Natural running

The word ‘natural’ invokes images of wholesomeness but also carries a hint that we are in danger of being hoodwinked by a snake-oil merchant.  In contrast, the word ‘technological’ has overtones of something lacking wholesomeness.  Nonetheless, on the whole, I am glad I belong to a species with the brainpower to develop technology.   Many inventions created by human wit, ranging from reading glasses to electronic devices, extend the range of activities that are accessible to me and make life more interesting.  But when it comes to running, there is good reason to ask whether we have lost our natural skill as a result of growing up a modern technological society.

Humans are in fact remarkably good endurance runners.  Although many species can outpace us in a short sprint, few can maintain a steady pace for such long distances. On the other hand, a very large proportion of us get injured each year.  In a comprehensive review of studies of injury rates among distance runners, van Ghent and colleagues found that the incidence of lower extremity injuries reported in published studies ranges from 20% to 79% (Br J Sports Med 41: 469-480, 2007).

Persistence hunting

Accumulating evidence suggests that humans became good endurance runners because evolution favoured the development of anatomical (and perhaps biochemical) adaptations that enabled our forebears to engage in persistence hunting – in which the hunter pursues his quarry to the point of exhaustion – on the African savannah around two million years ago. (Bramble and Lieberman, Nature, 432, 345-352  2004).   We do not know whether or not early persistence hunters were also prone to injury, though evolution would scarcely have favoured those who were as prone to injury as modern-day runners. Perhaps only an elite few in the tribe were able to run without injury.  Among the few remaining peoples of the Kalahari desert who continue to practice persistence hunting today, the huntsman who engages in the long chase is an elite member of the hunting group.  Nonetheless, an ability to run far with few injuries is likely to have been a fairly common skill among our early ancestors.

Bare feet v shoes

So if we wish to minimize injury, it is probably worthwhile to ask how did our forebears run.  Perhaps the first point to note is that they would not have worn shoes (though it is also noteworthy that  modern-day persistence hunters in the Kalahari do wear shoes).  The most striking difference between barefoot and shod runners is the nature of the foot-strike.  Hasegawa and colleagues demonstrated that about 75% of runners wearing modern running shoes heel-strike (J Strength Cond Res. 21(3):888-93, 2007).  In contrast, Lieberman and colleagues have demonstrated that bare-foot runners are much more likely to land on the forefoot and then transfer a portion of the load to the heel whilst on stance.  Lieberman and colleagues have demonstrated this is a systematic study comparing American habitual barefoot runners with shod runners, and confirmed it in a less systematic observation of Africans who had grown up never wearing shoes (Nature. 463(7280):531-5, 2010.   Landing on the forefoot minimises the initial sharp increase in vertical ground reaction force that is seen with heel striking.

Lieberman is firm in pointing out that there is no strong evidence that minimising this sharp increase in ground reaction force leads to lower injury risk. However, in general, the repeated application of a rapidly rising large force is stressful and might be expected to lead to stress fracture.  So it is plausible that injury risk is greater when wearing shoes. This plausibility is confirmed by Kerrigan’s demonstration of greater torques at hip and knee during shod running (PM &R: The Journal of Injury, Function and Rehabilitation, Vol. 1, pp 1058-1063, 2009).  Thus it appears that Bill Bowerman’s first experiments with a waffle iron that led to the modern running shoe, were a faltering mis-step based on the mistaken idea that putting padding between the runners foot and the ground would increase safety and efficiency.

Getting airborne

I have discussed the question of running shoes and foot-strike in a previous post, and I will probably return to it again in the future.  However my main interest today is in the question of how our forebears were equipped to deal with the cardinal challenge of running: exerting a strong enough force to get airborne.  Getting airborne is the essence of running.  It allows us to minimise the inefficient braking that is an inevitable consequence of maintaining a stationary foot on the ground during the stance phase.  To minimise braking we must spend as small a portion of the gait cycle on stance as is possible.  We can do this by landing with the foot only a short distance in front of our centre of gravity (COG), but that necessarily entails the exertion of a large push against the ground.   If we are on stance for only a third of the gait cycle, the average push against the ground during stance must be three times body weight.

A substantial part of this push is generated via elastic recoil.  But in fact, measurements suggest that at most about 50% of the required energy can be generated by elastic recoil (Alexander, R.M. Energy-saving mechanisms in walking and running. J.Exp.Biol.160,55–69,1991).  So an equally substantial portion of the work must be done by an active push.   What evolutionary development allowed early member of the homo genus to achieve this crucial push?  A clue can be found by examining the anatomical differences between ourselves and our nearest primate relative, the chimpanzee.  Chimps, like other non-human primates, are not capable of endurance running.

Differences between man and chimp

The most immediately apparent anatomical difference is man’s larger skull.  However, the larger skull is a feature of homo sapiens rather than early members of the homo genus.  Possibly we owe our large skull and brains  at least in part to a somewhat more subtle change at the lower end of the vertebral column that occurred earlier in homo evolution.  This subtle change, present in early members of the homo genus, such as homo erectus, is a substantial enlargement of the upper part of gluteus maximus.  Gluteus maximus is a hip flexor.  Although it acts with less mechanical advantage than the hamstrings, it is more massive   Could the enlargement of gluteus maximus have played a key role in the development of endurance running ability, thereby facilitating persistence hunting and providing the protein rich diet essential for the eventual development of homo sapiens’ large brain, over a million years later.

The roles of gluteus maximus

To address this question Lieberman and colleagues  examined the activity on gluteus maximus throughout the gait cycle, by recording the electrical signals from an electrode placed over the muscle, during both walking and running (Journal of Experimental Biology 209, 2143-2155, 2006.)  Their  first important  observation was that gluteus maximus is much more active during running than walking, consistent with it being an evolutionary development associated with the acquisition of capacity for endurance running.  During the running gait cycle, there are two main bursts of activity in gluteus maximus: one when the foot from the opposite side of the body is on stance and the other beginning shortly before the footfall of the foot on the same side as the muscle,  and continuing through early stance on that foot.  The activity when the opposite foot is on stance almost certainly reflects the action of arresting the forward motion of the swinging leg.   Interpretation of the role of the burst of activity in early stance on same-sided foot is more complex.  The magnitude of the activation increases with speed of running and is also correlated strongly with the velocity and timing of the forward pitch of the trunk that occurs at foot-strike.  Thus it is very likely that a major role of gluteus maximus is stabilizing the torso.

