RESISTIVE FORCES AND TECHNIQUE ANALYSIS IN FRONT
CRAWL SPRINT SWIMMING
By
Gina Belinda Dare Sacilotto
B. App. Sci. (Exercise and Sport Science)
B. Health Sci. (Hons.)
This thesis is submitted in fulfilment of the requirements for the completion of the degree of
Doctor of Philosophy in Health
Faculty of Health
University of Canberra
Locked Bag 1
University of Canberra
Australian Capital Territory, 2617
Australia
Submitted in June 2014
Abstract
Free swimming is the main component of a swim performance and is the most
complex to understand. Using the Assisted Towing Method (ATM), developed at the
Australian Institute of Sport, this thesis was designed to investigate resistive forces (active
drag) and how this method could be integrated into the assessment of free swimming
technique. Four investigations were conducted to assess the ATM protocol for integration as
an objective assessment tool for coaches, scientists and athletes in front crawl sprint
swimming. The aim of study 1 was to examine the reliability of the mean active drag values
collected using the ATM. Results indicated that towing the swimmers whilst permitting intrastroke fluctuations (ICC = 0.94, 0.88-0.97) allowed the swimmer to produce a higher level of
reliability values than when being towed with a constant velocity (ICC = 0.83, 0.57-0.94). A
considerably lower amount of percentage error was found in the fluctuating trials (CVTE % =
12.6) than the constant trials (CVTE % = 35.0). The aim of study 2 was to compare the stroke
mechanics (stroke lengths and rates) of a free swim and an assisted tow trial to determine
whether the ATM protocols alter stroke mechanics. Results indicated a significant increase in
stroke lengths and rates occurred when swimmers completed the ATM protocol. However,
the ratio of stroke length and stroke rate in the assisted condition was found to significantly
predict ratios of stroke mechanics in free swimming (r = 0.95). It was concluded that when
using the ATM protocol it is possible to transfer technique critiques found in assisted
swimming and apply them to free swimming. The aim of study 3 was to investigate
instantaneous active drag force-time profiles and deconstruct them into stroke phases. The
total sample was split into groups to identify differences between male and female, and elite
and sub-elite swimmers. A consistent biphasic curve was found between all elite swimmers
and stroke phases could be identified within a profile. The sub-elite group, however,
iii
produced multiphasic curves which could be linked to anthropometric differences, slower
velocities, or inconsistent propulsion generation during the propulsive stroke phases. The aim
of the final study, study 4, was to quantify technique through the use of a coach feedback
survey and identify whether a relationship existed between coach ratings of technique and
active drag force-time profiles. Eight coaches were provided with a survey which included
still images and video clips of thirty swimmers. Coaches were asked to rate technique on a
simplified Likert scale. The coach ratings were then correlated against force data at the same
time points. Findings revealed a range of low to moderate internal consistencies between
coach ratings in all swimmers, thus indicating a major limitation of using the coach feedback
survey in its current form. Alternatively, these low consistencies were maybe a direct result
of coaches not being able to quantitatively assess swim technique as the norm is to assess
qualitatively whilst on pool deck. The results found in this study should be interpreted with
caution as only weak correlations were found between coach ratings and force-time profiles.
Four main additions to the area of swimming biomechanics were concluded from this thesis:
1) the ATM protocol is a reliable tool to capture kinetic information for the assessment of
free swimming; 2) a consistent increase was observed in stroke length and stroke rate
between free swim and assisted towed trials which could enable the transfer of technical
assumptions from assisted tow trial outputs to free swimming; 3) the ATM active drag forcetime profiles revealed the possibility of an optimal profile being established within the elite
sprint swimmers thus enabling the ATM to be used in the future as an objective assessment of
technique; and 4) a novel tool was presented which has potential to quantify technique
proficiency with further investigation.
iv
Table of Contents
ABSTRACT ........................................................................................................................... III
STATEMENT OF AUTHORSHIP ....................................................................................... V
TABLE OF CONTENTS .................................................................................................... VII
LIST OF FIGURES ........................................................................................................... XIII
LIST OF TABLES ............................................................................................................... XV
LIST OF EQUATIONS .................................................................................................... XVII
LIST OF ABBREVIATIONS ........................................................................................... XIX
ACKNOWLEDGEMENTS .............................................................................................. XXI
PUBLICATIONS FROM DOCTORAL THESIS ......................................................... XXV
CHAPTER 1 ............................................................................................................................. 3
Introduction ................................................................................................................................ 3
Background .................................................................................................................... 3
Statement of Problem ..................................................................................................... 6
Literature Review: ............................................................................................. 6
Study 1: .............................................................................................................. 7
Study 2: .............................................................................................................. 7
Study 3: .............................................................................................................. 8
Study 4: .............................................................................................................. 9
Delimitations ................................................................................................................ 10
Limitations ................................................................................................................... 11
References .................................................................................................................... 11
vii
CHAPTER 2 ........................................................................................................................... 15
Literature Review..................................................................................................................... 15
Introduction .................................................................................................................. 16
Active Drag and Swim Performance ........................................................................... 17
Mechanical Power Output in Swimming ..................................................................... 18
Techniques of Drag Assessment .................................................................................. 19
Energetics Approach .................................................................................................... 19
Numerical Simulations................................................................................................. 21
Experimental Techniques............................................................................................. 23
Measuring Active Drag System ........................................................................ 23
Velocity Perturbation Method.......................................................................... 25
Assisted Towing Method .................................................................................. 28
Review Summary ......................................................................................................... 32
References .................................................................................................................... 33
CHAPTER 3 ........................................................................................................................... 43
Study 1: Reliability of Active Drag Values Using the Assisted Towing Method ................... 43
Introduction .................................................................................................................. 43
Methods........................................................................................................................ 46
Participants ...................................................................................................... 46
Testing Protocols ............................................................................................. 46
Data Processing ............................................................................................... 51
Statistical Analysis ........................................................................................... 53
Results .......................................................................................................................... 54
viii
Discussion .................................................................................................................... 56
Conclusion ................................................................................................................... 58
References .................................................................................................................... 59
CHAPTER 4 ........................................................................................................................... 63
Study 2: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted
Towed Swimming .................................................................................................................... 63
Introduction .................................................................................................................. 63
Method ......................................................................................................................... 65
Results .......................................................................................................................... 68
Discussion .................................................................................................................... 70
Conclusion ................................................................................................................... 72
References .................................................................................................................... 73
CHAPTER 5 ........................................................................................................................... 77
Study 3: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles
in Elite and Sub-Elite Sprint Swimmers .................................................................................. 77
Introduction .................................................................................................................. 77
Method ......................................................................................................................... 79
Results .......................................................................................................................... 82
Discussion .................................................................................................................... 85
Conclusion ................................................................................................................... 89
References .................................................................................................................... 89
ix
CHAPTER 6 ........................................................................................................................... 93
Study 4: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles
in Elite and Sub-Elite Front Crawl Sprint Swimmers. ............................................................. 93
Introduction .................................................................................................................. 93
Method ......................................................................................................................... 96
Results ........................................................................................................................ 100
Discussion .................................................................................................................. 103
Conclusion ................................................................................................................. 106
References .................................................................................................................. 107
CHAPTER 7 ......................................................................................................................... 111
Summary, Conclusions and Future Directions ...................................................................... 111
Summary .................................................................................................................... 111
Conclusions ................................................................................................................ 114
Study 1 ....................................................................................................................... 114
Study 2 ....................................................................................................................... 115
Study 3 ....................................................................................................................... 115
Study 4 ....................................................................................................................... 116
Future Directions ....................................................................................................... 116
APPENDIX A: PUBLICATIONS FROM DOCTORAL THESIS ................................. 121
APPENDIX B: PUBLISHED LITERATURE REVIEW................................................. 123
APPENDIX C: 2014 PROCEEDINGS FROM 32ND ISBS, TENNESSEE, UNITED
STATES ................................................................................................................................ 133
APPENDIX D: 2014 PROCEEDINGS FROM 12TH BIOMECHANICS AND
MEDICINE IN SWIMMING, CANBERRA, AUSTRALIA ........................................... 139
x
APPENDIX E: 2013 PROCEEDINGS FROM THE ASICS CONFERENCE OF
SCIENCE AND MEDICINE IN SPORT, PHUKET, THAILAND ................................ 145
APPENDIX F: 2012 PROCEEDINGS FROM THE 30TH ISBS, MELBOURNE,
AUSTRALIA ........................................................................................................................ 147
APPENDIX G: RIGHT AND LEFT SINGLE STROKE ACTIVE DRAG FORCETIME PROFILES AS USED IN STUDY 3 ....................................................................... 153
APPENDIX H: COACH FEEDBACK SURVEYS ........................................................... 187
APPENDIX I: RIGHT AND LEFT COACH EVENT RATINGS AND ACTIVE DRAG
EVENT VALUES AS USED IN STUDY 4 ........................................................................ 197
APPENDIX J: CONSENT FORMS AND PARTICIPANT INFORMATION ............. 207
xi
List of Figures
Figure 2.1: Adapted experimental setup of drag collection from Di Prampero et al.40 ........... 20
Figure 2.2: MAD-System setup for drag collection adapted from Hollander et al.22 .............. 24
Figure 2.3: Modification of the VPM approach as utilised in, and adapted from Wang et al.26
............................................................................................................................... 27
Figure 2.4: ATM technique to drag collection as shown in Sacilotto et al.55 .......................... 29
Figure 3.1: Assisted towing method set up ………………………………………………… 48
Figure 3.2: Placement of Velcro belt used in assisted tow trials …………………………… 48
Figure 3.3: Example of the grip required to hold towing rope in passive tow trials…….. … 50
Figure 3.4: Example of a processed data output for a fluctuating tow trial………………….52
Figure 3.5: Example of a processed data output for a constant tow trial…………………….53
Figure 3.6: Result summary of test retest reliability: a. Fluctuating trials: ICC = 0.94 (CIL =
0.88. CIU = 0.97); b. Constant trials: ICC = 0.83 (CIL = 0.57, CIU = 0.94). ◦ =
trials 1 and 2; ▪ = trials 1 and 3…………………………………………………..56
Figure 4.1: Example of a mixed image used to calculate assisted towed swim stroke rates ..67
xiii
Figure 4.2: Ratio (stroke length/stroke rate) comparison with line of best fit between assisted
towed swim and free swims stroke mechanics (r = 0.95)........…………...……. 70
Figure 5.1: Male elite (a) and sub-elite (b) breakdown of stroke phases within an active drag
profile in one-stroke cycle (right hand entry to subsequent right hand entry)
………………………………………………………………………..……….... 84
Figure 5.2: Female elite (a) and sub-elite (b) breakdown of stroke phases within an active
drag profile in one-stroke cycle (right hand entry to subsequent right hand entry)
…………………………………………………………………………..……… 84
Figure 6.1: An example of a coach feedback survey question at a stroke event ..………….. 98
Figure 6.2: Time-code matching with active drag force data (Time-code rounded to the
nearest 0.002 of a second) ……………………………………………………... 99
xiv
List of Tables
Table 3.1: Total sample mean (± SD) data for swim velocity, tow velocity, active drag, FINA
point score and 100 m performance time ……………………………………. 55
Table 3.2: Intra-class correlations coefficients between different groups of trials and the
average of all five trials within both constant and fluctuating active drag tow
velocity trials ....................................................................................................... 55
Table 3.3: Intra-class correlation confidence limits between different groups of trials and the
average of all five trials within both constant and fluctuating active drag tow
velocity trials ……………………………….…………………………………..... 55
Table 3.4: Summary of within-subject variation shown in percentages presented as
coefficients with 95 % confidence intervals …………………………………… 56
Table 4.1: Summary of differences between free swim and assisted tow trial variables …... 69
Table 4.2: Significance between stroke mechanics and swim velocities in both free swim and
assisted towed swimming ……………………………………………………...... 69
Table 5.1: Summary of performance variables shown as mean ± SD ..................................... 82
Table 5.2: Percentage of time spent in each stroke phase during assisted towed swimming
shown as mean ± SD percentages (%) .................................................................... 82
xv
Table 5.3: Maximum and minimum active drag forces for an individual stroke cycle shown as
mean ± SD ............................................................................................................... 83
Table 5.4: Time of maximum and minimum active drag forces, as a percentage of total stroke
time shown as mean ± SD ....................................................................................... 83
Table 6.1: Mean (± SD) of trial active drag and overall coach ratings …………………… 100
Table 6.2: Qualitative summary of common points at each stroke event made by all eight
coaches ……………………..……………………………………………...…… 101
Table 6.3: Trial active drag value correlation assessments between FINA, swim velocity, tow
velocity and coach overall values for elite and sub-elite groups …………….… 101
Table 6.4: Overall coach rating correlation assessments between FINA, swim velocity, tow
velocity and trial active drag values for elite and sub-elite groups …………... 102
Table 6.5: Summary of mean coach event ratings and mean active drag event (DAE) values
……………………………………………………………………………….…. 102
Table 6.6: Elite correlation analysis between coach event ratings and FINA scores and active
drag event values ………………………………………………………………. 103
Table 6.7: Sub-elite correlation analysis between coach event ratings and FINA scores and
active drag event values ………………………………………………………... 103
xvi
List of Equations
Equation 2.1: Drag Force …………………………………………………………………... 16
Equation 2.2: Mechanical Power Output …………………………………………………... 18
Equation 2.3: Propelling Efficiency ………………………………………………………... 19
Equation 2.4: Active Drag Force (Toussaint et al.48) ………………………………………. 24
Equation 2.5: Active Drag Force (Kolmogorov and Duplishcheva11) ……………………... 26
Equation 2.6: Active Drag Force (Alcock and Mason27) ………………………………...… 30
Equation 2.7: Propulsion using ATM protocol (Mason et al.54) …………………………… 31
Equation 3.1: Active Drag Force (Alcock and Mason2) ......................................................... 51
Equation 4.1: Swim Velocity .................................................................................................. 58
xvii
List of Abbreviations
3D
Three Dimensional
A
Frontal Surface Area
AIS
Australian Institute of Sport
ANOVA
Analysis of Variance
ATM
Assisted Towing Method
CD
Drag Coefficient
CFD
Computational Fluid Dynamics
CI
Confidence Intervals
CIU
Confidence Intervals Upper
CIL
Confidence Intervals Lower
CV
Coefficient of Variation
CVTE%
Typical Error Coefficient of Variation Percentage
ௗ
ௗ௧
Derivative
D
Swimming Drag
DA
Active Drag
DAE
Active Drag Event
Db
Drag Force with added resistance
DF
Drag Force
Fb
Drag Force with added assistance
FINA
The Federation Internationale de Natation Amateur
FS
Free Swimming
ICC
Interclass Correlations
xix
IdC
Index of Coordination
K
Constant value
MAD-System
Measuring Active Drag System
m
Mass
m/s
metres per second
N
Newtons
ɳP
Propelling efficiency
P
Propulsion
ρ
Density of fluid
PB
Personal Best
Pd
Useful Power
Po
Mechanical Power Output
SD
Standard Deviation
SL
Stroke Length
SR
Stroke Rate
TE
Typical Error
TS
Towed Swimming
v
Swim Velocity
v1
Swim Velocity
v2
Assisted Swim Velocity
vb
Resisted Swim Velocity
VO2net
Net Maximal Oxygen Consumption
VPM
Velocity Perturbation Method
xx
Acknowledgements
The support and guidance of many people enabled the completion of this body of work. The
following is an attempt to show my gratitude and appreciation for all involved in my making
this PhD possible.
Firstly, to my mentor and supervisor in the field of swimming biomechanics, Dr
Bruce Mason, without your guidance and wisdom, this journey would not have been a
successful one! Sincere and genuine thanks are due and I have appreciated the opportunity to
learn and work alongside one of the world’s best.
Sincere gratitude is due to Dr Nick Ball for your academic supervision over the last
three and a half years! Our hot chocolate meetings were where these study designs, testing
protocols, and final conclusions came to life. Thank you for allowing my creativeness to
come forward through my work as well as tolerating my other thinking brain!
Considerable acknowledgment is also due to Dr Peter Clothier, you were there right at
the beginning of my biomechanics journey and I thank you for honing my scientific skills and
being a sounding board for me, particularly in the latter stages of this work.
To the staff of the Aquatic Testing, Training and Research Unit, at the Australian
Institute of Sport, I thank you immensely. To Dr David Pease, thank you for always having
the time for a quick chat, I appreciate the time you did give to me. To the post-graduate
scholars over the years James Critoph, Cecilia Nguyen, Nicholas Smith, Renata Franco,
Allison Higgs and Rebecca Pahl – you have all helped me in some form or another over the
years collecting data or listening to me sort through my ideas. I sincerely thank you for your
xxi
assistance and patience. Renata, I will always cherish your friendship through what turned
into a pinnacle year both personally and professionally. In particularly, staying up and
helping me re-number reference lists! To Pendar Hazrati, thank you for helping with testing
protocols and working out processing issues! Good luck on the completion of your own
work!
I am deeply indebted to the Australian Institute of Sport who enabled me access to
their amazing facilities and scientists. The technology pool is easily the best in the world and
being able to use this facility for testing made sessions run smoothly. Because of this
resource, a more comprehensive data set was able to be collected. Sincere appreciation is due
to Col McKintosh and Leon Williams. Col, thank you for the hours upon hours you helped
Bruce and I work through the software required for collecting data and thank you for coming
up with new ways to present our findings. Leon, thank you for all the time you spent with me
working on the ‘dyno’ – any problem we had, even if it wasn’t your speciality area – I could
always count on Leon coming over to the pool with his toolbox to help in any way he could!
I would like to extend my gratitude to the University of Canberra, Faculty of Health
Science for providing me the opportunity to pursue a Doctoral degree. The Research Student
Office has a brilliant programme which offers extensive resources enabling the successful
completion of this degree. From the writing workshops to the financial aid, I have enjoyed
being a part of the UC research community and would like to extend my utmost appreciation.
Appreciation is also due to the University of Western Sydney, particularly the School
of Science and Health biostatistician Paul Faye for his assistance with analysis.
xxii
I wish to express thanks to Andrew Dingley who was a great friend in the early days
of learning and understanding the basic concepts of active drag and who was great to bounce
ideas off.
To all the coaches and swimmers who were a part of my testing I am sincerely
grateful for your time and efforts. In particular Haydn Belshaw (for being the person I’ve
always spoken about swimming too), Brett Winkworth, Greg McWhirter, Coleman Wong,
Matt Brown, Steve Critoph, Cameron McDonald, Yuriy Vdovychenko and Scott Talbot. I
would also like to acknowledge the swimmers from Lane Cove Swim Club, Trinity Grammar
Swim Club, Cherrybrook Carlile, Macquarie University Swim Club, Sydney University
Swim Club, Ginninderra Swim Club, Woden Swim Club, AIS Swim Programme, and Marion
Swim Club for the involvement within the studies.
To my good friend Andrew who supported me in taking on this journey. Thank you
for encouraging me and listening to me whilst I worked through my ideas. To my many other
friends (Sydney, Canberra and swimming) who helped me through this stage in my life I am
truly grateful for your love and support. To Ben, thank you for motivating me in the final
stages of my work, I really appreciate your understanding, patience and reassurance!
Finally, to my parents Lynelle and Marc Sacilotto thank you for allowing me to
pursue my goals. Without your love and support I would not be the person I am today or
working towards a career that I am passionate about. Thank you from the bottom of my heart!
Thank you also to my brother William for - sometimes - taking an interest and listening to my
work stories! I love you all.
xxiii
Publications from Doctoral Thesis
Sacilotto GB, Ball N, & Mason BR. A Biomechanical Review of the Techniques Used to
Estimate or Measure Resistive Forces in Swimming. Journal of applied
biomechanics. 2014;30:119-127
Sacilotto GB, Mason BR, Ball N, & Clothier PJ. Investigation of coach ratings of technique
and force-time profiles in elite male front crawl sprint swimmers. In: Proceedings
for the 32nd International Society of Biomechanics in Sport; Jul 12-16, 2014;
Tennessee, United States.
Sacilotto GB, Clothier PJ, Mason BR, & Ball N. (2014) Variability in coach assessments of
technique in front crawl sprint swimming. In: Proceedings for the 12th International
Symposium on Biomechanics and Medicine in Swimming; Apr 28-May 2, 2014;
Canberra, Australia. p. 222-226
Sacilotto GB, Franco R, Mason BR, & Ball N. Investigation of front crawl stroke phases
within force-time profiles in elite and sub-elite male sprint swimmers. Journal of
Science and Medicine in Sport. 2013;16(Supp 1).
Sacilotto GB, Mason BR, & Ball N. Intra-reliability of active drag values using the assisted
towing method (ATM) approach. In: Proceedings for the 30th International Society
of Biomechanics in Sport; Jul 2-6, 2012; Melbourne, Australia.
xxv
Resistive Forces and Technique
Analysis in Front Crawl Sprint
Swimming
Chapter 1: Introduction
Chapter 1
Introduction
Background
Competitive swimming is a highly popular sport worldwide in which many individuals
participate. With the difference between winning and losing in the current era sometimes
being as little as one-hundredth of a second, small improvements in performance are highly
sought after by swimming researchers. A competitive swim performance is comprised of four
phases: 1) start, 2) free swim, 3) turn, and 4) underwater kicking. The ability to reduce the
time spent in each of these phases is the main goal for all research investigations in
competitive swimming. As the free swim phase has the largest proportion of time spent
within total race time, improvements in free swimming will have a greater bearing on overall
performance.1 Unfortunately, the free swim phase is also the most difficult in which to
measure forces as a consequence of the dynamic actions of a swimmer being at the air and
water interface.
During the free swim phase the two most commonly identified factors that are
primarily responsible for swim performance are propulsion and resistive forces.2-7 Propulsion
is generated by the swimmer to move faster through the water and as a consequence of the
propelling movements, resistive forces are encountered.8 Resistive force (or active drag)
occurs as a consequence of the three components of drag force in swimming which include
wave, form, and friction drag. The ability of a swimmer to generate higher propulsive force
and overcome an increased amount of active drag generally results in faster swim velocities.93
Chapter 1: Introduction
13
Clarys14 concluded that increased active drag was influenced mainly by the changes in the
body’s shape and the movement of the body segments, or a direct result of faulty swimming
techniques. Findings outlined in Kolmogorov et al.15 supported this claim by demonstrating
that elite swimmers had the ability to reduce active drag over their sub-elite counterparts.
These findings were speculated as being a result of elite swimmers having superior stroke
mechanic ratios of stroke length and stroke rate, and a generally higher technical proficiency.
Therefore, swimming fast may be dependent on the ability of a swimmer to reduce active
drag through an efficient stroke technique which will assist in generating a greater forward
resultant force.16
Three experimental techniques used to measure or estimate active drag frequently
reported include the Measuring Active Drag System (MAD-System),17 the Velocity
Perturbation Method (VPM),18 and the more recently developed Assisted Towing Method
(ATM).19 The MAD-System was engineered in 1986 by a group of Dutch scientists and is the
only method of the three that actually measured active drag. This was because the MADSystem required the swimmer to progress by pulling on submerged force transducer paddles
as they swum down the pool.17 Whilst the MAD-System is the most frequently listed method
in literature, it is limited by measuring the arm pull forces as the swimmer’s legs are bound
and whether the pulling on the paddles replicates the normal swimming action, therefore
restricting normal stroke mechanics. Furthermore, although the inter-pad distance can be
altered, the restriction and reliability of the method is still yet to be confirmed. Schreven et al.
recently attempted to provide reliability data for the inter-pad distances and confirmed that
using a fixed distance could be used, however the study only involved 11 swimmers
swimming at sub-maximal speed.20 The VPM method was first published in 1992 by Russian
scientists.18 This method estimates the active drag by calculating the difference in velocity
between a free swim condition and a resisted swim condition, whilst also taking into account
4
Chapter 1: Introduction
the known resistant force (the resistance of a hydrodynamic body attached to the swimmer
and towed behind them). Such active drag values are considered as ‘estimations’ as they are
indirectly assessing the active drag value. Additionally, the VPM relies on two assumptions:
1) the athlete swims with a constant mean swim velocity, and 2) the athlete produces equal
power efforts in both conditions. These assumptions rely on the swimmers’ capability and
compliance to reproduce equal maximal efforts in both the free swim and resisted conditions
and are considered a limitation of the VPM. Despite the limitations of the VPM method,
researchers from the Australian Institute of Sport accepted the two assumptions and modified
the testing protocol to calculate active drag using the mean velocities in a free swim and in an
assisted tow swim condition.19 The ATM has advanced over recent years by investigating
testing protocols allowing for intra-stroke velocity fluctuations, which in turn could permit
athletes to swim with closer to normal stroke mechanics.21 In addition, the ATM method has
the capability to capture instantaneous active drag force-time profiles whilst allowing the
intra-stroke velocity fluctuations, adding a new approach to the assessment of technique.21-23
However, until such time as an accurate free swim force and velocity profile can be sampled,
the need to explore alternate methods in the kinetic assessment of free swimming, is
important. The ATM method has already been demonstrated to be a reliable assessment
system for estimating active drag within preliminary work, therefore further investigation into
how it can be used to benefit coaches and athletes as an objective assessment tool is
warranted.
In summary, the free swim phase is the main determinant of race performance;
however it is the area in which scientists are more limited by their capacity to accurately
measure factors which affect performance. Despite previous methods using outdated
assumptions, the ATM has been shown within the literature to move the topic of free swim
kinetic analysis forward. With this perspective, the use of the ATM approach to estimate
5
Chapter 1: Introduction
active drag is justified along with exploring the use of active drag force-time profiles as a
means for assessing front crawl swimming technique.
Statement of Problem
The purpose of the present research was to assess the ATM protocol for integration as an
objective assessment tool for coaches, scientists and athletes in front crawl sprint swimming.
Currently, there is a gap in the literature with regard to methods of examining active drag
forces utilising a swimmer’s intra-stroke velocity fluctuations as well as limited research
presenting active drag force-time profiles as an objective assessment of technique. Therefore,
within the scope of this thesis, the performance reliability of swimmers using the ATM will
be assessed. Moreover, a comparison of stroke mechanic values between free swim and
assisted tow trials will be examined. Additionally, the establishment of stroke phases within
an active drag force-time profile will be outlined. Finally, the feasibility of utilising these
force profiles in the assessment of technique will be investigated. This was undertaken
through a series of progressive studies presented as publishable work with the following
outlines.
Literature Review:
The aim of this review was to assess the currently used methods to estimate or measure active
drag in swimming to progress this area of investigation. An evaluation of the methods
currently used to measure or estimate active drag was undertaken to provide scope for this
thesis. The following research questions were assessed within this review:
•
What is the relationship between active drag and swim performance?
6
Chapter 1: Introduction
•
What are the current and previous methods used to assess active drag?
•
What are the strengths and limitations of these methods?
Study 1:
The aim of the study was to assess the performance reliability of swimmers using the ATM
method. The swimmer population used was a mixture of elite and sub-elite mixed gender
swimmers that represented a range in competitive swimming capability. The following subproblems were examined within this investigation:
•
How many trials are required for swimmers to obtain a high level of performance
reliability in both constant and fluctuating tow velocity conditions using the ATM?
•
What is the level of performance reliability in both the constant and fluctuating tow
velocity conditions?
Study 2:
The results from study 1 revealed that the ATM protocol established a high level of swimmer
performance reliability within the sample tested. It was evident that the protocol utilising the
fluctuating velocity tow trials enabled swimmers to reproduce more consistent active drag
values. Moreover, it was established that a minimum of two trials is required to capture
reliable active drag values in both constant and fluctuating trials. The purpose of study 2 was
to investigate how the ATM alters stroke mechanics (stroke length and stroke rate) and
whether this comparison could enable validation for using this tool to objectively assess
technique in free swimming. A known limitation of the ATM protocol is the inclusion of the
7
Chapter 1: Introduction
‘equal power assumption’ which accepts that a swimmer applies the same effort in both
conditions. By assessing if there are indeed changes in stroke mechanics between the free
swim and the assisted swim condition, validation of the equal power assumption may be
established. The following sub-problems were outlined to answer the aims of this study:
•
Is there a significant increase between free swim and assisted swim stroke length and
stroke rate?
•
Is there a relationship between free swim stroke rate and stroke length ratios, and
assisted swim stroke length and stroke rate ratios?
•
Do these relationships justify the use of the ATM protocol to objectively critique free
swimming?
Study 3:
The results found in studies 1 and 2 demonstrated the reliability and justification for utilising
the ATM in critiquing free swimming technique. Findings from study 2 established a
significant increase in stroke length and stroke rate occurred when swimmers completed the
ATM protocol. In addition, the ratio of stroke length and stroke rate in the assisted condition
was found to significantly predict ratios of stroke mechanics in free swimming. Therefore,
using the ATM protocol, it is possible to transfer technique assumptions found in assisted
swimming and apply them to free swimming. However, these studies were limited to only
