Identifying highly talented athletes: Conception and design of an expert system

Table of Contents

1 Introduction

2 Track and Field Sprint
2.1 History, Overview and Classification
2.2 The Skill of Running
2.2.1 Stride Length and Stride Frequency
2.2.2 The Race Sections
2.2.3 The Start and Acceleration Phase
2.2.4 Running at Maximum Velocity
2.3 Performance Structure
2.3.1 Emotional and Mental Attributes
2.3.2 Physical Fitness
2.3.2.1 Strength
2.3.2.2 Endurance
2.3.2.3 Flexibility
2.3.2.4 Speed
2.3.3 Technique and Coordinative Abilities
2.3.4 Other Factors
2.3.4.1 Physique
2.3.4.2 Tactical Abilities

3 The Search for Talent
3.1 The Elements of Talent
3.1.1 Distinctive Performance
3.1.2 Rate of Progression
3.1.3 Utilization
3.1.4 Physical Tolerance
3.1.5 Conclusion and Consequences
3.2 Successful Talent Identification Schemes
3.3 Talent Search with the Help of an Expert System
3.3.1 Goal
3.3.2 Limitations
3.3.3 Testing Procedures
3.3.3.1 Vertical Jump
3.3.3.2 40m Sprint
3.3.3.3 50m Bounding
3.3.4 Evaluation and Selection

4 Introduction to Expert Systems
4.1 Artificial Intelligence and the Emergence of Expert Systems
4.2 What are Expert Systems?
4.3 How do Expert Systems Work?
4.4 Expert Systems in Sports
4.4.1 RunCoach
4.4.2 Expert System for Tactical Player Positioning in Soccer
4.4.3 TESSY

5 An Expert System for Talent Identification
5.1 Phase 1 – Requirements Analysis and Planning
5.1.1 Project Management
5.1.2 Requirements Definition
5.1.3 Requirements Specification
5.2 Phase 2 – System Components Definition
5.2.1 Knowledge Acquisition
5.2.2 Evaluation of the Vertical Jump
5.2.3 Evaluation of the 40m Sprint
5.2.4 Evaluation of the 50m Bounding
5.2.5 Diagnoses
5.3 Phase 3 – Design
5.3.1 Prototyping the Shell
5.3.2 Introduction to d3web
5.3.2.1 The Back-End
5.3.2.2 The Front-End
5.4 Phase 4 – Implementation
5.4.1 Setting Up the Knowledge Base
5.4.1.1 Diagnoses Set
5.4.1.2 Question Hierarchy
5.4.2 The Knowledge Implementation
5.4.2.1 The Rule Editor
5.4.2.2 Heuristic Decision Table
5.4.3 The User Interface
5.5 Phase 5 – Deployment and Maintenance

6 Summary

7 Conclusion and Outlook

References

Index of Figures

Index of Tables

Appendix A – The Rule Base

Appendix B – Abstract (English)

Appendix C – Abstract (Deutsch)

1 Introduction

Modern track and field events feature all the basic forms of human locomotion, from walking to running and jumping to throwing. Since the beginning of mankind, the individual's skills in these abilities were highly valued, at first purely for the sake of survival, but soon an interest in less perilous competitive comparisons arose. The modern Olympic Games, continuing a tradition that began in ancient Greece almost 3000 years ago, are one of the most popular and widely received sporting events today, featuring more than 300 different events in 28 sports^[1].

In many of the core sports, including track and field, the level of performance is world-class, and every athlete dreams of participating in these games and winning a gold medal. The global interest in these games and in high-performance sport has resulted in ever improving athletes and a rapid rise of performance standard. No longer can just anybody with diligence and determination train by himself. Today it takes the perfect combination of several endogenous as well as exogenous factors to reach the highest level.

One of these factors is talent. Talent is commonly regarded as an above average ability for a specific task or the potential to perform at a high level in a certain domain. In a domain that is solely focused on ever improving feats of strength, speed and endurance, only particularly suited competitors stand a chance to reach these elusive heights.

It is therefore in the interest of the athletes, their families, coaches, associations and general sports spectators that methods to determine the most talented athletes at an early stage are devised. This endeavor, which is commonly known under such names as talent search, talent identification or talent selection, is however much more difficult than is generally acknowledged, as sport talent in young athletes is comprised of more than just the ability to perform better than one's current peers.

This thesis details the possibilities and limitations connected to the use of an Expert System for the identification of young athletes who have the potential to become outstanding track and field sprinters.

Expert Systems, which originated from the field of Artificial Intelligence, are a kind of software that can solve complex problems by imitating human reasoning capabilities and drawing upon the knowledge of domain experts. This ability stems from the unique structure of Expert Systems, whereby the knowledge, which is usually represented in the form of rules, is strictly separated from the part of the program that interprets them. This technology has been successfully implemented in various high complexity domains such as medicine and economics for more than 40 years, but it has been found that the area of sports has so far largely been ignored. This assessment is supported by leading German researchers in sports and informatics, who note that concepts such as data-mining seem to have taken over the areas in which this technology could potentially come to use (J. Perl^[2], personal communication, January 17, 2007).

In spite of this apparent scarcity, Expert Systems and sports can form a potent combination. A difficult task such as the efficient and reliable search for talent combines all the characteristics of a problem that cannot be satisfactorily solved by conventional software.

The goal of this thesis is to develop a concept for the development of an Expert System that supports the person in charge of talent identification. It is examined how the development of such a system could be undertaken, and a first functional prototype is created. Based on discussions with domain experts and literature research, a test-based sprint talent assessment is devised and a simple knowledge base created with an Expert System shell called d3web. The acquisition of the domain knowledge, its subsequent transformation to the knowledge base specific syntax and its implementation in the system are described in-depth.

Through this combination of theoretical deliberation and practical realization, the author intends to highlight the possibilities that the use of such non-conventional computer technologies can offer the field of sport, as well as the difficulties and limitations of using such a system. Recognizing “both sides of the coin” is important if Expert Systems are to be successfully and purposefully realized on a broader scale to the benefit of researchers and practitioners alike.

The author hopes that this thesis can play a part in bringing these fields closer together. There is no doubt that many complex problems pertaining to the field of sport exist which could benefit from such collaborations. Talent search, the subject of the present work, is certainly one of these fields.

This thesis is comprised of two parts. The first part, ranging from chapter 2 through to chapter 4, gives a detailed account of the necessary background knowledge, and is compartmentalized in the following way:

Chapter two introduces the reader to track and field sprint. A brief overview of the history is followed by an in-depth analysis of the performance structure of the sprint. It is explained how mental, physical and technical abilities affect the performance. Special emphasis is placed on the major physical attributes of strength, endurance, speed and flexibility.