Mark Cucuzella’s resonant phrase ‘you can’t fire a cannon from a canoe’ powerfully expresses the importance of stabilization of the torso, but it also raises the question of what cannon is being fired.   Could gluteus maximus also contribute to generation of the vertical ground reaction force (vGRF) that launches the body forwards and upwards from stance?   Lieberman and colleagues  observed that the timing and magnitude of activity in gluteus maximus is also correlated with the timing and magnitude of contraction of another major hip extensor, biceps femoris, which is one of the hamstrings.    This suggests an active role in hip extension.  It is important to note this active hip extension is largely confined to the early part of the stance phase.  As the hip and knee are slightly flexed at that time, the main consequence of hip extension will be an increase in the downwards push against the ground.   Thus, this action would be expected to contribute to the initiation of the upward acceleration of the body commencing in mid-swing. Perhaps gluteus maximus also contributes to firing the cannon.

It is noteworthy that one of the early proponents of ‘natural’ running, Ken Mierke, recognised that combining contraction of gluteus maximus with the hamstrings would greatly increase the power of hip extension, thereby reducing fatigue of the relatively weak hamstrings and promoting endurance.  While I think that the essence of Ken’s proposal is sound, I would place a somewhat different emphasis on the effect of the hip extension.  Ken argues that the hip extension largely provides forward propulsion.  I do not think that fits well with the timing of the active contraction of either gluteus maximus or the hamstrings.  Even after allowing for the 40-50 millisecond delay between the electrical signal and the mechanical effect of a muscle contraction, the active contraction of gluteus maximus and the hamstrings is complete shortly after mid-stance.   I think that the main consequence of this powerful hip extension is to accelerate the body upwards thereby achieving a stance that is short – this is the key requirement for efficient running.

Other muscles also contribute, notably contraction of the gastrocnemius, which reaches its peak contraction a little later in stance.  This will generate a forward and upward GRF.  The upward component will add to the impulse that gets us airborne, while the forward component will help compensate for the braking that occurred in early stance.  Because the hamstrings cross both hip and knee, residual tension in the hamstrings in late stance might add to the upswing of the lower leg relative to the torso thereby facilitating the breaking of contact.  However it should be noted that the contribution of a hamstring to pulling the foot towards the torso cannot contribute to raising the centre of mass of the body (as is proposed in Pose theory).  That would be pulling oneself up by ones bootstraps.  The upwards acceleration of the mass of the body must be produced by a push against the ground.  (Added note: it should be acknowledged that Pose theory appears somewhat ambiguous regarding the mechanism by which the centre of mass is raised.  See the comments from Hans and Jeremy below.)

Other evolutionary developments

Development of gluteus maximus was not the only anatomical change occurring early in the evolution of the genus homo.  Freeing up of the tethering of head to shoulders that limits the independent rotation of upper torso and head in other primates, allows us to produce the counter rotation of the torso necessary to balance the swinging leg, while maintaining the head upright and forward-facing.  In addition, the development of a longer Achilles tendon that occurred at some point along the evolutionary path from our even earlier ancestor, australopithecus, to early homo species, is likely to have enhanced the efficiency of capture of impact energy as elastic energy.   But in my opinion, it was the development of gluteus maximus that was the decisive development that allowed us to get airborne efficiently.

Minimizing risk of injury

While these speculations might explain how our forebears came to be efficient endurance runners, it still leaves us with the question of how we might avoid injury in the face the inevitably large vertical ground reaction forces generated by the powerful push.  I think that Kerrigan’s  demonstration of greater torques at hip and knee during shod running is a key observation.  This suggests that the orientation of the foot on the ground during the period around mid-stance when vGRF is at its peak is likely to play a major part in determining how much torque is produced.  Drills that help develop a sharp contraction of gluteus maximus that is well coordinated with the down swing of the contralateral arm will ensure that the non-conscious brain is well appraised of just when the peak vGRF will occur.  In addition, an appropriate  sharing of weight between forefoot and heel at mid-stance facilitated by  shoes that are light enough to allow a good perception of the distribution of ground reaction forces might allow the non-conscious motor control system in our brain  to coordinate the application of the push in a way that minimises potentially damaging torque at the knee and hip.


We have grown to adulthood spending hours each day sitting at a desk or in an automobile seat, and even longer periods with our feet encased in rigid shoes.  If we are to run naturally, in a style similar to that which allowed our early homo ancestors to master the art endurance running, perhaps we should focus on re-training our bodies so that our non-conscious brains can once again integrate the sensory signals from the joints in our arms and legs, with those from the numerous sensory nerve terminals in our feet, to coordinate the delivery a powerful, well-timed but fairly safe push against the ground to get us airborne.

204 Responses to “Natural running”

  1. Hans Holter Solhjell Says:

    Great article. Although, where did you read that upward motion is produced by the hamstring pull, as you reference POSE as saying? I can’s see any real difference between what you are writing here, and most of POSE theory, maybe apart front that you seem to be promoting a slightly more active push than what is promoted by POSE. The purpose of the hamstring pull is, as I have understood POSE theory, not to create upward motion, but to break contact with the ground, after upward motion has been created by the other mechanisms you describe in your article. A too active, exaggerated or prolonged push is counterproductive in several ways, at least according to theory, and also my personal experience trying to learn better running, and as you seem to agree with as well.

    • canute1 Says:


      Thank you for your comments.

      I agree that most of what I describe is consistent with much of Pose theory. I think that the biggest difference is that Pose theory explicitly states that there should be no push against the ground. In Pose archive article 000229 Dr Romanov states that the first requirement of Pose ‘is not to push off’,

      We agree that in Pose theory the purpose of the hamstring pull is to break contact with the ground. I also agree that Pose warns against ‘overpulling’. However, it seems to me that in light of Dr Romanov’s statement that the first requirement of Pose is ‘not to push off’, your statement that ‘The purpose of the hamstring pull is, as I have understood POSE theory, not to create upward motion’ only makes sense if Pose theory denies that breaking contact with the ground involves upward motion (unless perhaps you consider that elasticity provides the push, but there is good evidence that elastic recoil only recovers about 50% of the required energy (see Alexander R.M. Energy-saving mechanisms in walking and running. J.Exp.Biol. 160,55–69,1991)

      While I accept that Pose theory minimises the importance of the upward motion, in fact Dr Romanov acknowledges in Pose archive article 000229 that a vertical oscillation of 4-6cm does occur. This is not trivial In light of Dr Romanov’s explicit statement that the first requirement of Pose ‘is not to push off’, what mechanism does Pose theory propose for achieving this elevation of 4-6cm?