using mean active drag values. Therefore, the main purpose of study 3 was to investigate the
ATM active drag force-time profile. This was achieved through the following sub-problems:
•
What was the percentage of time spent in each stroke phase?
8
Chapter 1: Introduction
•
Are there similarities or differences in force magnitudes between male and female
active drag force-time profiles within a stroke cycle?
•
Are there similarities or differences between elite and sub-elite active drag force-time
profiles?
•
Can the individual stroke phases be identified from within the active drag force-time
profiles?
Study 4:
The results found in study 3 revealed a common biphasic characteristic within an elite
swimmer’s single stroke cycle. This common biphasic curve enabled assessment of the
timing of force production in relation to the phases of the stroke cycle. Variations in the
magnitude of force within the force profiles found between elite and sub-elite swimmers
could be a result of the velocity differences. Swimmer technique differences were suggested
as being the reason behind the variations in force-time profiles within the sub-elite sample.
The aim of study 4 was to examine whether the ATM active drag force-time profiles could be
utilised in the objective assessment of technique by quantifying coach ratings of technique by
determining if a relationship exists between the quality of technique and the profiles. To
achieve this, front crawl technique was objectively rated at the beginning of each stroke
phase by a group of coaches and correlated with the corresponding ATM active drag value
along the swimmers active drag force-time profile. Coach ratings of technique were captured
using a newly developed coach feedback survey. The following sub-problems were examined
using a series of experienced coaches:
9
Chapter 1: Introduction
•
Can technique be quantified through coach ratings of technique and what is the level
of agreement between coach ratings?
•
Is there a relationship between these coach ratings and the ATM active drag values at
selected events throughout a front crawl stroke using the force-time profiles?
Delimitations
For the purpose of the current study a number of delimiters were imposed:
•
Swimming participants must be registered with their State Swimming organisation
•
Studies were limited to swimmers who had achieved a 100 m front crawl performance
time of over 500 FINA points which was achieved in the last 12 months.
•
Swimming participants who specialised in 50 m, 100 m, or 200 m, event distances.
•
Swimming participants were only allowed to participate is they were free from injury
and illness which may limit their completion of testing protocols.
•
Swimming participants were required to complete testing at the Australian Institute of
Sport (AIS) technology pool in Canberra, Australia.
•
All testing protocols were conducted using the same equipment located at the AIS
technology pool
•
Coach participants were required to complete the coach feedback survey on all 30
swimming participants. If all 30 swimming participants were not analysed, then the
coach surveys’ from that coach participant were not included in the final study.
•
Coach participants were required to be registered with the Australian Swimming
Coaching and Teaching Association (ASCTA) and hold a Bronze or higher licence in
swim coaching.
10
Chapter 1: Introduction
Limitations
The ATM is based upon the equal power assumption where outcomes are reliant upon
participants completing all phases of testing with the same maximal effort. Furthermore, due
to swimming commitments and accessibility of participants traveling to Canberra small
sample sizes resulted. Similarly, coach feedback surveys were large and time consuming,
therefore of the large number of coaches who agreed to partake (n = 22) only a small number
completed the survey in its entirety (n = 8).
References
1.
Pai YC, Hay J, Wilson B, Thayer AL. Stroking techniques of elite swimmers. Medicine
& Science in Sports & Exercise. 1984;16(2):159.
2.
Barbosa TM, Costa MJ, Marques MC, Silva AJ, Marinho DA. A model for active drag
force exogenous variables in young swimmers. Journal of Human Sport & Exercise.
2010;5(3):379-88.
3.
Benjanuvatra N, Blanksby BA, Elliott BC. Morphology and hydrodynamic resistance in
young swimmers. Paediatric Exercise Science. 2001;13:246-55.
4.
Chatard JC, Bourgoin B, Lacour JR. Passive drag is still a good evaluator of swimming
aptitude. European Journal of Applied Physiology. 1990;59:399-404.
5.
Saavedra JM, Escalante Y, Rodriguez FA. A multivariate analysis of performance in
young swimmers. Pediatric Exercise Science. 2010;22:135-51.
11
Chapter 1: Introduction
6.
Toussaint HM, Truijens M, Elzinga MJ, et al. Effect of a Fast-skin 'body' suit on drag
during front crawl swimming. Sports Biomechanics. 2002;1(1):1-10.
7.
Toussaint HM. An alternative fluid dynamic explanation for propulsion in front crawl
swimming. In: Proceedings for the 18th International Society in Biomechanics in Sport;
Jun 25-30, 2000; Hong Kong, China. p. 96-103.
8.
Toussaint HM, Truijens M. Biomechanical aspects of peak performance in human
swimming. Animal Biology. 2005;55(1):17-40.
9.
Greco GC, Pelargio JG, Figueira TR, Denadai BS. Effects of gender on stroke rates,
critical speed and velocity of a 30-min swim in young swimmers. Journal of Sports
Science & Medicine. 2007;6:441-7.
10. Marinho DA, Barbosa TM, Costa MJ, et al. Can 8-weeks of training affact active drag
in young swimmers? Journal of Sports Science & Medicine. 2010;9:71-9.
11. Mollendorf JC, Termin II AC, Oppenheim E, Pendergast DR. Effect of swim suit
desgin on passive drag. Medicine & Science in Sports & Exercise. 2004;36(6):1029-35.
12. Smith JE, Molloy JM, Pascoe DD. The influence of a compressive laminar flow body
suit for use in competitive swimming. Journal of Swimming Research. 2007;17:10-6.
13. D’Acquisto L, Berry J, Boggs G. Energetic, Kinematic, and Freestyle Performance
Characteristics of Male Swimmers. Journal of Swimming Research. 2007;17:31.
12
Chapter 1: Introduction
14. Clarys JP. Human morphology and hydrodynamics. In: Proceedings of the 3rd
International Symposium of Biomechanics in Swimming; 1979; Edmonton, Canada; p.
3-41.
15. Kolmogorov SV, Rumyantseva OA, Gordon BJ, Cappaert JM. Hydrodynamic
characteristics of competitive swimmers of different genders and performance levels.
Journal of Applied Biomechanics. 1997;13:88-97.
16. Barbosa TM, Fernandes RJ, Keskinen KL, Vilas-Boas JP. The influence of stroke
mechanics into energy cost of elite swimmers. European Journal of Applied
Physiology. 2008;103:139-49.
17. Hollander AP, De Groot G, Van Ingen Schenau GJ, et al. Measurement of active drag
during front crawl arm stroke swimming. Journal of Sports Sciences. 1986;4:21-30.
18. Kolmogorov SV, Duplishcheva OA. Active drag, useful mechanical power output and
hydrodynamic force coefficient in different swimming strokes at maximal velocity.
Journal of Biomechanics. 1992;25(3):311-8.
19. Alcock A, Mason B. Biomechanical analysis of active drag in swimming.
In:
Proceedings of the 25th International Society of Biomechanics in Sports; Aug 23-27,
2007; Ouro-Preto, Brazil; p. 212-5.
20. Schreven S, Toussaint HM, Smeets JBJ, Beek PJ. The effect of different inter-pad
distances on the determination of active drag using the Measuring Active Drag system.
Journal of Biomechanics. 2013;46: 1933-1937
13
Chapter 1: Introduction
21. Mason B, Sacilotto G, Menzies T. Estimation of active drag using an assisted tow of
higher than max swim velocity that allows fluctuating velocity and varying tow force.
In: Proceedings of the 29th International Society of Biomechanics in Sports; Jun 27-Jul
1, 2011; Porto, Portugal; p. 327-30.
22. Mason B, Sacilotto G, Dingley A. Computation of a swimmer's propulsive force profile
from active drag parameters with fluctuating velocity in assisted towing.
In:
Proceedings of the 30th International Society of Biomechanics in Sport; Jul 2-6, 2012;
Melbourne, Australia.
23. Sacilotto G, Franco R, Mason BR, Ball N. Investigation of front crawl stroke phases
within force-time profiles in elite and sub-elite male sprint swimmers. Journal of
Science and Medicine in Sport. 2013;16(Suppl 1).
14
Chapter 2: Literature Review
Chapter 2
Literature Review
Published Manuscript
“A biomechanical review of the technique used to estimate or measure resistive forces in
swimming” Journal of Applied Biomechanics, 2014;30, 119-127
Appendix B
As co-authors of this paper, we confirm that Gina Sacilotto has made the following
contributions:
•
Conception and design of the research aims
•
Research and interpretation of the literature
•
Writing the paper and critical appraisal of its content
•
Corresponding author for communication with the journal
Signature:
Date: 13th June 2014
Ball, N.
The University of Canberra, Australia
Signature:
Date: 13th June 2014
Mason, BR.
Australian Institute of Sport, Australia
15
Chapter 2: Literature Review
Introduction
Understanding the resistive forces encountered within the free swim phase of a performance
is difficult. This difficulty is largely due to controversy surrounding the ability to measure
this force. Only a limited number of reviews were identified which discussed resistive forces,
or drag forces, in swimming and only one of those reviews were published in the last ten
years years.1-3 With the advancements in technologies and techniques, an updated review is
required to ensure sport scientists and coaches can accurately and effectively incorporate drag
testing within their athlete preparation and performance analysis.
Drag force is the force component parallel to and in the same direction as the relative
fluid flow.1 The equation that is used to calculate an object’s drag force (DF) is:
DF = 1 CD ρ v 2 A
2
Equation 2.1
where CD is the drag coefficient, ρ is the density of the fluid, v represents the velocity of the
object and A indicates the frontal surface area of the object. As water is the main fluid of
interest in this review, the key focus of all fluid dynamic principles will be hydrodynamics.
Hydrodynamic drag force in human swimming can be identified as the total resistive force
experienced by a swimmer.4
The resistive forces that influence the swimmer in the water include form, wave and
frictional drag5-7 which are influenced by the swimmer’s velocity, boundary layer, shape, size
and the frontal surface area8 as noted as well in equation 1. In swimming, the resistive forces
are termed active drag and passive drag. Active drag is the water resistance associated with
the dynamic swimming motion5, 9 and passive drag is the water resistance that a human body
experiences in a fixed or unchanging posture.5,9,10 Kolmogorov and Duplishcheva11
16
Chapter 2: Literature Review
confirmed that active drag varies between individuals and seems to relate to swim technique
and anthropometry.2 As noted in equation 1, in the context of human swimming, drag force
represents the swimming drag which could be an active or passive drag.9 For the purpose of
this review only active drag in relation to front crawl swimming will be discussed as it relates
to the performance of an individual swimmer.
This review was researched by database search engines using the following search
terms: active drag, swimming, swimming propulsive forces, front crawl technique, swimming
resistive forces. A reference check for each paper found was also performed. Searches were
also conducted using popular author’s names’ in this area of study. Therefore, the aim of this
paper was to assess the techniques used to estimate or measure active drag in swimming in
order to progress the study in this area.
Active Drag and Swim Performance
In competitive swimming the two most commonly identified factors that are primarily
responsible for swim speed are propulsion and drag.5,12-15 The ability of a swimmer to reduce
the active drag encountered allows for propulsive forces to be efficiently applied and
therefore producing faster swim velocities.8,16-19 Clarys20 confirmed that active drag was
mainly influenced by the changes in the body’s shape and the movement of the body
segments or, in other words, a direct result of faulty swimming techniques. Kolmogorov et
al.9 supported this claim by demonstrating that elite swimmers were more able to reduce
active drag than non-elite swimmers. These findings were speculated to be a result of elite
swimmers having superior stroke mechanics. Therefore, swimming fast may depend on the
ability of the swimmer to reduce drag through an efficient stroke technique which will
generate a higher velocity and limit the power lost in wasted kinetic energy.21 However,
17
Chapter 2: Literature Review
contrary to this argument, a review was completed in 1992,2 which outlined results found in
Hollander et al.22 as demonstrating no significant correlation in active drag and swim velocity
values at a constant swim velocity. As a result of these findings, it was concluded that active
drag was not a determining factor of maximal swimming when swimming with a constant
velocity. The previous review continues on to speculate that perhaps a swimmer’s
anthropometry may explain the active drag ranges seen in literature.2 However, research in
determining correlations between anthropometry and active drag have been minimally
investigated within the last two decades in relation to adult swimmers (see studies,5,8,23,24 for
research involving anthropometry and active drag in young swimmers). Further research into
determining the relationship between free swim velocity and active drag, whilst allowing for
the natural intra-stroke fluctuations, is important for updating the current understanding of
active drag and swim performance.
Mechanical Power Output in Swimming
It has been suggested that swimming performance is defined by the relationship between the
useful mechanical power output, active drag, the hydrodynamic force coefficient (drag
coefficient) and the maximal free swim velocity.11 The mechanical power output is the power
delivered by swimmers to overcome drag (useful power) and the power wasted in giving
kinetic energy change to the water.16,25 Mechanical power output (PO) has, therefore, been
evaluated as the product of the swimming drag (D) and velocity (v)11:
PO = D ⋅ v
Equation 2.2
The ratio between the useful power and the wasted kinetic energy is defined as the propelling
efficiency of a swimmer25:
18
Chapter 2: Literature Review
ηP =
Pd
PO
Equation 2.3
where ηP is the propelling efficiency and Pd is the useful power. Understanding of mechanical
power output and active drag has been the basis for many studies measuring or estimating
active drag.11,16,25,26
Techniques of Drag Assessment
For many years in swimming research, attempts have been made to accurately measure active
and passive drag,11,22,26-29 however, there has been much controversy as different techniques
have produced varying drag values.5,26,30,31 Energetics approach,32 numerical solutions6,33,34
and experimental techniques11,22,27,35 have been developed and used to estimate or measure
drag forces in swimming. All techniques have been modified and criticised as the demand for
understanding the resistive forces in swimming increases.
Energetics Approach
The energetics approach, or otherwise termed theoretical calculations, investigates the
theoretical relationship between the energy costs of swimming, the velocity, the overall
mechanical efficiency of the swimmer and the body drag.29 Predominantly the calculations
for this approach are aimed at deciphering the mechanical power output the swimmer
produces whilst free swimming.29,30,32,36-39 In this technique, the swimmer is towed while
swimming at a set pace – which is maintained by a towing carriage as seen in Figure 2.1 –
with known additional weights to provide assistance/resistance. The maximal oxygen
consumption is also recorded throughout each trial in order to understand the swimmer’s
energy expenditure at a given average velocity.40 The body drag of the swimmer is
19
Chapter 2: Literature Review
determined by adding (or subtracting) extra loads to (or from) swimmers moving at a known
speed. The extra drag was measured and related to the swimmer’s energy expenditure in
order to calculate the drag as well as the swimmer’s mechanical efficiency.40
Figure 2.1: Adapted experimental setup of drag collection from Di Prampero et al.40
To calculate the drag, Di Prampero et al.40 identified a linear relationship between
drag and maximal oxygen consumption (VO2net) at constant swim velocities which lead to
this technique of determining drag as a function of VO2net.20 Therefore, as stated by
Clarys,20(p13) this technique “extrapolated the linear regression between VO2net and the added
propulsion and added drag to VO2net = 0”. At a constant mean velocity, the mean propulsive
force exerted by the swimmer will be equal and opposite to the active drag produced.
20
Chapter 2: Literature Review
Investigations, which utilise the energetics approach,38,40,41 found similar values of
active drag when comparing propelling efficiency values as a percentage. Although, intrastudy drag values were similar, it must be noted that these authors assumed that the
propelling efficiency did not change in experiments where active drag was calculated.42 It is
likely that propelling efficiency will change, even at a constant velocity, when external loads
are applied as is the case in the approach to estimating active drag. Also, small changes in
VO2net values due to small deviations in propelling efficiency will be amplified by the
extrapolations which are the basis for these studies.42 Van De Vaart et al.43 reiterated this
point by determining that indirect techniques of estimating active drag (by extrapolation)
appeared to overestimate these values. Furthermore, by introducing a snorkel to measure the
VO2net, the swimmers frontal surface area is altered, which could modify the results for active
drag. Although, this technique for estimating active drag, as we now understand, produced
questionable results, it was the first to describe the total active drag of a front crawl
swimmer20 and therefore the first step in the progression of all the current techniques used
today. Because of this technique, the investigation into active drag values between male and
female swimmers, swimmers with different swim velocities as well as the analysis of the
energetics in swimming were able to undertaken.40
Numerical Simulations
Numerical simulations utilise the computational modelling of the water flow surrounding the
swimming to determine the resistive forces.8 The main approach to drag force measurement
using numerical solutions is through Computational fluid dynamics (CFD). Computational
fluid dynamics solves and analyses problems involving fluid flow by means of computerbased simulations.44 Using this method, the investigator can analyse computer models, for
example a 3D-model of a swimmer, and can simulate desired movement patterns in order to
21
Chapter 2: Literature Review
give feedback about the alteration, for example a modification in stroke technique. By
manipulating a computer model instead of a human form, studying the drag values in
swimming through CFD, theoretically, limits the amount of stroke mechanic alterations due
to the constraints set in the experimental protocols.31 Bixler and Schloder45 introduced twodimensional CFD into the swimming world and then after a further six years, the first threedimension CFD study was published.46 Since these landmark studies in swimming CFD,
several authors have investigated hydrodynamic drag using this method.6,33,45,47
Computational fluid dynamic simulations allow the elimination of within-subject
variability, which is found in laboratory and field experimentation. Furthermore, another
benefit of CFD is that for the same input you always have the same output. Bixler, Pease and
Fairhurst33 presented an investigation on the study of the water flow and drag force
characteristics (acting around and upon a human body) while in a submerged streamlined
position. In this study a comparison of total drag force was performed between an actual
swimmer, a virtual CFD model of the swimmer and an actual mannequin based on the virtual
model. Although this study only investigated the effects of passive drag, the results were
positive as the aim of establishing a CFD model of a submerged human body and the effects
of passive drag were achieved.33 In addition, Bixler et al.33 demonstrated the accuracy of
using the CFD technique by comparing model values with a real human. Therefore, the
potential for this method in resistive force assessment is quite promising and the results of the
previous paper represents a necessary first step towards more complicated CFD analysis in
which active drag could be evaluated.33 For CFD to become a readily available method of
resisted force assessment, basic kinematic measures during free swimming need to be
collected. For example, the ability to collect instantaneous swim velocity, or knowing where
the centre of gravity is whilst a swimmer is swimming. A further limitation of this technique
is that generally speaking, CFD simulations require an enormous amount of computing time,
22
Chapter 2: Literature Review
which is particularly true in swimming analysis. In the analysis process of a human
swimming, one has to move boundaries in the simulation and also resolve the flow around a
complex-shaped, deforming object.31 These extra computing requirements can further
increase the complexity and computational costs of the simulation,31 therefore, making it
difficult for coaches and scientists to utilise this method, in its current state, effectively.
Experimental Techniques
Experimental techniques have been developed and applied to try and accurately determine the
resistive forces encountered by a swimmer. The techniques most frequently found include the
direct measurement of active drag through the measuring active drag system (MADSystem),22 and the indirect techniques of collecting active drag values, for example, the
velocity perturbation method (VPM)11 and the assisted towing method (ATM).27
Measuring Active Drag System
In order to directly measure active drag, Hollander et al.22 developed the MAD-System. This
device measures the drag force generated by a swimmer, which enables the calculation of the
propulsive force produced during the trial. In order to obtain propulsive force, the assumption
was made that the mean propulsive force would be equal to the mean active drag values when
the swim velocity is constant.22
The MAD-system, as seen in Figure 2.2, requires the swimmer to push-off fixed pads
underneath the water. In the original study using this technique ten trials were completed at
different, however, constant velocities. The swimmer’s legs were restricted by the use of a
small buoy. The depths of the pads were able to be adjusted for the swimmer’s height as well
as the distance between pads. For each trial, the registered output signal of the force
23
Chapter 2: Literature Review
transducer was transmitted telemetrically in order to determine mean force. The average
propulsive force was calculated by integration from the force registrations at a constant swim
velocity. The swim velocity was determined from the sample frequency and the pad distance
(between the second and final pad).22 In the initial investigation of this technique, each test
yielded ten data points of propulsive forces at ten different speeds, which ranged from
minimal to maximal swim velocity.22 The original function used to calculate active drag can
be found in Hollander et al.,22 however in Toussaint et al.48 presented the calculation for drag
as
DA = Kv 2
Equation 2.4
where DA represents total active drag, K is a constant (incorporating the density, coefficient
of drag and frontal surface area) and v equals swim velocity.
Figure 2.2: MAD-System setup for drag collection adapted from Hollander et al.22
The MAD-System has been used extensively in swimming research in determining
direct values of active drag.7,22,25,48-50 Although there has been extensive research undertaken
24
Chapter 2: Literature Review
utilising the MAD-System, there is much criticism surrounding this technique. For example,
the system limits a swimmer’s natural stroke mechanics26,28,51 and it can only be used at a
constant velocity. Therefore, the outcomes from this technique should only be compared
against itself or if a swim velocity is the same between techniques in order to critique front
crawl pull technique. Another criticism is that the MAD-System protocols only allow the
swimmer’s hands to be in contact with the pads and not react with water as a swimmer would
normally. Hollander et al.,22 made note that normal hand trajectories may be altered with this
technique, however it was justified by stating that at the same swim velocity a different hand
trajectory did not necessarily imply a difference in active drag. Poizat et al.51 concentrated on
testing the MAD-System for useability as a training device for biomechanical evaluation and
performance analysis. Swimmer feedback was recorded along with the active drag values.
The swimmers involved in this study described that it was difficult to make contact with the
pads, particularly at high velocities.51 This technique is, however, well established and an
effective way to directly measure the forces encountered and produced by a swimmer’s upper
body throughout a maximal effort.
Velocity Perturbation Method
The VPM approach is based on the assumption that a swimmer is capable of producing an
equal amount of useful mechanical power output and that the swimmer will swim at a
constant velocity.11 This technique is seen as a progression from the energetics approach in
estimating active drag, however, without the inclusion of the maximal oxygen consumption
element. In the VPM a swimmer must produce two equal maximal efforts. This technique is
most commonly used over a 25 m distance; however it can be used across other distances.
The first maximal effort, the swimmer must swim ‘freely’ – without any attachments – and
the second effort is swum with a hydrodynamic body attached to the swimmer, creating a
25
Chapter 2: Literature Review
known additional resistance. Both conditions must be swum across the same distance. The
maximal mean velocity when swimming with the hydrodynamic body was compared with the
maximal mean free swimming velocity, which along with the known additional resistance is
used to calculate active drag for free swimming:
Db vb v 2
DA = 3 3
v − vb
Equation 2.5
where Db is the additional resistance from the perturbation buoy and vb and v are the
swimming velocities with and without the hydrodynamic body, respectively.11
Although this method has been said to be worth pursuing in research,23 it is criticised
as it is an indirect technique of measuring active drag7 and therefore the drag values found
may be overestimated.16 Furthermore, the assumptions to which this technique adheres to are
very much reliant on the level of swimmer’s participating. The first assumption of equal
power is dependent on the swimmer’s skill level, the rest interval between trials and whether
or not the swimmer understands the experimental conditions.11 This suggests that the use of
this technique, and others which utilise the equal power assumption, should limit participants
to those who are of a semi-elite or an elite level of swimming capability. The second
assumption of constant velocity limits the study of the intra-stroke fluctuations normally
produced in a front crawl stroke cycle. Kolmogorov and Duplishcheva11 revealed a varying
velocity during the trials and these intra-stroke velocity fluctuations were approximated using
a computer simulated Strouhal number.11 It was found that the maximal error due to the
stroke cycle fluctuations was around six – eight %. As a result of the VPM being reliant using
a constant velocity throughout resisted trials, the fluctuations were diminished in order to
maintain a near constant velocity. The hydrodynamic body that was used was built to not
decrease the swimmer’s velocity by more than 10 %.11 The restricting of the intra-stroke
26
Chapter 2: Literature Review
velocity fluctuations is imperative in to use of Equation 4 when determining mean active drag
values.
As a result of the near constant velocity assumption, a series of hydrodynamic bodies
were developed, each with a different additional resistance, to eliminate the dependency on a
swimmer’s performance level. However, it has been noted that the additional resistance
created by a hydrodynamic body can be affected by the floating movements generated by the
hydrodynamic body. Wang et al.,26 therefore proposed to develop a simple and convenient
device to estimate the active drag at maximal velocity based on the equal power assumption
and the VPM approach. Modifications that were made from the original VPM approach and
this version of the VPM approach was how the additional resistance was applied to the
swimmer. Figure 2.3 illustrates the apparatus developed by Wang et al.26 which maintained
the additional resistance in a steady position. This system minimises the floating movement
of the hydrodynamic body and allowed changes in the amount of additional resistance.
Figure 2.3: Modification of the VPM approach as utilised in, and adapted from Wang et al.26
The force transducer (shown in Figure 2.3) was used to measure the variation in
tension of the thread when the gliding block was moved by the swimmer. The results
27
Chapter 2: Literature Review
revealed that the tension of the thread fluctuates and, as a result, the additional resistance in
the swimming direction is variable, not a constant value as Kolmogorov and Duplishcheva11
had assumed in their assumptions. Therefore, in the original VPM approach, even if the
velocities were set to be constant, active drag values may differ due to the velocity fluctuation
restrictions.
Despite the limitations of this technique for assessing active drag the VPM has several
strengths. For example, this technique can be set up in any pool facility as it is completely
portable. This means that testing could occur on away meets or camps which could provide
coaches and athletes with extra information. Similar the VPM could be easily integrated into
a normal training session, unlike the MAD-System which requires a long set up time. Also in
comparison with the MAD-System, the VPM requires little to no adaptation to complete
analysis on the other strokes (i.e. butterfly, backstroke or breaststroke).
A comparison of the VPM and MAD-System was conducted in 2004,48 which
revealed that the two techniques yielded significantly different active drag values. Given the
stroking limitations of both techniques, this significant difference does not imply that either
technique is wrong or measuring different aspects of free swimming.48 It was concluded that
the question of which aspect of free swimming is being assessed by either the VPM or MADSystem was not resolved.
Assisted Towing Method
The ATM technique for estimating active drag has only been introduced in recent years.
Currently, only few studies have been published using this technique for determining active
drag.27,28,52-55 In its current format, the ATM is essentially the reverse of the VPM approach
28
Chapter 2: Literature Review
in estimating active drag, as the swimmer is assisted, rather than resisted. The ATM is also
based on the equal power assumption and the constant velocity assumptions. However, as
was outlined by Wang et al.26 and recognised in Kolmogorov and Duplishcheva,11 a swimmer
will not be swimming at a constant velocity at any point throughout a maximal effort due to
the intra-stroke fluctuations in front crawl swimming. Such fluctuations are a result of the
intra-stroke forces that are generated during a natural arm stroke cycle. The kicking action in
the front crawl stroke will also add to these fluctuations.
To further understand these fluctuations, Mason et al.28 compared constant active drag
values with fluctuating active drag values (Figure 2.4).
Figure 2.4: ATM technique to drag collection as shown in Sacilotto et al.55
During the constant trials the difference in the mean velocity maximal free swims
(individual swimming without any attachments) and the mean velocity of the towed swims
(towed from the hip using the dynamometer) are used to calculate for drag with respect to the
drag force required to tow the swimmer. The same calculations as the VPM approach are
used; however, the final equation for estimating active drag using the ATM is altered to:
29
Chapter 2: Literature Review
DA =
Fbv2v12
v23 − v13
Equation 2.6
where Fb is the force required to tow the athlete at the increased speed as measured from the
force platform, v2 is the increased tow velocity, and v1 is the maximum free swim velocity.
Similar to the VPM approach, when using this system with a constant velocity, the
dynamometer is set to five-percent faster than the swimmer’s mean maximum free swim
velocity with a high force selection to allow for a near constant tow. In order for the tow to
allow the swimmer’s intra-stroke fluctuations the force setting on the dynamometer is
reduced and the velocity setting is increased to 120 % of the swimmers maximum free swim
velocity. Along with these changes in set force and velocity settings a parameter on the
towing dynamometer is altered so that when the force setting is reached it will fluctuate the
tow velocity to maintain that force setting, therefore allowing the intra-stroke fluctuations.
The force setting used is a predetermined fraction of the swimmer’s passive drag tow
(streamlined tow at the swimmer’s maximal free swim velocity) and is different for every
individual swimmer. Despite the increase in the velocity setting the mean tow velocity will
still equal between five to ten percent greater than the swimmer’s maximal mean free swim
velocity, however when calculated the velocity profile will demonstrate the intra-stroke
fluctuations.28 The results obtained from the fluctuating trials seem to demonstrate a
smoother drag profile, more repeatable results as well as most likely resembling more natural
stroke characteristics than what was illustrated in the constant trials.28 The use of the ATM
approach, when allowing for a fluctuating tow velocity in active drag estimation, is still in its
infancy. However, the results shown thus far are positive in being able to decipher exactly
what affects performance during free swimming. A recent review by Sanders et al.3 noted that
30
Chapter 2: Literature Review
the work presented in Mason, et al.28 indeed needed validation however appeared promising.
Since then two new papers were published using the ATM technique.54,55 When towing with
a constant velocity, the assumption has been that drag was equal but opposite in direction to
propulsion. However, when utilising a tow allowing for intra-stroke fluctuations this cannot
be true. A recent study using the ATM has developed calculations to obtain a swimmer’s
propulsive profile, net force, and acceleration curves54 whilst allowing their intra-stroke
velocity fluctuations. Mason et al.54 presents active drag as a negative value with the equation
for propulsion (P) as:
P=
d
( mv ) − DA
dt
Equation 2.7
where m is the passive drag force of the swimmer (as a substitute for the mass of a swimmer),
v is the velocity profile , DA is the active drag profile presented as a negative value (P is the
propulsion profile presented as a positive value). The practical applications using ATM and
this propulsion equation again appear promising; however the validation of this method is
still under question as there is no experimental gold standard to validate against. In an attempt
to somewhat validate this technique, an intra-reliability study was conducted utilising ATM,
however, with a constant velocity tow.55 Although this paper only included a small sample
size, the results revealed very good reliability value for within-subject mean active drag
values (interclass correlation of 0.91, at a confidence limit of 95% and a likely range of 0.58
and 0.98). Future studies need to be completed utilising a fluctuating tow velocity in order to
determine the intra-reliability of this technique. Further investigations into whether or not the
velocity/force profiles obtained in this technique actually mimic real stroke mechanics will
also need to be undertaken. However, research cannot be continued with this until researchers
31
Chapter 2: Literature Review
are able to accurately measure basic kinematics whilst a swimmer is submerged in water.
When basic kinematics become available a comparison between a fluctuating tow velocity
trial and a maximal free swim velocity trial can be investigated, at which point this technique
can be validated. Furthermore, similar to the MAD-System, the ATM technique has only
presented papers in the investigation into front crawl resistive and propulsive forces. It could
be assumed that research into backstroke, using ATM, could be undertaken following the
same protocols as is for front crawl, however investigations into butterfly and breaststroke
seems a while away given the very large fluctuations in intra-stroke velocities.
Review Summary
Swim performance is very much dependent on the free swimming component of an event.
Therefore, gains or losses in this aspect of a swim performance can significantly affect the
outcome of an event. The purpose of this paper was to evaluate the current techniques used to
estimate active drag in elite front crawl free swimming.
Techniques for experimental drag collection found in literature appear to be based on
the ideals from the energetics approach in relation to mechanical power output39,
40
even
though these were established in the 1970s. More recently, the experimental indirect
techniques, although limited, have been found most frequently in literature.11, 27 The MADSystem,22 which is the only system to directly measure active drag, has been shown to have a
large number of limitations in regards to maintaining the swimmer’s natural stroke
mechanics. Although, when compared to the ATM and VPM, the MAD-System is more
established and can be compared to itself. Alternatively, the VPM is a more cost effective
technique and can be transported easily, unlike the MAD-System and ATM where the set-up
time and cost of equipment is quite high. The preference of researchers to use indirect
assessment, however, allows the swimmer to perform more natural stroke mechanics whilst
32
Chapter 2: Literature Review
they swim, in particular, when allowing normal velocity fluctuations, which can be achieved
using the ATM. The main concern with the indirect techniques has been shown to be the
assumptions associated with the testing protocols. In order to overcome these assumptions,
the CFD method would seem to be the best option in determining resistive and propulsive
forces in free swimming. However, until such time as the basic kinematic measures can be
determined in free swimming, CFD research also relies on the experimental assumptions.
Although this review is not exhaustive of all the techniques used around the world in
the measurement or estimation of active drag, a clear outline of the most common techniques
used can be obtained. Therefore, as a guide, the choice of technique or method used to
estimate or measure active drag should be dependent on what researchers want to gain from
the testing. For the pursuit of understanding what active drag actually is, the assessment
technique which allows a swimmer to be most natural in the water, or is able to be simulated