The chapter further focuses on the sprinting technique. The two velocity determining factors, stride length and stride frequency, are introduced and it is analyzed how the optimal technique is devised to maximize the sprinter's velocity.

Chapter three details the talent issue. The author first attempts to define the term, then outline the difficulties encountered throughout the identification process. As the elements of talent are subsequently discussed, it becomes clear that physical ability evaluations, as they are routinely performed, are often less than adequate. Nonetheless, an overview of three real-world talent identification schemes shows that this approach, while far from perfect, can often lead to the desired results.

The chapter is rounded off by an overview of the talent identification approach that was chosen for this work. According to common practice, a test-based assessment is conducted, and the results are afterwards evaluated by the Expert System. The goal and the limitations of this approach are outlined, and the chapter is concluded by the presentation of the chosen physical tests.

In chapter four, a closer look at the Expert Systems is provided. A brief account of the development of this technology is followed by an analysis of how these systems function and what characteristics they feature.

To substantiate the previous claim that Expert Systems can be effectively used in the field of sports, three such systems are presented in the final section.

The findings of these chapters come together in the fifth chapter for the conception of the Expert System. According to standard software engineering practice, the development procedure is divided into distinct stages that determine the required activities and guide the process from the idea to the deployment of the finished product. These phases are introduced, and the tasks with which the software engineers are faced are described.

The first phase entails the acquisition and compilation of the product requirements. Furthermore an estimation of the time and effort necessary for the completion of the project is performed.

In the second phase, the acquisition of the domain specific knowledge is discussed. The tests that were previously described serve the purpose of determining certain physical attributes, and it is herein analyzed how these test results are evaluated.

The third phase is characterized by decisions pertaining to the design of the system structure. The shell kit used for this work is introduced, and relevant features and attributes are highlighted.

The conception of the Expert System is concluded by the implementation of the knowledge from the second stage in the previously presented shell kit. It is detailed how the system guides the user through a series of questions according to its knowledge to derive a solution for the given problem.

2 Track and Field Sprint

This chapter introduces the reader to track and field, in particular short distance sprint. In the first section, the author briefly describes how track and field athletics developed. The role of the sprint within these developments is presented and the events that are classified as sprint races are introduced.

In the second section, the optimal running technique is portrayed. Within this discourse, it is highlighted which attributes affect the sprint performance. The two most important phases of every sprint, the starting phase and the maximum velocity phase, are also presented.

In the final section, it is elaborated which mental, physical and technical attributes constitute prerequisites for eminent performances. It is obvious that such factors must be well-known to identify talented athletes and the knowledge of this underlying structure of sprinting is also vital to understanding the principal rules of the Expert System.

2.1 History, Overview and Classification

Track and field athletics dates back thousands of years. Running, jumping and throwing objects, the basic forms of human motion and the foundation of most athletics disciplines, are as old as man himself. This chapter focuses on the running events, in particular the sprint events.

Currently, the Olympic track and field events roster comprises of 25 different disciplines. In the first Olympic Games (776 BC) however, there was only one contest: the stade, a foot race around the length of the stadium, a distance of approximately 192m (cf. Zech, 1971, p. 304).

Soon the Olympic Games and the celebration of athletic achievements gained immense popularity all over Greece. By the year 500 BC, they had become the central event of ancient Greece, a position they held for more than 800 years! However, after the games were banned by the Roman emperor Theodosius in the 4th century AD, it took almost one and a half millennia for organized athletic competitions to return!

In the middle ages, physical culture and games differed greatly from what is considered as sports today. Historic chronicles from those times tell of various physical exercises and drills performed by men (and sometimes women), but these were commonly practiced by soldiers and knights as a preparation for warfare (cf. Meier, 2005, p. 7).

The return of organized athletics came in the 19th century when track and field resurfaced in Great Britain as well as in the United States. In England, schools competed in various athletic fields, while in America athletic competitions also re-emerged for a different purpose: prize money. The public interest for such professional racing declined in the second half of the 19th century, but the people's enthusiasm for contests of speed, strength, agility and stamina had been aroused and was transferred to school and amateur athletic club competitions (cf. Aaseng, 2002, p. 14). Clubs and colleges all over Great Britain and the United States began organizing athletic meets, and in 1895 the first international competition was held at Travers Island in New York. In Germany, the first public athletics competition was held in 1880 in Hamburg (cf. Bauersfeld & Schröter, 1992, p. 11).

At the same time, around the end of the 19th century, Baron Pierre de Coubertin was drawing up a sporting competition that would propel track and field onto the world stage: the modern Olympic Games. Inspired by collegiate sports in England, de Coubertin searched for a “way to celebrate the importance of exercise in the education of youth” (Aaseng, 2002, p. 15). He proposed the revival of the ancient Greek games in 1892, and four years later the first modern Olympic Games were held in Athens, Greece. And so, since the first games of 1896, athletics have been an integral part of every Olympics and continues to be one of the most popular events of the Olympic summer games program.

As stated above, there are 25 track and field events currently contested at the Olympic Games, of which seven can be classified as short distance running events:

- 100 meters
- 200 meters
- 400 meters
- 110-meter hurdles (100 meter for women)
- 400-meter hurdles
- 4 x 100 meter relay
- 4 x 400 meter relay

Evidently, short distance sprints are defined as races up to and including a distance of 400 meters (cf. Bauersfeld & Schröter, 1992, p. 107). This is based on the notion that races of up to 400 meters mainly utilize the alactic and the lactic anaerobic energy pathway, since these pathways are “called upon principally by athletes whose sports demand high energy expenditure for up to approximately 60 seconds” (Dick, 1989a, p. 84). Furthermore, these races, with the possible exception of the 400 meters, can be run by elite athletes at full speed and do not require particular race strategies (cf. section 2.3.4.2). Short distance races, or sprints as they are commonly referred to, therefore display a distinct performance structure and require athletes with very unique attributes, or in other words, talents.

Since the first modern Olympic Games more than a century ago, the race times have been lowered in every event, often drastically. At the Olympic Games in Stockholm in 1912, the winner of the 100m race finished in a time of 10.6 seconds. Currently, the world record (men) stands at 9.77 seconds, an improvement of 0.83 seconds or about 8%. Similar progressions can be found in the 200m and 400m race times, while the improvements in the relay events and in women's sprinting have been even greater.

Many factors are accountable for this progression, like increased and more efficient training, better equipment and facilities (e.g. the introduction of tartan running tracks) or favorable conditions (e.g. high altitude at the Olympic Games in 1968), but also a more systematic and structured approach to finding and training the 'right' athletes.

The previous account served to illustrate the status of the sprint events within athletics and highlighted characteristics and developments.

The next section presents the running technique. It is assessed how the two performance determining variables, stride length and stride frequency, are closely connected. This is followed by the division of the sprint race into multiple phases, of which the two most important, the starting and the maximum velocity phase, are detailed in depth.

