 0.4 1.3 ^ 0.5 0.003* 0.073 0.346
Amount of energy transfer (J/kg)
Shoulder adduction 0.36 ^ 0.14 0.35 ^ 0.14 0.26 ^ 0.24 0.29 ^ 0.25 0.407 0.678 0.362
Shoulder flexion 0.03 ^ 0.06 0.02 ^ 0.04 0.05 ^ 0.05 0.04 ^ 0.04 0.328 0.390 0.832
Shoulder internal rotation 0.32 ^ 0.12 0.34 ^ 0.10 0.21 ^ 0.09 0.24 ^ 0.11 0.052 0.056 0.734
Elbow varus 0.49 ^ 0.13 0.50 ^ 0.13 0.37 ^ 0.14 0.39 ^ 0.15 0.106 0.214 0.688
Elbow flexion 0.32 ^ 0.10 0.36 ^ 0.11 0.46 ^ 0.10 0.48 ^ 0.14 0.024 0.181 0.656
Wrist radial deviation 0.11 ^ 0.02 0.12 ^ 0.03 0.09 ^ 0.03 0.09 ^ 0.03 0.115 0.033 0.216

*Significance ( p , 0.01).

30% and 18% larger than those for the intermediate players against light backspin
(12.5 ^ 2.6 rad/s vs. 9.6 ^ 2.0 rad/s) and against heavy backspin (12.2 ^ 2.2 rad/s vs.
10.3 ^ 2.1 rad/s), respectively, although there was no significant effect of performance level
on the component ( p ¼ 0.03). Thus, it is possible that a larger angular velocity component
for the advanced players also contributed to the higher rate of energy transfer.
It remains unclear why the advanced players could exert a significantly larger shoulder
internal rotation torque than the intermediate players (Table II). Various factors such as
cross-sectional areas of muscles, moment arms of muscular forces, mechanical properties of
tendons, and neuromuscular systems affect the magnitude of joint torque exerted (Winter,
1990). The contraction velocity of a muscle also affects the magnitude because the velocity
affects the force exerted by the muscle according to its force – velocity relationship (Hill,
1938). The shoulder external rotation velocities at the time of the maximum shoulder
internal rotation torque for the advanced players tended to be larger than those for the
intermediate players (light backspin: 3.3 rad/s vs. 0.7 rad/s, heavy backspin: 3.2 rad/s vs. 2.1

Table V. Maximum rates of energy transfer by the joint forces and the amounts of energy transfer during the
forehands against light and heavy backspins (M ^ SD).

Advanced Intermediate ANOVA p-value

Light Heavy Light Heavy Performance Spin Interaction

Maximum rate of energy transfer (W/kg)


Shoulder 10.6 ^ 1.9 12.3 ^ 1.9 10.0 ^ 2.8 11.9 ^ 3.1 0.649 0.001* 0.756
Elbow 11.4 ^ 1.8 11.1 ^ 2.3 8.1 ^ 3.1 9.3 ^ 3.4 0.061 0.084 0.015
Wrist 13.0 ^ 2.1 13.1 ^ 2.2 9.5 ^ 2.9 10.3 ^ 3.0 0.020 0.206 0.285
Amount of energy transfer (J/kg)
Shoulder 0.90 ^ 0.13 0.89 ^ 0.19 0.86 ^ 0.24 0.99 ^ 0.24 0.756 0.080 0.039
Elbow 0.71 ^ 0.13 0.68 ^ 0.14 0.55 ^ 0.23 0.63 ^ 0.23 0.249 0.144 0.002*
Wrist 0.85 ^ 0.10 0.86 ^ 0.12 0.76 ^ 0.19 0.83 ^ 0.17 0.387 0.020 0.106

*Significance ( p , 0.01).
Table VI. Increase in mechanical energy of the racket arm, energy transferred to the racket arm, and energy transfer ratio of the racket arm during the forehands against light
and heavy backspins (M ^ SD).

Advanced Intermediate ANOVA p-value

Light Heavy Light Heavy Performance Spin Interaction

Increase in mechanical energy of the racket arm (J/kg) 1.96 ^ 0.21 1.99 ^ 0.22 1.77 ^ 0.43 1.95 ^ 0.40 0.473 0.004* 0.034
Mechanical energy transferred to the racket arm (J/kg) 1.61 ^ 0.24 1.60 ^ 0.29 1.38 ^ 0.49 1.56 ^ 0.45 0.454 0.066 0.040
Energy transfer ratio of the racket arm (%) 82 ^ 5 80 ^ 9 76 ^ 11 79 ^ 9 0.376 0.860 0.086

*Significance ( p , 0.01).
Kinetics of the upper limb
373
374 Y. Iino & T. Kojima

Figure 5. The regression of the square of the racket speed at impact on each of the independent variables: the
normalized energy transferred by shoulder joint force and torque (A); the normalized energy transferred by shoulder
joint torque (B); the normalized energy transferred by shoulder joint force (C); and the normalized energy generated
by the shoulder, elbow/forearm and wrist joint torques (D) for the advanced players during the forehands against
light and heavy backspins.
Kinetics of the upper limb 375