naturally, could be the best scientific step forward.
References
1.
Rushall BS, Holt LE, Sprigings EJ, Cappaert JM. A re-evaluation of forces in
swimming. Journal of Swimming Research. Fall 1994;10:6-30.
2.
Toussaint HM, Beek PJ. Biomechanics of competitive front crawl swimming. Sports
Medicine. 1992;13(1):8-24.
3.
Sanders R, Thow J, Alcock A, Fairweather M, Riach I, Mather F. How Can
Asymmetries in Swimming be Identified and Measured? Journal of Swimming
Research. 2012;19:1.
33
Chapter 2: Literature Review
4.
Lyttle AD, Blanksby BA, Elliott BC, Lloyd DG. The effect of depth and velocity on
drag during the streamlined glide. Journal of Swimming Research. 1998;13:15-22.
5.
Benjanuvatra N, Blanksby BA, Elliott BC. Morphology and hydrodynamic resistance in
young swimmers. Paediatric Exercise Science. 2001;13:246-255.
6.
Marinho DA, Reis VM, Alves FB, et al. Hydrodynamic Drag During Gliding in
Swimming. Journal of Applied Biomechanics. 2009;25(3):253-257.
7.
Toussaint HM, De Groot G, Savelberg H, Vervoorn K, Hollander AP, van Ingen
Schenau GJ. Active drag related to velocity in male and female swimmers. Journal of
Biomechanics. 1988;21(5):435-438.
8.
Marinho DA, Barbosa TM, Costa MJ, et al. Can 8-weeks of training affact active drag
in young swimmers? Journal of Sports Science & Medicine. 2010;9:71-79.
9.
Kolmogorov SV, Rumyantseva OA, Gordon BJ, Cappaert JM. Hydrodynamic
characteristics of competitive swimmers of different genders and performance levels.
Journal of Applied Biomechanics. 1997;13:88-97.
10. Havriluk R. Performance Level Differences in Swimming: A Meta-Analysis of Passive
Drag Force. Research Quarterly for Exercise and Sport. 2005;76:112-118.
11. Kolmogorov SV, Duplishcheva OA. Active drag, useful mechanical power output and
hydrodynamic force coefficient in different swimming strokes at maximal velocity.
Journal of Biomechanics. 1992;25(3):311-318.
34
Chapter 2: Literature Review
12. Barbosa TM, Costa MJ, Marques MC, Silva AJ, Marinho DA. A model for active drag
force exogenous variables in young swimmers. Journal of Human Sport & Exercise.
2010;5(3):379-388.
13. Chatard JC, Bourgoin B, Lacour JR. Passive drag is still a good evaluator of swimming
aptitude. European Journal of Applied Physiology. 1990;59:399-404.
14. Toussaint MH. An alternative fluid dynamic explanation for propulsion in front crawl
swimming. In: Proceedings of the 18th International Society of Biomechanics in Sports;
June 25-30, 2000; Hong Kong, China.
15. Toussaint HM, Truijens M, Elzinga MJ, et al. Effect of a Fast-skin 'body' suit on drag
during front crawl swimming. Sports Biomechanics. 2002;1(1):1-10.
16. D’Acquisto L, Berry J, Boggs G. Energetic, Kinematic, and Freestyle Performance
Characteristics of Male Swimmers. Journal of Swimming Research. 2007;17:31.
17. Greco GC, Pelargio JG, Figueira TR, Denadai BS. Effects of gender on stroke rates,
critical speed and velocity of a 30-min swim in young swimmers. Journal of Sports
Science & Medicine. 2007;6:441-447.
18. Mollendorf JC, Termin II AC, Oppenheim E, Pendergast DR. Effect of swim suit
desgin on passive drag. Medicine & Science in Sports & Exercise. 2004;36(6):10291035.
19. Smith JE, Molloy JM, Pascoe DD. The influence of a compressive laminar flow body
suit for use in competitive swimming. Journal of Swimming Research. 2007;17:10-16.
35
Chapter 2: Literature Review
20. Clarys JP. Human morphology and hydrodynamics. In: Proceedings fo the 3rd
International Symposium of Biomechanics and Medicine in Swimming; 1979;
Edmonton, Canada.
21. Barbosa TM, Fernandes RJ, Keskinen KL, Vilas-Boas JP. The influence of stroke
mechanics into energy cost of elite swimmers. European Journal of Applied
Physiology. 2008;103:139-149.
22. Hollander AP, De Groot G, Van Ingen Schenau GJ, et al. Measurement of active drag
during front crawl arm stroke swimming. Journal of Sports Sciences. 1986;4:21-30.
23. Marinho DA, Garrido N, Barbosa TM, et al. Can 8 Weeks of Training in Female
Swimmers Affect Active Drag? Open Sports Sciences Journal. 2010;3:36-37.
24. Kjendlie P, Stallman RK. Drag characteristics of competitive swimming children and
adults. Journal of Applied Biomechanics. 2008;24(1):35.
25. Toussaint HM. Differences in propelling efficiency between competitive and triathlon
swimmers. Medicine & Science in Sports & Exercise. 1990;22(3):409-415.
26. Wang X, Wang L, Yan W, Li D, Shen X. A new device for estimating active drag in
swimming at maximal velocity. Journal of Sports Sciences. 2007;25(4):375-379.
27. Alcock A, Mason B. Biomechanical analysis of active drag in swimming.
In:
Proceedings of the 25th International Society of Biomechanics in Sports; Aug 23-27,
2007; Ouro-Preto, Brazil; p. 212-5.
36
Chapter 2: Literature Review
28. Mason B, Sacilotto G, Menzies T. Estimation of active drag using an assisted tow of
higher than max swim velocity that allows fluctuating velocity and varying tow force.
In: Proceedings of the 29th International Society of Biomechanics in Sports; Jun 27-Jul
1, 2011; Porto, Portugal; p. 327-30.
29. Pendergast DR, Di Prampero PE, Craig AB, Wilson DR, Rennie DW. Quantitative
analysis of the front crawl in men and women. Journal of Applied Physiology.
1977;43(3):475-479.
30. Zamparo P, Gatta G, Pendergast DR, Capelli C. Active and passive drag: The roles of
trunk incline. European Journal of Applied Physiology. 2009;106:195-205.
31. Von Loebbecke A, Mittal R, Mark R, Hahn J. A computational method for analysis of
underwater dolphin kick hydrodynamics in human swimming. Sports Biomechanics.
2009;8(1):60-77.
32. Larsen OW, Yancher RP, Baer CLH. Boat design and swimming performance.
Swimming Technique. 1981;18(2):42.
33. Bixler B, Pease D, Fairhurst F. The accuracy of computational fluid dynamics analysis
of the passive drag of a male swimmer. Sports Biomechanics. 2007;6(1):81-98.
34. Marinho DA, Rouboa AI, Alves FB, et al. Hydrodynamic analysis of different thumb
positions in swimming. Journal of Sports Science & Medicine. 2009;8(1):58-66.
37
Chapter 2: Literature Review
35. Lyttle AD, Elliot BC, Blanksby BA, Lloyd DG. An instrument for quantifying the
hydrodynamic drag of swimmers. A technical note. Journal of Human Movement
Studies. 1999;37(5):261-270.
36. Clarys JP, Jiskoot J. Total resistance of selected body positions in the front crawl. In:
Proceedings of the 2nd International Symposium on Biomechanics and Medicine in
Swimming; 1974; Brussels; p. 110-117
37. Clarys JP, Jiskoot J, Rijken H, Brouwer PJ. Total resistance in water and its relation to
body form. In Nelson, R.C. and Morehouse, C.A. (ed.), Biomechanics IV, Baltimore,
University Park Press, 1974, p. 187-196.1973.
38. Kemper HC, Verschuur R, Clarys JP, Jiskoot J. Total efficiency and swimming drag in
swimming the front crawl. Biomechanics in Swimming, Waterpolo and Diving.
1983;119-125.
39. Rennie DW, Pendergast DR, Di Prampero PE. Energetics of swimming in man.
Swimming II. 1975;97-104.
40. Di Prampero PE, Pendergast DR, Wilson DW, Rennie DW. Energetics of swimming in
man. Journal of Applied Physiology. 1974;37(1):1.
41. Holmer I. Propulsive efficiency of breaststroke and freestyle swimming. European
Journal of Applied Physiology & Occupational Physiology. 1974;33(2):95-103.
42. Toussaint HM, Van Der Helm FCT, Elzerman JR, Hollander AP, De Groot G, van
Ingen Schenau GJ. A power balance applied to swimming. In, Hollander, A.P. (ed.) et
38
Chapter 2: Literature Review
al., Biomechanics and medicine in swimming: proceedings of the Fourth International
Symposium of Biomechanics in Swimming and the Fifth International Congress on
Swimming Medicine, June 21-25, 1982, Champaign, Ill., Human Kinetics Publishers,
c1983, p. 165-172. United States 1983.
43. Van De Vaart AJM, Savelberg HHCM, De Groot G, Hollander AP, Toussaint HM, Van
Ingen Schenau GJ. An estimation of drag in front crawl swimming. Journal of
Biomechanics. 1987;20(5):543-546.
44. Marinho DA, Rouboa AI, Barbosa TM, Silva AJ. Modelling Swimming
Hydrodynamics to Enhance Performance. Open Sports Sciences Journal. 2010;3:43-46.
45. Bixler B, Schloder M. Computational fluid dynamics : an analytical toll for the 21 st
century swimming scientist. Journal of Swimming Research. Fall 1996;11:4-22.
46. Bixler B, Riewald S. Analysis of a swimmer's hand and arm in steady flow conditions
using computational fluid dynamics. Journal of Biomechanics. 2002;35(5):713-717.
47. Machado L, Ribeiro J, Costa L, et al. The effect of depth on the drag force during
underwater gliding: A CFD appraoch. In: Proceedings of the 28th International Soceity
of Biomechanics in Sports; July, 2010; Michigan, United States of America.
48. Toussaint HM, Roos PE, Kolmogorov S. The determination of drag in front crawl
swimming. Journal of Biomechanics. 2004;37(11):1655-1663.
39
Chapter 2: Literature Review
49. Toussaint HM, DeLooze M, Van Rossem B, Leijdekkers M, Dignum H. The effect of
growth on drag in young swmimers. International Journal of Sport Biomechanics.
1990;6:18-28.
50. Toussaint HM, Truijens M. Power Requirements for Swimming a World-Record 50-m
Front Crawl. International Journal of Sports Physiology & Performance. 2006;1(1):6164.
51. Poizat G, Adé D, Seifert L, Toussaint HM, Gal-Petitfaux N. Evaluation of the
Measuring Active Drag system usability: An important step for its integration into
training sessions. International Journal of Performance Analysis in Sport.
2010;10(2):170-186.
52. Formosa D, Mason B, Burkett BJ. The force-time profile of elite front crawl swimmers.
Journal of Sports Sciences. 2011;29(8):811-819.
53. Mason B, Formosa D, Toussaint HM. A method to estimate active drag over a range of
swimming velocities which may be used to evaluate the stroke mechanics of the
swimmer. In: Proceedings of the 11th International Symposium of Biomechanics and
Medicine in Swimming; un 16-19, 2010; Oslo, Norway.
54. Mason B, Sacilotto G, Dingley A. Computation of a swimmer's propulsive force profile
from active drag parameters with fluctuating velocity in assisted towing. In:
Proceedings of the 30th International Society of Biomechanics in Sport; Jul 2-6, 2012;
Melbourne, Australia.
40
Chapter 2: Literature Review
55. Sacilotto G, Mason B, Ball N. Intra-reliability of active drag values using the assisted
towing method (ATM) approach. In: Proceedings of the 30th International Society of
Biomechanics in Sports; Jul 2-6, 2012; Melbourne, Australia.
41
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
Chapter 3
Study 1: Reliability of Active Drag Values Using the
Assisted Towing Method
Introduction
Successful swim performance is the ability of the swimmer to optimise the relationship
between the active drag encountered and the propulsion generated during free swimming.1
For many years in swimming research, attempts have been made to accurately assess active
drag.2-6 However, there has been much controversy as the methods used often restrict or alter
the natural stroke mechanics of the swimmer.5-10 The most frequently used experimental
techniques in measuring swimming forces include the Measuring Active Drag (MAD)
System,4 and the Velocity Perturbation Method (VPM).5 Both methods have been questioned
for restricting normal stroke mechanics, however as there are limited alternatives they are
accepted as methods for assessing free swimming active drag. A review of the two methods
concluded that although they reported different mean active drag values, both methods could
still be estimating the same element of free swimming.11 This conclusion was explained as
the participants were most likely violating the required assumptions by not reproducing equal
power between free swim and experimental conditions.11
A recent innovation for measuring active drag during swimming has been the
development of the Assisted Towing Method (ATM). However, without readily available
kinetic free swim velocity information the question of how reliable the methods used to
measure active drag are often asked. Therefore, examining the reliability of an active drag
43
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
value produced from the ATM is important in its development. Hopkins12 states that “better
reliability implies better precision of single measurements and better tracking of changes in
measurement in research or practical settings”.12(p1) Simply put, if there were measurement
systems that were reliable at reproducing active drag values, more extensive analysis could be
undertaken on the assessment of performance and how the relationship between active drag
and propulsion affects free swimming. Therefore, examining and establishing the reliability
of kinetic values obtained from these methods is important for determining the validity of the
information they produce.
Attempts have been made to examine the validity and reliability of the MAD-System
and VPM to prove the reliability performance of their active drag values. Hollander et al.(4)
assessed the reliability in the measurements obtained from the MAD-system by having one
participant complete five tests over five days across 10 different swim velocities. Findings
revealed excellent agreement was found between the propulsive force values from the MADSystem and a previous investigation.4, 13 Furthermore, stroke characteristics were assessed by
asking 140 coaches to select the participant they thought was using MAD-System from three
above water video files. On average, only 50 % of the coaches were able to successfully
identify which participant was using the MAD-System and which was swimming normally.
Kolmogorov and Duplishcheva5 assessed validity by comparing the passive drag forces of
participants being towed at their free swim velocity and at their resisted swim velocity as
determined using the VPM. As the reported error between free swim passive drag and
resisted velocity passive drag was only 1 % it was concluded that the VPM method was
considered valid.5 Additionally, Wang et al.6 slightly modified the VPM method, by altering
how the resistance was administered, and reported that the reliability of this method was
established by repeating the measurement four times in each condition (both free swim and
44
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
resisted swim) on every participant. However, no reliability statistics were presented to
demonstrate how accurate this modification of the VPM was.6
The newest method being investigated for estimating active drag has recently
produced two studies exploring the reliability of its values. The ATM was first published in
2007 and is also a modification of the VPM technique; however instead of resisting the
swimmer, the ATM assists the swimmer in the second condition.2 Originally the ATM
complied with both assumptions used in the VPM, however recently, the exploration into
capturing active drag whilst allowing the participant to swim with intra-stroke velocity
fluctuations has negated the need for the constant velocity assumption.14 To test the reliability
of this technique Sacilotto et al.15 conducted a pilot study comparing the constant velocity
tow active drag values of seven national level swimmers and concluded that these
participants were able to produce very consistent results (intra-class correlation coefficients
(ICC) = 0.91). Similarly, in 2013, Hazrati et al.16 demonstrated very high correlations
between nine participants using the ATM with fluctuating tow velocity trials across two nonconsecutive days of testing (ICC = 0.92 between days and 0.80 and 0.84 for days 1 and 2
respectively). As these studies produced reliable results with small sample sizes, a
determination of the reliability of the ATM with a larger sample was recommended.15, 16
According to Hopkins12 there are three important types of reliability to test: 1) within
subject variation; 2) change in mean; and 3) retest correlation. Within-subject variation refers
to the standard deviation or the technical error. It was suggested that this type of reliability is
the most important as the smaller the error, the more noticeable a change in performance
could be.12 The change in means is comprised of random change or systemic change (also
known as systemic bias) between two trials of a test. While retest correlation methods are
used to evaluate the stability or repeatability of a specific variable across repeated trials.17
45
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
James et al.18 reported that the number of trials obtained from a subject in an experiment
influences the stability of the final value, thus it is important to consider how many trials are
required to attain reliable values. As the ATM is based on the assumption that equal power is
produced by the swimmer between repeat efforts, a more adequate term for reliability in this
context would be performance reliability. Therefore, the main purpose of this study was to
investigate the performance reliability of the swimmers by the consistency of their mean
active drag values collected using the ATM. To achieve this, the present study aimed to: 1)
determine the number of trials required for consistent active drag values, and 2) determine the
level of performance reliability of the two assisted towing conditions (constant and
fluctuating).
Methods
Participants
Thirty-one state and national level swimmers (18 males and 13 females, 18.06 ± 2.66 years of
age, 650 ± 83 FINA point rank for 100 m front crawl performance time) participated in this
investigation. Eligibility criteria required participants to have a FINA point rank above 500
(range from State to World Championship level swimmers). All participants or guardians of
minors were required to sign a consent form approved by both the Australian Institute of
Sport and University of Canberra Ethics committees prior to participation.
Testing Protocols
Prior to testing, participants were required to complete their usual race warm-up focusing on
10 m front crawl sprints. All swim trials were performed in a 50 m, 10 lane, and 3 m deep
pool. Water temperature was maintained at 27 degrees Celsius throughout all testing. All
46
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
swimming experimental trials were completed using the front crawl technique with maximal
effort. Participants completed three 10 m free swim trials (to obtain a free swim velocity),
three passive tow trials (using a streamlined glide position to obtain values in the settings for
the assisted fluctuating trials) and assisted swim trials. Due to swimmer availability not all
swimmers completed both the five constant and five fluctuating tow velocity trials (nine
completed both constant and fluctuating trials, six completed constant only, and 11 completed
fluctuating only). Each participant was allowed a minimum of one familiarisation trial before
each phase of testing. At least one practice trial was required to ensure that the reliability of
the system was captured in the subsequent trials. Moreover, practice trials were essential for
determining whether the towed velocity was outside the expected 105 % (for constant trials)
or wasn’t between 105-110% (for fluctuating trials) more practice trials were permitted.
During the passive and all assisted tow trials, participants were towed by a flux vector
dynamometer mounted directly on a calibrated Kistler™ force platform (Kistler Instruments
Winterthur Switzerland Dimensions: 900 x 600 mm Type Z20916). Details of this system
have been descried previously by Mason et al.14 and the equipment set-up is shown in Figure
3.1. During all assisted towing trials the swimmers were towed via a Velcro belt attached
around their waist (Figure 3.2). The participants were also instructed not to breathe during the
10 m measurement interval in all trials.
47
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
Figure 3.1: Assisted towing method set up
Figure 3.2: Placement of Velcro belt used in assisted swimming tow trials
Free Swim Trials. To determine each swimmer’s maximal front crawl free swim velocity,
participants were required to complete three maximal effort free swim trials over a 10 m
interval. Participants started mid-pool at the 25 m mark, with their performance recorded as
48
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
their head passed between the 15 m and 5 m marks from the wall using two 50 Hz cameras
(Samsung model: SCC-C43101P).19 Free swim velocity was calculated by dividing the 10 m
distance by the time derived from the video footage between the 15 and 5 m marks.20 The
median velocity from the three trials was used as the participant’s maximum free swim
velocity. Each free swim trial commenced two minutes apart to allow the participant adequate
rest between efforts.
Constant Velocity Towing Trials. Following the free swim trials, participants completed five
maximal effort constant velocity towing trials. These trials were completed with an actual
mean tow velocity equal to five-per cent greater than the maximal free swim velocity.21
Passive Tow Trials. Participants completed two passive tow trials to obtain their mean
passive drag force.22 All were conducted with the same set up as the other towing trials
(Figure 3.1), with the swimmer holding the rope as shown in Figure 3.3 and maintaining a
static streamline position. The rope was also positioned to allow the swimmer to be towed
along the surface of the water. The mean passive drag force value, calculated over the same
10 m interval as the free swim trials, was used in the settings for the assisted fluctuating
velocity tows.23
49
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
Figure 3.3: Example of the grip required to hold towing rope in passive tow trials
Fluctuating Velocity Towing Trials. Participants completed five fluctuating velocity assisted
tow trials. To allow for intra-stroke velocity fluctuations, the tow velocity on the
dynamometer was set at 120 % of the participant’s maximal free swim velocity and a fraction
of the participant’s passive drag force was used as the force setting on the dynamometer
(ranged in the sample between 25 and 50 N). The settings on the dynamometer were altered
to force control with speed override (once the set force was reached the velocity fluctuated to
remain within the drag force range). This allowed the actual swimmer’s velocity during the
trial to fluctuate in accordance with the participant’s assumed normal intra-stroke velocity
fluctuations. The actual mean tow velocity for all accepted trials ranged between 105 to 110
% of the maximum swim velocity. Trials that resulted in a mean tow velocity of greater or
less than 105 to 110 % of the participant’s mean maximal free swim velocity were excluded
from the analysis. However, no extra trials were completed outside of the five assigned.
For all trials of active assisted tow swimming, a sample of four full maximal stroke
cycles was captured for analysis (right-hand entry until right-hand entry was considered a full
stroke cycle). Four full stroke cycles for these swimmers covered a distance of between 5 to
50
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
10 m, therefore allowing comparisons to be made between the free swim trials and assisted
tow trials. Participants were allowed a three-minute time cycle from the start of one trial to
the start of the next trial to avoid the effects of fatigue and to allow computer processing of
the data in all towed trials.
Data Processing
Force and video data was captured using Contemplas GmbH Motion Analysis (version
6.2.274) software and then processed using an export/import function in Contemplas linked
to an AIS customised analysis program.23 The Kistler platform amplifier sensitivity was set at
5000 pC and the data was captured using a 12 bit Analogue to Digital card, which sampled at
500 Hz. An 8 Hz Butterworth low pass digital filter was applied to the force data.2, 19 These
settings ensured a larger number of data points were captured to present more accurate
information, whilst still minimising the effect of noise by using the filter. Active drag (DA)
was computed using the same equation as outlined in Alcock and Mason2:
‫ܦ‬஺ =
‫ܨ‬௕ ‫ݒ‬ଶ ‫ݒ‬ଵଶ
‫ݒ‬ଶଷ − ‫ݒ‬ଵଷ
Equation 3.1
where Fb is the force needed to pull the athlete at the increased velocity as measured by the
force platform, v1 is the swimmer’s maximum free swim velocity and v2 is the mean tow
velocity taken from the dynamometer. To calculate mean active drag, equation 3.1 was used
on each force data point collected (every 0.002 of a second) and the mean for active drag was
calculated across all values in the sample interval (four full stroke cycles). Outputs vary in
time duration depending on how long it takes for swimmers to complete the four full stroke
cycles. Examples of processed data outputs can be seen in Figure 3.4 (fluctuating tow trial)
and Figure 3.5 (constant tow trial).
51
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
Performance reliability has been defined as the swimmers consistency whilst using
the ATM protocol. However, any variance found could be due to the error associated with the
towing device. Therefore, the measurement error was assessed through the reliability of the
passive drag trials from 38 swimmers at an average swim velocity of 1.77 m/s. A standard
error of 2.5 N was found over three trials and considered small in relation to the error
associated with the swimmers. Thus, the active drag values found from the participants
swimming during the active drag trials were considered to the performance reliability.
350
300
250
Force (N)
200
150
100
50
0
0
0.5
1
1.5
-50
2
2.5
3
Time (s)
Fluctuating Active Drag
Fluctuating Mean Active Drag
Figure 3.4: Example of a processed data output for a fluctuating tow trial
52
3.5
4
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
350
300
Force (N)
250
200
150
100
50
0
0
0.5
1
1.5
2
Time (s)
Constant Active Drag
2.5
3
3.5
4
Constant Mean Active Drag
Figure 3.5: Example of a processed data output for a constant tow trial
Statistical Analysis
Descriptive statistics were derived for active drag, swim velocity, tow velocity, FINA point
ranking, and 100 m front crawl performance time. To test for normality, the Shapiro-Wilk
assessment was used on the average of all five trials from the constant and from the
fluctuating tow velocity trials. Average active drag values across the five trials were found to
be normally distributed for both constant and fluctuating tow velocities, as assessed by
Shapiro-Wilk’s test (p > 0.05). Repeated measure analysis of variance (ANOVA) was
conducted using a Wilks’ Lambda multivariate test between all five trials within both
constant and fluctuating trials to determine if systemic bias was present within the sample. A
level of significance was set to p < 0.05. An intra-class correlation (ICC) two-way mixed
model absolute agreement single measures analysis was generated to determine the number
53
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
of trials required to produce reliable mean active drag values in both constant and fluctuating
tow velocity trials. Trial 1, the average of trials 1 and 2, the average of trials 1, 2 and 3, and
the average of trials 1, 2, 3 and 4, were assessed for their level of agreement against the
average of all five trials. An absolute agreement average measures two-way mixed ICC with
95 % confidence interval (CI) was used to evaluate the test-retest correlation of reliability in
each of the constant and fluctuating trial groups. Test retest reliability was conducted on trials
1 to 3 as an absolute agreement ICC with 95 % CI. Thresholds for all ICC statistics were
taken from Fulton et al.24 as < 0.40 represented poor reliability, 0.40 – 0.70 fair reliability,
0.71 – 0.90 good reliability, and > 0.90 represented excellent reliability or agreement. These
statistical analyses were performed using IBM Statistical Package for Social Sciences (SPSS)
Statistics for Windows, Version 19.
Within-subject variation was assessed between trials 1 to 3 at both constant and
fluctuating assisted tow velocities. To reduce the likelihood of heteroscedasticity from the
small sample size, the data were log-transformed before analysis.25 The typical error (TE),
representing a coefficient of variation (CV), was expressed as a percentage as outlined in
Hopkins12. From previous active drag and other swimming results, a very good CVTE % was
considered to be < 10 %, an acceptable range was between 10 – 20 % and a CVTE% of > 20
% was considered a poor typical error.17 Confidence intervals were set to 95 % as calculated
by Hopkins25 reliability windows excel spreadsheet.
Results
Descriptive statistics for the total sample are shown in Table 3.1. No significance was
observed between all five trials in both the constant tow velocity tows (p = 0.745) or
fluctuating velocity tows (p = 0.594). Good to excellent agreement was found across all trial
54
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
numbers and the average of all five trials (Table 3.2). Generally, a high level of stability was
observed within confidence limits when comparing a number of trials against the average of
all five trials in both towing conditions (Table 3.3).
Table 3.1: Total sample mean (± SD) data for swim velocity, tow velocity, active drag, FINA
point score and 100 m performance time.
n
Constant
Fluctuating
13
27
Swim
Velocity
(m/s)
1.76 ± 0.13
1.75 ± 0.13
Tow
Velocity
(m/s)
1.85 ± 0.14
1.88 ± 0.15
Active
Drag
(N)
179 ± 73
121 ± 35
FINA
point rank
653 ± 73
650 ± 87
100 m
Performance
(s)
56.21 ± 3.79
57.12 ± 4.27
Table 3.2: Intra-class correlations coefficients between different groups of trials and the
average of all five trials within both constant and fluctuating active drag tow
velocity trials.
Constant average of 5
Fluctuating average of 5
n
8
16
Trial 1
0.83
0.88
Trials 1 - 2
0.94
0.95
Trials 1 - 3
0.99
0.98
Trials 1 - 4
0.99
0.99
Table 3.3: Intra-class correlation confidence limits between different groups of trials and the
average of all five trials within both constant and fluctuating active drag tow
velocity trials.
Constant average of 5
Fluctuating average of 5
Trial 1
0.35 – 0.96
0.67 – 0.96
Trials 1 – 2
0.74 – 0.99
0.86 – 0.98
Trials 1 - 3
0.96 – 1.00
0.95 – 0.99
Trials 1 - 4
0.98 – 1.00
0.98 – 1.00
To investigate which protocol (constant or fluctuating) enables the collection of the
highest level of reliability only trials 1 to 3 were utilised in the following analysis to increase
the sample size and statistical power. Table 3.4 presents the mean CVTE % and Figure 3.6
illustrates ICC values for trials 1 to 3 in both the constant and fluctuating assisted towed
trials. Poor CVTE % was found within the constant tow velocity sample, whereas the CVTE %
established in the fluctuating sample were considered acceptable.
55
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
Table 3.4: Summary of within-subject variation shown in percentages presented as
coefficients with 95 % confidence intervals.
n
13
27
Constant
Fluctuating
300
Within-subject variation
CIL
24.0
9.8
CVTE %
35.0
12.6
500
a
CIU
64.1
17.6
b
400
200
300
200
100
100
0
0
0
100
200
300
0
100
200
300
400
Figure 3.6: Result summary of test retest reliability a. Fluctuating trials: ICC = 0.94 (CIL =
0.88, CIU = 0.97); b. Constant trials: ICC = 0.83 (CIL = 0.57, CIU = 0.94): ◦ =
trials 1 and 2; ▪ = trials 1 and 3.
Discussion
The main purpose of this study was to examine the performance reliability of the ATM by
assessing maximal effort front crawl swimming mean active drag values. No significant
difference was observed between all five trials in both the constant and fluctuating assisted
towing conditions. With no main effect for trial number established, the first aim of this
investigation was to determine the number of trials required to collect consistent active drag
values. The initial protocol in this study required participants to complete five assisted active
drag trials using both constant and fluctuating tow velocities. Given the nature of the testing
protocol used in the ATM (maximal swim efforts over at least 10 m), the need to reduce the
number of trials and total testing time without affecting performance reliability is important if
56
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
the ATM is to be used as a servicing tool for technique analysis. Good agreement was found
between trial 1 and the average of all five trials in both constant and fluctuating towing
conditions. Moreover, excellent agreement was found between the average of trials 1 and 2,
trials 1 to 3, and trials 1 to 4, and the average of all five trials. This finding implies that a
minimum of one trial could be used when reporting mean active drag values as has been used
in the past.5, 6 However, James et al.26 noted that by only conducting a single trial protocol,
the measure has the inability to represent an average performance as the result may be
atypical. Furthermore, Connaboy et al.17 concluded that a practical approach is required in
selecting the number of trials to represent reliable values and is a balance of ensuring high
test retest reliability and acceptable levels of within-subject variation. As stability is
necessary for both the reliability of the data and the ability to generalise to a greater
population of trials,26 to progress the current investigation to include a larger sample, trials 1
to 3 were used to complete further analysis for this study. Additionally it is recommended
that future protocols consider using two or more assisted tow trials to minimise the risk of
collecting atypical data.
The second aim of the present study was to determine the level of performance
reliability in both the constant tow velocity and the fluctuating tow velocity testing protocols.
Results were determined using trials 1 to 3 in order to obtain a larger sample size. Within the
constant tow velocity trials test retest ICC values demonstrated good reliability, however,
poor within-subject variation was found when assessing the CVTE %. The extent of the
magnitude in CVTE % confidence limits in the constant velocity trials could be due to the
testing protocol being set to a fixed towing speed which could be emphasising resistive forces
encountered by flaws in swimmer’s technique. Alternatively, the stroke mechanics (stroke
rates and stroke lengths) of the swimmers could be altered as a result of the constant towing
velocity. This may result in increased variation in swimmer performance which could in turn
57
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
have resulted in increased variation in active drag values. The dilemma of assuming constant
velocity and how it affects stroke mechanics has impacted active drag research for several
decades.6, 27 Furthermore, Girold et al.28 reported a significant increase in swimmers stroke
rates and a significant decrease in stroke length when testing assisted swimming protocols as
an over-speed training tool. In contrast, smaller within-subject variation and excellent test
retest reliability was established within the fluctuating tow velocity values. The moderately
small range of the CVTE % confidence limits demonstrates the increased stability of the
fluctuating velocity condition.
Conclusion
The findings of the present study indicate that the fluctuating tow velocity protocol allowed
swimmers to produce a higher level of performance reliability whilst using the ATM
protocol. The excellent agreement found when comparing trials 1 and 2, and trials 1 to 3, in
both constant and fluctuating towing conditions suggests that the testing protocol could be
reduced from five to two or three trials whilst maintaining reliable and stable mean active
drag values. As no main effect was observed for the number of trials required to obtain
reliable values in either constant or fluctuating tow velocity protocols to minimise the risk of
an atypical result a recommendation of two or three assisted trials was made. Such a
modification of the ATM protocol will reduce testing and analysis time which is likely to
assist in reducing potential fatigue effects and enable a larger number of swimmers to be
tested in each session.
58
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
References
1.
Toussaint HM. Biomechanics of propulsion and drag in front crawl swimming. In:
Proceedings of the 20th International Society of Biomechanics in Sport; Jul 1-6, 2002;
Caceres, Spain; p. 13-22.
2.
Alcock A, Mason B. Biomechanical analysis of active drag in swimming. In:
Proceedings of the 25th International Society of Biomechanics in Sports. Brazil; 2007.
p. 212-5.
3.
DiPrampero PE, Pendergast DR, Wilson DW, Rennie DW. Energetics of swimming in
man. Journal of Applied Physiology. 1974;37(1):1.
4.
Hollander AP, De Groot G, Van Ingen Schenau GJ, et al. Measurement of active drag
during front crawl arm stroke swimming. Journal of Sports Sciences. 1986;4:21-30.
5.
Kolmogorov SV, Duplishcheva OA. Active drag, useful mechanical power output and
hydrodynamic force coefficient in different swimming strokes at maximal velocity.
Journal of Biomechanics. 1992;25(3):311-8.
6.
Wang X, Wang L, Yan W, Li D, Shen X. A new device for estimating active drag in
swimming at maximal velocity. Journal of Sports Sciences. 2007;25(4):375-379.
7.
Kolmogorov SV, Rumyantseva OA, Gordon BJ, Cappaert JM. Hydrodynamic
characteristics of competitive swimmers of different genders and performance levels.
Journal of Applied Biomechanics. 1997;13:88-97.
59
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
8.
Benjanuvatra N, Blanksby BA, Elliott BC. Morphology and hydrodynamic resistance in
young swimmers. Paediatric Exercise Science. 2001;13:246-55.
9.
Von Loebbecke A, Mittal R, Mark R, Hahn J. A computational method for analysis of
underwater dolphin kick hydrodynamics in human swimming. Sports Biomechanics.
2009;8(1):60-77.
10.
Zamparo P, Gatta G, Pendergast DR, Capelli C. Active and passive drag: The roles of
trunk incline. European Journal of Applied Physiology. 2009;106:195-205.
11.
Toussaint HM, Roos PE, Kolmogorov S. The determination of drag in front crawl
swimming. Journal of Biomechanics. 2004;37(11):1655-63.
12.
Hopkins W. Measures of reliability in sports medicine and science. Sports medicine.
2000;30(1):1-15.
13.
Schleihauf R, Gray L, DeRose J. Three dimensional analysis of hand propulsion in
sprint front crawlstroke. Asian Journal of Physical Education. 1983;6(1):57-70.
14.
Mason B, Sacilotto G, Menzies T. Estimation of active drag using an assisted tow of
higher than max swim velocity that allows fluctuating velocity and varying tow force.
In: Proceedings of the 29th International Society of Biomechanics in Sports; Jun 27-Jun
1, 2011; Porto, Portugal; p. 327-30.
15.
Sacilotto G, Mason B, Ball N. Intra-reliability of active drag values using the assisted
towing method (ATM) approach. In: Proceedings of the 30th International Society of
Biomechanics in Sport; Jul 1-6, 2012; Melbourne, Australia.
60
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
16.
Hazrati P, Mason B, Sinclair PJ. Reliability of estimating active drag using the assisted
towing method (ATM) with fluctuating velocity. In: Proceedings of the 31st
International Society of Biomechanics in Sport; Jul 7-11, 2013; Taipei, Taiwan.
17.
Connaboy C, Coleman S, Moir G, Sanders R. Measures of Reliability in the Kinematics
of Maximal Undulatory Underwater Swimming. Medicine and science in sports and
exercise. 2010;42(4):762-70.
18.
James CR, Herman JA, Dufek JS, Bates BT. Number of trials necessary to achieve
performance stability of selected ground reaction force variables during landing.
Journal of Sports Science and Medicine. 2007;6(1):126-34.
19.
Formosa D, Mason B, Burkett BJ. Measuring active drag within the different phases of
front crawl swimming. In: Proceedings of the 11th International Symposium of