Based on this discourse, the last sections focus on the factors that ascertain sprint performances. Identifying these factors and their relationships is absolutely essential to specifying training goals and methods and to determining whether an athlete is suitable (in other words, talented) for sprinting. As such, the following sections must be regarded as the core of this thesis.

2.2 The Skill of Running

The goal of sprinting is actually very simple: to run a given distance in the shortest possible time. Most people associate outstanding physical attributes like speed and strength with this task, and it is easily forgotten that sprinting is also a highly complex sport which requires technically gifted athletes. Only with the optimal running technique can the outstanding sprinter fully utilize his physical attributes and achieve velocities of 11 m/s and more.

The following elaborates on the different phases a sprint race consists of, and describes the technique and its most characteristic traits. This is preceded by a presentation of the two most important variables in running: stride length and stride frequency.

2.2.1 Stride Length and Stride Frequency

The goal of every sprinter is to run as fast as possible, this means the greatest possible velocity has to be achieved. Physically, velocity (v) is expressed as the distance (s) covered in a certain time (t).

illustration not visible in this excerpt

In running, the distance is necessarily covered in a number of strides of a certain length (l). The time required is inversely proportional to the frequency of these strides (f). Hence the velocity of a runner can be expressed as the length of each stride multiplied by the frequency (cf. Ballreich & Gabel, 1975, p. 346).

illustration not visible in this excerpt

Consequently, the race time can be improved by increasing the stride length, stride frequency or both. Ballreich and Gabel (1975) conducted a study to determine the magnitude of influence of these two variables. They found that the majority of their subjects realized higher velocity gains by increasing the stride frequency, not the length.

This is however only possible to a certain, individual degree, and a further increase of one variable inevitably leads to the reduction of the other, and a resultant lesser overall velocity (cf. Mann, 1999, p. 27). As such, the runner and the coach must aim to develop the optimal relation of length and frequency for the particular case, as someone with very powerful legs may be able to reach greater stride lengths, but can perhaps not keep up with the frequency set by another athlete with stronger coordinative abilities, even though both of them achieve the same race times.

A study conducted at the Olympic Games in 1988 (cf. Jonath et al., 1995, p. 86) showed that Carl Lewis achieved a maximum stride length of 2.65 meters and a frequency of 4.84 strides a second! The women's 100m champion that year, Florence Griffith-Joyner, recorded similarly impressive rates (2.4 meters stride length and a frequency of 4.68 strides per second) that were only slightly below those of her male counterparts.

Now it must be examined how each stride needs to be executed in order to optimize stride length, rate and subsequently velocity.

2.2.2 The Race Sections

Every race can be divided into different segments according to various characteristics, of which the most important one is undoubtedly velocity. In the past, experts used to separate a 100 meter race into four phases (cf. Jonath et al., 1995, p. 61):

- the reaction phase, including the start and the first couple of steps
- the positive acceleration, until the achievement of maximum velocity
- the steady velocity phase, in which the maximum speed could be sustained
- the deceleration phase, in which the speed inevitably decreased

Bauersfeld and Schröter (cf. 1992, p. 112) opt for a similar, four part segmentation, but express the different velocities of each segment through an overview of split times for each phase. They divide a typical 100 meter race into start (0-30m), acceleration (30-60m), top speed (60-80m) and forced deceleration (80-100m), and specify targeted split times for each segment. Such split times are listed in so-called pace tables (cf. Joch, 1996, p. 104).

More recent findings however indicate that a short sprint race can be broken up into more than just four stages:

Tab. 1: The key phases of a 100m race (cf. Mouchbahani, 2002, p. 1)

illustration not visible in this excerpt

As one can see, the four core phases as identified above have not been changed: rather they have been complemented by the addition of transitional phases (in bold). The integration of the start and the acceleration phase into a single segment is also no longer regarded as accurate enough, and has therefore been split up. This had earlier already been noticed by Dick (cf. 1989b, p. 3475), who advocates a similar five-phase separation.

A short overview of the characteristics of each phase will highlight the reasons for this division:

Phase 1 is determined by the quality of the starting technique (crouch start, cf. section 2.2.3) and the quick planting of the first step.

The first transitional phase, titled phase 1.1, is in great part determined by the (elastic) strength of the athlete. Fast, powerful strides with increasing length serve to set up the acceleration phase.

The acceleration itself begins after the first 10 meters. In this phase, the runner gradually straightens up and brings the pelvis in the correct position for an optimal stride length. Every stride must be longer than the preceding one.

The next transitional phase is called pick-up. The central nervous system (coordination) replaces strength as the major determining factor. The runner is now upright and attempts to reduce the ground contact times to a minimum (high frequency) while maintaining the optimal stride length. The runner approaches maximum velocity, with top international sprinters reaching their greatest velocity between 50 to 60 meters, while women and weaker runners hit top speed sooner, around the 40m mark (cf. Mouchbahani & Seagrave, 2006, p. 32).

Phase 3 is characterized by the frequency orientated upright posture and the active, supporting arm motion. The perfect coordination between all elements is required.

Now the athlete must attempt to sustain the speed for the remainder of the race (phase 4 and 5). Mouchbahani (cf. 2002, p. 3) remarks that the stride length can not be kept up and the sprinter must instead focus on securing the rhythm to avoid overly losses of velocity.

This overview served as a summary of the changed conceptions in recent years regarding the segmentation of sprint races. It was kept as brief as possible as it does not directly affect the subject of this thesis, which is identifying athletes who have the potential to become good sprinters. As such, in the next two sections, the starting technique will be described together with the acceleration phase, followed by the running technique at maximum speed. Key elements of these two phases are highlighted, as this knowledge is necessary in order to determine deficits of individual athletes.

2.2.3 The Start and Acceleration Phase

At the beginning of a race, the athlete is in a stationary position and must, upon hearing the signal, leave the starting blocks and accelerate to the fastest possible speed in the shortest amount of time. This task is facilitated by the use of starting blocks and the technique known as crouch start. The crouch start, which is compulsory for all races of up to 400m, can be divided into four phases.

The first phase begins with the command “on your marks” and ends on “get set”. As the first command is given, the athlete steps out from behind the starting block and into it from the front, assuming the crouched starting position. The purpose of this position is to create favorable conditions for a strong, predominantly horizontal push off during the start. The body posture is largely determined by the setup of the starting blocks and the distance between the front and rear block. Among elite sprinters, the medium-long setup has been established as the most advantageous starting block position for a quick and powerful push off. In this setup, the front block (against which the sprinter places his stronger and more agile leg) is about 1 ¾ to 2 foot lengths and the rear block about 3 to 3 ½ foot lengths behind the starting line (cf. Bauersfeld & Schröter, 1992, p. 132), with the distance between the two blocks measuring about 35 to 42 centimeters (cf. Jonath et al., 1995, p. 21). A study by Schot and Knutzen (cf. 1992, p. 138) showed that the optimal distance between the starting line and the front foot is about 60% of the athlete's leg length.

