Archive for April, 2008

A Puzzle and a Tribute

April 19, 2008

The puzzle

The proposed third section of my series on Dancing with the Devil will deal with the mental state necessary to allow us to achieve the movements required by the principles of biomechanics to achieve efficient running – the psychodynamics of running. However, I am delaying tackling the psychodynamics because there is one large question of the biomechanics of running that still puzzles me. The problem is deciding on the most efficient muscle action to achieve an adequately long stride when running at fairly high speed.

Evidence strongly suggests that once cadence reaches 200 steps per minutes, further increases in cadence would require a rate of muscle contraction that becomes increasingly inefficient. Therefore, efficient increase in speed required an increase in stride length. A longer stride length at fixed cadence requires a faster average speed of the leg as it moves forwards from stance to stance to support the body, because the distance from footfall to footfall is greater but the duration of the stride is unchanged.

One way to achieve a higher average speed of the leg is to use more vigorous hip flexion in early swing. However, this action will tend to bring the foot forwards rather than upwards at lift-off, resulting in the foot being carried at the end of a longer lever arm In contrast, if lift-off is largely via hamstring contraction with minimal hip flexion, the knee flexes more, the foot rises behind the line for point of lift-off to hip, the lever arm is shorter, and hence the leg naturally swings faster under the influence of gravity while the other leg is on stance.

In fact the swing is a power assisted pendulum rather than a simple pendulum driven by gravity. To the extent that we use muscle power rather than gravity to drive the swing, it makes sense to drive the foot along the shortest path from lift-off to foot fall. The shortest trajectory is a low arch in which the highest point of the foot trajectory is achieved at approximately mid-swing, rather than the pear-shaped trajectory that includes the initial high upward loop to a position behind the buttocks.

So to resolve the issue, we need either a complex mechanical model which is impractical, or alternatively, we need to examine the evidence from observation of elite runners. Many elite runners, such as Haile Gebrselassie, do lift the heel quite high behind the buttocks, so that the foot trajectory is not a simple symmetrical curve, but it does nonetheless appear that the foot is pulled forwards at lift off as well as upwards. So the answer appears to be a compromise.

While I have a fairly clear mental image that promotes a strong hamstring contraction and a different mental image that promotes hip flexion, I have not yet been able to decide on the mental image that promotes the optimum compromise. So for the immediate future, I am postponing part three of ‘’Dancing with the Devil’ until I have had a greater opportunity to try different mental images.

The tribute

The tribute is to the ‘Efficient Running’ thread on the Fetcheveryone website ( ). This thread was started on 20th August 2007 with a question about heel striking, and reached its 10,000th entry a few days ago. That reflects an average of about 50 entries a day, so keeping up with the torrent of material has been enough to leave one breathless, or maybe just sleepless. The entries have included many thought provoking ideas about efficient running, in addition to much social chit-chat.

At times people have complained that very little is ever decided on the thread because too many individuals are fixed in their thinking and unprepared to change. My own view is that most people hold fairly strong opinions largely because a particular way of running ‘feels right’ and/or is based on concepts which appear to make a great deal of sense: concepts such as the notion that it is best to land with the foot under the centre of gravity (COG). This seems to make sense because it minimizes the inevitable braking effect when the point of support is in front of the COG. However, as discussed previously in this blog, unless the foot lands in front of the COG, the body must continually accelerate or suffer a face down crash after a few strides. It is impossible to land under the COG when running at constant speed on a level surface in the absence of wind resistance

Despite some skepticism regarding the laws of physics, illustrated by the statement of one Pose coach that arguments based on physics ‘matter diddly squat’, the discussions on the thread have often sparkled and have challenged me to examine carefully what I am doing when I run, and also to think deeply about biomechanics. I hope this has led to some useful conclusions.