Biomechanics and Medicine in Swimming; Jun 16-19, 2010; Olso, Norway.
20.
Tor E. Quantifying the underwater trajectory of a swimming start. In: Proceedings of
the 31st International Society of Biomechanics in Sport; Jul 7-11, 2013; Taipei, Taiwan.
21.
Formosa D, Mason B, Burkett BJ. The force-time profile of elite front crawl swimmers.
Journal of Sports Sciences. 2011;29(8):811-9.
22.
DeVellis RF. Scale Development - Theory and Applicants. 2nd ed. United States of
America: Sage Pulications, Inc; 2003.
23.
Mason B, Sacilotto G, Dingley A. Computation of a swimmer's propulsive force profile
from active drag parammeters with fluctuating velocity in assisted towing.
61
In:
Chapter 3: Reliability of Active Drag Values Using the Assisted Towing Method
Proceedings of the 30th International Society Biomechanics in Sport; Jul 2-6, 2012;
Melbourne, Australia.
24.
Fulton SK, Pyne DB, Burkett B. Validity and reliability of kick count and rate in
freestyle
using
inertial
sensor
technology.
Journal
of
sports
sciences.
2009;27(10):1051-8.
25.
Hopkins W. A New View of Statistics: Reliability Spreadsheet. A New View of
Statistics
2002
September
2010
[cited;
Available
from:
http://www.sportsci.org/resource/stats/index.html
26.
James CR, Herman JA, Dufek JS, Bates BT. Number of trials necessary to achieve
performance stability of selected ground reaction force variables during landing.
Journal of sports science & medicine. 2007;6(1):126.
27.
Sacilotto G, Ball N, Mason BR. A Biomechanical Review of the Techniques Used to
Estimate or Measure Resistive Forces in Swimming. Journal of applied biomechanics.
2014;30:119-127
28.
Girold S, Calmels P, Maurin D, Milhau N, Chatard JC. Assisted and resisted sprint
training in swimming. Journal of Strength & Conditioning Research. 2006;20(3):54754.
62
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
Chapter 4
Study 2: A Comparison of Front Crawl Stroke Mechanics
between Free Swim and Assisted Towed Swimming
Introduction
Basic stroke mechanics, such as stroke time, the optimal ratio between stroke lengths and
stroke rates, and mean swim velocities, are currently used to assess swim performances.1-5
Other free swimming kinetic parameters, for example active drag, propulsion and
instantaneous swim velocity profiles, are also used, however there are limitations in the
protocols used to collect kinetic values.6-8 Given that swim velocity (v) is defined as equal to
the product of the stroke rate (SR) and stroke length (SL),5 this relationship is key to
understanding the fundamentals in free swimming. The relationship is outlined as:2
‫ܴܵ ∙ ܮܵ = ݒ‬
Equation 4.2
where SL is the distance the body travels per stroke cycle and SR is the stroke frequency in
strokes per second. Previous research has suggested that in maximal swimming conditions
the relationship between swim velocity, stroke lengths, and stroke rates reflects the ability of
a swimmer to swim with a higher efficiency, which emphasises the effective use of stroke
technique and propulsive force.9 However, several authors have found that there is no optimal
ratio of stroke length and stroke rate for all swimmers,2, 10-12 and is therefore, idiosyncratic in
nature. Individual stroke length and stroke rate ratios during competitive swim performance
(e.g. 50 m, 100 m, and 200 m freestyle events) have been studied and reported extensively .2-
63
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
4, 13-15
Other studies examining different parameters of swim performance report findings
using stroke length and stroke rate signifying the importance of this relationship to free
swimming.16-18 Given the importance of the relationship between stroke mechanics and free
swimming, researchers have started to consider these when investigating protocols in assisted
and resisted swimming.1, 13, 19
Previous research which studied the degree of change in stroke mechanics whilst
using assisted swimming have indicated that stroke lengths are maintained between free and
assisted towed swimming, whereas the stroke rates have increased.19 The significant increase
in stroke rates as seen in William et al.19 was suggested to be a result of a modified stroke
pattern as a consequence of the over speed swimming. Similar findings were observed in
Maglischo et al.20, however, these over speed, or assisted, trials were completed with a
constant velocity and not a tow velocity allowing for intra-stroke velocity fluctuations.
Theoretically, an assisted tow allowing for intra-stroke velocity fluctuations should reduce
the changes in stroke mechanics. The Assisted Towing Method (ATM) estimates active drag
by comparing the velocity difference between a free swim condition and assisted towed swim
condition whilst taking into account a known assisted force.21 Recent work utilising the ATM
has shown promising results in estimating active drag with a tow velocity which allows near
normal intra-stroke velocity fluctuations.8,
22
Additionally, the ATM is based on an
assumption which suggests that power developed in the free and towed swims are equal. By
comparing the stroke mechanics between ATM and free swimming, the results could provide
validation of the ATM protocol in the assessment of front crawl technique. Therefore, an
investigation is needed to identify to what extent the stroke mechanics are altered during
fluctuating velocity assisted towing and whether technique assumptions can be made from the
data captured during the towed trials and transferred to free swimming.
64
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
Hence, the purpose of this study was to compare the stroke mechanics between a free
swimming condition and an assisted towed swimming condition with fluctuating velocity
using the ATM. To achieve this, the present study aimed to: 1) determine the extent of
difference between free swim and assisted towed swim stroke mechanics (stroke lengths and
stroke rates); and 2) examine the relationship between free swim stroke length and stroke rate
ratios, and assisted towed swim stroke length and stroke rate ratios.
Method
Twenty-six state and national level swimmers (14 males and 12 females, 18.74 ± 2.49 years
of age, 642.56 ± 95.35 FINA points for 100 m Freestyle, 56.57 ± 4.32 s 100 m performance
times) participated in this investigation. All participants were required to sign a consent form
approved by both the Australian Institute of Sport and the University of Canberra Ethics
committees. In the case of minors participating, parent or guardian consent was acquired.
Participants were required to complete their personal race warm-up and were asked to
focus on 10 m front crawl sprints. The swimmers completed the same protocol as outlined in
Sacilotto et al.23, which involved three free swim trials, three passive and five assisted tow
trials. However, during the assisted trials the tow velocity allowed intra-stroke velocity
fluctuations to occur. Swimmers were instructed to swim with maximal effort (not breathing)
throughout all testing protocols. Free swim mean velocities were calculated across a 10 m
interval, by dividing the 10 m distance by the time that it takes for the head to pass from the
15 m mark through to the 5 m mark. The video tracking system is the same as outlined in Tor
et al.24. Passive drag force was calculated by passively towing the swimmer in a streamline
position at their maximal free swim velocity across a 10 m interval. A fraction of the passive
drag force was utilised in the settings for the assisted tow trials and this ranged between 20
65
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
and 50 N within the sample. The passive drag force is influenced by the morphology and free
swim velocity of each participant, hence the range of values between swimmers. The
swimmers were then asked to swim with maximal effort whilst being towed with assistance
for all five assisted tow trials. For the assisted tow trials, swimmers were asked to start at the
35 m mark and build to maximum speed and hold pace between the 25 m and 15 m mark
(roughly four full strokes cycles). Swimmers were towed from the waist during the assisted
tow trials.
The median free swim and assisted towed trial were used in the final analysis of
comparing stroke length and stroke rate. Stroke rates were calculated for each free swim trial
using a stop watch (Seiko ST MORITZ) as these trials were not captured in a format which
could be digitised later. Stroke rates were collected over three full stroke cycles following the
first right hand entry after the 15 m mark. All assisted trials were video recorded with
cameras positioned in the head-on and side-on (Figure 4.1) (head-on camera: JVC DV
Camcorder GY-DV550, side-on cameras: Applied Microvideo variable zoom lens JRNL
26056 (above) and a SwimPro Basic 8.0 mm (below)). The side-on above and underwater
camera images were mixed with an Edirol video mixer (EDI-V8) to produce a single moving
above/below image. Video and velocity data were collected using Contemplas GmbH
Templo Motion Analysis (version 6.2.274) software in all passive and assisted tow trials
only.
66
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
Figure 4.1: Example of a mixed image used to calculate assisted towed swim stroke rates
Stroke rates for the assisted tow trials were calculated post testing using the side-on
footage. The video footage was played back in real-time and the same stopwatch used in the
free swim trials was used to measure stroke rates across a sample of three stroke cycles from
the first right hand entry after the 20 m mark. To ensure accuracy in obtaining stroke rates, a
single tester measured both free swim and towed swim stroke rates. The reliability of the
tester was measured by calculating a total of 40 stroke rates on a single trial of a swimmer,
which was not included in this study. Two stroke rates were measured from video footage
played in real-time every 10-minutes. The mean of all 40 stroke rates was 47.88 ± 0.67 rates
per minute (rpm) and high measurement reliability was established by a technical error of
measurement of less than 1 % (0.38 %). Stroke lengths were calculated by dividing the mean
(free or towed) swim velocity by the stroke frequency (SR/60). Ratios for free swim and
assisted towed swimming stroke mechanics were calculated by dividing the stroke length by
the stroke rate as a novel way to compare free swim and assisted tow trial stroke mechanics.
Statistical analyses were performed using IBM SPSS (Statistical Package for Social
Sciences) Statistics for Windows, Version 19.0. Descriptive statistical information (mean and
standard deviations) were derived to characterise the participant sample. Paired t tests were
67
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
used to compare free swim and assisted tow trials stroke mechanics. The level of significance
was fixed at p < 0.05. Effect size was established between free swim and assisted tow trials
for both stroke rates and stroke lengths. Effect size (ES) was calculated by dividing the
difference between the mean towed variables and the free swim variable by the standard
deviation of the free swim variable.25 The magnitude of the effect sizes were outlined by
Rhea26 as: trivial < 0.25, small = 0.25 – 0.50, moderate = 0.50 – 1.0, and large > 1.0, for a
highly trained population (individuals who have trained for at least five-years).
Pearson’s product moment correlation coefficients with a level of significance set at p
< 0.05 were used to determine the relationship between stroke rate and stroke length and free
and assisted tow velocities. Magnitudes of all correlations were interpreted using the
following thresholds: low r = 0.10 – 0.30; moderate r = 0.30 – 0.50; and high r = > 0.50.25, 27
The relationship between free swim and assisted towed swim stroke mechanics was assessed
using linear regression analysis between the ratios of stroke lengths and rates.
Homoscedasticity was found between the two variables and the data appeared to be normally
distributed upon visual inspection of histograms and a normal P-P plot. The magnitude of
effect size was outlined through a scatterplot (R2) and an analysis of variance provided
statistical significance at a level of p <0.05.
Results
Significant increases were found between free swim and assisted tow trials in all variables
(Table 4.1). A large magnitude of difference was found between free swim and assisted tow
trial velocities, a moderate effect was found between free swim stroke length and assisted tow
trial stroke length, and a small effect was found between the two conditions in stroke rate.
68
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
Table 4.1: Summary of differences between the mean free swim and assisted tow trial
variables
Swim Velocity (m/s)
Free Swim
1.76 ± 0.13
Assisted Swim
1.89 ± 0.15
% diff
107
p
0.000**
ES
1.00
Stroke Length (m)
1.53 ± 0.20
1.70 ± 0.23
111
0.000**
0.85
Stroke Rate (rpm)
52.16 ± 3.75
53.77 ± 4.15
103
0.006**
0.43
*level of significance at the 0.01 level; m/s = metres per second; m = metres; rpm = rates per minute;
% diff = percentage difference between free swim and assisted tow trials; ES = effect size)
Significant correlations were revealed between free swim velocity and both stroke
length and stroke rate. Similarly, significant correlations were found between assisted tow
velocity trials and in both stroke length and stroke rate (Table 4.2). A linear regression
established that assisted tow ratios of stroke length and stroke rate could significantly predict
free swim ratios of stroke length and stroke rate, F(1, 24) = 862.022, p < 0.0005 and a
moderate effect size was found, ES = 0.97. Figure 4.2 presents significant linear relationship
between free swim and assisted tow ratios.
Table 4.2: Significance between stroke mechanics and swim velocities in both free swim and
assisted towed swimming
Free Swim Velocity
Assisted Tow Velocity
Stroke Length
0.90**
0.85**
Stroke Rate
0.58**
0.46*
* Correlation is significant at the 0.05 level; ** Correlation is significant at the 0.01 level
69
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
0.035
Free Swim Ratio
0.033
0.031
y = 0.878x + 0.002
R² = 0.973
0.029
0.027
0.025
0.025
0.027
0.029
0.031
0.033
0.035
Assisted Tow Ratio
0.037
0.039
Figure 4.2: Ratio (stroke length/stroke rate) comparison with line of best fit between assisted
towed swim and free swims stroke mechanics (r = 0.95).
Discussion
The purpose of this investigation was to explore the relationship between stroke rates and
stroke lengths at free swim and assisted towed swim velocities. The two central findings to
emerge from this study were: 1) an increase occurred between free swim and assisted tow
trials in stroke length, stroke rate and swim velocity; and 2) assisted tow ratios of stroke
mechanics could significantly predict the ratios of stroke mechanics for free swimming.
The first aim was to determine the extent of the difference between free swim and
assisted tow trials’ stroke mechanics. Results revealed that the stroke mechanics were
significantly higher in the assisted tow trials than in free swimming which is consistent with
other findings.1, 19 William et al.1 sampled a group of ten junior elite females and compared
stroke mechanics between free and assisted towed swimming. Results indicated that there
was a significant increase in stroke rates (p = 0.002) and stroke lengths (p = 0.000) at an
assisted tow velocity of 1.72 ± 0.04 when comparing to free swim stroke mechanics at a free
70
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
swim velocity of 1.48 ± 0.11.1 William et al.1 concluded the increase between stroke
mechanics as the hand slipping through the water, rather than the hand shortening or
lengthening the stroke cycle. Girold et al.28 also found a significant increase in the stroke rate
during assisted swimming during free swimming and conversely they also reported a
significant decrease in stroke length. This decrease in stroke length was considered to be
proportionally less than the increase in stroke rate.28 Within the current study the increase in
stroke mechanics was thought to be as a result of allowing the swimmers to swim intra-stroke
velocity variations rather than a constant velocity. However, given the larger percentage of
change between the stroke length for the free swim and for the assisted tow swim it is
suggested that the swimmers may have let the assist tow pull them along, therefore
lengthening their distance per stroke. This finding could indicate a limitation in the ability to
use ATM kinetic outputs as a representation of free swimming or a limitation in the
swimmer’s ability to comply with testing protocols. Despite these limitations, the stroke
length was considered to only moderately increase according to its effect size. Similarly,
despite the significant increase in stroke rate values between the free swim and the assisted
tow swims, only a small effect size was found. This small effect size in conjunction with the
low change in percentage difference in stroke rate values could imply that the swimmers are
indeed complying with protocols and have attempted to produce the same effort in both the
free swim and assisted tow trials. As stroke rate and stroke length were found to be
significant larger during ATM towed swimming, compared to free swimming, the
relationship between the stroke mechanics and free swim or assisted tow velocity velocities
was undertaken. A very high correlation was identified between stroke length and velocity in
both the free swim (r = 0.90) and the assisted towed swimming (r = 0.85). This is consistent
with the results of previous research findings that concluded that stroke lengths are the most
significant contributor to higher velocities in competitive swimming.2, 11 Additionally, a high
71
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
correlation was found between free swim stroke rates and free swim velocity (r = 0.58) and a
moderate correlation was found between assisted towed stroke rates and velocity (r = 0.46).
Although previous conclusions have been made regarding the importance of stroke
mechanic variations between free swimming and assisted swimming, none found have
reported a numerical ratio. The inclusion of this ratio in the current study will enable a novel
assessment of the effectiveness of an assisted protocol in representing a free swim trial in
order to validate the output from the ATM protocol. Albeit a novel approach in assessing the
stroke mechanic variability between free swimming and assisted tow swimming through a
linear regression between stroke mechanic ratios, it was revealed that assisted tow trials had
the ability to predict free swim stroke mechanics. In addition, a significantly high correlation
was found between the free swim and the assisted towed swim ratios (r = 0.95). These
findings suggest that a consistent increase occurred between the two conditions which could
allow assumptions to be made from the kinetic ATM output and transferred into the
assessment of free swimming.
Conclusion
Although significant differences in stroke mechanics were found between free swim velocity
and assisted tow swim velocity, the ratios of stroke length and stroke rate revealed a
consistent change between two velocity conditions. Despite a difference of stroke mechanics,
consistency in performance between free swimming and assisted swimming was found.
Therefore, the ATM protocol has the potential to transfer kinetic outputs of technique into the
assessment of free swim kinetics.
72
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
References
1.
Williams BK, Sinclair P, Galloway M. Changes in stroke mechanics during resisted and
assisted freestyle swimming.
In: Proceedings of the 25th International Society of
Bioemchanics in Sport; Aug 23-27, 2007; Ouro Preto, Brazil.
2.
Pelayo P, Sidney M, Kherif T, Chollet D, Tourny C. Stroking Characteristics in
Freestyle Swimming and Relationships With Anthropometric Characteristics. Journal
of Applied Biomechanics. 1996;12(2):197-206.
3.
Craig A, Pendergast D. Relationships of stroke rate, distance per stroke, and velocity in
competitive swimming. Medicine and Science in Sports and Exercise. 1979;11(3):27883.
4.
Seifert L, Chollet D, Chatard JC. Kinematic changes during a 100-m front crawl:
Effects of performance level and gender. Medicine and Science in Sports and Exercise.
2007:1784-93.
5.
Greco GC, Pelargio JG, Figueira TR, Denadai BS. Effects of gender on stroke rates,
critical speed and velocity of a 30-min swim in young swimmers. Journal of Sports
Science & Medicine. 2007;6:441-7.
6.
Wang X, Wang L, Yan W, Li D, Shen X. A new device for estimating active drag in
swimming at maximal velocity. Journal of Sports Sciences. 2007;25(4):375-379.
7.
Toussaint HM, Roos PE, Kolmogorov S. The determination of drag in front crawl
swimming. Journal of Biomechanics. 2004;37(11):1655-63.
73
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
8.
Mason B, Sacilotto G, Dingley A. Computation of a swimmer's propulsive force profile
from active drag parammeters with fluctuating velocity in assisted towing.
In:
Proceedings of the 30th International Soceity of Biomechanics in Sport; Jul 2-6, 2012;
Melbourne, Australia.
9.
Keskinen KL, Tilli LJ, Komi PV. Maximum velocity swimming: Interrelationships of
stroking characteristics, force production and anthropometric variables. Scandinavian
Journal of Medicine & Science in Sports. 1989;11(2):87-92.
10.
Pai YC, Hay J, Wilson B, Thayer AL. Stroking techniques of elite swimmers. Medicine
& Science in Sports & Exercise. 1984;16(2):159.
11.
Kennedy P, Brown P, Chengalur SN, Nelson RC. Analysis of Male and Female
Olympic Swimmers in the 100-Meter Events. International Journal of Sport
Biomechanics. 1990;6(2):187-97.
12.
Craig AB, Pendergast DR. Relationships of stroke rate, distance per stroke, and
velocity in competitive swimming. Medicine and Science in Sports and Exercise.
1979;11(3):278-83.
13.
Toussaint HM, Carol A, Kranenborg H, Truijens MJ. Effect of fatigue on stroking
characteristics in an arms-only 100-m front-crawl race. Medicine and Science in Sports
and Exercise. 2006;38(9):1635.
14.
Chollet D, Chalies S, Chatard JC. A new index of coordination for the crawl:
description and usefulness. International Journal of Sports Medicine. 2000;21(1):54-9.
74
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
15.
Craig A, Skehan PL, Pawelczyk JA, Boomer WL. Velocity, stroke rate, and distance
per stroke during elite swimming competition. Medicine Science in Sports and
Exercise. 1985;17(6):625-34.
16.
Đurović M, Beretić I, Dopsaj M, Pešić M, Okičić T. A comparison of kinematic
variables between European elite, national elite and regional elite male 100 m freestyle
swimmers. Facta Universitatis: Series Physical Education & Sport. 2012;10(4):339-46.
17.
Barbosa T, Silva AJ, Reis AM, et al. Kinematical changes in swimming front Crawl
and Breaststroke with the AquaTrainer® snorkel. European Journal of Applied
Physiology. 2010;109(6):1155-62.
18.
Kleshnev V. Method of analysis of speed, stroke rate and stroke distance in aquatic
locomotions. In: Proceedings of the 24nd International Society of Biomechanics in
Sports; Salzburg; Jul 14-18, 2006; Salzburg, Austria.
19.
Williams BK, Sinclair P, Galloway M. The effect of resisted and assisted freestyle
swimming on stroke mechanics. In: Proceedings for the 19th International Society of
Biomechanics in Sport; 2001; San Francisco, United States
20.
Maglischo EW, Maglischo CW, Zier DJ, Santos TR. The effect of sprint-assisted and
sprint-resisted swimming on stroke mechanics. Journal of Swimming Research.
1985;1(2):27-33.
21.
Alcock A, Mason B. Biomechanical analysis of active drag in swimming.
In:
Proceedings for the 25th International Society of Biomechanics in Sports; Aug 23-27,
2007, Ouro-Preto, Brazil; p. 212-5.
75
Chapter 4: A Comparison of Front Crawl Stroke Mechanics between Free Swim and Assisted Towed
Swimming
22.
Mason B, Sacilotto G, Menzies T. Estimation of active drag using an assisted tow of
higher than max swim velocity that allows fluctuating velocity and varying tow force.
In: Proceedings of the 29th International Society of Biomechanics in Sports; Jun27-Jul
1, 2011; Porto, Portugal; p. 327-30.
23.
Sacilotto G, Mason B, Ball N. Intra-reliability of active drag values using the assisted
towing method (ATM) approach. In: Proceedings of the 30th International Society of
Biomechanics in Sport; Jul 2-6, 2012; Melbourne, Australia.
24.
Tor E. Quantifying the underwater trajectory of a swimming start. In: Proceedings for
the 31st International Society of Biomechanics in Sport; Jul 7-11, 2013; Taipei, Taiwan.
25.
Cohen J. Statistical power analysis for the behavioral sciences: Psychology Press; 1988.
26.
Rhea MR. Determining the magnitude of treatment effects in strength training research
through the use of effect size. Journal of Strength & Conditioning Research (Allen
Press Publishing Services Inc). 2004;18(4):918-20.
27.
Anderson ME, Hopkins WG, Roberts A, Pyne DB. Ability of test measures to predict
competitive
performance
in
elite
swimmers.
Journal
of
Sports
Sciences.
2008;26(2):123-30.
28.
Girold S, Calmels P, Maurin D, Milhau N, Chatard J. Evaluation of an assisted sprint
training period in swimming. Isokinetics and Exercise Science. 2003;11(1):72-.
76
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
Chapter 5
Study 3: Investigation of Front Crawl Stroke Phases
within Active Drag Force-Time Profiles in Elite and SubElite Sprint Swimmers
Introduction
Free swimming is a cyclical movement where propulsion is generated to overcome resistive
forces or active drag.1 Swimming kinematic performance measures such as stroke rates,
lengths, and counts can easily be measured by coaches from the side of the pool. However,
kinetic measures such as active drag, propulsion and acceleration remain difficult to measure
and validate.2 Researchers have developed different techniques for measuring or estimating
these components. The most frequently used techniques include the Measuring Active Drag
System (MAD-System)3 and the Velocity Perturbation Method (VPM)4. The MAD-System
measures the propulsive force generated by a swimmer via a series of pads underneath the
water which the swimmer pushes against as they swim past. Alternatively, the VPM method
compares the velocity difference between a free swim and a resisted swim at maximal effort
with a known resistance attached to the swimmer. More recently a method has been
developed which uses the same assumptions as the VPM, however, active drag is calculated
by comparing the free swim velocity against an assisted towed swim.5 This Assisted Towing
Method (ATM) has been investigated using protocols which allow for a swimmer’s intrastroke velocity variations.6, 7 In addition, the ATM protocol allows collection of instantaneous
force-time profiles, which as Wang et al.8 acknowledged, was a recommendation when
77
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
investigating free swimming as a mean value does not provide information within a stroke
cycle.
To date, the literature on front crawl active drag force-time profiles is limited as most
methods present mean active drag values.3-5,
8
Although a mean value provides holistic
information on a full stroke cycle, force-time profiling will enable a greater understanding of
active drag at instantaneous points within the stroke cycle thus providing more information to
coaches which they can interpret for possible technique modification. Formosa, et al.9 were
the first to investigate active drag force-time profiles utilising the ATM protocol. These
force-time profiles were captured using a constant tow velocity, whereas recent work on the
ATM has utilised a fluctuating tow velocity.6,
7
However, Formosa et al.9 reported on
asymmetries found within their sample of elite males rather than examining the curves and
outlining whether there was potential for an optimal curve to be established. By investigating
active drag force-time profiles utilising an assisted tow which allows for intra-stroke velocity
variations, the need to establish normative data is warranted.
To characterise force-time profiles, four phases have been outlined to deconstruct the
front crawl stroke:1, 10
1.
Entry and Catch, the time from the hand’s entry into the water until the
beginning of its backwards movement;
2.
Pull, the time from the beginning of the hand’s backward movement to the
hand’s arrival directly underneath the shoulder;
3.
Push, the time from the hand’s position under the shoulder until its exit from the
water; and
4.
Exit and Recovery, the time from the hand’s exit from the water until its
following entry into the water.
78
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
Chollet et al.10 also identified that the propulsive phases of the front crawl stroke were the
pull and push phases, whilst the non-propulsive phases were the entry and catch, and the exit
and recovery phases.
As the recent ATM protocol collects instantaneous active drag, research into
identifying where these four phases in front crawl technique occur in the profile is warranted.
Therefore, the overall purpose this study was to investigate the stroke phases within ATM
swimming. This was achieved by 1) outlining the percentage of time spent in each stroke
phase; 2) examining the force magnitude of male and female active drag profiles; 3)
determining if there is a difference between elite and sub-elite active drag force-time profiles;
and 4) establishing whether the profiles have the capacity to indicate the individual stroke
phases.
Method
Fourteen elite (19.79 ± 2.69 years; 52.18 ± 3.13 s 100 m front crawl performance best time)
and seventeen sub-elite (18.24 ± 2.75 years; 58.71 ± 3.62 s 100 m front crawl performance
best time) swimmers participated in this investigation. Elite status was characterised by a
Federation Internationale de Natation Amateur (FINA) point scores of over 700 and sub-elite
by a score under 700. All participants were required to a sign consent form approved by both
the Australian Institute of Sport (AIS) and the University of Canberra Ethics committees. In
the case of minors, a parent or guardian were also required to sign consent forms.
Participants were instructed to complete their own race warm but to focus on short 10
m front crawl sprints prior to testing. Experimental testing was conducted in accordance with
the protocol outlined in Mason et al.6. Briefly, the protocol included three maximal free swim
79
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
trials (maximal effort was collected over a 10 m distance), three passive tows (a mean passive
drag force was collected over a 10 m distance) and three maximal effort assisted tows, with
adjustment for intra-stroke velocity variations, was completed by each participant. All
swimming trials were completed using maximal effort front crawl swimming and participants
were instructed not to breath within the trial capture distance. Data was captured across a
period of four full stroke cycles during the assisted tow trials.7 Participants were given up to
three minutes rest between trials to minimise fatigue affects and to allow for adequate
computing time.
All force data was recorded using Contemplas GmbH Templo Motion Analysis
(version 6.2.274) software and processed using an export/import function in Templo linked to
an Australian Institute of Sport customised analysis programme.6 A Kistler™ force platform
(Kistler Instruments Winterthur Switzerland Dimensions: 900 x 600 mm Type Z20916) was
used for all data capture. The amplifier’s sensitivity was set at 5000 pC and the data was
collected using a 12 bit A to D card, sampled at 500 Hz. An 8 Hz Butterworth low pass
digital filter was applied to the raw force.5 Instantaneous and mean active drag values were
computed using the method outlined in Mason et al.7. The assisted tow trial which produced
the median mean active drag score was used for the stroke phase analysis.11 All trials were
video recorded using three genlocked cameras. The cameras were positioned head-on and
side-on of the swimmer (head-on camera: JVC DV Camcorder GY-DV550, side-on cameras:
Applied Microvideo variable zoom lens JRNL 26056 (above) and a SwimPro Basic 8.0 mm
(below)). The side-on image was produced from two images which were mixed with an
Edirol video mixer (EDI-V8) to produce a single moving image.12 Although the video data
collected force data over four full stroke cycles, only the second stroke was used for analysis
and was used in preliminary work presented in Sacilotto et al.11. A digital time-code was
applied to both camera inputs to allow the visual deconstruction of a stroke cycle into the four
80
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
phases. The four stroke phases were manually determined through synced video and force
data on both right and left stroke cycles, using the guidelines from Seifert et al.1, as 1) Entry
and Catch; 2) Pull; 3) Push; and 4) Exit and Recovery. The percentage of time spent in each
stroke phase was also determined. The analysis of instantaneous force-time data was
conducted for a single right and left arm stroke cycle. The four variables examined were: 1)
maximum active drag; 2) time of maximum active drag as a percentage of total stroke time;
3) minimum active drag; and 4) time of minimum active drag as a percentage of total stroke
time.
Statistical analysis was completed using IBM Statistical Package for the Social
Sciences (SPSS) Statistics for Windows, Version 19.0. Homogeneity of variance was
assessed by Levene’s test for equality of variance in elite and sub-elite swimmers. A violation
was found between the elite and sub-elite right stroke cycle maximum active drag values. To
correct this and be assured of equal variance, elite and sub-elite variables were split into male
and female categories. Further Levene’s tests for equality revealed no significant violations
between the new sub groups (male elite and sub-elite, and female elite and sub-elite). Mean
and standard deviations (SD) were generated and are presented for all measured variables.
Independent t tests were performed on all measured variables to determine if significant
differences exist between male elite and sub-elite and female elite and sub-elite swimmers.
Paired t tests were conducted to determine if a) any significant differences were present in
percentage of time spent in each stroke phase in all sub-groups; and b) if significant
differences existed between right and left stroke cycles in all variables. The level of
significance was fixed at p < 0.05 for all statistics.
81
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
Results
Significant differences in performance variables between elite and sub-elite swimmers were
present for male and female, elite and sub-elite groups (Table 5.1). No significant differences
between mean percentages of time spent in each stroke phase were found (Table 5.2).
Table 5.1: Summary of performance variables shown as mean ± SD
Male
Elite
Sub-Elite
Female
Elite
Sub-Elite
n
Swim Velocity (m/s)
Tow Velocity (m/s)
FINA point score
Active Drag (N)
11
7
1.89 ± 0.05
1.83 ± 0.08
2.03 ± 0.07
1.98 ± 0.08
787 ± 77*
607 ± 65
138 ± 24
136 ± 24
3
10
1.68 ± 0.03*
1.61 ± 0.04
1.83 ± 0.05*
1.73 ± 0.05
756 ± 74*
616 ± 51
91 ± 14
93 ± 18
*Significant (p < 0.05) difference between elite and sub-elite groups
Table 5.2: Percentage of time spent in each stroke phase during assisted towed swimming
shown as mean ± SD percentages (%)
Entry and Catch
Male
Elite
Sub-Elite
Female
Elite
Sub-Elite
Male
Elite
Sub-Elite
Female
Elite
Sub-Elite
Pull
Push
Right arm cycle %
22.5 ± 3.9
21.0 ± 2.4
Exit and Recovery
28.1 ± 4.2
25.3 ± 2.0
18.4 ± 0.9
20.3 ± 2.9
31.0 ± 4.7
33.4 ± 4.3
35.3 ± 3.1
35.1 ± 7.2
15.7 ± 4.2
21.7 ± 8.0
18.3 ± 3.2
19.7 ± 2.5
Left arm cycle %
27.3 ± 11.0
26.9 ± 5.3
29.5 ± 7.0
28.4 ± 1.7
19.4 ± 3.6
18.1 ± 2.0
21.8 ± 3.7
20.1 ± 4.6
29.4 ± 4.5
33.3 ± 4.9
30.7 ± 6.7
32.9 ± 8.7
18.7 ± 5.5
17.4 ± 3.8
21.7 ± 1.2
20.3 ± 3.2
29.0 ± 6.0
29.4 ± 5.2
The maximum and minimum active drag forces for all groups are outlined in Table
5.3. No significant differences were found when comparing right and left active drag values
in any variable.
82
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
Table 5.3: Maximum and minimum active drag forces for an individual stroke cycle shown
as mean ± SD
Male
Elite
Sub-Elite
Female
Elite
Sub-Elite
Maximum (N)
(right)
Maximum (N)
(left)
Minimum (N)
(right)
Minimum (N)
(left)
315 ± 68*
240 ± 49
311 ± 69
260 ± 58
19 ± 25
33 ± 26
25 ± 34
30 ± 39
185 ± 21
194 ± 34
184 ± 17
184 ± 34
25 ± 15
20 ± 14
21 ± 12
22 ± 13
*Significant (p < 0.05) difference between elite and sub-elite groups
The percentage of time at the occurrence of maximum and minimum active drag
values in all groups are outlined in Table 5.4. No significant differences were found when
comparing right and left stroke active drag values at maximum and minimum time points in
any group.
Table 5.4: Time of maximum and minimum active drag forces, as a percentage of total
stroke time shown as mean ± SD
Male
Elite
Sub-Elite
Female
Elite
Sub-Elite
Time of maximum
(%) (right)
Time of maximum
(%) (left)
Time of minimum
(%) (right)
Time of minimum
(%) (left)
68 ± 23
70 ± 31
50 ± 25
54 ± 22
57 ± 36
43 ± 30
52 ± 27
44 ± 22
63 ± 28
69 ± 25
66 ± 36
42 ± 23
30 ± 26
48 ± 33
48 ± 45
40 ± 30
Figures 5.1 and 5.2 present typical breakdowns of the percentage of time spent in each
stroke phase for an elite and sub-elite male and female respectively. Visual assessment
concluded that the elite sample in both males and females produced a biphasic profile with
two distinctive smooth peaks in active drag. Although some participants in the sub-elite
category produced this biphasic pattern, the majority produced a multiphasic curve with no
distinct peaks (Appendix G).
83
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
a)
Peak 1
Peak 2
Push
Pull
Entry and Catch
Exit and Recovery
b)
Figure 5.1: Male elite (a) and sub-elite (b) breakdown of stroke phases within an active drag
profile in one-stroke cycle (right hand entry to subsequent right hand entry)
a)
Peak 2
Peak 1
Entry and Catch
Pull
Push
Exit and Recovery
b)
Figure 5.2: Female elite (a) and sub-elite (b) breakdown of stroke phases within an active
drag profile in one-stroke cycle (right hand entry to subsequent right hand entry)
Given the consistent biphasic characteristics within the elite sample, specific
indicators of each stroke phase can be identified in the elite sample only. Specifically, during
a right stroke cycle the following can be observed: 1) Entry and catch: gradual increase in
active drag; 2) Pull: continued ascent to first peak of active drag followed by gradual descent;
3) Push: small descent of active drag to slightly plateau state before starting to increase again;
and 4) Exit and recovery: continued ascent to second peak phase and then the final descent
84
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
before repeated for the next full right stroke cycle. This series of identifiers can also be
observed during the left stroke cycle. The varied profiles seen in the sub-elite group limited
the ability to generalise stroke phases. Within the female sample, although all elite swimmers
produced the same biphasic curve, several sub-elite swimmers also produced this common
biphasic curve, whilst others produced the multiphasic curve seen in Figure 5.2b.
Discussion
The main purpose of the present study was to establish stroke phases within ATM protocol
active drag force-time profiles. The mean active drag values were supported by previous