The knee of the rear leg rests on the ground, and the athlete strives to relax the body as much as possible. The weight is distributed equally on the hands and feet, the head is in line with the upper body and the eyes are fixed on the ground in front of the line. The IAAF Coaches Education & Certification System textbook also details aspects of the optimal starting technique. It is stated that the medium starting block setup is favorable to the other variations, as the bunched setup results in a “pronounced forward body lean [and] short, almost stumbling strides” (Mouchbahani & Seagrave, 2006, p. 4), while the elongated setup leads to the straightening of the upper body too soon after the start. Both instances yield a lower initial acceleration than the medium setup.

When all athletes have assumed the starting position, the starter gives the “set” command. This “set” position signals the beginning of the second phase of the crouch start, which ends on the “go” command or the firing of the gun. The set posture serves the following purposes:

- to shift the weight of the body and the center of gravity forward, thereby assuming a better stance for a horizontal push off,
- to assume a position at an optimal knee angle for a powerful leg extension,
- and to generate an initial tension in the leg muscles to facilitate the fastest possible push off (cf. Bauersfeld & Schröter, 1992, p. 131).

From the starting posture, the athlete raises his hips slightly above the level of the shoulders to assume the set position. In this stance, the knee angle of the front leg is approximately 90° and the back leg has an angle of about 110-120°. The arms remain fully extended and the head stays in prolongation of the upper body, the sight trained on the starting line. Schot and Knutzen (cf. 1992, p. 146) found that an arm position perpendicular to the ground is preferable to a forward leaning posture with a ground to arm angle of about 80°. This position allows for a push off in the optimal angle of 45°. Although the characteristics of the set position are strictly defined, Jonath et al. (cf. 1995, p. 25) stress the importance of adjusting the posture to the individual physiological attributes. They also note that in the end, the crucial aspect is the athlete's comfort, and a posture that is biomechanically optimized but does not feel comfortable for the athlete will seldom result in great race times.

The set posture is succeeded by the actual starting motion, which per definition (cf. Bauersfeld & Schröter, 1992, p. 131) begins at the “go” command and ends when the front foot has left the blocks. The aim of the starting action is the greatest possible acceleration in the shortest possible time of the thus far stationary body through an explosive extension of the legs in the optimal (approx. 45°) direction. According to Jonath et al. (cf. 1995, p. 24, fig. 3), the rear block registers a strong but relatively short burst of force at the beginning of the push off, with the front leg contributing the main impulse (⅔ of the overall force), which lasts longer but does not reach the peak values of the rear leg push off, a split second later. This is contrasted by dynamographic studies by Bauersfeld and Schröter (cf. 1992, p. 133, fig. 31), which show a simultaneous begin of force exertion on both blocks.

The rear leg is the first to leave the blocks and does so before becoming fully extended. As the knee of the rear leg passes the anterior block, the front leg begins to extend explosively and fully. The upper body remains tilted forward during the entire starting action and only straightens well into the acceleration phase. This results in a lower center of gravity and a greater horizontal propelling impulse (cf. Schot & Knutzen, 1992, p. 146). The runner's eyes remain trained on the ground in front of him in order to prevent a premature straightening up of the body. As the foot leaves the front block, the sprinter has reached the typical posture as illustrated in figure 1.

illustration not visible in this excerpt

Fig. 1: Body posture at the end of the starting action (cf. Jonath et al., 1995, p. 26)

The starting motion is concluded when the foot that was resting on the posterior block strikes the ground for the first time. In order to maintain a high acceleration during the start, this ground contact shall occur quickly after the push off (as the body can only be accelerated on ground contact, not while in the air), and the emphasis thus lies on a quick rather than a long first stride. The early footstrike also leads to the foot touching the ground behind the vertical projection of the runner's center of gravity, thus avoiding forces that would counteract the acceleration.

In the following transition, acceleration and pick-up phase and the elite sprinter reaches about 92-95% of his maximum velocity. These first 30 meters are good indicator of the overall time, whereby the rule of thumb is that a time of four seconds for 30m will yield an overall result of 10.4 seconds for the 100 meters (cf. Joch, 1996, p. 104). The velocity must be continuously increased with every stride, which is accomplished by maintaining a forward tilted posture for as long as possible and only gradually straightening the upper body, as the horizontal component of force applied to the ground from this posture is greater. It is further desirable to execute a high number of smaller strides instead of a few long strides (high frequency) to avoid falling into a skipping-like motion.

Bauersfeld and Schröter (cf. 1992, p. 135) note that in this part of a race, the run is more frequency accentuated, and the stride length only gradually increases. The explosive leg push is the major impulse generator, whereas at maximum velocity the “forward extension of the hips ... is the real key to good sprinting” (Mouchbahani & Seagrave, 2006, p. 22). This will be discussed in more detail in the following section.

2.2.4 Running at Maximum Velocity

Although walking and running constitute the most basic forms of human locomotion, the perfect execution of the athletic sprinting technique requires an extremely well developed coordination of the individual elements and a high degree of technical expertise. The quest for maximum speed necessitates the total mobilization of one's physical attributes and the extraordinarily high frequency of limb movements make sprinting one of the most challenging sports.

As the sprinter reaches his maximum traveling speed, his body posture is upright and the strides have reached their optimal length and rate. Each stride displays certain characteristics that permit a breakdown into four distinct phase, which shall be discussed in the following.

illustration not visible in this excerpt

Fig. 2: The stride cycle at maximum velocity (cf. Bauersfeld & Schröter, 1992, p. 121)

A stride can be divided into support or ground phases and flight or swing phases, and depending on whether the leg is in front of or behind the center of gravity as either rear or front. The four distinct phases are thus the rear flight phase (figure 2/8-2/11), hereafter referred to as recovery phase, followed by the front flight phase (figure 2/1-2/4) – also called transition or ground preparation phase –, the front support phase (figure 2/5-2/6) and the rear support phase (figure 2/6-2/7), before beginning anew. The support phases play the major role in the generation of propulsion force, as the body can only be accelerated when the foot is in contact with the ground, but the flight phases are equally important as the “quality of each phase of running is determined by the quality of the phase that immediately precedes it” (Mouchbahani, 2006, p. 2).

The recovery phase (the left leg in figure 2/8-2/11 and the right leg in figure 2/3-2/6) begins when the foot takes off from the ground and ends when the leg travels past the center of gravity (CG) underneath the body, also called mid stance. During this phase, the muscle has time to relax while the flexion of the leg stretches the quadriceps in preparation for the next step. To ensure a sufficiently long period of time for the recovery of the muscle, the leg swings relatively far backward (figure 2/4) and the lower leg is brought close to the upper leg (almost touching the gluteus) during the transition phase. This shortens the pendulum and secures a high angular speed. The flexion also serves to ensure a high knee lift during the front flight phase.