Looking back, I think that the Fetcheveryone thread has been the second strongest influence (after Gordon Pirie’s book ‘Running Fast and Injury Free’) on my thinking about efficient running. The scene was set by an entry by Cabletow, posted 5 hours after the initiation of the thread, which provided a succinct summary of the five principles espoused by modern schools of thought about efficient running. In his words, the principles are:

‘increase cadence to 90 per leg
Land under your cog with a bent knee to release plyometric energy
Land with a rearward moving foot and relaxed ankle
Do not push off but lean forward into the run
Limit arm movement, bend the elbows and do not reach too far forward.’

These principles have exerted a huge influence on much of the discussion on the thread. I think they are excellent principles, though I would take issue with two features. As already discussed, landing under the COG is impossible when running at constant velocity on a level surface in the absence of wind resistance. The other issue is the recommendation to lean forward. This is a confusing and sometimes confused issue. At times, people take it to mean a forward inclination of the long axis that passes through point of support, hips and shoulders at mid-stance, as illustrated in the spectacular picture of a muscular young man running on a beach in the PoseTech advertising literature. At other times, as emphasized by Cabletow himself in recent posts, it is taken to mean the forward lean of the line from point of support to COG that occurs as a result of hip extension in the second half of stance.

Both forms of lean result in a gravitational torque that tend to rotate the body in a face forwards and downwards direction. More confusing is the notion that lean promotes forward propulsion by means of this gravitational torque. This is a concept that arises from the questionable biomechanics proposed by Dr Nicholas Romanov in the theoretical underpinning of the Pose Method. It is a concept that ignores the law of conservation of angular momentum. If a face forwards and downwards torque is applied at some point in the gait cycle, it must be counteracted by an oppositely directed torque at some other point in the gait cycle if a face down crash is to be avoided. Nonetheless, the concept that gravitational torque provides useful forward propulsion continues to exert a strong influence on the discussions on the Fetcheveryone thread.

In relation to my personal priorities, the main limitation of the five principles proposed by Cabletow are that they do not adequately address the question of how to run with maximal mechanical efficiency, in the sense of using the minimum amount of energy per unit distance at a fixed speed. The five principles encourage a safe running style that minimises risk of injury. Minimising risk of injury is crucial to promoting good performance, but it leaves unanswered the question of the most efficient way to get the legs forward quickly enough to allow running at a constant high speed.

In my opinion, three of the modern schools of efficient running do make a coherent attempt to address the issue of how to run efficiently at high speed. These are Evolution Running developed by Ken Mierke (; Stride Mechanics developed by Jack Cady ( ); and the BK method developed by Frans Bosch and Ronald Klomp ( The concepts of Stride Mechanics have been presented occasionally on the Fetcheveryone thread by Jack Cady himself. However, as far as I can discern, neither Evolution Running nor BK method have been presented so far by individuals who are experts in those schools of efficient running, and I hope that they might be at some time in the future.

This tribute has not been a eulogy to a departed friend, but an account that reflects my own personal priorities, and includes an expression for hope for a broader perspective in the future. The Fetcheveryone Efficient Running thread has been a major source of inspiration to me. Thank-you to all who have contributed.

The Dance with the Devil: putting the steps together

April 5, 2008


In the last three posts, I have attempted to describe what happens in the various stages of the gait cycle. However, the cycle is one integrated sequence, so this post will focus on how it all fits together, and including some detail about the torso and arms


Gordon Pirie recommends upright torso; Pose recommends a forward lean maintaining a straight-line from point of support via hips to shoulders, at mid-stance. The theory behind the Pose lean is based on what I believe to be false biomechanics. The proposal by Dr Romanov that gravitational torque can generate forward propulsion even when running at constant velocity is tempting, but violates the law of conservation of angular momentum. Subjectively, the lean can feel helpful – but I think that is a misleading perception based on the experience of starting from a stationary position. There is no doubt that a lean promotes acceleration that is helpful for a sprinter driving from the blocks, but acceleration is only a very minor part of longer distance running. So whose advice is more helpful: Pirie or Romanov? In a previous post I came down favouring Romanov, but that was after a session running into a strong wind. Maybe lean helps when you need continuing reinforcement of the horizontal drive on account of wind, but I am now inclined to think that Pirie’s advice is best on a level surface when the wind is not too strong.