research using the ATM with a fluctuating tow velocity as reported by Hazrati et al.13 with
mean values of 140 N at 1.86 m/s for males and 83 N at 1.61 m/s for females. Within the
current study mean values reported of 138 ± 24 N at 1.89 ± 0.05 m/s in the elite male sample
and 93 ± 18 N at 1.61 ± 0.04 m/s in the sub-elite females sample. The mean active drag
values were also noticeably lower than those reported in previous studies previous literature
using the ATM with a constant tow velocity. Formosa et al.9 reported mean active drag
values of 262 ± 33 N at 1.92 ± 0.06 m/s within a study of eight male swimmers. This
difference may be attributed to the difference in tow velocity protocol as constant towing has
been shown to produce higher mean active drag values. Additionally, ATM protocols
enabling intra-stroke velocity variations have demonstrated higher performance reliability.
The first aim of the present study was to outline the percentage of time spent in each
stroke phase during ATM swimming. There were no significant differences between male
right and left, and female right and left, stroke phases. Previous work conducted by Toussaint
and Truijens15 revealed significant differences in percentage of time spent in each stroke
phase between elite and sub-elite swimmers. These differences occurred in the entry and
85
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
catch and push phases.15 However, Toussaint and Truijens14, observed these differences
whilst testing only the arm pulls of swimmers (legs were bound), whereas in the ATM
protocols unrestricted arm and leg movements were allowed. Chollet, et al.10 also presented
percentage of time spent in each phase within a stroke cycle, however within their sample of
14 elite swimmers (10 males and 4 females) results revealed that 25.2 ± 5 % of a stroke time
is spent in the entry and catch, 23.4 ± 2.4 % in the pull, 25.2 ± 3.5 % in the push, and 26.2 ±
2.7 % in the exit and recovery phase. These values were more evenly spread across the four
phases than the current study values which could be attributed to the assisted swimming
instead of normal free swimming. Additionally, the differing results may be due to more
varied stroke patterns within the current study’s larger sample (n = 31). Seifert, et al.1,
however, reported varying means of percentage of time in each stroke phase. Percentages of
seven elite and seven sub-elite males were reported as elite entry and catch 24.5 %, pull 20.5
%, and push 27.3 %; and sub-elite entry and catch 27.7 %, pull 21.3 %, and push 22.7 %. The
variance between Seifert, et al.1 and the current ATM protocol in the elite sample could be
again due to the difference between free swim technique and assisted tow technique as the
participants’ average 100 m performance times are quite similar in both samples (Seifert, et
al.1 51.3 ± 2.3 s, current study 50.80 ± 1.53 s). The sub-elite group, however show similar
values for time spent in each stroke phase when comparing the male sample only. Although
there are reported differences within the literature the percentage of time spent in each stroke
could be individualised to the ATM fluctuating tow velocity protocol.
The second aim of this investigation was to examine the magnitude of the male and
female active drag profiles. Of the four variables assessed, the maximum active drag value
during the right stroke cycle was significantly higher in the male elite group than the male
sub-elite group (p = 0.024), however no differences were found on the left stroke. The male
elite sample tended to have a greater active drag magnitude (mean range of 296 N) than the
86
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
male sub-elite group (mean range of 207 N). However, this concept of active drag varying
between swimmers of different performance levels is well supported in literature as it is
thought that as the swimming velocity increases so too does the active drag.7,
8, 14, 15
Moreover, active drag magnitudes were also found to be significantly greater in males versus
females for both left and right cycles. This difference of magnitude could also be explained
by the significant difference in swim velocity between the genders. Active drag has also been
determined by anthropometry dimensions, such as body cross-sectional area or height, within
groups of elite swimmers producing similar free swim performance.16, 17 However, this area
of study has not been investigated in recent years. Furthermore, research involving a larger
sample of female front crawl swimmers, as the current study only had three elite females, is
warranted to provide a better understanding of the active drag magnitudes and force-time
profiles within this sample.
A noticeable difference in curvature was found between the elite and sub-elite
instantaneous single stroke active drag profiles. The third aim of the current study was to
determine if there was a difference between elite and sub-elite active drag force-time profiles.
Visual assessment concluded that the elite sample in both male and female groups produced
smooth biphasic profiles with two distinctive peaks in active drag. In contrast, the sub-elite
group demonstrated a mix of biphasic and multiphasic curves. Specifically, when
qualitatively observing the male elite and sub-elite profiles a distinction between the two
groups is achievable without prior knowledge of performance characteristics given the
significant difference in peak active drag values and multiphasic appearance of the sub-elite
curve. Formosa et al.9 reported active drag force-time profiles using the ATM with a constant
velocity tow and quantitatively found that within the sample of eight elite male sprinters, half
of the swimmers produced an asymmetrical stroke pattern through inspection of maximum
and minimum absolute forces. In contrast, the current study revealed no significant
87
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
asymmetrical differences between right and left stroke maximum and minimum peak forces.
The observed difference may be attributed to the velocity method whereby Formosa et al.9
used constant assisted towing while the current method used a fluctuating assisted tow.
Within a constant velocity tow any minor asymmetries could be exaggerated as the tow does
not permit for intra-stroke velocity fluctuations and therefore may emphasise a stronger or
weaker arm cycle thus potentially explaining this variation between ATM protocols.
The fourth aim was to establish whether a profile can indicate individual stroke
phases. As a common biphasic curve was observed in the elite sample, the ability to judge
where the swimmer is in their individual stroke phase is therefore more easily assessed in
conjunction with synced video footage within this group of swimmers. Kolmogorov’s2
investigation into an elite male swimmer’s steady-state non-stationary motion in water
provided an example of what a force-time profile in front crawl swimming could look like.
Despite both Kolmogorov’s2 and current study involving different techniques of producing
these active drag profiles, the biphasic nature of the profiles are similar. However, the
distribution of stroke phases shown in Kolmogorov’s2 study varies to what is revealed in the
current study. This could be due to Kolmogorov2 displaying a different phase breakdown
during the front crawl stroke. The four phases are outlined with simple labels and diagrams in
Kolmogorov’s2 study include, 1) gripping with pull, 2) drawing with over-arm motion, 3) pull
off with over-arm motion, and 4) pull off with support. In addition, Kolmogorov2 presents a
figure demonstrating a medium distance swimmer’s force-time profile which contains a more
asymmetrical and multiphasic curve. With this visual comparison with the current study,
perhaps the distinct biphasic curvature is representative of pure sprint front crawl swimmers,
whereas a more multiphasic curve exists for other distance swimmers or sub-elite sprinters.
Although research in this area is still in its infancy the ATM protocol which allows for intrastroke velocity fluctuations is a novel approach in the investigation of free swimming. Future
88
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
work on these profiles should be aimed towards establishing a larger female cohort and
further examining the male sprint profile in an attempt to determine if an optimal sprint
profile exists. Additionally, research should be carried out to determine whether these forcetime profiles can evaluate front crawl technique.
Conclusion
In conclusion, the current investigation sought to examine the fluctuating ATM protocol
active drag force-time profiles. The main finding of this paper was the consistent biphasic
pattern found within the elite male sample. This pattern could be a precursor to identifying an
optimal force-time profile in elite swimmers. Swimmers who did not prefer sprint events
within the sub-elite group generally produced the multiphasic pattern. Therefore, identifying
an optimal force-time profile for elite sprinters could enable training or technique
interventions to attempt to replicate the profile to enhance performance.
References
1.
Seifert L, Toussaint HM, Alberty M, Schnitzler C, Chollet D. Arm coordination, power,
and swim efficiency in national and regional front crawl swimmers. Human Movement
Science. 2010;29(3):426-39.
2.
Kolmogorov S. Kinematic and dynamic characteristics of steady-state non-stationary
motion of elite swimmers. Russian Journal of Biomechanics. 2008;12(4):56-70.
3.
Hollander AP, De Groot G, Van Ingen Schenau GJ, et al. Measurement of active drag
during front crawl arm stroke swimming. Journal of Sports Sciences. 1986;4:21-30.
89
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
4.
Kolmogorov SV, Duplishcheva OA. Active drag, useful mechanical power output and
hydrodynamic force coefficient in different swimming strokes at maximal velocity.
Journal of Biomechanics. 1992;25(3):311-8.
5.
Alcock A, Mason B. Biomechanical analysis of active drag in swimming. In:
Proceedings of the 25th International Society of Biomechanics in Sports. Brazil; 2007.
p. 212-5.
6.
Mason B, Sacilotto G, Dingley A. Computation of a swimmer's propulsive force profile
from active drag parammeters with fluctuating velocity in assisted towing.
In:
Proceedings of the 30th International Society of Biomechanics in Sports; Jul 2-6, 2012;
Melbourne, Australia.
7.
Mason B, Sacilotto G, Menzies T. Estimation of active drag using an assisted tow of
higher than max swim velocity that allows fluctuating velocity and varying tow force.
In: Proceedings of the 29th International Society of Biomechanics in Sports; Jun 27-Jun
1, 2011; Porto, Portugal; p. 327-30.
8.
Wang L, Yan W, Li D, Shen X. A new device for estimating active drag in swimming
at maximal velocity. Journal of Sports Sciences. 2007;25(4):375-379.
9.
Formosa D, Mason B, Burkett BJ. The force-time profile of elite front crawl swimmers.
Journal of Sports Sciences. 2011;29(8):811-9.
10.
Chollet D, Chalies S, Chatard JC. A new index of coordination for the crawl:
description and usefulness. International Journal of Sports Medicine. 2000;21(1):54-9.
90
Chapter 5: Investigation of Front Crawl Stroke Phases within Active Drag Force-Time Profiles in Elite and
Sub-Elite Sprint Swimmers
11.
Sacilotto G, Franco R, Mason BR, Ball N. Investigation of front crawl stroke phases
within force-time profiles in elite and sub-elite male sprint swimmers. Journal of
Science and Medicine in Sport. 2013;16(Supplement 1).
12.
Formosa D, Mason B, Burkett BJ. Measuring active drag within the different phases of
front crawl swimming. In: Proceedings of the 11th International Symposium of
Biomechanics and Medicine in Swimming; Jun 16-19, 2010, olso, Norway.
13.
Hazrati P, Mason B, Sinclair PJ. Reliability of estimating active drag using the assisted
towing method (ATM) with fluctuating velocity. In: Proceedings of the 31st
International Society of Bioemchanics in Sport; Jul 7-11, 2013; Taipei, Taiwan.
14.
Toussaint HM, Truijens M. Biomechanical aspects of peak performance in human
swimming. Animal Biology. 2005;55(1):17-40.
15.
Toussaint HM, Roos PE, Kolmogorov S. The determination of drag in front crawl
swimming. Journal of Biomechanics. 2004;37(11):1655-63.
16.
Huijing PA, Toussaint HM, Mackay R, et al. Active drag related to Body Dimensions:
Human Kinetics; 1988.
17.
Toussaint HM. Biomechanics of propulsion and drag in front crawl swimming. In:
Proceedings of the 20th International Society of Biomechanics in Sport; Jul 1-5, 2002;
Caceres, Spain.
91
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
Chapter 6
Study 4: Investigation of Coach Ratings of Technique and
Active Drag Force-Time Profiles in Elite and Sub-Elite
Front Crawl Sprint Swimmers.
Introduction
Numerous studies have been conducted with the aim of improving swim performance either
through kinematic analysis, intervention studies, or kinetic analysis. Although research into
accurately measuring kinetic variables in free swimming is still in its infancy, extensive work
has been completed in the area of estimating active drag.1-7 Originally, work investigating
active drag was focused on mean values,5,
6, 8
however, in recent years more attention has
been given to generating active drag force-time profiles with the aim to utilise them as
objective assessments of free swimming performance.9-10 Previous investigations also
assumed that swimmers would complete test trials at a constant velocity, thus allowing an
inverse interaction between generating propulsion and encountering active drag to be
determined.5, 6, 11 However, as new methods were developed with the ability to capture the
changes in the form of a swimmer during the stroke cycle, i.e. intra-stroke velocity
fluctuations, dynamic interaction effects of propulsion and active drag could be observed and
studied.11 By understanding this ever changing interaction between how a swimmer generates
propulsion to overcome active drag will enable scientists, coaches and swimmers to more
objectively evaluate their technique to minimise resistance and maximise propulsion.
93
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
Recent publications have indicated that the ATM protocols which allow for intrastroke velocity fluctuations are reliable in capturing propulsion and active drag force-time
profiles.
8, 10, 12, 13
Sacilotto et al.14 produced preliminary work on the propulsive force-time
profiles using the ATM method on elite and sub-elite male swimmers. It was concluded that
the elite sample produced comparable curves which suggests the possibility of an optimal
force-time profile existing in elite male front crawl sprint swimmers.14 As this previous study
only incorporated propulsion force-time profiles of eight elite and six sub-elite swimmers,
future work was recommended to include a larger sample size and incorporating female
profiles.14 Moreover, investigation into whether comparable force-time profiles are found in
elite active drag profiles should also be a focus of future work. Therefore, by analysing
propulsion and active drag force-time profiles further distinctions could be investigated
within elite and sub-elite swimmers. This in turn could determine whether force-time profiles
could be used to appraise swimming technique or outline technical flaws. However, further
examination is required before the overall goal of integrating this form of biomechanical
analysis into the assessment of front crawl technique can be achieved. As the ATM is a
reliable method in capturing instantaneous active drag data whilst allowing for intra-stroke
velocity fluctuations and can already identify differences between elite and sub-elite male
propulsive force-time profiles, the inclusion of this method for further examinations seems
warranted.
Recent work completed by Waters et al.15 outlined that a coach’s ability to diagnose
the strengths and weaknesses of an athlete’s performance is critical for optimal development
of the athlete. With this in mind, the opportunity could exist to utilise coach ratings of
technique as a quantifiable measure to assess force-time profiles. Preliminary work was
conducted on a new coach feedback survey, which examined the degree of variability
between coach ratings of front crawl technique.17 Initial findings indicated that coach overall
94
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
ratings of technique (coaches observed a video of swimmers completing the ATM protocol
with a fluctuating tow velocity and rated their performance) were in statistical agreement
with 100 m performance time (r = -0.75, p = 0.005) indicating that the coaches rated the
faster swimmers with higher overall scores.17 An attempt was also made to assess coach
ratings at different phases throughout the front crawl stroke (e.g. pull and push17), however
only a moderate level of internal consistency was found (α = 0.615). As Sacilotto et al.16 only
utilised a sample of seven coaches who assessed 12 swimmers, recommendations were to
include more coaches and swimmers to determine if the survey could allow coaches to
identify technique proficiency. Additional preliminary work was conducted using the same
coach feedback survey ratings of technique, however analysis was conducted by assessing the
relationship between force data and the coach ratings.18 Sacilotto et al.18 concluded that the
correlations ranged randomly in strength of association and in direction, therefore an unclear
understanding of within stroke force-time profiles was outlined and further investigation was
encouraged with the inclusion of more coaches and a greater range of swimmers.
This novel approach of examining active drag could be achieved by first quantifying
stroke technique using coach ratings and then, using these ratings, attempt to identify
relationships between stroke technique ratings and force-time profiles. Therefore, the aim of
this investigation was to assess the effectiveness of using coach ratings of technique when
determining characteristics within ATM active drag force-time profiles. To achieve this, the
first aim was to examine the quantification of technique via coach ratings and secondly to
determine if a relationship exists between coach technique ratings and active drag force-time
profiles.
95
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
Method
Eight coaches (2 Gold and 6 Silver License (Australian swim coaching and teaching
association (ASCTA) standards); 18.88 ± 7.22 years of coaching experience) completed
technique assessment surveys on all elite and sub-elite swimmers. Elite swimmers were
categorised as over 700 FINA points, and sub-elite as under 700 FINA points. Previously
recorded ATM trials from thirty State and National level swimmers (12 elite swimmers with
a mean age 19.42 ± 2.7 years, 771 ± 58 FINA points; and 18 sub-elite swimmers with a mean
age 18.39 ± 2.81 years, 612 ± 54 FINA points) were included in the analysis for this
investigation. Ethical approval was granted through both the Australian Institute of Sport and
University of Canberra. Consent was obtained from each swimmer, and in the case of minors,
consent was obtained from a parent or guardian.
All swim trials used in this investigation were performed, recorded and analysed
using the protocol outlined in Sacilotto et al.14,
18
. Briefly, this involved the swimmers
performing a modified race warm-up which focused on 10 m front crawl sprints, followed by
three free swim trials across a 10 m interval to obtain a mean maximal free swim velocity.
Participants were then towed twice in a passive state (streamline position) at their mean
maximal free swim velocity. A fraction of their mean passive drag force was then used to
generate an individualised fluctuating tow velocity protocol. All participants then performed
three maximal effort assisted towing swim trials with fluctuating velocity. The assisted tow
trial with the median active drag value was selected and the second single stroke cycle from
within this trial was chosen for analysis to ensure the whole stroke to be digitised.
Instantaneous active drag values were calculated from the equations described in previous
work by Mason el al.13 to produce the force-time profiles. All trials were video recorded
using three genlocked cameras (head-on camera: JVC DV Camcorder GY-DV550, side-on
96
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
cameras: Applied Microvideo variable zoom lens JRNL 26056 (above) and a SwimPro Basic
8.0 mm (below)). The side-on above and underwater camera images were mixed with an
Edirol video mixer (EDI-V8) to produce a single moving above/below image. A digital timecode (MicroImage Video Systems, Video Timer VT300) was applied to both camera inputs
to allow accurate deconstructions of a stroke cycle into the four phases used in analysis. The
breakdown of stroke events were the first frame observed for each of the four phases outlined
in Sacilotto et al.18 and were defined as: 1) Entry (first frame of the hand entering the water);
2) Pull (first frame of the hand moving backwards); 3) Push (first frame of the hand being
directly underneath the shoulder); and 4) Exit (first frame of the hand exiting the water).
Each coach independently assessed and rated the overall technique of each swimmer
from video footage and at designated events within the front crawl stroke cycle. They
performed this by completing a survey (Figure 6.1 and Appendix H) which required them to
rate the overall technique from sagittal and frontal plane video of the assisted tow trial. Each
event was presented to coaches as left and right side still images captured from frontal and
sagittal plane video from the selected single stroke cycle. Coaches were not made aware of
the swimmer’s name or performance ability, however given the specific aims instructed to the
coaches, coaches were not excluded from critiquing their own swimmers if they within the
sample of swimmers.
All coach ratings were collected using a mark on a Likert Scale ranging from 0 (poor
technique) to 20 cm (excellent technique), coaches were asked to indicate their opinion of the
quality of technique with an ‘X’ which was later measured in centimetres to obtain a
numerical rating.
97
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
Figure 6.1: An example of a coach feedback survey question at a stroke event.
Trial active drag values used were the mean scores taken from the four full cycles.
Deconstructed active drag event values were obtained from instantaneous active drag force
data and selected from the same time point as the images at each stroke event (Figure 6.2
example of one stroke event). The variables assessed were: 1) coach event ratings for entry,
pull, push, and exit; and 2) active drag event values for entry, pull, push, and exit. All eight
variables were assessed for both right and left stroke cycles. In addition, all overall and event
variable scores were correlated against FINA point scores as a measure of performance. As
FINA points normalise performance irrespective of whether it is a male or a female swimmer.
98
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
Figure 6.2: Time-code matching with active drag force data (Time-code rounded to the
nearest 0.002 of a second)
Statistical analysis was completed using IBM Statistical Package for Social Sciences
(SPSS) Version 19.0. Descriptive statistical parameters (means, standard deviations) were
calculated for the overall variables (overall ratings and trial active drag values). Examination
of the survey was completed using Cronbach’s Alpha for coach ratings for each swimmer to
determine the internal consistency of the survey. Good internal consistency was considered as
α > 0.7.19 An anecdotal repeatability analysis was conducted using one coach and the
percentage difference between measures was reported. Time difference between measures
from this one coach was nine-months after initial completion of the coach feedback survey
and was conducted on two of the subjects. Coaches were also asked to list their opinion on
what technical characteristics they look for when assessing swim performance. A qualitative
summary was used to assess coach opinion of technique and common points were highlighted
if two or more coaches made note of the technique characteristic. The relationship between
coach ratings and active drag values was assessed by Pearson’s product moment correlation
coefficients with a level of significance set at p < 0.05. Magnitudes of all correlations were
interpreted using these thresholds: low r = 0.10 – 0.30; moderate r = 0.30 – 0.50; and high r =
> 0.50.20, 21 Where r is positive, there is a trend for both variables to increase together, and
where r is negative there is a trend for one variable to increase as the other decreases.
Independent t tests, with a level of significant set at p < 0.05, were utilised to determine if
99
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
there was a significant difference in coach ratings between elite and sub-elite swimmers.
Paired t tests, with a level of significance set at p < 0.05, were utilised to determine if
variance exists between right and left stroke cycle variables.
Results
The mean and standard deviations (± SD) values for the sample are presented in Table 6.1.
Significantly faster velocities were found in the elite compared to the sub-elite swimmers in
both the mean swim and assisted towed trials. Additionally, as expected, the FINA point
scores were also found to be significantly higher in the elite sample when compared to the
sub-elite sample.
Table 6.1: Mean (± SD) of trial active drag and overall coach ratings
Swim
Tow
FINA
Velocity
Velocity
point
m/s
m/s
scores
Elite (n = 12
1.84 ± 0.11* 1.98 ± 0.11*
771 ± 58*
Sub-elite (n = 18)
1.71 ± 0.13
1.84 ± 0.15
612 ± 54
*significant (p < 0.05) difference between elite and sub-elite
Active
Drag
N
126 ± 32
115 ± 34
Coach
Overall
Rating
11.7 ± 2.1
10.5 ± 2.0
Cronbach’s alpha tests conducted across all eight coaches for their ratings on all 30
swimmers revealed that in only six occasions the coach’s produced alphas over 0.7. In the
remaining 24 participants, alphas ranged between 0.119 and 0.684, signifying that poor
consistency was present between coaches. Anecdotal repeatability analysis of one coach from
this sample measuring two subjects on two separate testing days revealed a 9 % and a 13 %
variation between their first and second attempt. Table 6.2 shows common qualitative
technique characteristics outlined by the coaches. Common points were defined as being
stated by two or more coaches.
100
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
Table 6.2: Qualitative summary of common points at each stroke event made by all eight
coaches.
Stroke Phase
Entry
Pull
Push
Exit
Commonly mentioned technique factors
1. High elbows (90̊ at the time of catch)
2. Shoulder in line with hand
1. Weight on front end (or hand on forearms)
2. Hand leads elbow
3. Trunk rotation is important
1. Controlling/maintaining hand depth
2. Other arm starting entry phase
3. Body rotation to allow a ‘snap’ exit to finish the stroke (suggests 30̊)
1. Elbow leads hand
2. Arm is never full extended upon exit
3. Fingers relaxed
4. Hand, wrist, elbow, shoulder alignment is important
A high correlation was found to exist between the elite overall active drag value and
coach overall rating (Table 6.3). In contrast, an almost negligible correlation were found
between the sub-elite’s overall active drag value and coach overall rating of technique.
Additionally, both elite and sub-elite overall active drag values were poorly correlated with
FINA points. Results outlined in Table 6.3 demonstrate that within the sub-elite sample,
overall mean active drag is strongly influenced by their swim and tow velocity, which is not
the case in the elite sample. Higher mean overall active drag values correlated highly with
higher overall coach ratings within the elite sample, however not in the slower sub-elite
sample. Weak relationships were found between coach overall ratings and performance
variables (Table 6.4), with the exception of the positive significant relationship between the
sub-elite swimmers and their FINA points.
Table 6.3: Trial active drag value correlation assessments between FINA, swim velocity, tow
velocity and coach overall values for elite and sub-elite groups.
Trial Active Drag
Elite (N)
Sub-elite (N)
FINA
0.043
0.230
Swim Velocity
0.547
0.800**
*significant (p < 0.05) relationship between variables
101
Tow Velocity
0.383
0.732**
Coach Overall
0.588*
0.044
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
Table 6.4: Overall coach rating correlation assessments between FINA, swim velocity, tow
velocity and trial active drag values for elite and sub-elite groups.
Coach Overall
Elite
Sub-elite
FINA
0.303
0.562*
Swim Velocity
0.367
0.068
Tow Velocity
0.198
0.077
Trial Active Drag
0.588*
0.044
*significant (p < 0.05) relationship between variables
When establishing whether coach event ratings were a predictor of stroke event active
drag values, visual inspections of scatter plots suggested no linear relationship existed
between coach event ratings and event active drag values (see Appendix I). The right and left
stroke event coach ratings and active drag event values for each swimmer’s stroke cycle are
presented as means and standard deviations in Table 6.5. The data revealed significantly
higher active drag values in the elite sample at the right and left pull, and left exit events than
in the sub-elite. Similarly, the elite sample produced a significantly lower coach event rating
at the left exit. The left active drag event value was also revealed to be significantly higher
than the right active drag event.
Table 6.5: Summary of mean coach event ratings and mean active drag event (DAE) values.
R
Entry
R DAE
Entry
(N)
R
Pull
R DAE
Pull
(N)
R
Push
R DAE
Push
(N)
R
Exit
R DAE
Exit
(N)
Elite
11.8 ± 2.3
81 ± 87
11.3 ± 2.2
221 ±
87*
11.6 ±2.2
80 ± 48
12.5 ±1.8
103 ±
46^
Subelite
12.7 ± 2.3
98 ± 49
12.4 ± 1.8
137 ± 73
12.4 ± 1.9
102 ± 70
13.6 ± 1.7
89 ± 51
L
Entry
L DAE
Entry
(N)
L
Push
L DAE
Push
(N)
L
Pull
L DAE
Pull
(N)
L
Exit
L DAE
Exit
(N)
Elite
12.0 ± 2.2
72 ±73
11.1 ± 1.7
57 ± 45*
10.8 ± 2.6
231 ±69*
11.6 ±
2.0*
128 ±
54*
Subelite
12.4 ± 2.1
115 ± 62
11.9 ± 2.2
117 ± 85
12.0 ± 2.0
140 ± 70
13.3 ± 1.9
79 ± 36
R = Right stroke cycle, L = Left stroke cycle, DAE = Active drag event value, *Significant (p
< 0.05) difference between elite and sub-elite; ^Significant (p < 0.05) difference between
right and left arm stroke variables
102
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
Very low to moderate correlations were observed between the elite sample’s active
drag event values and coach event ratings (Table 6.6). Similarly, very low to moderate
relationships were found between the coach event ratings and the FINA point scores.
Table 6.6: Elite correlation analysis between coach event ratings and FINA scores and active
drag event values
Elite
R
Entry
L
Entry
R
Pull
L
Pull
R
Push
L
Push
R
Exit
L
Exit
FINA
-0.331
-0.252
0.201
0.044
-0.21
-0.375
0.011
0.079
DAE
-0.207
0.495
0.598*
0.210
0.028
-0.215
-0.464
0.096
R = Right stroke cycle, L = Left stroke cycle, DAE = Active drag event, *Significant (p <
0.05) correlation between coach event rating and active drag event value
Within the sub-elite sample, as shown in Table 6.7, very low to moderate correlations
were found in both directions (positive and negative) between active drag event values and
coach event ratings. Further, almost negligible correlations were found between FINA point
scores and coach event ratings.
Table 6.7: Sub-elite correlation analysis between coach event ratings and FINA scores and
active drag event values
Subelite
R
Entry
L
Entry
R
Pull
L
Pull
R
Push
L
Push
R
Exit
L
Exit
FINA
0.074
-0.005
-0.009
0.052
0.063
0.108
-0.031
-0.155
DAE
0.095
0.401
-0.087
0.102
0.092
-0.159
0.245
-0.336
R = Right stroke cycle, L = Left stroke cycle, DAE = Active drag event
Discussion
The two central findings to emerge from this study were: 1) the coach feedback survey in its
current format and with a low sample of coaches does not produce accurate quantitative
103
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
information regarding a swimmers technique; and 2) in its current state the coach feedback
survey should not be utilised in determining if relationships exist between coach ratings and
ATM active drag force-time profiles.
The first aim was to examine quantification of technique via coach ratings of coach
ratings of overall technique and ratings of technique at selected events within a stroke cycle.
Although all eight coaches have produced elite level swimmers and revealed similar
qualitative opinions of technique characteristics, technique was inconsistently identified
within the surveys. The variation in Cronbach alpha scores warrants further exploration
before future work utilises the coach feedback survey as an effective method in quantifying
technique. Anecdotal analysis of one coach from this sample was tested for repeatability in
two of the participants and revealed a 9 % and a 13 % variation between their first and second
attempt. This limitation of the current study could be a clear indication of a learning effect
whilst completing this survey, adding to the internal inconsistency. Results from a recent
study involving the assessment of visual search behaviours of coaches when they assess
swimming performance could be incorporated to allow for a higher level of examination into
what coaches look for when they complete the coach feedback survey.15 Waters et al.15 found
that verbal responses by expert coaches (> 10 years coaching experience) on front crawl
technique were more in-depth and the coaches’ provided a greater number of technique
points than their developing counterparts. Leas and Chi22 also established this when
comparing qualitative technical feedback from swimming coaches. Within the current study,
which only includes coaches defined as being ‘expert’ provided individual qualitative
feedback which was inline in both Waters et al.15 and Leas and Chi22 thus demonstrating
coach ability in being consistent with verbal assessment, however once this critique is
quantified mixed results were found. Therefore, two conclusions can be made from the
current coach feedback survey: 1) the survey itself is ineffective at allowing coaches to
104
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
consistently critique technique; and 2) the survey is effective, however coaches are not
consistent in how they assess technique in this formatting.
Coaches from the previous study were required to predict a swimmer’s 50 m
performance time from a five-second video clip.15 Water et al.15 concluded that there was no
difference between expert and developing coaches’ ability to predict a performance time,
however neither groups were accurate in their predictions. This result contradicted previous
findings from Leas and Chi22 who reported that two expert coaches were accurate at
predicting performance time. Similar to Waters et al.16 in the current study it was found that
higher coach ratings of technique did not necessarily correlate highly with the faster
swimmers. For example, the swimmers with the higher FINA points did not obtain the higher
coach ratings. Although in the current study the coaches aren’t being critiqued on their ability
to predict performance times it was thought that coaches would identify the elite swimmers as
having better technique than the sub-elite swimmers. A reason for this could be similar to that
found in Water et al.15, that predicting swim time from a five-second video clip could be too
novel a task to properly assess a skill and therefore determine a relevant rating of technique.
Within the current study an average of a five-second clip and still images were used by
coaches to assess technique. To establish a more reliable study of how coaches assess
technique perhaps an investigation of the length of exposure to the trial or multiple viewings
of a video clip may enable a more accurate technique rating. In addition, previous work
assessing how coaches critique performance has only utilised small samples of coaches
(current study: n = 8 expert; Waters et al.15: n = 4 expert level; Leas and Chi22: n = 2 expert)
therefore, a larger cohort of coaches should be recruited to minimise the error associated with
a small sample size as relationships between variables are harder to demonstrate with fewer
data points.23
105
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
Although statistical analysis was generated for determining if relationships existed
between coach ratings and ATM active drag values, the data presented may not be a true
indication of the effectiveness of the system in critiquing performance. As only minor
relationships were found between the ratings and the ATM active drag force-time profile
events, the process used in this study could be used as a blueprint for future investigations
once the coach feedback survey has been assessed for accuracy. Therefore, the establishment
of relationships between coach ratings and ATM active drag was not met the current study.
Moreover, the third aim of this investigation was to conclude whether the ATM could be
implemented as an objective measure of technique for future training interventions. The
results obtained in the current investigation are limited in providing an answer for this given
the novel approach to technique assessment. Further investigation is required in to how
coaches critique technique before reassessing coach ratings against active drag events.
However, as previous work has outlined, the reliability of the ATM and the active drag
profiles have indicated the potential of an optimal curve, future experimentation could utilise
the ATM as a tool to measure performance during an intervention study. Additionally, the
need to commence research using active drag as a performance variable in intervention
studies is warranted as intervention studies over the last three decades have utilised
performance measures (e.g. race distances, 50 m, 100 m, etc), physiological indicators, and
stroke mechanics (stroke lengths, rates and depths) however none have utilised active drag.
Conclusion
Given the coach survey did not provide an accurate method of technique assessment, limited
conclusions as to whether a relationship exists between the coach ratings and ATM active
drag force-time values can be made. In general, only minor relationships were found between
the coach ratings and the active drag values. However, this does not negate the future use of
106
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
the ATM active drag force-time profiles when using it as a feedback tool in athlete
development. There are a limited number of investigations which quantify coaches’
proficiency of assessing technique and this avenue of study should be examined further as
well as how the ATM protocol can be incorporated into athlete servicing. Other methods of
examining how technique impacts force-time profiles should also be explored.
References
1.
Alcock A, Mason B. Biomechanical analysis of active drag in swimming. In:
Proceedings of the 25th International Society of Biomechanics in Sports. Brazil; 2007.
p. 212-5.
2.
Benjanuvatra N, Dawson G, Blanksby BA, Elliott BC. Comparison of buoyancy,
passive and net active drag forces between Fastskin and standard swimsuits. Journal of
Science and Medicine in Sport. 2002;5(2):115-23.
3.
Bixler B, Schloder M. Computational fluid dynamics : an analytical toll for the 21 st
century swimming. Journal of Swimming Research. 1996 Fall;11:4-22.