Figure 6. The regression of the square of the racket speed at impact on each of the independent variables: the
normalized energy transferred by shoulder joint force and torque (A); the normalized energy transferred by shoulder
joint torque (B); the normalized energy transferred by shoulder joint force (C); and the normalized energy generated
by the shoulder, elbow/forearm and wrist joint torques (D) for the intermediate players during the forehands against
light and heavy backspins. A regression equation is presented using y as the dependent variable (the square of the
racket speed at impact) and x as the independent variable when the regression was significant. R*2 is the coefficient
of determination adjusted.
376 Y. Iino & T. Kojima

rad/s), although the performance level had no significant effect ( p ¼ 0.07). The muscle
contraction velocity may be possibly one of the factors contributing to a larger internal
rotation torque. Future studies using more detailed models of the body, electromyography,
or both would be necessary to elucidate the reasons for the difference in maximum shoulder
internal rotation torque.
More than 76% of the increase in the energy of the racket arm was due to the energy
transfer for both groups of players and both ball spins (Table VI). Thus, the first hypothesis
that a large portion of the mechanical energy of the racket arm at ball impact would be due to
energy transfer through the shoulder joint was verified. Furthermore, the results of the
regression analyses (Figure 6) suggest that increasing the energy transfer to the upper arm by
the shoulder joint force and torque may be an important factor for intermediate table tennis
players to generate a high racket speed at impact in the forehands. For the advanced players,
the regressions for the energy transfer by the shoulder joint torque and force almost reached
significance ( p ¼ 0.07 and 0.04, respectively, Figure 5). Further investigation is necessary to
determine the significance of the regressions of the racket speed for advanced players.
The maximum elbow flexion torque was significantly larger against heavy backspin than
against light backspin (Table II). The result may be related to a tendency of higher racket
speeds at ball impact against heavy backspin than against light backspin (Iino & Kojima,
2009; p ¼ 0.02). The maximum rate of energy transfer by the shoulder joint force was also
significantly larger against heavy backspin than against light backspin. The reason for this is
unknown.
A significant interaction of performance level and ball spin on the amount of energy
transferred by the elbow joint force was observed (Table V). Post-hoc comparison tests revealed
that in the intermediate players, the amount of energy transfer by the force against heavy
backspin was significantly larger than that against light backspin; however, in the advanced
players, no significant difference was observed in this regard. A larger amount of energy
transferred by the elbow joint force against heavy backspin in the case of the intermediate players
may be related to the fact that the racket speed at impact was larger against heavy backspin than
against light backspin for the players (Iino & Kojima, 2009; 18.4 m/s vs. 17.9 m/s).
As a limitation of this study, energy transfer was discussed without considering the actions
of multi-joint muscles, because a joint power analysis (Robertson & Winter, 1980) implicitly
assumes that the joint torques are produced by one-joint muscles. Two-joint muscles have
been reported to play an important role in the mechanical energy transfer in human
movements (Prilutsky & Zatsiorsky, 1994). Clarification of the functions of muscles,
including multi-joint muscles, in table tennis forehands was beyond the scope of this study
and requires further investigation.

Conclusion
In table tennis topspin forehands, the advanced players exerted a significantly larger
normalized internal rotation torque at the shoulder joint of the racket arm than the
intermediate players did. Owing to a larger torque, the advanced players could transfer
normalized mechanical energy from the trunk to the upper arm at a higher rate than the
intermediate players could. More than 76% of the increase in mechanical energy of the
racket arm from the beginning of pelvic forward rotation to ball impact was due to the energy
transfer to the upper arm in the case of both the groups of players and in both the ball spins.
Regression analyses showed that the racket speed at impact in the case of the intermediate
players increased with the amount of the energy transfer to the upper arm by the shoulder
joint force and torque irrespective of the ball spins.
Kinetics of the upper limb 377

Acknowledgements
This study was funded by the Japan Table Tennis Association. We thank Dr T. Ohtsuki and
Dr K. Kudo for their support in the experiment.

References
Abdel-Aziz, Y. I., & Karara, H. M. (1971). Direct linear transformation from comparator coordinates into object
coordinates in close-range photogrammetry. Proceedings of the ASP Symposium on Close-Range Photogrammetry,
American Society of Photogrammetry (pp. 1 –18). Falls Church, VA.
Ae, M., Tang, H., & Yokoi, T. (1992). Estimation of inertia properties of the body segments in Japanese athletes.
Biomechanism, 11, 23–33. [in Japanese].
Geske, K. M., & Mueller, J. (2010). Table tennis tactics: Your path to success. Maidenhead: Meyer & Meyer.
Hill, A. V. (1938). The heat of shortening and dynamic constants of muscle. Proceedings of the Royal Society B, 126,
136– 195.
Iino, Y., & Kojima, T. (2009). Kinematics of table tennis topspin forehands: Effects of performance level and ball
spin. Journal of Sports Sciences, 27, 1311–1321.
Kasai, J., & Mori, T. (1998). A qualitative 3D analysis of forehand strokes in table tennis. In A. Lees, I. Maynard, M.
Hughes, and T. Reilly (Eds.), Science and racket sports II (pp. 201 –205). London: E & FN Spon.
Lee, K. T., & Xie, W. (2004). Comparative studies between the techniques of Singapore and Thailand male elite table tennis
players for sea games 2001 (pp. 330 –333)
Neal, R. J. (1991). The mechanics of the forehand loop and smash shots in table tennis. The Australian Journal of
Science and Medicine in Sport, 23, 3 –11.
Prilutsky, B. I., & Zatsiorsky, V. M. (1994). Tendon action of two-joint muscles: Transfer of mechanical energy
between joints during jumping, landing, and running. Journal of Biomechanics, 27, 25– 34.
Robertson, D. G. E., & Winter, D. A. (1980). Mechanical energy generation, absorption and transfer amongst
segments during walking. Journal of Biomechanics, 13, 845–854.
Seemiller, D., & Holowchak, M. (1997). Winning table tennis: Skills, drill, and strategies. Champaign, IL: Human
Kinetics.
Winter, D. A. (1990). Biomechanics and motor control of human movements. New York: Wiley.