The reason I think an upright torso is best is that it promotes a greater eccentric stretch of the hip flexors during late stance, and this will facilitate hip flexion after lift off, thereby bringing the leg forwards to overtake the torso by mid-swing. Upright torso also likely to promote effective deceleration of the leg by hamstrings and gluteus maximus in late stance, and good coordination of hamstrings and quads to achieve the required flexion of hips and knee at footfall. Both the quads and hamstrings cross hip and knee, and appear to have evolved so that despite being mutual antagonists, simultaneous contraction of both can produce well coordinated movements at both joints when the torso is upright.

Similarly, keeping the hips forward (i.e. avoiding ‘sitting in the bucket’) promotes more efficient hip flexor stretch in late stance which helps get the legs forward quickly; and promotes good coordination of the hip and knee at footfall.


When the neurosurgeon Wilder Penfield used electrodes applied directly to the brain to stimulate muscle contraction, as part of pre-surgical exploration of brain function in patients needing surgery for epilepsy at the Montreal Neurological Institute in the 1930’s, he demonstrated that a much larger area of the motor cortex in the brain is devoted to controlling the upper limb than the lower limb. This fits with the observation that most people are more dextrous with their hands than their feet. However, the upper and lower limbs automatically work in synchrony when running. So it is plausible that conscious focus on what we do with the upper limb will be more effective than focussing on the legs and feet. In particular, focus on the backward movement of the arm on the side of the leg that is beginning to swing in early swing phase is likely to help bring the leg forward in the optimum direction at the beginning of swing. Even more importantly, making sure that the subsequent forward swing of that arm does not go too far forward will promote the required (non-conscious) braking of the swing leg in late swing and minimize risk of over-striding. The harness proposed by Jack Cady of Stride Mechanics achieves this. This arm swing should be purposeful and controlled but not too tense to avoid wasteful isometric contraction in the shoulders. I find it helpful to form a lightly held ring-shape by resting the index finger of each hand against the adjacent thumb. I think this fairly delicate action encourages a controlled but relaxed arm. This recommendation for hand posture is attributed to Emil Zatopek, the great distance runner of 1950’s, who was famous for his contorted neck and facial features while running, but nonetheless, managed to maintain remarkable relaxation of his limbs (see Wikipedia entry for Emil Zatopek).

Rotation about the long axis of the body

As the hip extends in late stance, the hip rotates externally, thereby producing eccentric stretching of the internal rotators. At lift off, the hip rotates internally bring the leg around and forwards, thereby lengthening the stride and ensuring that the foot is in the midline by foot-fall. This rotation is facilitated by a balancing rotation of the upper torso produced by arm swing.


Assembling the main features from the previous postings, together with these principles regarding the upper body leads to the following integrated picture:

1) Cadence should be high (e.g. in the range 180-200 strides per minute) to minimise work required to overcome gravity in the airborne phase.

2) Time on stance should be short, though there is a balance between peak mechanical efficiency achieved with a time on stance around 50-60 milliseconds) and minimization of risk of tissue damage (maybe best achieved at around 100-120 milliseconds on stance. Short time on stance minimizes braking in early stance

3) Torso should be held near to upright, with perhaps a slight forward lean if needed to counter wind resistance.

4) Arms should swing in a relaxed but controlled manner, avoiding swinging too far forwards.

5) At foot fall, the hip and knee should be slightly flexed and the ankle near neutral, but with very sight plantar flexion to that the initial point of contact is on the outside edge just forward of mid-sole. As speed increases, the degree of flexion of the hip and knee should decrease making the leg stiffer, leading to a more rapid recoil and shorter time on stance. However, this will increase stress on musculo-skeletal tissues. It should be noted that many elite athletes actually land on the heel. This will result in an even stiffer leg, which may enhance mechanical efficiency, but the risks of over-striding and of musculo-skeletal damage are likely to be higher.