4.
Formosa D, Mason B, Burkett BJ. Measuring active drag within the different phases of
front crawl swimming. In: Proceedings of the 11th International Symposium of
Biomechanics and Medicine in Swimming; Jun 16-19, 2010; Olso, Norway.
5.
Hollander AP, De Groot G, Van Ingen Schenau GJ, et al. Measurement of active drag
during front crawl arm stroke swimming. Journal of Sports Sciences. 1986;4:21-30.
107
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
6.
Kolmogorov SV, Duplishcheva OA. Active drag, useful mechanical power output and
hydrodynamic force coefficient in different swimming strokes at maximal velocity.
Journal of Biomechanics. 1992;25(3):311-8.
7.
Wang X, Wang L, Yan W, Li D, Shen X. A new device for estimating active drag in
swimming at maximal velocity. Journal of Sports Sciences. 2007;25(4):375-379.
8.
Sacilotto G, Mason B, Ball N. Intra-reliability of active drag values using the assisted
towing method (ATM) approach. In: Proceedings of the 30th International Society of
Biomechanics in Sport; Jul 1-6, 2012; Melbourne, Australia.
9.
Kolmogorov S. Kinematic and dynamic characteristics of steady-state non-stationary
motion of elite swimmers. Russian Journal of Biomechanics. 2008;12(4):56-70.
10.
Mason B, Sacilotto G, Menzies T. Estimation of active drag using an assisted tow of
higher than max swim velocity that allows fluctuating velocity and varying tow force.
In: Proceedings for the 29th International Society of Biomechanics in Sports; Jun 27-Jul
1, 2011; Porto, Portugal; p. 327-30.
11.
Toussaint HM, Truijens M. Biomechanical aspects of peak performance in human
swimming. Animal Biology. 2005;55(1):17-40.
12.
Hazrati P, Mason B, Sinclair PJ. Reliability of estimating active drag using the assisted
towing method (ATM) with fluctuating velocity. In: Proceedings of the 31st
International Society of Biomechanics in Sport; Jul 7-11, 2013; Taipei, Taiwan.
108
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
13.
Mason B, Sacilotto G, Dingley A. Computation of a swimmer's propulsive force profile
from active drag parammeters with fluctuating velocity in assisted towing.
In:
Proceedings of the 30th International Society Biomechanics in Sport; Jul 2-6, 2012;
Melbourne, Australia.
14.
Sacilotto G, Franco R, Mason BR, Ball N. Investigation of front crawl stroke phases
within force-time profiles in elite and sub-elite male sprint swimmers. Journal of
Science and Medicine in Sport. 2013;16(Supplement 1).
15.
Waters A, Lay B, Tidman S, Benjanuvatra N. Visual search behaviour an information
extraction differences between high-level and developing swimming coaches.
In:
Proceedings for the 12th International Symposium of Biomechanics and Medicine in
Swimming; Apr 28-May 2, 2014; Canberra, Australia.
16.
Sacilotto G, Clothier PJ, Mason BR, Ball N. Variability in coach assessment of
technique in front crawl sprint swimming. In: Proceedings for the 12th International
Symposium of Biomechanics and Medicine in Sport; Apr 28-May 2, 2014; Canberra,
Australia; 2014. p. 222-6.
17.
Seifert L, Chollet D, Allard P. Arm coordination symmetry and breathing effect in front
crawl. Human Movement Science. 2005;24(2):234-56.
18.
Sacilotto G, Ball N, Mason B, Clothier PJ. Investigation of coach ratings of technique
and force-time profiles in elite male front crawl sprint swimmers. In: Proceedings for
the 32nd International Society of Biomechanics in Sport; Jul 12-16, 2014; Tennessee,
United States.
109
Chapter 6: Investigation of Coach Ratings of Technique and Active Drag Force-Time Profiles in Elite and
Sub-Elite Front Crawl Sprint Swimmers
19.
DeVellis RF. Scale Development - Theory and Applicants. 2nd ed. United States of
America: Sage Pulications, Inc; 2003.
20.
Cohen J. Statistical power analysis for the behavioral sciences: Psychology Press; 1988.
21.
Anderson ME, Hopkins WG, Roberts A, Pyne DB. Ability of test measures to predict
competitive
performance
in
elite
swimmers.
Journal
of
Sports
Sciences.
2008;26(2):123-30.
22.
Leas R, Chi M. Analysing diagnostic expertise of competitive swimming coaches. In:
Starkes J, Allard F, editors. Cognitive Issues in Motor Expertise; 1993.
23.
MacCallum RC, Widaman KF, Zhang S, Hong S. Sample size in factor analysis.
Psychological methods. 1999;4(1):84.
110
Chapter 7: Summary, Conclusions and Future Directions
Chapter 7
Summary, Conclusions and Future Directions
Summary
Over the last several decades scientists have attempted to understand the free swimming
phase of a swim performance. Thus far, methods have been created and examined for the
assessment of the propulsive and resistive forces with much criticism due to the inability to
develop testing protocols which minimise stroke alterations. However, until such time as
kinetic variables can be accurately measured, these methods are accepted in order to progress
the research in free swimming. In recent years the assisted towing method (ATM) has been at
the forefront of estimating active drag whilst allowing for intra-stroke velocity fluctuations.
Additionally, the inclusion of active drag force-time profiles enables new research to be
conducted on free swimming kinetics. This thesis is comprised of a series of studies which
examined the potential of the ATM for its integration as an objective assessment tool in
technique analysis of front crawl free sprint swimming.
The first study in this series evaluated the reliability of the mean active drag values
collected using the ATM. The reliability was assessed using both a constant tow velocity and
a fluctuating tow velocity. Originally in literature when calculating a mean active drag value,
velocity was considered to be stationary, however, as acknowledged in literature, swimming
is a dynamic change in body form and thus non-stationary velocity changes should be
investigated. Previous investigations utilising the ATM have used both the constant velocity
assumption and have examined assisted tow trials which allow for intra-stroke velocity
111
Chapter 7: Summary, Conclusions and Future Directions
fluctuations, or non-stationary velocity. Study 1 was therefore designed to assess these two
protocols to determine the level of performance reliability of swimmers to produce consistent
mean active drag values. It was revealed that a minimal number of trials were required to
produce a high level of retest reliability in both constant and fluctuating tow velocities. It was
concluded that a minimum of two or three assisted tow trials be incorporated in future work
to minimise the risk of producing atypical generalisation. Conversely, the difference between
the two protocols was most pronounced when determining within-subject variation. The
constant velocity tows (CVTE % = 35.0) revealed very large error when compared to the
fluctuating tow velocities (CVTE % = 12.6). Therefore it was concluded that future work
utilising the ATM should focus on adopting the protocol allowing for intra-stroke velocity
fluctuations.
As a result of the high level of performance reliability using the fluctuating assisted
tow trials, an attempt was made to validate or add further justification as to why this method
should be used in the assessment of free swimming. Therefore, study 2 aimed to compare the
stroke mechanics commonly assessed in free swimming (stroke length and stroke rate) and
determine if there was a significant difference between a free swimming trial and an assisted
towed trial. This study revealed a significant difference between the free swim and assisted
tow trial stroke mechanics, which was consistent with previous literature. This increase was
due to the significantly increased swim velocity between the free and assisted swimming.
Despite this finding, it was then thought to attempt a novel reporting of the ratio between
stroke length and stroke rate in numerical form to compare between the two velocity
conditions. It was revealed that assisted tow trials had the ability to predict free swim stroke
mechanics and a significantly high correlation was found between the two ratios (r = 0.95).
Therefore, it was concluded that the consistent increase found between the free swim and
112
Chapter 7: Summary, Conclusions and Future Directions
assisted tow trial could allow assumptions to be made from the kinetic ATM output and
transferred into the assessment of free swimming.
As the ATM protocol produces instantaneous force-time profiles as well as mean
active drag values, the assessment of how this profile could be integrated into normal
servicing for athletes and coaches was examined in study 3. This study had the purpose of
investigating the stroke phases within ATM swimming. This was achieved by determining
that no significant difference existed with respect to percentage of time spent in each stroke
phase, establishing the magnitude of the different groups of swimmers (elite and sub-elite
male and female groups), establishing whether a common force-time profile existed within
groups, and finally it was established that with the use of synced video footage a common
biphasic curve was observed in the elite sample. This was a promising finding, as there is
potential to identify an optimal force-time profile.
Study 4 aimed to explore whether coach ratings of technique are an acceptable
method of outlining technique characteristics via the ATM active drag force-time profiles. As
previous reliability, validity and the establishment of normative profiles were conducted in
this thesis, the need to apply this method for integration to be used as an effective assessment
tool was warranted. Although the coach feedback survey was shown to be limited in its
current state to provide accurate coach ratings of technique, future work can incorporate the
survey to determine if the survey is flawed or the capability to quantitatively convert coaches’
measures of technique is unobtainable. The outline of this study can therefore be utilised as a
blueprint for future work when determining if a relationship exists between technique ratings
and active drag force-time profiles.
113
Chapter 7: Summary, Conclusions and Future Directions
The work conducted in this thesis aimed to assess the reliability and justify the use of
the ATM in technique analysis in front crawl sprint swimming. Particularly in the male elite
sample, it was found that there may be an optimal force-time profile present within the area
of active drag. This finding has not been examined before in literature and poses a few
questions into how this can be incorporated into regular sport science services for enhancing
performance. A consistent profile was also found within the elite female group, however with
only three participants categorised into this group, further work is required to determine if
this is a common pattern in both male and female elite swimmers. Further investigation is
required into how these profiles could be used in future work given the important findings
discussed within this thesis.
Conclusions
This thesis produced four main additions to the area of swimming biomechanics: 1) the ATM
protocol is a reliable tool to capture kinetic information for the assessment of free swimming;
2) the consistent increase observed in stroke mechanics enabled a transfer of technique
characteristics between assisted tow trials and free swimming; 3) the ATM active drag forcetime profiles revealed the possibility of an optimal profile being established within the elite
sprint swimmers thus enabling the ATM to be used in the future as an objective assessment of
technique; and 4) a novel tool was presented which has potential to quantify technique
proficiency. The main findings were as follows:
Study 1: Reliability of active drag values using the assisted towing method
1. Swimmers are able to consistently produce reliable mean active drag values using both
the constant and fluctuating tow velocity protocols.
114
Chapter 7: Summary, Conclusions and Future Directions
2. No main effect was observed for the number of trials required to obtain reliable values in
either constant or fluctuating tow velocity protocols. However, to minimise the risk of an
atypical result a recommendation of two or three assisted trials was made.
3. Assisted tow trials which allowed for fluctuating intra-stroke velocity variations were
found to produce a more stable protocol to use when estimating kinetic variables utilising
the ATM.
Study 2: A comparison of front crawl stroke mechanics between free swim and assisted
towed swimming
4. When towing a swimmer with an assisted fluctuating tow velocity, their stroke
mechanics will significantly increase as a result of the significantly increased mean swim
velocity.
5. The alteration in stroke mechanics is consistent throughout the sample as shown when
comparing stroke length and stroke rate ratios, which suggests that assisted tow trials can
be used to transfer kinetic output of technique into the assessment of free swim kinetics.
Study 3: Investigation of front crawl stroke phases within force-time profiles in elite and
sub-elite sprint swimmers
6. No significant differences were found between percentage of time spent in each stroke
phase within any of the groups ass (elite and sub-elite male and female groups)
7. Significant differences were found between elite and sub-elite magnitude in peak active
drag values. This was thought to be as a consequence of a faster swim velocity
115
Chapter 7: Summary, Conclusions and Future Directions
8. A common biphasic profile was found within the elite sample and is thought to represent
an optimal force-time profile. Although some sub-elite swimmers produced the same
biphasic profile, some sub-elite swimmers’ profiles were multiphasic and distinction
between stroke events was not possible.
9. Front crawl stroke phases were able to be identified with synced video footage alongside
the force-time profiles. This was particularly so in the elite sample where distinctive
curve features were observed.
Study 4: Investigation of coach ratings of technique and active drag force-time profiles
in elite and sub-elite front crawl sprint swimmers
10. The coach feedback survey in its current format paired with a low sample of coaches,
does not produce accurate quantitative information regarding a swimmers technique.
11. In its current state, the coach feedback survey should not be utilised in determining if
relationships exist between coach ratings and ATM active drag force-time profiles.
Future Directions
The following points represent general recommendations for future research.
1. The first study in this series investigated the reliability of the assisted towing method and
what was the best protocol for collecting consistent active drag values. Given the high
instance of reliability in the fluctuating tow velocity trials, investigation into what
protocols produce reliable results in the other competitive strokes is warranted. Of the
three other strokes, backstroke should be the easiest to determine protocols for as the
116
Chapter 7: Summary, Conclusions and Future Directions
asymmetrical arm movements and kicking patterns are quite similar to front crawl.
Breaststroke and Butterfly will pose a challenge as there is a higher variation in intrastroke velocity due to the technical aspects of each of these strokes.
2. Given the current progression in aquatic research in building a velocity transducer, the
possibility to validate the assisted towing method is nearly a reality. Despite study 2
revealing the consistent increase in stroke length and stroke rate ratio between free
swimming and towed swimming, the need to validate this system is necessary to confirm
that the equal power assumption is either true or false. Once this transducer is available,
future work could compare the towed velocity profile to a free swim velocity profile and
if these profiles produce similar characteristics, such as shape or magnitude, the equal
power assumption could be validated.
3. The present study demonstrated that a similar biphasic active drag profile was found
within the elite sample, although, only three elite females were included in this study.
Future work should incorporate a larger number of female sprinters to allow a more
thorough cross section to be observed and comparisons made.
4. Further, once reliable protocols are established for the three other competitive strokes
(Backstroke, Breaststroke and Butterfly) identification of stroke phases should be
investigated. It would be expected that as a common shape of an active drag profile in
front crawl swimming has been found, that a common shape will be observed in the other
strokes as well.
5.
Additionally, a delimitation of this study was the restriction of the participants not to
breathe during testing trials. During the sprint races (50 m, 100 m and 200 m) breathing
117
Chapter 7: Summary, Conclusions and Future Directions
patterns change dramatically. Examinations into if there are variances between breathing
and non-breathing active drag profiles should be investigated.
6. The present analysis method could be used to investigate how a technique intervention
could affect the shape of a profile. An intervention study would be the next phase from
this body of work in applying the ATM into becoming a regular servicing tool for
athletes and coaches.
7. All participants in the present study were considered free from injury. Examination into
how swimmer’s active drag profiles are shaped who are returning from injury may
enable this protocol to be used in their rehabilitation by monitoring their progress
through the profiles. Further, general observation into whether asymmetries can be found
between the right and left arm strokes could be used to prevent injuries from happening.
8. The surveys used by coaches to rate technique, although were considered very thorough
and provided a large amount of information, given the small return from original
distribution; future work should develop a survey which requires less work from the
coaches. This would also enable examination into how a larger cross section of coaches
rate technique and why. Further insight could be gained by determining what factors
influences coaches to assess technique the way they do, for example, does their opinion
change if they completed a sport science degree or whether they were a competitive
swimmer.
9. Although comparing the coach surveys and active drag values was insightful, comparing
active drag values with an already established method of quantifying technique should
118
Chapter 7: Summary, Conclusions and Future Directions
also be investigated. Therefore, an examination of how active drag values correlate with
the IdC approach could be undertaken.
10. This thesis focused solely on how the active drag values obtained from the ATM
approach related to technique analysis, however the ATM also produces a propulsive
force-time profile and as a consequence of these two force traces, a resultant force
profile. Therefore, investigation into the characteristics common in these profiles, as well
as how they can be used in technique analysis could be considered for future work.
119
Appendix A: Publications from Doctoral Thesis
Appendix A: Publications from Doctoral
Thesis
121
Appendix A: Publications from Doctoral Thesis
Sacilotto GB, Ball N, & Mason BR. A Biomechanical Review of the Techniques Used to
Estimate or Measure Resistive Forces in Swimming. Journal of applied
biomechanics. 2014;30:119-127
Sacilotto GB, Mason BR, Ball N., & Clothier PJ. Investigation of coach ratings of technique
and force-time profiles in elite male front crawl sprint swimmers. In: Proceedings
for the 32nd International Society of Biomechanics in Sport; Jul 12-16, 2014;
Tennessee, United States.
Sacilotto GB, Clothier PJ, Mason BR, & Ball N. Variability in coach assessments of
technique in front crawl sprint swimming. In: Proceedings for the 12th International
Symposium on Biomechanics and Medicine in Swimming; Apr 28-May 2, 2014;
Canberra, Australia. p. 222-226
Sacilotto GB, Franco R, Mason BR, & Ball N. Investigation of front crawl stroke phases
within force-time profiles in elite and sub-elite male sprint swimmers. Journal of
Science and Medicine in Sport. 2013;16(Suppl 1).
Sacilotto GB, Mason BR., & Ball N. Intra-reliability of active drag values using the assisted
towing method (ATM) approach. In: Proceedings for the 30th International Society
of Biomechanics in Sport; Jul 2-6, 2012; Melbourne, Australia.
122
Appendix B: Published Literature Review
Appendix B: Published Literature Review
123
Journal of Applied Biomechanics, 2014, 30, 119-127
http://dx.doi.org/10.1123/jab.2013-0046
© 2014 Human Kinetics, Inc.
An Official Journal of ISB
www.JAB-Journal.com
REVIEW
A Biomechanical Review of the Techniques
Used to Estimate or Measure Resistive Forces in Swimming
Gina B.D. Sacilotto,1,2 Nick Ball,2 and Bruce R. Mason1
1Australian
Institute of Sport; 2University of Canberra
Resistive or drag forces encountered during free swimming greatly influence the swim performance of elite
competitive swimmers. The benefits in understanding the factors which affect the drag encountered will
enhance performance within the sport. However, the current techniques used to experimentally measure or
estimate drag values are questioned for their consistency, therefore limiting investigations in these factors.
This paper aims to further understand how the resistive forces in swimming are measured and calculated. All
techniques outlined demonstrate both strengths and weaknesses in the overall assessment of free swimming.
By reviewing all techniques in this area, the reader should be able to select which one is best depending on
what researchers want to gain from the testing.
Keywords: active drag, biomechanics, front crawl, assisted swimming, resisted swimming
Understanding the resistive forces encountered
within the free swim phase of a performance is difficult.
This difficulty is largely due to controversy surrounding
the ability to measure this force. Only a limited number of
reviews were identified that discussed resistive forces, or
drag forces, in swimming and only one of those reviews
was published in the last ten years.1–3 With the advancements in technologies and techniques, an updated review
is required to ensure that sport scientists and coaches can
accurately and effectively incorporate drag testing within
their athlete preparation and performance analysis.
Drag force is the force component parallel to and in
the same direction as the relative fluid flow.1 The equation that is used to calculate an object’s drag force (DF) is
DF = 1 2 C D r v 2 A (1)
where CD is the drag coefficient, ρ is the density of the
fluid, v represents the velocity of the object, and A indicates the frontal surface area of the object. As water is the
main fluid of interest in this review, the key focus of all fluid
dynamic principles will be hydrodynamics. Hydrodynamic
drag force in human swimming can be identified as the
total resistive force experienced by a swimmer.4
Gina B.D. Sacilotto (Corresponding Author) is with the Aquatic
Testing, Training and Research Unit, Australian Institute of
Sport, Bruce, ACT, Australia, and with the Faculty of Health,
University of Canberra, Bruce, ACT, Australia. Nick Ball is
with the Faculty of Health, University of Canberra, Bruce,
ACT, Australia. Bruce R. Mason is with the Aquatic Testing,
Training and Research Unit, Australian Institute of Sport, Bruce,
ACT, Australia.
The resistive forces that influence the swimmer in the
water include form, wave, and frictional drag,5–7 which
are influenced by the swimmer’s velocity, boundary
layer, shape, size, and the frontal surface area,8 as noted
as well in Equation 1. In swimming, the resistive forces
are termed active drag and passive drag. Active drag is
the water resistance associated with the dynamic swimming motion5,9 and passive drag is the water resistance
that a human body experiences in a fixed or unchanging
posture.5,9,10 Kolmogorov and Duplishcheva11 confirmed
that active drag varies between individuals and seems to
relate to swim technique and anthropometry.2 As noted
in Equation 1, in the context of human swimming, drag
force represents the swimming drag, which could be an
active or passive drag.9 For the purpose of this review,
only active drag in relation to front crawl swimming
will be discussed as it relates to the performance of an
individual swimmer.
This review was researched by database search
engines using the following search terms: active drag,
swimming, swimming propulsive forces, front crawl
technique, swimming resistive forces. A reference
check for each paper found was also performed.
Searches were also conducted using the names of author
popular in this area of study. Therefore, the aim of this
paper was to assess the techniques used to estimate or
measure active drag in swimming to progress the study
in this area.
Active Drag and Swim Performance
In competitive swimming, the two most commonly identified factors that are primarily responsible for swim speed
are propulsion and drag.5,12–15 The ability of a swimmer
119
120 Sacilotto, Ball, and Mason
to reduce the active drag encountered allows for propulsive forces to be efficiently applied, therefore producing
faster swim velocities.8,16–19 Clarys20 confirmed that active
drag was mainly influenced by the changes in the body’s
shape and the movement of the body segments or, in other
words, a direct result of faulty swimming techniques.
Kolmogorov et al9 supported this claim by demonstrating
that elite swimmers were more able to reduce active drag
than nonelite swimmers. These findings were speculated
to be a result of elite swimmers having superior stroke
mechanics. Therefore, swimming fast may depend on the
ability of the swimmer to reduce drag through an efficient
stroke technique, which will generate a higher velocity
and limit the power lost in wasted kinetic energy.21 However, contrary to this argument, a review was completed
in 1992,2 which outlined results found in Hollander et al22
as demonstrating no significant correlation in active drag
and swim velocity values at a constant swim velocity. As
a result of these findings, it was concluded that active
drag was not a determining factor of maximal swimming
when swimming with a constant velocity. The preceding
review continues on to speculate that perhaps a swimmer’s anthropometry may explain the active drag ranges
seen in the literature.2 However, research in determining
correlations between anthropometry and active drag have
been minimally investigated within the last two decades
in relation to adult swimmers (see studies,5,8,23,24 for
research involving anthropometry and active drag in
young swimmers). Further research into determining the
relationship between free swim velocity and active drag,
while allowing for the natural intrastroke fluctuations, is
important for updating the current understanding of active
drag and swim performance.
Mechanical Power Output in Swimming
It has been suggested that swimming performance is
defined by the relationship between the useful mechanical power output, active drag, the hydrodynamic force
coefficient (drag coefficient) and the maximal free swim
velocity.11 The mechanical power output is the power
delivered by swimmers to overcome drag (useful power)
and the power wasted in giving kinetic energy change to
the water.16,25 Mechanical power output (PO) has, therefore, been evaluated as the product of the swimming drag
(D) and velocity (v):11
PO = D ⋅ v (2)
The ratio between the useful power and the wasted
kinetic energy is defined as the propelling efficiency of
a swimmer:25
Pd
(3)
PO
where ηP is the propelling efficiency and Pd is the useful
power. An understanding of mechanical power output and
active drag has been the basis for many studies measuring
or estimating active drag.11,16,25,26
hP =
Techniques of Drag Assessment
For many years in swimming research, attempts have
been made to accurately measure active and passive
drag;11,22,26–29 however, there has been much controversy,
as different techniques have produced varying drag
values.5,26,30,31 The energetics approach,32 numerical solutions6,33,34 and experimental techniques11,22,27,35 have been
developed and used to estimate or measure drag forces
in swimming. All techniques have been modified and
criticized as the demand for understanding the resistive
forces in swimming increases.
Energetics Approach
The energetics approach, or otherwise termed theoretical calculations, investigates the theoretical relationship
between the energy costs of swimming, the velocity,
the overall mechanical efficiency of the swimmer, and
the body drag.29 Predominantly, the calculations for
this approach are aimed at deciphering the mechanical
power output the swimmer produces while free swimming.29,30,32,36–39 In this technique, the swimmer is towed
while swimming at a set pace—which is maintained
by a towing carriage as seen in Figure 1—with known
additional weights to provide assistance/resistance. The
maximal oxygen consumption is also recorded throughout each trial to understand the swimmer’s energy expenditure at a given average velocity.40 The body drag of the
swimmer is determined by adding (or subtracting) extra
loads to (or from) swimmers moving at a known speed.
The extra drag was measured and related to the swimmer’s energy expenditure to calculate the drag as well as
the swimmer’s mechanical efficiency.40
To calculate the drag, Di Prampero et al40 identified
a linear relationship between drag and maximal oxygen
consumption (VO2net) at constant swim velocities, which
led to this technique of determining drag as a function of
VO2net.20 Therefore, as stated by Clarys,20(p13) this technique “extrapolated the linear regression between VO2net
and the added propulsion and added drag to VO2net = 0.”
At a constant mean velocity, the mean propulsive force
exerted by the swimmer will be equal and opposite to the
active drag produced.
Investigations that use the energetics approach38,40,41
found similar values of active drag when comparing
propelling efficiency values as a percentage. Although
intrastudy drag values were similar, it must be noted
that these authors assumed that the propelling efficiency
did not change in experiments where active drag was
calculated.42 It is likely that propelling efficiency will
change, even at a constant velocity, when external loads
are applied, as is the case in the approach to estimating
active drag. In addition, small changes in VO2net values
due to small deviations in propelling efficiency will be
amplified by the extrapolations that are the basis for
these studies.42 Van de Vaart et al43 reiterated this point
by determining that indirect techniques of estimating
active drag (by extrapolation) appeared to overestimate
Evaluation of Active Drag Assessment Techniques 121
Figure 1 — Adapted experimental setup of drag collection from Di Prampero et al.40
these values. Furthermore, by introducing a snorkel to
measure the VO2net, the swimmers frontal surface area
is altered, which could modify the results for active
drag. Although this technique for estimating active
drag, as we now understand, produced questionable
results, it was the first to describe the total active drag
of a front crawl swimmer20 and therefore the first step
in the progression of all the current techniques used
today. Because of this technique, the investigation into
active drag values between male and female swimmers,
swimmers with different swim velocities, as well as the
analysis of the energetics in swimming were able to be
undertaken.40
Numerical Simulations
Numerical simulations use the computational modeling
of the water flow surrounding the swimming to determine the resistive forces.8 The main approach to drag
force measurement using numerical solutions is through
computational fluid dynamics (CFD). Computational
fluid dynamics solves and analyses problems involving
fluid flow by means of computer-based simulations.44
Using this method, the investigator can analyze computer models, for example, a 3D model of a swimmer,
and can simulate desired movement patterns to give
feedback about the alteration, such as a modification
in stroke technique. By manipulating a computer
model instead of a human form, studying the drag
values in swimming through CFD, theoretically, limits
the amount of stroke mechanic alterations due to the
constraints set in the experimental protocols.31 Bixler
and Schloder45 introduced two-dimensional CFD into
the swimming world and then after a further six years,
the first three-dimensional CFD study was published.46
Since these landmark studies in swimming CFD, several
authors have investigated hydrodynamic drag using this
method.6,33,45,47
Computational fluid dynamic simulations allow
the elimination of within-subject variability, which is
found in laboratory and field experimentation. Furthermore, another benefit of CFD is that for the same
input you always have the same output. Bixler, Pease,
and Fairhurst33 presented an investigation on the study
of the water flow and drag force characteristics (acting
around and upon a human body) while in a submerged
streamlined position. In this study, a comparison of total
drag force was performed between an actual swimmer, a
virtual CFD model of the swimmer, and an actual mannequin based on the virtual model. Although this study
only investigated the effects of passive drag, the results
were positive, as the aim of establishing a CFD model of
122 Sacilotto, Ball, and Mason
a submerged human body and the effects of passive drag
was achieved.33 In addition, Bixler et al33 demonstrated
the accuracy of using the CFD technique by comparing
model values with a real human. Therefore, the potential
for this method in resistive force assessment is quite
promising and the results of the previous paper represents
a necessary first step toward more complicated CFD
analysis in which active drag could be evaluated.33 For
CFD to become a readily available method of resisted
force assessment, basic kinematic measures during free
swimming need to be collected, for example, the ability
to collect instantaneous swim velocity, or knowing where
the center of gravity is while a swimmer is swimming.
A further limitation of this technique is that, generally speaking, CFD simulations require an enormous
amount of computing time, which is particularly true in
swimming analysis. In the analysis process of a human
swimming, one has to move boundaries in the simulation and also resolve the flow around a complex-shaped,
deforming object.31 These extra computing requirements
can further increase the complexity and computational
costs of the simulation,31 thereby making it difficult for
coaches and scientists to use this method, in its current
state, effectively.
Experimental Techniques
Experimental techniques have been developed and
applied to try to accurately determine the resistive
forces encountered by a swimmer. The techniques most
frequently found include the direct measurement of
active drag through the measuring active drag system
(MAD-system),22 and the indirect techniques of collecting active drag values, for example, the velocity
perturbation method (VPM)11 and the assisted towing
method (ATM).27
Measuring Active Drag System. To directly measure
active drag, Hollander et al22 developed the MAD-system.
This device measures the drag force generated by a swimmer, which enables the calculation of the propulsive force
produced during the trial. To obtain propulsive force, the
assumption was made that the mean propulsive force
would be equal to the mean active drag values when the
swim velocity is constant.22
The MAD-system, as seen in Figure 2, requires
the swimmer to push off from fixed pads underneath
the water. In the original study using this technique,
ten trials were completed at different yet constant
velocities. The swimmer’s legs were restricted by the
use of a small buoy. The depths of the pads were able
to be adjusted for the swimmer’s height as well as the
distance between pads. For each trial, the registered
output signal of the force transducer was transmitted
telemetrically to determine mean force. The average
propulsive force was calculated by integration from the
force registrations at a constant swim velocity. The swim
velocity was determined from the sample frequency and
the pad distance (between the second and final pad).22
In the initial investigation of this technique, each test
yielded ten data points of propulsive forces at ten different speeds, which ranged from minimal to maximal
swim velocity.22 The original function used to calculate
active drag can be found in Hollander et al;22 however,
Toussaint et al48 presented the calculation for drag
as
Figure 2 — MAD-system setup for drag collection adapted from Hollander et al.22
DA = Kv 2 (4)
Evaluation of Active Drag Assessment Techniques 123
where DA represents total active drag, K is a constant
(incorporating the density, coefficient of drag, and frontal
surface area), and v equals swimming speed.
The MAD-system has been used extensively in
swimming research in determining direct values of
active drag.7,22,25,48–50 Although there has been extensive
research undertaken utilizing the MAD-system, there is
much criticism surrounding this technique. For example,
the system limits a swimmer’s natural stroke mechanics26,28,51 and it can only be used at a constant velocity.
Therefore, the outcomes from this technique should only
be compared against itself or if a swim velocity is the
same between techniques to critique front crawl pull technique. Another criticism is that the MAD-system protocols only allow the swimmer’s hands to be in contact with
the pads and not react with water as a swimmer would
normally. Hollander et al22 made note that normal hand
trajectories may be altered with this technique; however,
it was justified by stating that at the same swim velocity a different hand trajectory did not necessarily imply
a difference in active drag. Poizat et al51 concentrated
on testing the MAD-system for usability as a training
device for biomechanical evaluation and performance
analysis. Swimmer feedback was recorded along with
the active drag values. The swimmers involved in this
study described that it was difficult to make contact with
the pads, particularly at high velocities.51 This technique
is, however, well established and an effective way to
directly measure the forces encountered and produced
by a swimmer’s upper body throughout a maximal effort.
Velocity Perturbation Method. The VPM approach is
based on the assumption that a swimmer is capable of
producing an equal amount of useful mechanical power
output and that the swimmer will swim at a constant
velocity.11 This technique is seen as a progression from
the energetics approach in estimating active drag, but
without the inclusion of the maximal oxygen consumption element. In the VPM, a swimmer must produce two
equal maximal efforts. This technique is most commonly
used over a 25 m distance; however, it can be used across
other distances. The first maximal effort, the swimmer
must swim “freely”—without any attachments—and the
second effort is swum with a hydrodynamic body attached
to the swimmer, creating a known additional resistance.
Both conditions must be swum across the same distance.
The maximal mean velocity when swimming with the
hydrodynamic body was compared with the maximal
mean free swimming velocity, which along with the
known additional resistance is used to calculate active
drag for free swimming:
DA =
Db vb v 2
(5)
v 3 − vb3
where Db is the additional resistance from the perturbation
buoy and vb and v are the swimming velocities with and