      • jhuff Says:


        This seems to answer your question:

        “”””””From this short list we have to choose what we’ll be doing, where our focus will be. Vertical displacement in running happens by utilizing muscle/tendon elastic property, which lifts the body just 4-6 centimeters above the ground, just enough to shift the body weight from one support to the other. This data was found in all ranges of speed and distances of elite runners. Any efforts over this need would mean rising the GCM against gravity and wasting energy. R.Margaria (1976) found that the GCM vertical displacement in running doesn’t go higher than an erect standing position of the body, which could be interpreted as the best position of the body to fall forward and down, get muscles/tendons loaded and stretched, then return this energy for moving body up to the same height.”””””

        Comes from this article:

      • canute1 Says:


        As I pointed out, this statement denies the evidence that elastic recoil only recovers about 50% of the energy required.

      • jhuff Says:


        He seems to think that it does. You are free to disagree.

      • canute1 Says:


        Yes it does appear that Dr Romanov believes this. He appears ignorant of both the observational evidence (see the review by Alexander in Journal of Experimental Biology in 1991) and also to lack an appreciation of how a musculo-tendinous unit works.

    • jhuff Says:


      Is the reflexive push you are refering to in addition to elastic recoil?

  2. jhuff Says:

    Canute, I don’t think your comment about pose theory is accurate in regards to pulling. Where did you get that theoretical idea? What literature? In this article here: it seems clear that it isn’t pulling that is responsible for vertical lift.

    • canute1 Says:


      I got that idea from reading articles on the Pose tech site and from discussion with Pose coaches. In the article you quote (Pose archive 000229) Dr Romanov states ‘One of the first requirements in this new paradigm of running is not to push off’ From the context of the article, the new paradigm is Pose.

      Dr Romanov describes the vertical motion of the COG in the following terms. ‘ This height, as shown by some different sources, has a small range of 4-6 centimeters between the best runners (sprinters and marathoners). In other words, the general center of mass of the body travels evenly in the horizontal direction.’

      I agree that 4 to 6 cm is a good estimate of the range of vertical oscillation. However this is not trivial. The essential feature that distinguishes running from walking is getting airborne when running. An elevation of 4-6 cm is all that it required, but it is crucial.

      In light of Dr Romanov’s statement that the first requirement of Pose ‘is not to push off’, how does Pose theory propose that the elevation of 4-6cm is achieved?

      As noted in my similar response to Hans (above) elasticity not an adequate explanation because elastic recoil recovers only about 50% of the required energy.

      • jhuff Says:

        Canute, the answer is above.

      • canute1 Says:


        I accept that Pose theory is inconsistent on this issue and I have added a modifying sentence to the text of my post. In fact it is not credible that elastic recoil provides more than about 50% of the energy required for recoil. Dr Romanov’s statement that one of the first requirements of Pose is not to push off is inconsistent with his acceptance that the COG is elevated by 4 to 6 cm.

      • jhuff Says:


        I don’t think that there is an inconsistentency either. It is clear to me that he is refering to the general accepted term of “active push off”. Do not actively push off is what he is saying. The vertical movement will happens “passively”. I understand that you don’t are with him but he is consistent imo.

      • canute1 Says:


        The push does not have to be conscious but it must occur.

        The reason I am continuing to discuss this point is not merely to be antagonistic. I believe that the under-estimation in Pose theory of the forces that are actually involved in running is seriously misleading.

      • jhuff Says:


        Which force is under estimated? From all the information I have viewed there seems to a meticulous amount of thought put into the forces Involved in running.

      • canute1 Says:


        The force is under-estimated in the statement by Dr Romanov that there is no need to push off, (in Pose archive article 000229), and in many statements in chapter 12 of ‘Pose Method of Running’

        Furthermore, his statement that elastic recoil is adequate to achieve elevation of the COG, under-estimates the force required. In light of the evidence reviewed by Alexander in his article written in 1991, demonstrating that elastic recoil is not adequate, It does not appear that Dr Romanov has put a meticulous amount of thought into the question of the forces involved in running.

  3. jhuff Says:


    Just curious… most/many of your topics it is clear that you believe strongly in a distingishable and recognized “push off” to be employed in both theory and application when running. However I also get the distinct feeling based on your writings that you don’t really know how to best implement this “push off” that you believe so strongly in. Forgive me if you have written this in another text somewhere but can you tell me some of the ways you recommend practicing the “push off”? I’d like to see and understand if and what things are different about your actual training based on your clear difference in IMO of theory of running technique. Thanks in advance.

    For me personally perception and experience takes precedes over equations. When I run I do perceive running exactly as prescribed by Pose Method. I chose and continue to choose it as my means of understanding and implementing running because it most closely defines my running both intellectually as well as perceptually. Given my previous experiences in running and my current experiences in running it is very doubtful that I will change my POSITION in regards to technique. However I really enjoy seeking out people who don’t agree with me. That way if I am missing something in my approach I will find a better solution or at the very least understand others view points. This is why I like to follow your blog. Anyway most go tend to son….but thanks for having a dialogue with me.

    • canute Says:


      Thank you for your comments.

      With regard to how the push is best implemented in practice, I leave the details to non-conscious brain processes. Provided we have developed the appropriate motor programs by past practice, it is best to let the non-conscious motor control system organize the recruitment of the appropriate muscles at the appropriate time.

      I believe that the most important thing when actually running is aiming for a short time on stance. I do this partly by visualization of running with rapid lift-off from stance and partly by monitoring the sensation of getting airborne. Achieving a short time on stance will ensure a strong push. When sprinting, I focus on high cadence and on driving my arms forcefully back and down in a sharp economical movement. I focus on my arms when running because I devote a substantial amount of time to drills that reinforce the motor program linking a down and backward drive of the arm with a down and back drive of the opposite leg.