The following phase begins as the leg swings through the mid stance (figure 2/6 and 2/11) and ends on ground contact. The characteristic trait of this phase is the knee lift, whose purpose is to guarantee an optimal stride length. As the knee is raised until it is approximately 15º below the horizontal line, the lower leg swings forward (figure 2/2-2/3 and 2/8-2/9), and the almost fully extended leg then drives into the ground in a backward and downward directed motion, triggered by an explosive action of the hip extensors. A slight forward shift of the respective side of the pelvis ensures a good stride length. The position of the pelvis plays a decisive role in the quality of the ground preparation phase and touchdown, because a forward titled (figure 3, right) posture handicaps a far reaching forward stride and forces the sprinter to plant the foot too early, limiting the speed of the limb.

illustration not visible in this excerpt

Fig. 3: The posture of the pelvis and the resultant recovery phase (cf. Mouchbahani & Seagrave, 2006, p. 22).

The forward swing is completed by the active, forceful planting of the foot (downward and backward motion). This is assisted by flexing the foot before the touchdown, thereby tightening the foot and calf muscles and creating an optimal state for an active footstrike through the extension of the foot in a “claw like action” (Dick, 1989b, p. 3477).

The front support phase (or amortization phase) begins on ground contact and ends at mid stance. Here, the runner must first of all absorb the shock of the landing while reducing the momentum as little as possible. The front support phase is inevitably a phase where the forward momentum is reduced, so the runner must ensure to keep it as short as possible and overcome it effectively. This is partly accomplished by planting the foot as close to the vertical projection of the CG as possible (cf. Bauersfeld & Schröter, 1992, p. 118), thereby ensuring that the hip moves above the point of ground contact rapidly and beginning the rear support phase.

Much of the success of keeping this phase short also depends on the preparation of the landing in the previous phase. A forceful, whip-like downward and backward motion of the leg prior to and immediately after the landing will help to move the CG – namely the hips – above the point of ground contact quickly. High tension in the leg muscles prior to the landing, followed by bending the leg to not more than 160º on impact, also ensure that the front support phase can be overcome quickly. Ground contact is made with the ball of the foot only and then the heel is lowered slightly towards the ground to absorb the force of the landing. The calf muscles are thereby already contracted, which intensifies the force of the push off in the rear support phase.

The last and arguably most important phase is the rear support phase (figure 2/6-2/7). The actions herein determine the magnitude and the direction of the impulse generated by the footstrike and, as such, are the direct and only positive influence on the speed of motion. The athlete aims to generate the greatest amount of horizontal propulsion force whilst keeping the ground contact time to a minimum. The emphasis on horizontally directed force is of utmost importance, as the forces generated on ground contact inevitably also contain a vertical component. While the explosive extension of the leg is the major driving force during the start and acceleration, the phase of maximum velocity is largely determined by the rapid extension of the hip, resulting in a kind of pulling motion of the almost fully straightened leg during the support phase (cf. Jonath et al., 1995, p. 67). An overly strong leg instead of hip extension in this phase will result in the run becoming a series of small jumps instead of forward strides, whose downward force has to be cushioned and compensated in the following front support phase, which leads to longer ground contact times and an overall reduction of speed. The phase ends on take-off.

The forward momentum during the entire run is further enhanced by the forces generated through the purposeful, straight-forward actions of the arms. The elbow angle is kept at approximately 90º. As the arms beat forward and reach the turn around point (hand at shoulder level), they are quickly decelerated leading to a transfer of momentum to the upper body (cf. Bauersfeld & Schröter, 1992, p. 117; Dick, 1989b, p. 3477).

As can be deduced from figure 2 as well as the above description, the recovery phase of the left leg largely coincides with front support phase of the right leg, and vice versa. The same naturally applies to the rear support phase and the front flight phase.

2.3 Performance Structure

The previous sections presented the technical elements that determine a sprinter's performance. It was addressed how stride length and stride frequency form the basic variables of running speed, and the optimal technique was presented. This knowledge makes it possible to deduce which physical or coordinative attributes the athlete needs to possess in order to fully utilize the technique and maximize his velocity.

A sporting performance is the result of a great number of factors, and the entirety of these factors and the relationships between them is called performance structure, or Leistungsstruktur^[3] in German. The performance structure of a sport defines the goals and the methods of training, because only with the knowledge of the determinants of a high-level achievement can purposeful and goal-oriented training procedures be devised. Whereas in the past training concepts were based on the experience of the coaches and methods were mainly devised by imitating the training plans of world-class athletes (cf. Hohmann, Lames & Letzelter, 2007, p. 11), modern exercise science incorporates insights from various different base sciences like biology and psychology. Such collaborations have yielded great understanding of how the human body functions, how techniques can be optimized and performances improved. Exercise science is therefore a highly interdisciplinary venture that strives towards supporting and substantiating practical training methods with scientific facts. This quest has led to the development of various conceptual models of sports performance.

Models are theoretical constructs which represent specific aspects of the real world. The complexity of reality is reduced to the core variables (factors) and relationships involved in accomplishing certain goals: in this case, high-level sports performances. This idealized reduction enables the user (of the model) to determine connections and dependencies between different factors that would be too difficult to observe within the actual, full complexity situation. The development of such models is regarded as one of the central tasks of exercise science by Hohmann, Lames and Letzelter, and various models are presented in their book (cf. 2007, pp. 41- 49). They differentiate between two kinds of models: those with a performance criterion^[4] and those without.

Models without a performance criterion restrict themselves to naming the various components which influence a performance, whereas models with a performance criterion attempt to explicitly determine the magnitude of each component's influence on the result.

The performance model by Bauersfeld and Schröter (cf. 1992, p. 25) belongs to the former category and is aligned towards elite performance. Stating the goal is as important for the correctness of the model as the identification of the factors, as two performance models can have completely different components depending on the nature of the goal (e.g. elite performance vs. weight-loss training). The model entails both internal and external factors.

External factors are defined as those that are not tied to the athlete directly, but nevertheless affect the performance more or less considerably, such as weather conditions, the state of the facilities and training equipment as well as the opponents and spectators. These factors concern all competitors alike and thus only play a minor role in the search for talents. The external factors are represented as rectangles in figure 4.

illustration not visible in this excerpt

Fig. 4: The general performance model (cf. Bauersfeld & Schröter, 1992, p. 25)

Another external factor that is not represented in this model and whose influence, according to Treutlein and Stork (cf. 1976, p. 416), is more often than not greatly underestimated is the social factor. This entails all aspects from social class to financial security and sporting background of the athlete and his family, and often determines whether an athlete will at all have access to talent search programs, let alone be identified as one. However, as the focal point of this work is the identification of talented athletes based on their performance, this factor is out of the control of the coach or talent scout whose task is solely the assessment of a person's potential based on his attributes and skills and can not be considered in further detail. As such, the interest of the coach, and therefore the Expert System, must rest on the internal factors.