without the hydrodynamic body, respectively.11
Although this method has been said to be worth
pursuing in research,23 it is criticized, as it is an indirect
technique of measuring active drag7 and therefore the
drag values found may be overestimated.16 Furthermore,
the assumptions to which this technique adheres are very
much reliant on the level of swimmers participating.
The first assumption of equal power is dependent on the
swimmer’s skill level, the rest interval between trials,
and whether the swimmer understands the experimental
conditions.11 This suggests that the use of this technique, and others that use the equal power assumption,
should limit participants to those who are of a semielite
or an elite level of swimming capability. The second
assumption of constant velocity limits the study of the
intrastroke fluctuations normally produced in a front
crawl stroke cycle. Kolmogorov and Duplishcheva11
revealed a varying velocity during the trials and these
intrastroke velocity fluctuations were approximated using
a computer-simulated Strouhal number.11 It was found
that the maximal error due to the stroke cycle fluctuations
was around 6–8%. As a result of the VPM being reliant
using a constant velocity throughout resisted trials, the
fluctuations were diminished to maintain a near constant
velocity. The hydrodynamic body that was used was
built to not decrease the swimmer’s velocity by more
than 10%.11 The restricting of the intrastroke velocity
fluctuations is imperative in the use of Equation 4 when
determining mean active drag values.
As a result of the near constant velocity assumption,
a series of hydrodynamic bodies were developed, each
with a different additional resistance, to eliminate the
dependency on a swimmer’s performance level. However,
it has been noted that the additional resistance created
by a hydrodynamic body can be affected by the floating
movements generated by the hydrodynamic body. XinFeng et al26 therefore proposed to develop a simple and
convenient device to estimate the active drag at maximal
velocity based on the equal power assumption and the
VPM approach. Modifications that were made from the
original VPM approach and this version of the VPM
approach consisted of how the additional resistance was
applied to the swimmer. Figure 3 illustrates the apparatus, developed by Xin-Feng et al,26 that maintained the
additional resistance in a steady position. This system
minimizes the floating movement of the hydrodynamic
body and allowed changes in the amount of additional
resistance.
The force transducer (shown in Figure 3) was used
to measure the variation in tension of the thread when the
gliding block was moved by the swimmer. The results
revealed that the tension of the thread fluctuates and, as
a result, the additional resistance in the swimming direction is variable, not a constant value as Kolmogorov and
Duplishcheva11 had assumed. Therefore, in the original
VPM approach, even if the velocities were set to be constant, active drag values may differ due to the velocity
fluctuation restrictions.
Despite the limitations of this technique for assessing
active drag, the VPM has several strengths. For example,
this technique can be set up in any pool facility, as it is
completely portable. This means that testing could occur
124 Sacilotto, Ball, and Mason
Figure 3 — Modification of the VPM approach as used in, and adapted from Xin-Feng et al.26
on away meets or camps, which could provide coaches
and athletes with extra information. Similarly, the VPM
could be easily integrated into a normal training session,
unlike the MAD-system, which requires a long setup
time. Also in comparison with the MAD-system, the VPM
requires little to no adaptation to complete analysis on the
other strokes (ie, butterfly, backstroke, or breaststroke).
A comparison of the VPM and MAD-system was
conducted in 2004,48 which revealed that the two techniques yielded significantly different active drag values.
Given the stroking limitations of both techniques, this
significant difference does not imply that either technique is wrong or measuring different aspects of free
swimming.48 It was concluded that the question of which
aspect of free swimming is being assessed by either the
VPM or MAD-system was not resolved.
Assisted Towing Method. The ATM technique for
estimating active drag has only been introduced in recent
years. Currently, only few studies have been published
using this technique for determining active drag.27,28,52–55
In its current format, the ATM is essentially the reverse
of the VPM approach in estimating active drag, as the
swimmer is assisted, rather than resisted. The ATM is also
based on the equal power assumption and the constant
velocity assumptions. However, as was outlined by XinFeng et al26 and recognized in Kolmogorov and Duplishcheva,11 a swimmer will not be swimming at a constant
velocity at any point throughout a maximal effort due
to the intrastroke fluctuations in front crawl swimming.
Such fluctuations are a result of the intrastroke forces
that are generated during a natural arm stroke cycle. The
kicking action in the front crawl stroke will also add to
these fluctuations.
To further understand these fluctuations, Mason et
al28 compared constant active drag values with fluctuating
active drag values (Figure 4).
During the constant trials, the differences in the
mean velocity maximal free swims (individual swimming
without any attachments) and the mean velocity of the
towed swims (towed from the hip using the dynamometer) are used to calculate for drag with respect to the drag
force required to tow the swimmer. The same calculations
as the VPM approach are used; however, the final equation for estimating active drag using the ATM is altered to
DA =
Fb v2 v12
(6)
v23 - v13
where Fb is the force required to tow the athlete at the
increased speed as measured from the force platform, v2
is the increased tow velocity, and v1 is the maximum free
swim velocity. Similar to the VPM approach, when using
this system with a constant velocity, the dynamometer
is set to 5% faster than the swimmer’s mean maximum
free swim velocity with a high force selection to allow
for a near constant tow. In order for the tow to allow the
swimmer’s intrastroke fluctuations, the force setting on
the dynamometer is reduced and the velocity setting is
increased to 120% of the swimmer’s maximum free swim
velocity. Along with these changes in set force and velocity settings, a parameter on the towing dynamometer is
altered so that when the force setting is reached it will
fluctuate the tow velocity to maintain that force setting,
therefore allowing the intrastroke fluctuations. The force
setting used is a predetermined fraction of the swimmer’s
passive drag tow (streamlined tow at the swimmer’s
maximal free swim velocity) and is different for every
individual swimmer. Despite the increase in the velocity,
setting the mean tow velocity will still equal between 5%
and 10% greater than the swimmer’s maximal mean free
swim velocity; however, when calculated, the velocity
profile will demonstrate the intrastroke fluctuations.28 The
results obtained from the fluctuating trials seem to demonstrate a smoother drag profile, more repeatable results,
as well as most likely resembling more natural stroke
characteristics than what was illustrated in the constant
trials.28 The use of the ATM approach, when allowing
Evaluation of Active Drag Assessment Techniques 125
Figure 4 — ATM approach to drag collection as shown in Sacilotto et al.55
for a fluctuating tow velocity in active drag estimation,
is still in its infancy. However, the results shown thus
far are positive in being able to decipher exactly what
affects performance during free swimming. A recent
review by Sanders et al3 noted that the work presented
in Mason et al28 indeed needed validation; however, it
appeared promising. Since then, two new papers were
published using the ATM technique.54,55 When towing
with a constant velocity, the assumption has been that
drag was equal but opposite in direction to propulsion.
However, when utilizing a tow allowing for intrastroke
fluctuations this cannot be true. A recent study using the
ATM has developed calculations to obtain a swimmer’s
propulsive profile, net force, and acceleration curves54
while allowing their intrastroke velocity fluctuations.
Mason et al54 presents active drag as a negative value,
with the equation for propulsion (P) as
d
( mv ) - DA (7)
dt
where m is the passive drag force of the swimmer (as a
substitute for the mass of a swimmer), v is the velocity
profile, and DA is the active drag profile presented as a
negative value (P is the propulsion profile presented as
a positive value). The practical applications using ATM
and this propulsion equation again appear promising;
however, the validation of this method is still under question, as there is no experimental gold standard to validate
against. In an attempt to somewhat validate this technique,
an intrareliability study was conducted utilizing ATM,
however, with a constant velocity tow.55 Although this
paper only included a small sample size, the results
P=
revealed very good reliability value for within-subject
mean active drag values (interclass correlation of 0.91,
at a confidence limit of 95% and a likely range of 0.58
and 0.98). Future studies need to be completed utilizing
a fluctuating tow velocity to determine the intrareliability
of this technique. Further investigations into whether
the velocity/force profiles obtained in this technique
actually mimic real stroke mechanics will also need to
be undertaken. However, research cannot be continued
with this until researchers are able to accurately measure
basic kinematics while a swimmer is submerged in water.
When basic kinematics become available, a comparison
between a fluctuating tow velocity trial and a maximal
free swim velocity trial can be investigated, at which point
this technique can be validated. Furthermore, similar to
the MAD-system, the ATM technique has only presented
papers in the investigation into front crawl resistive and
propulsive forces. It could be assumed that research into
backstroke, using ATM, could be undertaken following
the same protocols as for front crawl; however, investigations into butterfly and breaststroke seem a while
away given the very large fluctuations in intrastroke
velocities.
Review Summary
Swim performance is very much dependent on the free
swimming component of an event. Therefore, gains or
losses in this aspect of a swim performance can significantly affect the outcome of an event. The purpose of
this article was to evaluate the current techniques used to
estimate active drag in elite front crawl free swimming.
126 Sacilotto, Ball, and Mason
Techniques for experimental drag collection found
in literature appear to be based on the ideals from the
energetics approach in relation to mechanical power
output39,40 even though these were established in the
1970s. More recently, the experimental indirect techniques, although limited, have been found most frequently in literature.11,27 The MAD-system,22 which is
the only system to directly measure active drag, has been
shown to have a large number of limitations in regards
to maintaining the swimmer’s natural stroke mechanics.
Although, when compared with the ATM and VPM, the
MAD-system is more established and can be compared
with itself. Alternatively, the VPM is a more cost effective technique and can be transported easily, unlike the
MAD-system and ATM, where the setup time and cost of
equipment is quite high. The preference of researchers to
use indirect assessment, however, allows the swimmers to
perform more natural stroke mechanics while they swim,
in particular, when allowing normal velocity fluctuations, which can be achieved using the ATM. The main
concern with the indirect techniques has been shown to
be the assumptions associated with the testing protocols.
To overcome these assumptions, the CFD method would
seem to be the best option in determining resistive and
propulsive forces in free swimming. However, until such
time as the basic kinematic measures can be determined
in free swimming, CFD research also relies on the experimental assumptions.
Although this review is not exhaustive of all the
techniques used around the world in the measurement
or estimation of active drag, a clear outline of the most
common techniques used can be obtained. Therefore,
as a guide, the choice of technique or method used to
estimate or measure active drag should be dependent on
what researchers want to gain from the testing. For the
pursuit of understanding what active drag actually is, the
assessment technique that allows a swimmer to be most
natural in the water, or is able to be simulated naturally,
could be the best scientific step forward.
References
1. Rushall BS, Holt LE, Sprigings EJ, Cappaert JM. A reevaluation of forces in swimming. Journal of Swimming
Research. 1994;10:6–30.
2. Toussaint HM, Beek PJ. Biomechanics of competitive front
crawl swimming. Sports Med. 1992;13(1):8–24. PubMed
doi:10.2165/00007256-199213010-00002
3. Sanders R, Thow J, Alcock A, Fairweather M, Riach I,
Mather F. How Can Asymmetries in Swimming be Identified and Measured? Journal of Swimming Research.
2012;19:1.
4. Lyttle AD, Blanksby BA, Elliott BC, Lloyd DG. The effect
of depth and velocity on drag during the streamlined glide.
Journal of Swimming Research. 1998;13:15–22.
5. Benjanuvatra N, Blanksby BA, Elliott BC. Morphology
and hydrodynamic resistance in young swimmers. Pediatr
Exerc Sci. 2001;13:246–255.
6. Marinho DA, Reis VM, Alves FB, et al. Hydrodynamic
Drag During Gliding in Swimming. J Appl Biomech.
2009;25(3):253–257. PubMed
7.Toussaint HM, De Groot G, Savelberg H, Vervoorn K,
Hollander AP, van Ingen Schenau GJ. Active drag related
to velocity in male and female swimmers. J Biomech.
1988;21(5):435–438. PubMed doi:10.1016/00219290(88)90149-2
8. Marinho DA, Barbosa TM, Costa MJ, et al. Can 8-weeks
of training affact active drag in young swimmers? J Sports
Sci Med. 2010;9:71–79. PubMed
9. Kolmogorov SV, Rumyantseva OA, Gordon BJ, Cappaert
JM. Hydrodynamic characteristics of competitive swimmers of different genders and performance levels. J Appl
Biomech. 1997;13:88–97.
10. Havriluk R. Performance Level Differences in Swimming:
A Meta-Analysis of Passive Drag Force. Res Q Exerc
Sport. 2005;76:112–118. PubMed doi:10.1080/0270136
7.2005.10599273
11. Kolmogorov SV, Duplishcheva OA. Active drag, useful
mechanical power output and hydrodynamic force
coefficient in different swimming strokes at maximal
velocity. J Biomech. 1992;25(3):311–318. PubMed
doi:10.1016/0021-9290(92)90028-Y
12. Barbosa TM, Costa MJ, Marques MC, Silva AJ, Marinho
DA. A model for active drag force exogenous variables
in young swimmers. Journal of Human Sport & Exercise.
2010;5(3):379–388. doi:10.4100/jhse.2010.53.08
13. Chatard JC, Bourgoin B, Lacour JR. Passive drag is still a
good evaluator of swimming aptitude. Eur J Appl Physiol.
1990;59:399–404. PubMed doi:10.1007/BF02388619
14. Toussaint MH. An alternative fluid dynamic explanation
for propulsion in front crawl swimming. Paper presented
at: 18th International Society of Biomechanics in Sports;
June, 2000; Hong Kong, China.
15.Toussaint HM, Truijens M, Elzinga MJ, et al. Effect
of a Fast-skin ‘body’ suit on drag during front crawl
swimming. Sports Biomech. 2002;1(1):1–10. PubMed
doi:10.1080/14763140208522783
16. D Acquisto L. Berry J, Boggs G. Energetic, Kinematic, and
Freestyle Performance Characteristics of Male Swimmers.
Journal of Swimming Research. 2007;17:31.
17. Greco GC, Pelargio JG, Figueira TR, Denadai BS. Effects
of gender on stroke rates, critical speed and velocity of
a 30-min swim in young swimmers. J Sports Sci Med.
2007;6:441–447. PubMed
18.Mollendorf JC, Termin AC, II, Oppenheim E, Pendergast DR. Effect of swim suit desgin on passive drag.
Med Sci Sports Exerc. 2004;36(6):1029–1035. PubMed
doi:10.1249/01.MSS.0000128179.02306.57
19.Smith JE, Molloy JM, Pascoe DD. The influence of a
compressive laminar flow body suit for use in competitive
swimming. Journal of Swimming Research. 2007;17:10–16.
20. Clarys JP. Human morphology and hydrodynamics. Paper
presented at: Third International Symposium of Biomechanics in Swimming; 1979; Edmonton, Canada.
21.Barbosa TM, Fernandes RJ, Keskinen KL, Vilas-Boas
JP. The influence of stroke mechanics into energy cost of
elite swimmers. Eur J Appl Physiol. 2008;103:139–149.
PubMed doi:10.1007/s00421-008-0676-z
22.Hollander AP, De Groot G, Van Ingen Schenau GJ, et
al. Measurement of active drag during front crawl arm
stroke swimming. J Sports Sci. 1986;4:21–30. PubMed
doi:10.1080/02640418608732094
23.Marinho DA, Garrido N, Barbosa TM, et al. Can 8
Weeks of Training in Female Swimmers Affect Active
Drag? Open Sports Sciences Journal. 2010;3:36–37.
doi:10.2174/1875399X01003010036
Evaluation of Active Drag Assessment Techniques 127
24.Kjendlie P, Stallman RK. Drag characteristics of competitive swimming children and adults. J Appl Biomech.
2008;24(1):35. PubMed
25.Toussaint HM. Differences in propelling efficiency
between competitive and triathlon swimmers. Med Sci
Sports Exerc. 1990;22(3):409–415. PubMed
26.Xin-Feng W, Lian-Ze W, We-Xing Y, De-Jian J, Xiong
S. A new device for estimating active drag in swimming
at maximal velocity. J Sports Sci. 2007;25(4):375–379.
PubMed doi:10.1080/02640410600812021
27.Alcock A, Mason B. Biomechanical analysis of active
drag in swimming. Paper presented at: 25th International
Society of Biomechanics in Sports; July, 2007; Ouro Preto,
Brazil.
28. Mason B, Sacilotto G, Menzies T. Estimation of active drag
using an assisted tow of higher than max swim velocity that
allows fluctuating velocity and varying tow force. Paper
presented at: 29th International Society of Biomechanics
in Sports; July, 2011; Porto, Portugal.
29. Pendergast DR, Di Prampero PE, Craig AB, Wilson DR,
Rennie DW. Quantitative analysis of the front crawl in men
and women. J Appl Physiol. 1977;43(3):475–479. PubMed
30. Zamparo P, Gatta G, Pendergast DR, Capelli C. Active and
passive drag: The roles of trunk incline. Eur J Appl Physiol.
2009;106:195–205. PubMed doi:10.1007/s00421-0091007-8
31. Von Loebbecke A, Mittal R, Mark R, Hahn J. A computational method for analysis of underwater dolphin kick
hydrodynamics in human swimming. Sports Biomech. 2009;
8(1):60–77. PubMed doi:10.1080/14763140802629982
32. Larsen OW, Yancher RP, Baer CLH. Boat design and swimming performance. Swimming Technique. 1981;18(2):42.
33. Bixler B, Pease D, Fairhurst F. The accuracy of computational fluid dynamics analysis of the passive drag of a male
swimmer. Sports Biomech. 2007;6(1):81–98. PubMed
doi:10.1080/14763140601058581
34. Marinho DA, Rouboa AI, Alves FB, et al. Hydrodynamic
analysis of different thumb positions in swimming. J
Sports Sci Med. 2009;8(1):58–66. PubMed
35. Lyttle AD, Elliot BC, Blanksby BA, Lloyd DG. An instrument for quantifying the hydrodynamic drag of swimmers.
A technical note. J Hum Mov Stud. 1999;37(5):261–270.
36. Clarys JP, Jiskoot J. Total resistance of selected body positions in the front crawl. In Proceedings—International symposium on biomechanics in swimming, 2d, Brussels, 1974;
Baltimore, University Park Press, 1975, p. 110-117.1974.
37. Clarys JP, Jiskoot J, Rijken H, Brouwer PJ. Total resistance
in water and its relation to body form. In Nelson, R.C.
and Morehouse, C.A. (ed.), Biomechanics IV, Baltimore,
University Park Press, 1974, p. 187–196.1973.
38. Kemper HC, Verschuur R, Clarys JP, Jiskoot J. Total efficiency and swimming drag in swimming the front crawl.
Biomechanics in Swimming, Waterpolo and Diving;
1983:119–125.
39. Rennie DW, Pendergast DR, Di Prampero PE. Energetics
of swimming in man. Swimming. 1975;II:97–104.
40.Di Prampero PE, Pendergast DR, Wilson DW, Rennie
DW. Energetics of swimming in man. J Appl Physiol.
1974;37(1):1. PubMed
41.Holmer I. Propulsive efficiency of breaststroke and
freestyle swimming. Eur J Appl Physiol Occup Physiol.
1974;33(2):95–103. PubMed doi:10.1007/BF00449511
42.Toussaint HM, van der Helm FCT, Elzerman JR, Hollander AP, de Groot G, van Ingen Schenau GJ. A power
balance applied to swimming. In, Hollander, A.P. (ed.) et
al., Biomechanics and medicine in swimming: proceedings
of the Fourth International Symposium of Biomechanics
in Swimming and the Fifth International Congress on
Swimming Medicine, June 21–25, 1982, Champaign, Ill.,
Human Kinetics Publishers, 1983, p. 165–172. United
States 1983.
43. Van de Vaart AJM, Savelberg HHCM, de Groot G, Hollander AP, Toussaint HM, Van Ingen Schenau GJ. An
estimation of drag in front crawl swimming. J Biomech.
1987;20(5):543–546. PubMed doi:10.1016/00219290(87)90254-5
44. Marinho DA, Rouboa AI, Barbosa TM, Silva AJ. Modelling Swimming Hydrodynamics to Enhance Performance. Open Sports Sciences Journal. 2010;3:43–46.
doi:10.2174/1875399X01003010043
45. Bixler B, Schloder M. Computational fluid dynamics: an
analytical toll for the 21 st century swimming scientist (La
dynamique des fluides informatisee: un outil d’analyse
pour le scientifique de la natation du 21eme siecle). Journal
of Swimming Research. 1996;11:4–22.
46. Bixler B, Riewald S. Analysis of a swimmer’s hand and
arm in steady flow conditions using computational fluid
dynamics. J Biomech. 2002;35(5):713–717. PubMed
doi:10.1016/S0021-9290(01)00246-9
47. Machado L, Ribeiro J, Costa L, et al. The effect of depth
on the drag force during underwater gliding: A CFD
approach. Paper presented at 28th International Society
of Biomechanics in Sports; July, 2010; Michigan, United
States of America.
48.Toussaint HM, Roos PE, Kolmogorov S. The determination of drag in front crawl swimming. J Biomech.
2004;37(11):1655–1663. PubMed doi:10.1016/j.jbiomech.2004.02.020
49. Toussaint HM, DeLooze M, Van Rossem B, Leijdekkers
M, Dignum H. The effect of growth on drag in young
swmimers. Int J Sport Biomech. 1990;6:18–28.
50.Toussaint HM, Truijens M. Power Requirements for
Swimming a World-Record 50-m Front Crawl. Int J Sports
Physiol Perform. 2006;1(1):61–64. PubMed
51. Poizat G, Adé D, Seifert L, Toussaint HM, Gal-Petitfaux N.
Evaluation of the Measuring Active Drag system usability:
An important step for its integration into training sessions.
International Journal of Performance Analysis in Sport.
2010;10(2):170–186.
52. Formosa D, Mason B, Burkett BJ. The force-time profile of
elite front crawl swimmers. J Sports Sci. 2011;29(8):811–
819. PubMed doi:10.1080/02640414.2011.561867
53. Mason B, Formosa D, Toussaint HM. A method to estimate
active drag over a range of swimming velocities which may
be used to evaluate the stroke mechanics of the swimmer.
Paper presented at: Biomechanics and Medicine in Swimming XI; 2010; Oslo, Norway.
54. Mason B, Sacilotto G, Dingley A. Calculation of a Swimmer’s Whole Body Propulsive Force Profile from Active
Drag Parameters which were Computed using a Fluctuating Velocity in Assisted Towing. Paper presented at: 30th
International Society of Biomechanics in Sports; July 4,
2012; Melbourne, Australia.
55. Sacilotto G, Mason B, Ball N. Intra-reliability of active
drag values using the assisted towing method (ATM)
approach. Paper presented at: 30th International Society
of Biomechanics in Sports; July 4, 2012; Melbourne,
Australia.
Appendix C: Proceedings from the 32nd ISBS, Tennessee, United States
Appendix C: 2014 Proceedings from 32nd
ISBS, Tennessee, United States
133
Appendix C: Proceedings from the 32nd ISBS, Tennessee, United States
INVESTIGATION OF COACH RATINGS OF TECHNIQUE AND FORCE-TIME
PROFILES IN ELITE MALE FRONT CRAWL SPRINT SWIMMERS
Gina B.D. Sacilotto1,2, Nick Ball2, Bruce R. Mason1 and Peter J. Clothier3
Australian Institute of Sport, Canberra, Australia1
University of Canberra, Canberra, Australia2
University of Western Sydney, Sydney, Australia3
The purpose of this investigation was to examine the relationship between assisted
towing method (ATM) force-time profiles and coach ratings of front crawl technique.
Nine elite male swimmers completed the ATM sprint swimming protocol to obtain active
drag and propulsion values. Six coaches each rated overall technique from video
footage and technique at each of four stroke events (entry, pull, push, and exit) from
images captured throughout the ATM trials. Mean coach technique rating scores were
then correlated against four performance measures (FINA point score, 100 m
performance best time, active drag value and propulsion value). Results demonstrated
weak to strong relationships between the ratings and performance variables for each
stroke event.
KEY WORDS: active drag, propulsion, coach assessments
INTRODUCTION: The Assisted Towing Method (ATM) is a technique used to
estimate active drag and propulsion in free swimming. This is achieved by
comparing the velocity differences between a free swim maximal effort and an
assisted (towed) maximal swim with regard to the added assistance used to tow the
swimmer. A recent advancement to the ATM involves towing the swimmer with
fluctuating velocity to match the naturally occurring velocity fluctuations present
during the front crawl stroke cycle (Mason, Sacilotto, & Menzies, 2011). To date,
literature on the ATM using fluctuating tow velocities have only presented whole
stroke mean force-time parameters which have been shown to be reliable within
participants (Hazrati, Mason, & Sinclair, 2013). To further the development of the
ATM as a tool for assessment of swimming proficiency, greater understanding of
the within stroke force-time profiles is warranted. One approach considered
appropriate for interpreting the within stroke force-time profiles is to examine the
relationships between technique proficiency and force-time measures at key
stroking events. Therefore, the aim of this study was to quantify technique
proficiency via coach ratings of technique and correlate these ratings against ATM
active drag and propulsion values at selected events of the stroke.
METHODS: Nine elite male national level front crawl sprint swimmers (20.38 ± 2.88
yrs, 776 ± 57 FINA points, and 50.99 ± 1.17 s 100 m performance best time)
completed active drag testing using the ATM protocol described by Mason,
Sacilotto, and Menzies (2011). This involved performing a modified race warm-up
which focused on short front crawl sprints, followed by three free swim trials across
a 10 m interval to obtain a mean maximal swim velocity. Participants were then
towed in a passive state (streamline position) at their mean maximal free swim
velocity. A fraction of their passive drag force was then utilised to generate an
individualised fluctuating tow velocity protocol. All participants performed three
maximal swim effort assisted towing trials. The assisted tow trial with the median
active drag value was selected and the second single stroke cycle from within this
trial was chosen for analysis. Active drag and propulsion values were calculated
from the force-time profiles using the equations described in previous work by
134
Appendix C: 2014 Proceedings from 32nd ISBS, Tennessee, United States
Mason, Sacilotto and Dingley (2012). Following completion and analysis of the ATM
trials, six coaches (two Gold and four Silver Australian coaching licenses) assessed
and rated the technique of each swimmer. The overall assisted swim performance
was assessed by a survey which required the coaches to rate technique from the
sagittal and frontal plane video. Additional ratings were required from still images at
each of the four events within the front crawl stroke. These events represented the
start of the entry, pull, push, and exit stroke phases and were defined as: 1) Entry
(first frame of the hand entering the water); 2) Pull (first frame of the hand moving
backwards); 3) Push (first frame of the hand being directly underneath the
shoulder); and 4) Exit (first frame of the hand exiting the water). Each event was
presented to coaches as left and right side still images captured from frontal and
sagittal plane video from the selected single stroke cycle. Coach ratings were made
by indicating a mark on a Likert Scale ranging from 0 – 20 cm with 0 representing
poor technique and 20 representing excellent technique for overall swim
performance. Coach ratings were also collected on selected technique elements
within each stroke event on a scale of -10 to +10, with 0 being classified as
excellent technique and -10 and +10 indicating poor technique. This rating scale
range allowed for classification of technique where performance could be less than
ideal with regard to direction. For example, hand entry following recovery could be
too narrow, ideal or too wide. These ratings were transformed post-hoc to a scale of
0 to 10, with 0 representing poor technique and 10 being classified as excellent
technique. These elements included: hand position (HP), entry length (EL), trunk
rotation (TR), depth of hand (DH), elbow positioning (EP), and exit length (EL).
Each swimmer was critiqued on a total of 29 different technique parameters. Mean
coach ratings for overall swim performance were then calculated and correlated
against four performance variables (FINA point scores, 100 m performance best
times (PB), active drag, and propulsion). For the stroke events the active drag and
propulsion values used to correlate against coach ratings were instantaneous
measures that corresponded to each stroke event. Pearson’s product moment
correlations were used to determine the relationship between coach ratings and
performance variables. Magnitudes of all correlations were interpreted using the
following thresholds: low r = 0.10 – 0.30; moderate r = 0.30 – 0.50; and high r = >
0.50. The level of significance was set at p < 0.05 and indicated when p < 0.01.
RESULTS AND DISCUSSION: Nine swimmers were tested using the ATM protocol
(1.89 ± 0.06 m/s swim velocity, 2.03 ± 0.08 m/s tow velocity, 150 ± 31 N propulsion,
-150 ± 33 N active drag) and then assessed on their front crawl technique by six
coaches. Results from correlational analyses demonstrate a range of relationships
between coach’s ratings of technique and performance variables. Table 1 shows
the mean coach overall ratings correlated against the performance variables. A
significantly high correlation was found between the FINA point score and the 100
m performance best time. This was an expected finding given the FINA point score
is a ranking based on personal best time and the current World Record. Another
expected significant correlation was found between the mean propulsion and the
mean active drag values. As Mason, Sacilotto and Dingley (2012) outlined, the
formula used to calculate propulsion includes the active drag value and therefore a
high degree of relatedness exists. Aside from these expected findings, however, the
overall coach ratings had negligible correlation to all four performance variables.
This could be a function of the small sample of coaches and swimmers whereby
relationships between variables are harder to demonstrate with fewer data points
135
Appendix C: 2014 Proceedings from 32nd ISBS, Tennessee, United States
(MacCallum, Widaman, Zhang, & Hong, 1999). Alternately, this finding could be a
reflection of the lack of relationship between coach rating of technique and the
performance variables. If so, this demonstrates notable differences in opinion and
high variability in coach ratings of good and poor technique.
Table 1
Correlations between overall coach rating of technique and the four performance measures
Overall Swim Performance
Overall
FINA
100 m PB
Propulsion
FINA
.075
**
100 m PB
-.074
-.965
Propulsion
.138
-.435
.262
**
Active Drag
-.134
.431
-.250
-.993
**= statistically significant at p < 0.01 level
Tables 2, 3, 4, and 5 identify the relationships between the performance variables
and the coach ratings at each of the four stroke events. Within all four stroke events
it was shown that all coach overall ratings when regressed against FINA point
scores presented negative correlations. This is an unexpected finding and suggests
that the higher the FINA point score, which is indicative of greater swimming ability,
the lower the overall rating of technique. A similar relationship trend was observed
between the overall rating at each stroke event and 100 m PB suggesting that the
faster swimmers were rated to have poorer technique. A possible explanation for
these findings could be that coach perception of what constitutes good technique is
not consistent with performance. Or, as previously stated, large variability in opinion
between coaches regarding their perception of good and poor technique existed
which could have confounded the relationships. In addition, the small spread of
performance best times and FINA point scores between participants could have
also confounded these relationships. This is likely given the relatively small sample
of coaches and swimmers who participated in this study.
Table 2
Correlations between overall and selected technique elements and the four performance
measures at entry
Right Entry
Left Entry
Entry L
TR
Entry L
TR
Overall
HP
Overall
HP
**
FINA
-.630
.336
.080
.640
-.822
.425
-.383
.654
*
100 m PB
.628
-.478
-.019
-.502
.740
-.555
.293
-.551
Propulsion
-.359
.583
-.355
-.104
-.051
.718*
-.192
-.248
Activedrag
-.051
.377
-.267
-.038
-.555
.367
.067
.473
HP = Hand Positioning; Entry L = Entry Length; TR = Trunk Rotation; * = statistically significant at
p<0.05 level; **= statistically significant at p<0.01 level
Mixed relationships were observed between coach ratings for the technique
elements within in each stroke event and the performance variables. These
correlations ranged randomly in strength of association and direction (positively or
negatively correlated). However, TR was consistently positively correlated with
FINA points (range r = .539 to .721) and negatively correlated against 100 m PB
(range r = -.432 to -.561). This trend could indicate that a real relationship exists
between the quality of TR and performance in all three underwater stroke events.
When relating TR against the force data, no recognizable trends or patterns were
found across the three underwater stroke events.
136
Appendix C: 2014 Proceedings from 32nd ISBS, Tennessee, United States
Table 3
Correlations between overall and selected technique elements and the four performance
measures at the pull
Right Pull
Left Pull
DH
TR
DH
TR
Overall
HP
Overall
HP
*
FINA
-.292
.827**
.411
.721
-.348
.019
.472
.685*
*
100 m PB
.224
-.767
-.233
-.557
.330
-.039
-.308
-.549
Propulsion
.210
-.090
-.437
.052
.287
.496
-.311
-.109
**
Activedrag
-.886
.022
.341
.417
-.164
-.041
.009
.210
HP = Hand Positioning; DH = Depth of Hand; TR = Trunk Rotation; * = statistically significant at
p<0.05 level; **= statistically significant at p<0.01 level
The results in Table 4 show the highest number of moderate to high correlations for
all performance variables. This could be due to greater observable variation
between swimmer techniques or less variation in coach perception of poor and
good technique. The push stroke event represents the mid component of the
underwater stroke and is a common point of focus for coaches. Therefore, the
coaches may have been more likely to rate similarly which would result in stronger
relationships between measures.
Table 4
Correlations between overall and selected technique elements and the four performance
measures at the push
Right Push
Left Push
EP
TR
EP
TR
Overall
HP
Overall
HP
*
**
FINA
-.342
-.091
.572
.601
-.773
.398
.825
.539
*
*
100 m PB
.344
.074
-.576
-.561
.775
-.539
-.716
-.432
Propulsion
-.527
-.388
.258
.093
-.114
.029
.367
.656
*
Activedrag
-.341
-.059
.145
.403
-.329
-.018
.673
.631
HP = Hand Positioning; EP = Elbow Positioning; TR = Trunk Rotation; * = statistically significant at
p<0.05 level; **= statistically significant at p<0.01 level
Table 5
Correlations between overall and selected technique elements and the four performance
measures at the exit
Right Exit
Left Exit
Overall
Exit L
Overall
Exit L
FINA
-.568
-0.64
-.321
.162
100 m PB
.505
-.020
.224
-.266
**
Propulsion
.455
.467
.861
.433
Activedrag
.181
.174
-.069
.063
Exit L = Exit Length; * = statistically significant at p<0.05 level; **= statistically significant at p<0.01
level
Comparison of left and right side correlations between coach ratings and
performance scores identified mixed associations within the stroke events and
technique elements. These paired comparisons reveal correlation values that also
vary randomly in strength of association and direction (positively or negatively
correlated). Assuming there was consistent rating between sides by coaches, this
lack of similarity between sides could indicate asymmetry in stroke technique within
this participant sample.
CONCLUSION: This work represents a novel approach to interpreting the ATM
force-time profiles in front crawl spring swimming. Correlations that ranged randomly
137
Appendix C: 2014 Proceedings from 32nd ISBS, Tennessee, United States
in strength of association and direction did not provide for a clearer understanding of
within stroke force-time profiles. Further investigation is required with a larger sample
size of coaches and a greater range of swimmers to further explore the relationship
between technique proficiency and ATM force-time profiles.
REFERENCES:
Hazrati, P., Mason, B., & Sinclair, P. J. (2013). Reliability of estimating active drag using the assisted
st
towing method (ATM) with fluctuating velocity. Paper presented at 31 International Society of
Biomechanics
in
Sports,
Taiwan.
Retrieved
from
https://ojs.ub.unikonstanz.de/cpa/article/viewFile/5560/5054
MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor analysis.
Psychological methods, 4(1), 84.
Mason, B., Sacilotto, G., & Dingley, A. (2012). Computation of a swimmer's propulsive force profile
from active drag parammeters with fluctuating velocity in assisted towing. Paper presented at
th
30
International Society of Biomechanics in Sports, Australia. retrieved from
https://ojs.ub.uni-konstanz.de/cpa/article/view/5274/4848
Mason, B., Sacilotto, G., & Menzies, T. (2011). Estimation of active drag using an assisted tow of
higher than max swim velocity that allows fluctuating velocity and varying tow force. Paper
th
presented at the 29 International Society of Biomechanics in Sports, Portugal. Retrieved
from https://ojs.ub.uni-konstanz.de/cpa/article/view/4839
Acknowledgements
The authors would like to thank the coaches and swimmers who were involved in
this investigation. Further acknowledgement is due to the staff in the Aquatic Testing,
Training and Research Unit within the Australian Institute of Sport for their
assistance in data collection.
138
Appendix D: 2014 Proceedings from 12th Biomechanics and Medicine in Swimming, Canberra, Australia
Appendix D: 2014 Proceedings from 12th
Biomechanics and Medicine in Swimming,
Canberra, Australia
139
'
E
T
P
>
H
D
$
?
E
T
#
J
I
K
@
%
&
'
D
E
(
i
_
r
`
s
t
`
o
,
)
'
-
+
)
&
.
+
-
/
0
1
2
3
.
(
&
4
2
.
+
5
&
.
2
,
6
7
-
+
&
6
(
&
+
0
8
&
2
2
&
+
9
A
F
G
3
3
/
H
P
!
"
H
N
&
J
,
&
>
R
R
H
D
K
J
I
R
+
I
I
H
R
-
7
J
(
6
O
6
P
,
D
+
K
-
2
7
S
.
H
6
N
&
N
5
D
I
/
K
3
W
2
S
3
S
G
.
J
I
D
5
H
F
H
K
6
N
'
R
I
,
Y
X
K
V
(
I
)
H
L
'
;
P
.
,
3
P
0
T
&
Q
V
(
6
O
)
J
0
N
'
N
;
N
.
N
,
I
3
0
Q
F
E
M
)
;
Q
,
;
&
6
&
+
,
&
+
0
8
A
&
2
2
&
+
N
9
*
N
Z
E
J
[
^
+
A
[
]
*
)
<
+
E
*
T
@
N
N
E
H
B
J
M
L
=
>
"
C
H
J
@
N
<
H
N
#
A
K
$
I
J
U
H
L
>
#
J
#
<
T
I
J
"
B
N
(
N
H
!
<
!
9
7
H
+
#
6
H
C
&
,
#
;
B
,
2
,
.
6
&
.
5
E
:
B
^
a
k
`
b
k
u
v
"
`
`
w
u
x
$
c
d
e
s
@
`
e
f
g
^
f
\
=
h
^
>
i
i
j
?
i
i
E
k
j
e
J
c
i
g
l
c
j
f
h
e
`
m
n
j
`
e
l
_
g
e
c
f
_
^
o
b
i
p
_
`
e
c
q
y
z
{
{
z
|
}
~