      The main reason I am interested in the equations that provide estimates of the actual forces is that it gives me an estimate of the risks of injury. For example, calculation of the relationship between vGRF and time on stance confirms the wisdom of lowering the tension in my leg muscles when running down hill (except when I am specifically engaging in three quarter speed downhill running on a gently sloping grassy surface, with the goal of improving the coordination of the the muslce sinvolved in rapid lift -off from stance.

      The reason why I am interested in which muscles are engaged in getting airborne is to ensure that I put an adequate focus on strengthening the muscles that make a major contribution, and in particular, in ensuring that they function powerfully when acting as they are require to do during running. In fact a few years ago I targeted the hamstrings – and still do, but I have added an increased emphasis on gluteus maximus as a result of my speciation about the mechanism of the push. My understanding of mechanism also provides the impetus to perform exercises that promote elastic recoil. Because of recent episodes of inflammatory arthritis, I am cautiously experimenting with trampolining and at this stage am unclear how beneficial it is.

      With regard to your own running I accept that it is reasonable to be guided by your own experience, especially as you must have been doing something right when you ran a mile in less than 4 minutes. However, any athlete wishing to perform at the best level they can needs to continue to explore the possibility that there are ways in which they might improve. Furthermore, because of the tendency for muscle strength to diminish with age, it is important for an athlete who wishes to continue to perform well in middle age to be aware of which aspects of strength and power are most important for running safely and efficiently.

      I am delighted that you are willing to engage in debate about running technique. I do not aim to convert you to any predetermined style of running. In fact my major motivation for engaging in debate is to clarify my own ideas. As I have remarked before, one of the reasons I have focused a substantial amount of attention on Pose over the past 8 years is because I think it has many potentially useful features even though I consider that some aspects of the theory are contrary to the laws of physics and to what is known about how the body works. Perhaps ironically, despite Dr R’s statement that we should not push, one of the features of Pose is a short time on stance. Achieving this actually ensures a strong push. I have been intrigued by the possibility that some Pose drills might actually promote a strong push. It is plausible that achieving a strong push without actually focussing on pushing against the ground might allow the non-conscious brain to coordinate the timing of the push well. However I do not believe it is good to be too mentally detached from the push. That is why I focus on driving the arms back and down when sprinting.

      • jhuff Says:


        Thx for your explanations and the discussion. Until we speak again…enjoy your running 🙂

  4. Klas Says:

    I agree with most of what you wrote here. But I don’t understand why you attribute vGRF as a major role of gluteus maximus. I do agree that it seems to play an important role in stabilizing the torso, but I don’t see how it would significantly generate vGRF.

    Here is how I see it. In the braking part of stance, the torso would tip forwards without hip extensors stabilizing. But the actual hip extension is mainly an effect of momentum.

    If we are born to run, there are good reasons to suspect that we are optimized to run safely. We are probably also optimized to run while carrying stuff like weapons, children, food, water. Both running safely and with cargo suggest running with minimal oscillation of the torso. Vertical oscillation of the COM should be confined to the range of motion of the limbs that is a consequence of speed and relaxed running, as demonstrated by sprinters.

    • canute1 Says:


      I agree that it is easier to visualise how gluteus maximus stabilises the torso by arresting the forward pitch at foot fall, than to see how it exerts a downward push. However the main point of my post was to address the question of what muscles to play the major role in developing the vGRF that gets us airborne. I believe this is a crucial question if we are to understand the mechanics of running.

      You emphasise the importance of minimising vertical oscillation of the torso. I agree that it is possible to achieve a smaller vertical oscillation of the torso than the COG if at mid-flight the hip flexion of the leading leg, together with the knee flexion of the trailing leg, results in the COG being located at a somewhat higher point in the torso than is the case at mid-stance when the stance leg is only moderately flexed. Thus, as you imply, the range of motion of the legs can diminish the vertical oscillation of the torso relative to the oscillation of the COG.

      However, whether or not the torso oscillates less than the COG, the inescapable fact is that work must be done to elevate the COG. While the magnitude of vertical oscillation of the COG at first sight seems rather small (typically about 5 cm elevation from mid-stance to mid-flight) the calculations I presented on 16th January demonstrate that the energy required to achieve this is an appreciable fraction of the energy cost of running. Furthermore, those calculations demonstrated that we can reduce the combined cost of elevation and overcoming braking by exerting a greater push against the ground. Although this does increase the cost of elevation it produces a greater reduction in braking costs so there is a net saving. Thus, the inescapable realty is that if we wish to run efficiently we must elevate the COG. My post was largely about how we achieve this.

      As far as possible, we utilise elastic recoil to recover impact energy, but as is clearly demonstrated by the evidence reviewed by Alexander, this can provide at most 50% of the energy. So muscles must do mechanical work. This presents us with what is known as the extensor paradox. Recording of the electrical signals from muscles indicate that apart from gastrocnemius, the major muscles of the leg are only minimally active after mid-stance. So it appears that the relevant muscular contraction must be occurring mainly in the early part of stance. The electrical recordings suggest that gluteus maximus is the most active of the large muscles at this time. While this activity is in part associated with stabilization of the torso, it is mechanically plausible that gluteus maximus also exerts a downward push.

      Stand on one leg, with that leg slightly flexed at hip and knee. Rest your hand on gluteus max while you elevate your torso by straightening hip and knee. You will feel gluteus max contract. The quads and hams also contribute, but there is no doubt that there is a strong contraction of gluteus max. Thus, I consider that when the electromyographic evidence is taken together with the evidence that gluteus max appear to have undergone a major evolutionary development around the time that our forebears appear to have developed the capacity for endurance running, it is plausible that gluteus maximus is a major contributor to the work required to get us airborne. However I must emphasise that this is speculation.

      • Klas Says:

        I believe we should have the perception of running with hardly any vertical oscillation, since our perception is based on the torso.

        But I totally agree with you that the range of motion of the legs require more elevation of the COM the faster we run. This will also result in less braking the faster we run, which is important for efficiency in faster running, as your calculations show.

        I totally agree that this requires stronger and stronger push. We need to generate a greater effective impulse and in less time, so peak force increases significantly despite increased cadence. This happens reflexively when running in a relaxed way with good posture. Forcing a short stance time results in overdoing it, harmful and probably inefficient.