6) In mid-stance, contraction of the hip abductors prevents the hip dropping on the unsupported side, allowing the leg to swing freely and avoiding sideways slanting of the torso.

7) In late stance, recoil aided by contraction of the quadriceps will generate the vertical Ground Reaction Force that provides the impulse required to lift the body. Extension of the hip will preload the hip flexors.

8 ) A conscious pull using hamstring promotes a well-timed lift-off, and an associated concentric contraction of hip flexors in early swing brings the leg forwards to overtake the torso by the time the other leg is at mid-stance.

9) Rotation about the vertical axis of the body, produced by synchronised concentric contraction of the hip rotators with arm swing, will help open up the stride, and ensure that the support foot lands on the midline.

10) In late swing, gluteus maximus and hamstrings decelerate the leg so that horizontal velocity relative to the ground is near zero at footfall.

The next section of the Dance with the Devil will tackle the issue of the perceptions that allow us to achieve these actions effectively, and the mental state that prepares us for peak performance.

The steps of the dance: 3. Swing Phase

April 2, 2008


The goal of early swing is to get airborne and accelerate the leg forwards on a trajectory that will allow it to overtake the torso by mid-swing. While it is essential that the foot should accelerate in early swing, it should be borne in mind that it must decelerate in late swing if it is to have zero horizontal velocity relative to the ground at foot-strike. It might seem at first sight that the need to match an energy consumptive acceleration with a deceleration that will also consume energy should encourage us to be conservative in the generation of acceleration. However, this would be a very misleading conclusion. Our ability to generate adequate forward acceleration of the foot in early swing determines our ability to maintain a particular target speed.

The crucial role of acceleration of the leg in early swing
To understand why forward acceleration of the leg in early swing is crucial, we need to return to basic biomechanical principles. In the earlier posts in this series in which we considered the implications of Newtonian physics we reached the conclusion that cadence should be high and time on stance should be short. Except at very slow speeds, cadence should be near the limit determined by the optimum speed of contraction of muscles. Observation of elite runners suggests the optimum is a cadence in the range 180-200 strides per minute. Elite athletes employ a cadence in this range for all except very slow paces.

Furthermore, time on stance should be as short as can be tolerated, after allowing for the fact that ground reaction forces and risk of tissue damage increase dramatically as time on stance becomes very short. Elite athletes tend to spend only about 50-100 milliseconds on stance, with the longer times being applicable in long events where protection of muscles from damage due to repetitive impacts in important. Apart from these relatively small variations, cadence and time on stance are fairly consistent over a range of paces extending from 1500K pace to marathon pace. Therefore, over this range of paces, the major variable that increases as pace increases is stride length.

As shown on the calculations page accessed via the side bar, the work that must be done against gravity (per unit of time) is determined by cadence, time on stance and body weight. The energy required to lift the body is not directly influenced by stride length. However, increase in stride length must be matched by an increase in the amount of acceleration required to bring the foot forward fast enough to support the body at foot fall. Thus, it is ability to accelerate the leg in early swing phase (and then decelerate it again in late swing phase), that is the main determinant of our ability to maintain a high pace. So how should we do this?

Breaking contact with the ground
In late stance the elastic recoil of quadriceps, augmented by concentric contraction, has imparted an upward impulse to raise the centre of gravity and hip extension has preloaded the hip flexors (e.g. psoas). As the body rises, an active contraction of hamstrings lifts the foot from the ground. Contraction of the hamstring alone, when the hip is already extended, will produce flexion at the knee, pulling the foot up wards behind the line from foot to hip. While this is the path of the foot observed in many athletes, if the main goal is to accelerate the leg forwards, the hamstring contraction should be accompanied by hip flexion.

Accelerating the leg
Fortunately, the preloading of the hip extensors (i.e stretching associated by eccentric contraction) during hip extension in late stance can be utilized to facilitate a powerful recoil associated with concentric contraction of the hip flexors that accelerates the leg forwards.