The internal or personal factors play a vital role. They are directly connected to the athlete and vary from person to person. Bauersfeld and Schröter (cf. 1992, p. 25) divide them into five areas: physical fitness, physique, psychological characteristics as well as technique and coordinative and tactical abilities, whereby each category consists of further subcategories. These will be extensively examined in the following sections.

The model in figure 4 has a very general quality and can be applied to most sports. Models of the second category on the other hand, those with a performance criterion, are more specific to a certain sport or performance. One such model is the so-called deduction chain. Here, the performance is first divided into several sub-criteria, which are then further broken up into attributes that determine them. These models are commonly used in the field of biomechanics to simulate complex performances. Such a deduction chain model for the long jump performance is presented in figure 5.

illustration not visible in this excerpt

Fig. 5: The deduction chain model for the long jump performance (cf. Hohmann, Lames & Letzelter, 2007, p. 46, translation by author)

The figure above exemplifies how the deduction chain model works from the top (the actual performance) to the bottom to identify the relevant factors. The long jump distance for example can be divided into three partial distances, of which the second one is determined by the horizontal and vertical velocity of the jumper as well as the difference between the CG level at take-off and landing. The other partial distances are similarly influenced by other factors that are not illustrated. An extensive study of how deduction chain models can be used to analyze the role of different variables for the high jump was conducted by Killing (1999).

For the sprint, the performance can likewise be broken up into partial factors. It was explained in section 2.2.1 how the velocity is determined by the two variables stride length and stride frequency. According to the governing logic of the deduction chain model, these variables must now be further divided until the foundations of velocity are discovered. That is the case when (mainly physical) attributes that can be trained and influenced are found.

This means it is necessary to identify the attributes that determine the stride length on the one side, and those which affect the stride frequency on the other side. The length of each stride is determined by two physical factors, the athlete's strength and flexibility, whereas the frequency is regulated by his ability to activate and relax his muscles rapid succession.

In accordance with the performance model by Bauersfeld and Schröter (figure 4), these attributes will be presented in the succeeding sections. Even though this model does not have a performance criterion like the deduction chain model, it is obvious that if they correctly reflect the real-world performance they must yield similar conclusions about the variables, since both attempt to model the same subject.

The great advantage of models and the main reason for using them, which is the reduction of the real-world complexity, can however also be their greatest drawback and various models exhibit different advantages and disadvantages.

A performance model such as the one by Bauersfeld and Schröter (figure 4) does not reflect the magnitude of each factor, and the order in which the elements are placed around the goal (the center) seems to be random. Models which try to rectify this issue, e.g. by Parsons (cf. Hohmann, Lames & Letzelter, 2007, p. 43), easily become too complex or abstract.

Deduction chain models attempt to solve this problem, but have the problem that they can only indicate the physically observable attributes. The influence of mental or emotional capacities cannot be inferred through such deduction. Hohmann, Lames and Letzelter (cf. ibid., p. 46) further criticize that such models postulate an independent relation between the variables which is seldom the case. As highlighted in section 2.2.1 for example, a sprinter cannot increase his velocity by simply increasing one or even both variables, as a stride length increase usually results in a frequency reduction and vice versa.

In the following, the author will present the internal factors as illustrated in the model by Bauersfeld and Schröter (figure 4). The sections detailing the physical attributes strength (section 2.3.2.1), flexibility (section 2.3.2.3), speed (section 2.3.2.4) and coordination (section 2.3.3) will moreover illustrate how these influence the stride length and frequency, thereby following the logic of the deduction chain model.

In this way, the author hopes to unify the characteristics of both models to comprehensively depict the prerequisites for peak sprinting performances.

2.3.1 Emotional and Mental Attributes

The value of psychological and mental characteristics for elite level achievements in sports is often underestimated, particularly in sports that appear to be dominated by physical components like strength or endurance. This is reflected in the way most people use the term talent to describe someone who exhibits outstanding physical attributes, but seldom consider psychological characteristics. Experts however agree that strong mental and emotional skills are basic prerequisites for all elite-level sports (cf. Bauersfeld & Schröter, 1992; Brown, 2001).

Of the five internal performance factor categories illustrated in figure 4 above, the mental characteristics represent one of the three major influences, together with physical fitness (section 2.3.2) and coordination and technique (section 2.3.3). This factor is characterized by qualities like a strong mindset and a healthy attitude towards regular, high intensity training. Only athletes empowered with these mental attributes will be able to reach the highest level.

Sprinters in particular, like other athletes in highly specialized domains, need to be mentally strong and exhibit a positive attitude towards the strenuous and monotonous training, a firm belief in the necessity of such training and a good relationship with the coach. Apart from these training traits, stellar performances in competitions require mental properties like self-confidence, focus and will power. The mindset of the elite athlete is rounded off by behavioral patterns that conform with a lifestyle committed towards achieving sports excellence.

However, it is important to confine the moral education, especially of young athletes, to a strictly sport specific context, and not misuse the opportunities that organized training presents in order to influence the students ideologically, as this was done in the former German Democratic Republic (cf. Deutscher Schwimmsport-Verband der Deutschen Demokratischen Republik, n.d., p. 5).

This is the theory, defining which specific skills are ultimately relevant is a different issue. As Brown (2001) laments, one can “talk with 10 sport psychologists and ... get 10 different lists of the emotional characteristics of exceptional athletes” (p. 29). Therefore, various experts, from sports psychologists to experienced track coaches, must be consulted to identify the commonly stated traits.

Based on interviews with experts from different fields, Brown names ten characteristics that separate talented athletes from others. They are drive, passion, stability, mental toughness, positive attitude, realism, focus, effort, persistence and competitiveness (cf. 2001, pp. 29-35).

Ruoff (cf. 1979, p. 168, table 1) conducted a study among coaches from three different sports to find out, among other things, what they considered to be the mental and emotional attributes of the “ideal child”. Concentration, focus, ambitiousness, will power and the ability to overcome defeats and setbacks were the most frequently stated traits.

Several of these characteristics are also mentioned by Bauersfeld and Schröter and placed in a sprint specific context (cf. 1992, p. 113). Before a race, the athlete must have full confidence in his own abilities, show an urgent desire to compete and develop a strong fighting spirit. He must also be able to block out any external sources of irritation while at the same time building up a positive tension and state of readiness. These attributes equate to the factors positive attitude, realism, competitiveness, drive, passion and stability stated by Brown.