        I have no problem seeing how gluteus maximus can generate vGRF in that body position. But in running, they have another important job to do. And there is a very rapid hip extension by momentum at the same time, so it seems hard for the glutes to pull the thigh backwards in running.

        Nevertheless, it does not really matter. The glutes are very important for running, more so the faster we run. Training on a kickbike sounds like a good idea to me.

        I belive the vGRF is generated mainly by quads and calfs. They are active around mistance which is when vGRF is high.

      • canute1 Says:


        Thanks for your comment.

        I appears we agree on many of the important issues.

        However, I do not regard the range of motion as the primary determinant of the elevation of the COM (COG). The elevation is determined by the vertical impulse (average vGRF x stance time). Range of motion increases at high speed. However, the elevation of the COM does not increase at high speed (i.e. above 6 m/sec) because the decrease in stance time is associated with a lower vertical impulse despite the increased vGRF.

        I think the push exerted by G max is mainly a downwards push when the hip is moderately flexed in early stance. In this respect my view differs from Ken Mierke, who places greater emphasis on a backwards push. I accept that because G max inserts fairly high in the thigh it does not appear to have much mechanical advantage, though the fact that it inserts into the iliotibial band gives it greater mechanical advantage.

        I am quite willing to acknowledge that the full role of G max remains uncertain. It is falry clear that it plays a role in arresting the forward swing after mid-swing and in stbailizing the torso agaist forward pitch at foot-strike, but I think the timing of its action indicates an addtional role in generating vGRF vearly and mid-stance

        I accept that quads and gastrocnemius also play a role in elevation.

      • Klas Says:

        I think we agree in essence. We just look at it from different angles.

        The elevation is of course determined by the impulse. My point is that we create this impulse in order to balance the legs in the range of motion. The reduced braking is an important side-effect, cleverly designed by four million years of evolution.

        I strongly believe the reason why Weyand’s data shows a decrease in aerial time after 6 m/s is because a typical runner will have too high impulse until they approach top speed, when they are no longer able to make that error.

      • canute1 Says:


        Yes we are looking at the elevation of the COM from different viewpoints. I am attempting to look at the mechanism from the point of view of causal influences because I think this helps determine what things we should attempt to change in order that we might improve, either by conscious attention while actually running, or by appropriate exercises in training,
        I do accept that it is sometimes better to avoid focussing attention on the mechanism of running while actually running because in many circumstances conscious control of motor action is less effective than non-conscious control provided we have developed the correct motor patterns.

        While I do not disagree with making an effort to develop an efficient swing I believe that is less important than developing an efficient push to get off stance quickly. At low speed, the swing cost is only a minor fraction of the energy cost of running. Because swing cost increases in proportion to the square of running speed, the swing cost does become a major factor at high speed. However, the fact that elite sprinters achieve only a slightly shorter swing tie than poor sprinters, I do not think the gain from improving the swing are relatively minor.

    • jhuff Says:


      Where do you think the extra propulsion comes from besides elastic recoil and external forces?

      • Klas Says:

        The external forces are just the reaction to our contraction of muscles and tendons. (Every force has an equal and opposite reaction force.)

        The extra propulsion comes from our muscle contractions. That happens reflexively. When we have the perception of no active push-off, we still make a small reflexive push-off that we developed around the age of 3 when learning how to run without falling over. This reflexive push-off becomes stronger with higher speed. This is partly because stance time becomes shorter due to momentum, but also because we need to swing the knee up higher in order to balance the legs and have stable torso. This follows reflexively as a consequence of the desire to maintain speed with good posture and a folding knee during the first half of swing.

        The biggest propolsion we need is vertical, to get the flight phase that is the difference between running and walking, and to avoid inefficient braking. The only net horizontal propulsion we need in constant speed running is to overcome air resistance, which is small at the speed of running.

  5. Ewen Says:

    Canute, I’ve been thinking about purchasing a ‘kickbike’ for use in cross-training (although they’re expensive compared to regular bikes). It appears the kickbike action would work well to develop strong glutes and develop a good feel for hip extension. Your thoughts?

    • canute1 Says:


      I think the kick bike is a potentially a good form of cross training. It is also a practical means of local transportation –not as efficient for long distance travelling as the bicycle but maybe better in town.

      With regard to the cross training benefits, I think that strengthening G max is one of the big advantages. The range of hip motion is greater than running; I think that is good for hip flexibility. I suspect it is hard on the soles of your shoes but not too hard on your joints.

      For several years I have used an elliptical for cross training largely because if gives a reasonable cardiovascular workout without the impact of running. I also regard it as moderately good exercise for the G max , but I think the larger range of hip motion with the kick bike makes the kick bike a much better option than the elliptical for working G max.

      The other issue is whether you want to do your cross training outdoors. A big disadvantage of an indoor elliptical is that it is rather monotonous – in contrast the elliptigo is fun but terribly expensive. Maybe when the days are short and snow is deep it is better to be indoors, but that is not usually a big issue in Canberra. So I think a kick bike is a good idea – better for G max than an elliptical or an elliptigo, and less expensive – though not a trivial cost.

  6. Klas Says:

    I don’t mean to imply that efficient swing is more important.

    But most runners in the western world develop bad habits of too much tension in the run, for various reasons. We need to pay attention to a relaxed swing, allowing the knee to fold, until that is automated.

    We also need to pay attention to good posture, although that is mostly the result of training core stability and good posture in daily life.

    None of those things are harmful. Together they will also lead to a stronger push at higher speed, which is important for efficiency as well as to avoid dangerous overstriding.

    Focusing on a stronger push per se is dangerous.

    Training strength and plyometrics is a good thing, but a mental focus on a strong push in my opinion very bad advice.

    • Klas Says:

      One of the benefits of the Pose method is the focus on perception rather than mechanics. 95% of the population don’t understand mechanics anyway. The drawback is that those of us who want to really understand get confused.

      I want to fix that by describing efficient running in a way that is mechanically sound and does not lead people to the wrong conclusions.

    • canute1 Says:


      Thank you for your comments.

      I agree that there is a serious danger that focus on the push will lead to a mistimed push and also to counter-productive tension in other muscles. Very talented athletes like Usain Bolt can focus successfully on the push while also avoiding unproductive tension in other muscles. However I do not generally advise that approach.