Deceleration of the leg
However, the price paid for this powerful forward acceleration is the need for a powerful deceleration in late swing, provided by an eccentric contraction of the hip extensors. This is stressful for the hamstrings, and suggest that exercises such as hip swings might play a useful role in conditioning the body during training.

As the hip extensors decelerate the leg, the lower leg and foot should be allowed to swing down to that the knee is only mildly flexed, in preparation for footfall. The combination of contraction of hip extensors and relaxed un-flexing of the knee present a challenge. Because the hamstrings cross both hip and knee joint, pure hamstring contraction to decelerate the leg would prevent the relaxed swinging of the knee. Therefore it is essential to use gluteus maximus to assist in the deceleration of the leg. In addition, some contraction of the quadriceps might also be used to un-flex the knee, but this should be done very sparingly, as vigorous contraction of quadriceps at this stage is likely to result in over-striding.

In summary

Contraction of the hamstrings will help break contact with the ground as the body rises under the influence of the upwards impulse generated by recoil and quadriceps contraction in late stance. However, the ability to accelerate the leg forwards in early swing phase (and then decelerate it again in late swing phase), is the main determinant of our ability to maintain a high pace. Rapid forward acceleration of the leg in early swing might be achieved by employing the preloading of the hip flexors (e.g. psoas) that occurred during late stance to facilitate a powerful contraction of the hip flexors. However, this must be matched by a deceleration produced by contraction of hamstrings and gluteus maximus in late swing, allowing the foot to drop to the ground with the knee slightly flexed and travelling with approximately zero horizontal velocity relative to the ground.

The Steps of the Dance: 2. Stance

April 1, 2008


The early articles in this series examined the constraints that Newton’s laws of motion place on the way in which we run. The most recent article, posted on 31st March examined the question of how we should orient the joints and tension the muscles at foot fall in order to capture the energy of impact as elastic potential energy, while minimizing the risk of damage to muscles and other tissues. This article examines the actions that occur during stance.


Following foot fall the foot remains stationary on the ground, anchored by friction, while the COG passes over the point of support. Then as the COG continues forwards the hip extends until lift-off. We will use the term early stance to describe the period from foot fall to point where COG passes over point of support, and late stance for the period from the passing of the COG over point of support to lift off.

The preceding article in this series discussed the mechanisms by which a substantial portion of the energy of impact at footfall is converted to elastic potential energy, stored in quadriceps, calf muscles and the connective tissues of the foot. During late stance the elastic potential energy is recovered and contributes to the generation of the impulse required to get airborne at lift off.


While the main task that must be performed during stance is the generation of the upwards impulse required to get airborne, there are also a number of important subsidiary actions, including:

– generation of horizontal GRF.

– subjecting the hip flexors to eccentric contraction preparing them for the task of accelerating the leg forwards in early swing phase

– subjecting the hip rotators to eccentric loading in preparation for the rotation around the long axis of the body during swing phase

– preventing the unsupported hip from dropping and tilting the pelvis.

These actions are mostly performed automatically when running, but understanding them is important to allow us to develop strength and resilience in the muscles and other connective tissues, and to identify faults that might cause injuries. In particular, it is important to identify the muscles that undergo a large change in length during the gait cycle, as flexibility exercises might profitably be employed to maintain as these muscles and their tendons in a pliant state. In contrast, those muscles which exert large forces over relatively short distances are generally better maintained in a stiffer state.


Generation of the upwards impulse.
The only upwards directed external force acting on the body is the vertical component of Ground Reaction Force (vGRF) and hence this force is responsible for lifting the body. vGRF is a reaction by the ground as it resists compression by a downwards directed force exerted though the foot. It should be noted that pulling the foot towards the hip (like pulling on ones own boot-straps) cannot lift the body.

About 50% of the energy required to become airborne might be derived from elastic recoil of quads and calf muscles. The knee had been initially slightly flexed at footfall and then flexed even more as the impact energy was absorbed in early stance; as this flexion occurs the quadriceps undergo a moderate eccentric contraction , then in late stance, recoil augmented by the concentric contraction of the quadriceps straightens the leg, and imparts an upwards impulse to the body.