During the race start, highest concentration and focus is required. Strong will power is also necessary to produce maximum effort on command. Lastly, during the race itself, the sprinter needs to be able to release his or her full potential in order to produce a top performance. Mental toughness and perseverance, expressed in the ability to push oneself to the absolute limit, are the key factors for this.

The emotional and mental skills do not only surface during competitions, but also play an even more important role in the years of training that precede participation in elite competitions. A great passion for the sport and a strong drive to succeed can be observed in even young athletes, and these attributes coupled with the willingness to invest time and effort into training often separate the drop-outs from the successful athletes.

Perseverance and focus are also two of the relevant mental factors cited by Jonath, Krempel, Haag and Müller (cf. 1995, p. 102). Furthermore included are the readiness for exertion (encompassing the will power to suppress exhaustion and push the body to its limit) as well as the ability to relax the body and the mind before a race in order not to tense up.

While Brown's assertion that ten experts will produce ten different lists of attributes may be true, it seems that there are certain characteristics that the experts agree on. The world's best sprinters possess them, and being able to reliably determine them in young aspiring athletes is vital to the talent identification process.

2.3.2 Physical Fitness

The physical attributes of an athlete are the most obvious factor and biggest contributor to sports performances. To determine their influence, the deduction chain model is better suited than the general performance model.

In the following sections, the four physical attributes strength, endurance, flexibility and speed will be portrayed. The characteristics and components of these attributes are defined, and it is elaborated how certain traits play a role in the composition of the previously introduced variables stride length and stride frequency.

2.3.2.1 Strength

Strength is defined as the ability of the neuromuscular system to express force or exert the greatest possible resistance against external forces (cf. Kent, 1998, p. 487). Strength appears in three classifications, namely maximum strength, strength endurance and power, and plays a vital role for all physical activity.

Maximum strength is “the greatest force the neuromuscular system is capable of exerting in a single maximum voluntary contraction” (Dick, 1989a, p. 171). As such, maximum strength is more important to weight lifters or hammer throwers than to sprinters, but a certain level of this component is nonetheless required and positively influences power and strength endurance.

Power, or elastic strength, is the ability to “exert forces quickly and to overcome resistance with a high speed of muscle action” (Kent, 1998, p. 162) and develop the greatest possible momentum within a given time (cf. Güllich & Schmidtbleicher, 2001, p. 17). Most track and field events are so-called explosive sports in which the performances are greatly determined by this strength component.

Strength endurance, the third component, is defined as the ability to “withstand [muscle] fatigue while performing repeated muscle actions” (Kent, 1998, p. 488) against sub maximal resistance that requires more than 30% of the individual maximum strength.

These three components do not stand alone: rather maximum strength acts as the base component and determines strength endurance and elastic strength. It is easier for a body builder than for a golfer to produce a fast muscle contraction against a resistance simply because he has much greater maximum strength and the resistance is relatively lighter for him than it is for the golfer. This also explains why weight lifters are almost as fast as sprinters over short distances (30m), but have no chance in actual races over 100 meters or more (cf. Bauersfeld & Schröter, 1992, p. 114).

Strength endurance is similarly influenced by maximum strength, as a greater base strength level will allow the athlete to uphold an activity longer than a weaker athlete. Strength or muscular endurance “can be determined by the maximum number of repetitions performed at a given percentage of an individual's one-repetition maximum” (Kent, 1998, p. 338).

Strength has been established as an important component for sprinting because it greatly influences the stride length. It is a requisite for a strong push-off in every single stride and plays a decisive role during the start and acceleration phase.

Here, the ground contact time is longer than during the phase of maximum velocity, and a strong and fast extension of the legs enables the sprinter to leave the blocks and achieve their maximum running speed faster (cf. Bauersfeld & Schröter, 1992, p. 114).

Apart from the maximum strength level, the elastic strength of an athlete is determined by the muscle fiber type. There are two types of muscle fibers, the red, small, slow-twitching fibers (also known as Type I) and the white, large, fast-twitching fibers (Type II). The fast-twitching fibers can be further divided into intermediary (Type IIc), fast twitch oxidative (Type IIa) and fast twitch glycolytic (Type IIb). These four types of muscle fibers have vastly different characteristics ranging from size to rate of innervation. For the development of the greatest amount of force in a short time, the speed and frequency of innervation as well as the force output of a contraction are of interest.

The slow twitching Type I fibers have an innervation frequency of about 10 to 30 impulses per second, whereas the Type IIb fibers can be stimulated up to 150 times per second. They also transmit the stimulus faster, at about 5.5 meters per second, opposed to the 2.5 meters per second with which the Type I fibers can relay the signal. Aside from being activated faster and more often, the force of each Type IIb contraction (approx. 100mg) is also considerably higher than that of a Type I contraction, approximately 70mg (cf. Weineck, 2000, p. 83).

This overview of the characteristics of the different muscle fiber types shows that a high percentage of fast-twitching fibers is advantageous for sprinters and other athletes from explosive sports. However, the disposition of these fibers is genetically determined and cannot be changed. This gives some ground to the claim that certain athletes are born to be sprinters – Carl Lewis was found to have a 90 to 10 distribution in favor of fast-twitching fibers in his legs (cf. Jonath et al., 1995, p. 55) - while others are born to be distance runners. A favorable genetic set-up is only a very small piece of the puzzle however, and even born sprinters have to train hard in order to fulfill their potential.

Jonath et al. (cf. 1995, p. 101) give an overview of all the physical fitness components that are essential for sprinters. This overview is called a demand or requirement profile. It states that elastic strength is the most important strength component for sprinters and that a high level of maximum strength is a prerequisite for achieving this degree of elastic strength. An athlete's stride length is directly related to his level of strength, and tests that measure the stride length like the 50m bounding (section 3.3.1.3) are therefore seen as a primary indicator for strength.

Furthermore, strength endurance is required to maintain the maximum running velocity for the longest possible time, and also to enable the athlete to engage in frequent and intensive training. For this purpose, not only the lower extremities need to be strong and powerful, but also the upper body, arms, shoulders, abdominals and back muscles must be adequately trained. Although it is the legs that drive the sprinter forward, the upper body muscles are responsible for maintaining the correct and most energy efficient posture during the race.

Recent studies have shown that plainly increasing the athlete's maximum strength level does not necessarily result in greater performances, rather the focus of strength training sessions must lie on involving the central-nervous system by selecting exercises that function in the same movement pattern and resemble the target motion (e.g. the active footstrike) as closely as possible. Effective sprint specific strength work therefore requires training at 90 to 95% of the maximum strength (cf. Mouchbahani, 2002, p. 11).