      When I am sprinting I do focus on a strong, sharp and economical drive of the arms because i think we can control our arms more precisely than our legs – partly because the arms are lighter and partly because they are controlled by a much more intricate neural representation in the motor cortex of the brain. I have established a good automatic linkage between a down swing of each arm and its contralateral leg. However when running a moderate speeds I do not even think about driving the arms. I do not focus on a strong push at all, but I do monitor time on stance.

      With regard to Pose I think it is generally quite a good running style for non-elite athletes though I do get annoyed with Pose coaches who deny the forces that are involved in efficient running. I have encountered situations where it appears to me that Pose coaches have ignored warning signs of potentially serious injury, apparently because of a reluctance to acknowledge the forces involved. However I think that the situation is changing.

      Seven or eight years ago, in the advice given on Pose forums, any hint of injury was blamed on the runner’s poor technique without consideration of whether the forces associated with correct technique might contribute. I think this was both unhelpful with regard to avoiding physical injury and I believe also potentially demoralising. I have not been a regular reader of Pose Tech forums in recent years, but on occasional visits to the site nowadays, I find the general attitude is much less focussed on blaming of the novice. Although a great deal has improved, I still think it would be good if the technical pages on Pose Tech included a realistic discussion of GRF and also a balanced description of potential risks associated with with forefoot strike.

      One the positive side, the Pose emphasis on decreasing training volume and performing drills during the early stages is very good, though it presents a problem for the serious competitive athlete. This might be minimised by a more realistic acknowledgement of the muscular requirements and the risks, allowing the atlhete to address these in a well informed manner.

      Another feature for which Dr Romanov deserves credit is drawing attention to the benefits of addressing running technique. The recommendation to increase cadence rather than striving for a longer stride is of great value. I have been pleased to observe many recreational athletes enjoy a reduced injury rate since adopting Pose, and I think this is the recommendation that has helped most. This recommendation is not unique to Pose. Of the more recent approaches to ‘efficient running’ or ‘natural running’ I think that Mark Cucuzella’s ‘Principles of natural running’ offers the best overall package of advice. However the organization with which he is associated does not offer such a rich body of information nor the network of coaches that Pose offers.

      • Klas Says:

        Thanks. I don’t see any danger in focusing on a strong push at top sprint speed. It is almost by definition impossible to push off too much at top speed.

        Sprint running is stressful to the body, but luckily it does not last for very long.

        The danger lies in focusing on push off or short time on stance at distance running speeds.

      • Klas Says:

        And thanks for pointing me to Mark Cruzella’s site. I had not seen that before. Looks pretty good. I disagree with some of the statements about form, and some important points are missing, but overall it looks like a good attempt. And it seems to have an open scientifically minded approach.

      • Klas Says:

        Taking a closer look, I have to say that Cucuzella’s running looks a bit tense. A bit like an overpulling pose-runner.

  7. Hans Holter Solhjell Says:

    POSE theory and method is certainly not perfect (even if I have seen no better atempt at a theory of running and movement that combines physics and biomechanics) but it’s a big difference to state that pulling is responsible for vertical lift, as you stated, and to be incomplete in describing all elements contributing to vertical lift. Romanov says in his book on triathlon technique that muscular activity is involved various degree in all parts of running movement, but I can agree with you that he could be more precise in terms of describing the mechanism of vertical lift, and the muscular effort involved. But, and I think this is more important as my interest is to improve practice, and might in some way be Romanovs point though not well explained, the term push off might not be a good description of what is going on, and might, and can very often be seen, to create an image in the runner of an active, exesive push. Most likely with an idea to propel oneself forward by pushing, rather than just providing a quick vertical motion to create the potential for the next fall. This can even be seen in very accomplished runners. Finding an other term than push off is not easy, but I think nescessary as it creates a lot of confusion both in theory and in practical efforts. The movement involved is only a few centimeters, 30-50% of the effort is ideally done by elastic recoil, and the movement has to be quick and light, most likely it is also more of a reflex action than a conscious, active movement. Any active efforting can easily disturb both the ideal vertical movement and feed forward and disturb the whole coordination of running movement. The faster one tries to run, the worse this error and the feeding forward gets (the same happens with overly active “driving” of the arms) A good term should reflect all of these qualities and contribute to a better image of light and quick ground contact. My guess is that this is why Romanov downplays the role of push off and does not use this or any other term, but still has various drills that helps to great a well timed, quick and light vertical lift.

    • canute1 Says:


      Thank you for your reply

      I think we both agree that a deliberate push is often unhelpful. As stated elsewhere, I believe this is because control of motor activity by conscious effort is less precise than non-conscious control by automatic programs that have been developed by good practice. Maybe the word push is not a good word to have in mind when actually running. However, there is no doubt that a quite strong push must occur. Although 35-50% of the energy for this can be provided by elastic recoil, the calculations presented in my previous posts (eg Jan 16th) demonstrate that the energy cost of getting airborne is an appreciable cost of the total energy cost of running. Therefore, it is unwise to underestimate its importance.

      With regard to Dr Romanov’s description of running, I believe that he has presented a good intuitive account of some aspects of running but some parts of Pose theory simply violate the laws of physics (eg the theory of the fall following the Pose at mid-stance) and do not describe what happens in practice. While no scientific theory is correct –even Newton’s laws break down at both sub-atomic and astronomical levels – I think there is a high probability of drawing bad conclusions from a theory that is wrong about key facts of the situation.

      With regard to my statement about the Pose pull, I accept that I did not accurately describe the confused account presented by Pose theory on this topic. On the one hand Dr R says we must not push. On the other hand he says that all of the required energy comes for elastic recall, which is contrary to the evidence. Furthermore, he appears to conclude from the fact that the elevation of the COG is only in the range 4-6cm that it does not matter very much. Insofar as the cost of getting airborne is an appreciable fraction of the energy cost, this is simply wrong.

      With regard the role of the Pose pull, in Pose tech article 000157, Dr R states: ‘The Pull is really the only useful action in running. …Finally, the Pull does two things, first it allows for the change of support. Second, it attenuates vertical oscillation which accompanies the effect of the stretch reflex.’ I think his statement that the Pull acts to allow change of support does imply that it achieves a lifting of the weight of the body from the ground, but I also agree that attenuating the vertical oscillation associated with the stretch reflex Implies the opposite. Overall, in my view it is almost impossible to construct a coherent biomechanical theory from Dr Romanov’s statements.