Similarly, the recoil in the calf abolishes the mild dorsiflexion of the ankle that had developed by midstance and re-establishes the mild degree of plantar flexion present at footfall. The pronation of the foot in mid-stance is replaced by slight supination.


Generation of horizontal impulse
As the hip extends in late stance, the predominantly downwards directed forces exerted by the leg on the ground inevitably have a small backward directed component, which is resisted by friction, thereby generating a forward directed horizontal GRF. This will propel the body forwards against wind resistance – but unless there is a strong head wind, this forward propulsion is usually excessive and the excess must be matched by braking in early stance (as discussed when we addressed the question of where the foot must land at footfall). Apart from the contribution that overcomes wind resistance, the impulse due to horizontal GRF does not achieving any useful purpose. Therefore this impulse should be minimized as far as is feasible by lift-off from stance which is as rapid as possible (bearing in mind that stance must be long enough to allow the generation of adequate vertical impulse without necessitating very high and potentially damaging vGRF


Preloading the hip flexors
Immediately after foot fall the hip extensors (mainly gluteus maximus and the hamstrings) will undergo a degree of eccentric contraction as impact forces are absorbed. In mid and late stance, the elastic energy stored during eccentric contraction will be released in conjunction with moderate concentric contraction of the hip extensors. In addition there will be a strong impetus to passive extension of the hip as momentum carries the COG forwards beyond the anchored foot. Thus there is an almost effortless extension of the hip which stretches the hip flexors, priming them for a powerful concentric contraction in early swing phase that will help propel the foot and leg forwards to overtake the torso. Because the hip flexors, predominantly psoas undergo a major change in length, these muscles should be maintained in flexible, pliant state. The hip flexors are perhaps the most important muscles for a runner to maintain in a flexible, pliant state.

Preloading the hip rotators
In late stance, as the torso moves forwards leaving the support foot behind, not only is there passive extension of the hip but there is also a passive (external) rotation of the hip about the long axis of the body because the hip on the supported side is tethered via the leg to the ground while the other hip is free to move with the torso. Thus the internal rotators of the hip are subjected to an eccentric contraction which primes them for an internal rotation after lift-off that will help increase stride length.


Supporting the pelvis
The unsupported hip tends to drop, causing the pelvis to tend swing inwards towards the supporting leg. This adduction of the hip must be opposed by the hip abductor muscles of the supporting leg. The main abductors are gluteus medius and gluteus minimus, assisted by tensor fascia lata (TFL), a long muscle running down the outside aspect of the thigh from the rim of the pelvis to attach to the tibia via the iliotibial band (ITB). If gluteus medius is weak, TFL is called upon to bear too much of the load in prevent the pelvis from tilting. Excessive tension of the iliotibial band creates a risk of friction at the point where the ITB passes adjacent to the bony protrusion (femoral condyle) on the lateral aspect of the knee. Excessive eversion of the ankle, which tilts the tibia (shin bone) outwards at the knee also increases the friction on ITB. Increased friction may lead to painful inflammation (ITB syndrome).



In summary

The quads and calf muscles undergo eccentric contraction in early stance thereby storing much of the energy of impact as elastic energy. In late stance, recoil of these muscles, aided by moderate concentric contraction provides the impulse required to accelerate the body upwards against gravity. The other major muscle action during stance is hip extension. Some of the impact energy is absorbed in an eccentric contraction of the hip extensors (gluteus maximus and hamstrings) which subsequently recoil while undergoing concentric contraction in late stance. Hip extension is further promoted by passive extension generated by the momentum of the torso. The resulting hip extension pre-loads the hip flexors preparing them for a powerful hip flexion to accelerate the leg forwards after lift-off. Other important actions are external rotation of the hip that preloads the internal rotators, and hip abduction that prevents tilt of the pelvis.

In the next article in this series we will examine the actions occurring at lift-off and during swing phase.