€

{

‚
ƒ
„
y
z
{
s

‚
†

v
‚
‡
x

ˆ
‰
u
Š
z
t
|

‹
s
x
Œ
v
u
y
y
~

‘
’
¡
“
¢
{
‡
Ÿ
x
z
š
s
x
ª
‡
‚

s
„
ª
u
u
›
u
t
s
Ÿ
º
x
s
t
œ
Š
u
‚
s
À
®
t
I
I
‚
H
Ž
š
=
s
t
ž
t
>

Š

x
z
{

s

‚

º
s
‚
x
š
Ÿ
y
x
z
‚
š
s
Š

z
u
x
@
s
t
ž

{

{
Š
‚

s

H
Ÿ
‚

{
s
z
s

P

H
Š
{

=
y
s
Š
Š

>
x
u
‚
«
œ
s
¯
t
‚
s
{
u
x
{
«
¬

s
®
Š
u
s
s
y


‚
ž
u
{
s

{
s
Š
{
s
x
E
H
E

Š
s
z
t

s
Š

‚
Š

z
t
x
{
u
z
x
Š
‚
Š
x
Š
u
u
š
t
s
x
Š
{

s

Š
‚
{
‚
x
‚

x

{
t
{

s
ª
{
s

y
ž
t

z
Š
‚

«
y
z

t
t
{
z
x
y
ª
z
¼
u
Š
t
‡
s
‚
u
›
{
z
u

ª
t
Š
z
{
z
u
y
Š
œ
{
¾
{
Š
u
z
¬
t
Š
Ÿ
{
Œ

t
y
{


‡
‚
œ
Ÿ
t
z
z

s
z
z
‚
Ÿ
›
u


s
ž
ª
ž
‡
Š

‚
º
t
Š
t
‡
›
s
s
s

‚
œ
{

‚
‚
ž
Ÿ
‚
s
u
t
Ÿ
Š
‡
ž

z

s
s

š
‚
z
{
s

z
µ
t
‚
‚
‚


š
y
s
y
{
‡
Ÿ
¹
›
x
«
u
›

{
z

Ÿ
Š
u
œ
‚
z
{
Ÿ
ª
x
z
‚

w
u

{
{
›
{
z
t
u
{
{
t

z

u
Š
‚
s

{

z

y
{

u
u
{
‚
x

‚
ž
u
Š
{
Š
ª
u
‚
y
{

š
‚
{
s
{
z
t
Š
s
t
x
‚

ž
z
Š
{
Š
ª
š
{
‚

z
‚

s
s
t

Ÿ

t
»

‡
Š
t

z
¾
{
y
ª
{
y
‰
¬
z
u

u
{

‚

z
Š

y

z
u
¬
‡
Š
s
{
ž
š
{
›
t
‚
u
z

›
ž
{
{
Š
u
µ
x
u
Š
u
x
u
´
s

‡

Š

Š
{
ž
{


¯
y


Š
ª

u
{
‡
‚
Ÿ
‚
ª
s
t
z
u
u
z
u
Š
u
u
{
›
‚

x
t
s
y
Š

š
t
Š
z

y
z
s

x
{
x
{
ˆ
{
{
«
{

±
›
x

x
‡
š
s
‚
ž
‡
Š
ž
Š
s

Š
x
š
›
œ
u
z
š
‡
u
Ÿ
x
{
s
x


›


t

{
u
Š

{
z

z
Ÿ
‡
›
z
Š
z
y
Š
ª
s
‚
µ
u
z

š
š
x

‚
œ

¿
x
ž
ª

x

Œ

Š

z
Š
z
‚
{
ž
²
š
z

‡

ž
z
‚

t
z
{
Š
›
š
z
u
x
¬
{
{

x
u
x
{
x
x
‡
s


t

y
t
u

{

‡

Œ
{
Š
µ
Ÿ
y
‚
s
{
u
Ÿ

t
°
¶

š

š
³
u
t
Š
‚

¯

Š

{
t
ª
u
³
x
{

ª

‡
s
t
«
z
ª
y
{
t
²
‚
s
‚
t
{
š

ª
u
u
t
š



s
{
œ
{
«
›

t
y
Ÿ
{

u

t
Š
Š
z
±
t
u

ª

y
t

u
«
x
u
u
‚
{


t
x
y
y
Ÿ

u
«
‚
‡
y
‚
«

z
Š
x
ª
u


œ
y
‚
š



u
{
u
x
x
s
‚
ª
{
‚
x
‡
z
ž
{
{
{
s
z

t
y

t
{
«
s
y
›
u
Ÿ
Š
t
ª
{
t
ª
{

u
u
{

t
Š
‡
t
t

z

›
¯
‡
‚
ª
ª
š

œ
u
š
›
y
‡
t
s
‚
Ÿ
{
t
š
t
u
t
s
‚
s

{

s
u
t

ž
u
t
š
„
y

{

ž
µ
{
š
Š
s
‚
‡
ª
¸
‡
z
s
t
·
Ÿ
s
s
š
¯
‚
z
­
ž
u
x

u
s
Ÿ
ª

¬
t
u
{
ª
š
‚
s
u
z


s
ª
Š
Š
‡
‚

‡
y

¯
z
y
u

s
t
{
x
t

Ÿ

u
z

›
x
z
y
‚
x
Š
ž

y
x
Š
‡
{
u
t

s
{
y
‚
‚
t
ž

u
x
z
t

{
œ

Ÿ
‚
u
u
ª
ž
s
‚

{
z
ž
ª
‚
t
µ
›
t
{
z
y


›
s
œ
s
y
‡
u
Š
s
š
„
{

{

³

š
ž
{
ª

x
u
ž
š

‚
š
{
‚
t
s

s
{
s
ª
ª
t
{
ª
u

s

z
Š
š

Š
u
y

³
x
s

³

›
‡
{
z
{
u
ª
x
{
{
z
s
²
Š
{
u
‚
{
š
‚
{


s
x

±
‡

u
ª
t
Š
Ÿ

Ÿ

›
{
š

‚
›
s
Š
t
u
š
x
Š
Š
«
š
š
t
s
ª
s
z
z



{

‚
s
u
ª

y
Ÿ
Š
s
s

z
‚
Ÿ

w
Š
{
x
ª
y
{
t


z
Š

x
u
z


t
Š
Š
t
Š
y
x
Ÿ
x
u
u
‚
¬
Š
‚
t
Ÿ
µ
y

{
‚

y
‚
´
x
„
Ÿ
t

u
®

u
³
Š
{
¬
t
ž

‡
t
x
‚
u



u
u
{
{
Œ
s
­
³
Š
Š

Š
t
u

š
¬
ª
z
s
{
s
z
u
{
{
t
Š
{
{
u
z
š
Š
s
{
²
z

‚
u
Š
œ
t

›
{
{
š
t

x
t
s

Ÿ
›
{
{

s
u
t

‚
‚
u
s
{

z
{
Š

Š

s
Š
s
{
‚
y
°
y
s
x
š
x
x
{
š
{
Š
š
‚

s
s
t
Š
t



z
œ
u
u
‚
u
x
y
‚
‚
t

Š
­
{
P
š
š
s
š
z
ž
z
¾
s
‚
Œ
t
¬
{
x
¬

y
u
{

z
Š
²
µ
t
ª
t
Å
u
Ÿ
t
‚
s
‚
Ÿ
Š

z
u
Ÿ
u
x
‚
‚
‚
„
z
Š
z
z
u
Ÿ


x
t
z
š
y
ž

Š
s
z

›
¯
‚
‡
ª
ª
s

t
{
¬
¦
z
Š
s
‚
Ÿ
H


‚
y
u
¬
Š
s
‚
Ÿ


¥
‚
‚
{
‚
{
´
¶

u
t
s
Š
Ÿ
z
u

{
µ
‚
z



{

›
y
t
Š
¹
º
ª
Š
z
Â
{

ž
š
Ÿ
Š
{
s
u
Ã
¸
‡

›
z
£
¬
¬
¿
I
‚

½
¹
{
š
Á
¸
y

Ÿ
‰
{
Œ
«

œ

š
Š

z
½

u
y
z
ª
s
Š
s
­
‚
u