      In conclusion, I do accept that there is value in encouraging runners to avoid unnecessary muscle action. This requires a proper mental approach. However, I do not believe that this is best done by under-estimating necessary muscle action. I believe that people like Moshe Feldenkrais made a good attempt to integrate realistic principles of engineering and physiology with psychology in formulating proposals for healthy human movement. I do not believe that Dr Romanov has succeeded in his attempt to avoid the pitfalls of overly ‘muscular’ approaches to running. As stated in my response to Klas, I think Mark Cucuzella has been more successful.

      With regard to driving the arms, as you say ‘overly active driving’ is likely to be unhelpful. I hope that by focussing on a sharp movement that is economical in its range of movement, I might facilitate a light, quick ground contact.

      • Hans Holter Solhjell Says:

        Could you elaborate on how you think the laws of physics are violated by Romanov’s theory, or point me to somewhere you have written about this previously? I guess you are refering to the falling stick model, where forward motion in running is seen as created by a combination of gravity and friction, similar to how the top of a falling stick, with one end fixed to the ground would fall forward? Of course the human runners also have legs that can be moved under the body to both create a new support to fall from, as well as to create some vertical lift. If you disagree with this model, which is the core of POSE theory, I find it very interesting to read more about. 

        In my current thinking I seperate between to different models of, ways to understand, running. Either one sees forward motion in running as a result of an effort to “push and drive to counquer gravity”, or as “flow with gravity”. Push and drive models mostly talk of active pushing into the ground, and about active knee drive/lift and/or active arm drive. 
        Most, or all, of the push and drive models does not really look at physics at all, and just seem to imply that pushing forward is something that has to occur. 

         I have no problem in seeing your point that the vertical lift is underexplained and maybe underemphasied in theory in POSE, and that there over time has been statements that has contributed to confusion. Although I do think you are overcharacterising this as a fundamental flaw, and I see these statements more as a problem with pedagogy, as I see as a more serious problem with POSE, than a problem of basic theoretical understanding. This problem of pedagogy I see even more so in Romano’s model of swimming than running. The Total Immersion model of swimming I see as vastly superior as a practical and pedagogical model of swimming, and much closer to a model like the Feldenkrais method, that I find it very interesting that you mention. But in TI there is no mention, or very little at least , of physics. 

        I am actually a Feldenkrais practitioner myself, and have found the Feldenkrais method very helpful in learning POSE, and I think POSE’ers would benefit greatly both in pedagogy and in the finer details of technique, awareness, proprioception, balancing tension/relaxation and so on from learning the Feldenkrais method, but have not found models of running in the Feldenkrais community that I find supperior to the POSE model. Moshe himself did not say much about running to my knowledge, and most models of running in the Feldenkrais community are variations of a push and drive model. Personally I see many similarities in Feldenkrais theory and POSE theory, but for running, and some aspects of most atlethic endevours, building balance, strength and mobility from the forefoot, I find POSE to have usefull ideas not emphasised or well understood in the Feldenkrais community, and Feldenkrais to be better for learning processes, awareness and many other aspects of movement. I actually see POSE running as being more in tune with the basic ideas of the Feldenkrais methods understanding of movement in relation to gravity than several ideas of running promoted within the Feldenkrais community. But then again, the Feldenkrais method is not primarily about running, and there is no problem in incorporating the ideas of POSE into the Feldenkrais method, with a reference of course.

        I did not solve my own problems with running until I, with considerable time and an amount of problems involved, got at least some of the elements of POSE in place  Or at least closer to POSE standard (partly with the help of some skype coaching from Jhuff). Most important for me was to reduce pushing, and also knee drive. And I have spent a lot of time figuring out how to reduce those. Both of these I did in a way that created serious problems. And in my experience now I really don’t see any need for or reason to activly drive the knee forward. Mark Cucuzella seem to emphasis both of these, and seem to have more of a push and drive approach both in theory and as shown in his running video, and while his speed and barefootnees is impressive, I did not particulary like the finer details of technique shown and explained. Personally I can now run a lot longer and faster than in my previous push and drive injury prone state. I am also a lot more relaxed in my running when I don’t have to deal with the excessive tension created by pushing and driving. I can now also run at least 5k in five fingers, and also some in barefoot, so I am pretty happy with what I have learnt from POSE so far. I will continue to experiment and see if I can build some distance in barefoot as well. 

    • canute1 Says:


      The statement (in ‘Pose Method of Running’) that by virtue of the fall after mid-stance, gravity provides free energy is either meaningless or false. There is no fall after mid-stance.

      The statement that no push is required (in Post Tech article 000229) is false. I accept that for many runners, especially for recreational runners, it might be unhelpful to focus on a deliberate push.

      I have discussed these issues at various times on my blog. My post of 14th Feb 2010 summarises the major issues, though I have developed these thoughts a little in the subsequent two years.

      As for push and drive models, while I accept that conscious push or drive might be unhelpful, I think that if we are to develop the best method of training and also run safely, we need to identify the forces that are involved in running. This is what I attempted to do in my post of Jan 16th and subsequent posts. A strong vertical push is essential to get airborne. A moderate forward push is required to overcome braking, though the need for horizontal push can be diminished by spending a shorter time on stance. One of the main points I have been making in recent posts is that even though shorter stance time demands a stronger vertical push there is a net saving because the reduction in work required overcome braking is greater than the increase in work to elevate the body, under typical circumstances. However the greater vGRF might increase the risk of injury. Because improved efficiency can be associated with greater risk of injury, I think it is useful to try to understand the mechanics.

      I consider that Moshe Feldenkrais integrated a sound understating of physics and physiology with an understanding of mental processes. From a runner’s point of view, it is disappointing that he did not address running technique. As I explained above, Dr R’s writings suggest that he does not have a good understanding of physics. However, I think that due to a mistaken under-estimation of the forces involved in running, Dr R has developed a method of running that diverts attention away a conscious push, and this has allowed many runners to reduce their injury rate. I believe that the three most important contribution to injury reduction are:

      1) Increasing cadence rather than trying to increase stride length. This is very sound but is not unique to Pose.