2.3.2.2 Endurance

Endurance is characterized as the ability to resist fatigue. This entails sustaining a given intensity for the longest possible time, keeping the decline of intensity as marginal as possible and upholding a high technical quality of movement throughout the entire game or race (cf. Hohmann, Lames & Letzelter, 2007, p. 50).

Like strength, endurance is divisible into different subcategories. These are not as clearly defined as the three components of strength and the classification depends on the perspective. For example, “sports scientists ... have found it useful to divide endurance into short-term endurance (35s-2min), medium-term endurance (2-10min), and long-term endurance (longer than 10min)” (Kent, 1998, p, 168). From the perspective of energy production, endurance can be categorized as either mainly aerobic or anaerobic (cf. Weineck, 2000, p. 141). It can also be viewed from the perspective of muscle activity and divided into dynamic and static muscle endurance, or divided according to the percentage of muscle mass used (general vs. local endurance).

For sprinters, anaerobic endurance is more important than aerobic, dynamic more than static and short-term more than long-term. Even short-term endurance, which is classified as the decisive endurance component for efforts of 35 to 120 seconds duration (cf. Hohmann, Lames, Letzelter, 2007, p. 62), only becomes relevant in the long 400m sprint race. Therefore, the crucial endurance component for sprinters is speed endurance. Bauersfeld and Schröter (cf. 1992, p. 114) define speed endurance as the organism's ability to resist a decrease in intensity and velocity, which is caused by both muscular and central nervous fatigue and is revealed by a breakdown of coordination and reduction of stride frequency.

The ability to produce energy under anaerobic conditions (that is without the presence of oxygen) is limited by the size of the Adenosine Triphosphate (ATP), creatine phosphate and glycogen depots in the muscles, as well as the capacity of the organism to withstand the side effects of lactate. The different race distances pose different demands on the athlete's endurance, and one can therefore not talk of the endurance that every sprinter requires. The ATP and creatine phosphate depots of an elite sprinter enable him to sustain a maximum intensity for up to ten seconds, long enough to finish a 100m race. The 200m and 400m on the other hand are predominantly determined by anaerobic lactic energy production and a reduction in intensity is attributed to lactate accumulation. Regular exposure to the “lactic anaerobic stressors in training will increase the athlete's ability to utilise this pathway” (Dick, 1989a, p. 85), the so-called lactate threshold.

The requirement profile by Jonath et al. (cf. 1995, p. 101) also qualifies speed endurance as the most important endurance component for short distance runners. However, as with strength, a good general aerobic endurance is equally indispensable in order to tolerate the high stresses that intensive sprint training calls for. Aerobic endurance also increases the ability of the body to recover from high lactate concentrations. Special caution is required for planning and conducting the endurance training of sprinters. Mouchbahani (cf. 2002, p. 4) points out that careful attention must be paid to the quality of movement especially during endurance work, since (flawed) stride patterns easily sink into the motor memory due to the high number of repetitions necessarily connected to such training. It is further stated that speed endurance can only be trained by running in the targeted zone (>95% of the maximum speed), and it must therefore be ensured that the sprinter has mastered the technique completely and has developed a sound aerobic base endurance. As such, speed endurance work should only be the last step in the training process and does not belong into the training regime of children and young teenagers. For this reason, the tests conducted within the scope of the talent search (section 3.3) do not include endurance examinations.

2.3.2.3 Flexibility

Flexibility is defined as “the capacity to perform joint actions through a wide range of movement” (Dick, 1989a, p. 216). A certain extent of this capacity is needed not only sports, but also for overall health in life. It is important to note that the definition emphasizes a wide range of movement (ROM), not the widest range. This is to say that the athlete must develop an optimal degree of flexibility for the respective sport, and not strive towards a maximal but ultimately unnecessary and potentially dangerous ROM. Kent (1998) notes that “an athlete who concentrates on flexibility exercises at the expense of strength training may reduce joint stability and increase the risk of dislocations” (p. 195).

Several factors limit an athlete's flexibility, among them the classification and structure of the joint as well as the ligaments, tendons and muscles. Stretching exercises usually involve decreasing the muscles' resistance to being extended and should not target “elongating the joint capsule and ligaments ... of normal length” (Alter, 1996, p. 51), because this may also increase the risk of injury.

Flexibility can be differentiated into active static, active dynamic, passive static and passive dynamic flexibility. Active flexibility is displayed when the force that causes a muscle to stretch is being exerted by its antagonist, while passive flexibility is called upon when external forces such as gravity cause the muscle to stretch. The stretch can be either static or dynamic, whereby dynamic flexibility is always greater than static and passive is greater than active (cf. Weineck, 2000, p. 488).

The velocity with which a runner can cover a distance has already been identified as simply a product of stride length and frequency. As the stride length is directly dependent on the runner's dynamic flexibility, an optimal ROM is an absolute necessity for elite sprinting performances. If the athlete can increase the stride length by as little as two inches without decreasing the frequency, he will run about 0.2 seconds faster over the distance of 100 yards (cf. Alter, 1996, p. 294), a time that can make the difference between the first and the last place of a race.

An increased flexibility also results in a longer range over which muscle force can be applied. For example, “if the runner has very short calf muscles, ... the force applied to the ground by contraction of the calf muscles will be through a shorter range of motion” (ibid., p. 297). Similarly, optimal ROM in the hip flexors allows for a longer ground contact as the body is moved forward past the leg, while flexibility of the gluteals facilitates the forward swing of the leg and the knee lift, which in turn results in a longer stride.

It must also not be forgotten that a flexible muscle offers less internal resistance and the force exerted by the antagonistic muscles which is needed to assume a certain (dynamic) position is lower. As a consequence, each step requires a smaller energy expenditure and the running technique becomes more efficient.

Jonath's requirement profile (cf. 1995, p. 102) indicates the necessity for a good, overall flexibility, especially of the leg and hip muscles. This enables the athlete to run more economically and also serves as a measure of injury prevention.

[...]

---

^[1] Source: http://olympic.org/uk/news/olympic_news/full_story_uk.asp?id=1797, accessed on February 2, 2007.

^[2] Jürgen Perl is a professor of applied informatics and sports at the University of Mainz, Germany and worked on sports related Expert Systems such as TESSY.

^[3] “Die Leistungsstruktur ist die Organisation einer sportlichen Leistung, die Darstellung ihrer Elemente, ihrer miteinander verknüpften Relationen – also ihres Wirkungsgefüges, der Wertigkeit, Abhängigkeit und Verflochtenheit ihrer Leistungsfaktoren” (Bauersfeld & Schröter, 1992, p. 24).

^[4] I contacted Professor Hohmann for the correct translation of the term “Modell mit/ohne Kriteriumsleistung” and he suggested “model with/without a performance criterion” (A. Hohmann, personal communication, January 31, 2007). As such, these terms are used.