      2) Encouraging less conscious focus on pushing. Insofar as this minimizes unnecessary or mistimed muscle action this is very beneficial. I am inclined to speculate that Moshe Feldenkrais might have approached this issue by identifying the unnecessary or mistimed pushing rather than denying the need for any pushing, as Dr R did in article 000229.

      3) By discouraging a large volume training and focus on drills in the early stages. This might be acceptable for non-competitive recreational athletes. However, for competitive athletes I think it would be better to have a correct understanding of the forces involved and to train in a way that addresses these.

      I am pleased to hear that you are running better since changing to Pose. However be aware that the vertical ground reaction forces are as large with Pose as with any other technique that advocates a similar time on stance, and therefore a forefoot landing without allowing the heel to take some of the weight, is risky. I believe that this risk can be diminished by transferring a moderate proportion of the weight to the heel while on stance. Increasing distance very gradually helps, but I do not believe the risk is entirely abolished as indicated by Jack Becker’s metatarsal stress fracture long after he commenced Pose.

      • Hans Holter Solhjell Says:

        So do you completely deny the relevance of the stick model and that gravity in this way provide at least part of forward motion? Reading some other posts on your blog, and the lengthy comments back and forth, you seem to do so. Personally I have not been convinced, and have not found any writing that has been able to disprove this, and in my experience the gravity/stick model clearly does provide forward movement.
        A simple movement experiment shows this. If I stand on one leg, lean forward into a fall, and let the other leg fall forward and land under my COG creating a new support, I have now moved one step forward, with no pushing forward involved. If I repeat this several times, in a kind of funny walk, I have now both initiated movement and kept it up without any pushing at all. If I now do the appropriate movements of pulling and allowing vertical lift trough elastic recoil and undisturbed reflex action I am now running, and no active pushing involved.
        So, my experience tells me I can both start and continue both walking and running movement without pushing forward, and without active vertical pushing. I have also found during my speed training that I can accelerate a lot simply by adjusting my lean, allowing for a sharper fall, and allowing my arms and legs to coordinate around my weight shifting, rather than actively driving/pushing harder to run faster. My personal observation of myself, subjective of course, is that I can run faster and maintain speed longer as well as feel better this way, compared to a more effort demanding push and drive style.
        The reason I am interested in clarifying this topic is to further illuminate my own thinking and division of the various ideas on running into push and drive models, and flow with gravity models, and then further to develop my thinking about what language, underlying ideas, and practical implementations that go with each model, and also to see when a model are blending, or confusing the two basic ideas, and/or basically have no clear idea about physics at all.
        We, and you and others previously, have discussed the vertical lift issue at length, and seem to in some way agree, but clearly like to talk about this issue in different ways. I have found your explanation and your pointing out that elastic recoil might not be enough to provide all of the vertical lift clarifying and useful in my own thinking, even though I do not agree on certain specifics of what this means, and use of words to describe it. I believe the word push is not a good choice, and especially when combined with strong. But you have inspired my thinking in this regard, and I have found that i like the term “allowing light vertical lift”, and in the last couple of runs I have been trying to observe subtle differences in my running that create conditions for a quick light ground contact and light vertical lift, and I have found that a small adjustment in where I let my foot fall, closer to the mid line compared to what I used to do, helps, and that even subtle over striding reduces quickness and lightness.
        That there is risks involved also in POSE and other forefoot oriented models of running (as with all movement. Even not moving at all carries risks) I clearly agree with, and anyone who have spent some time on the POSE forum have seen several posts on calf and other issues. Calf issues was for a long time a big problem for me in learning the new style, but I am happy to say that they got solved, mostly, after my coaching sessions with Jhuff, where he through video analysis clearly showed me that I was pushing a lot more than I thought, even when standing still and moving my weight to the balls of my feet, as well as pulling late, and showed me various ways to work with this.
        But to solve the push problem I really had to invent my own Feldenkrais inspired ways of working with this, in a lot more detail than what I could find in the POSE drills. Although to be fair I only had a few lessons with Jhuff, so I am not sure what he would have told me going forward, so my comment is more on POSE drills in general. Overall most of my injury issues (I still have problems with soreness, and some muscle pain at times, but nothing close to the previous problems, and nothing that lasts more than some days, or stops me from running) that has appeared both before and during learning POSE has been solved by better movement quality, better technique, and moving closer to POSE standard. I am still pretty far of standard, but even so I have improved a lot, in several ways, injuries, speed, endurance, economy and enjoyment.
        If you are interested in the Feldenkrais method and running there are several options available. The classic is Jack Heggie’s running with the whole body, which is clearly a push and drive model, but there are several good ATM’s in his book and CD set and anyone can benefit from doing those, and many has also found, as I did, it helps their running, even though I at this point can see it led me in the wrong direction in a search for a solution to my problems. Nowaday’s I think Jae Gruenke,, is one of the most knowledgeable and experienced Feldenkrais practitiones dealing with running, and she is also familiar with POSE, but I am not entirely clear about her current position on the issues we are discussing here. Here is an article by her posted on,

      • Klas Says:

        Hans, my 2c on this debate of gravitational torque is that the burden of proof should lie on Pose.

        It is not enough to claim that gravitational torque might provide forward propulsion. The forward rotation must be cancelled somehow, or the runner will fall on his head. Pose has not explained how. It seems likely that it is cancelled by the backward rotation caused by the push that happens at the same time.

        In my experience, a simpler model is always better if it can explain the same thing. The push must happen, or there would be no flight. We can’t get away from that.

        Until there is proof that GT provides propulsion in running, I think we are better off assuming that it does not.

        The Pose method does not need GT.

      • Hans Holter Solhjell Says:


        I do not agree with you that the fall has to be canceled completely, if you by this means bringing the stick back into vertical, or stopped or canceled completely somehow by an opposite force. The fall only has to be balanced. We do this by letting the next leg fall to the ground to provide support, and the nervous system initiates various muscle action to provide both stability, dynamic balance and vertical lift, while momentum carries us over the sticky spot and into a new balanced fall. So this is more of a finely tuned continues balanced fall, rather than a complete start/stop/start/stop fall, a flat on the nose fall, or a push, push, push action that seem to be the alternative.

        Regarding proof, I guess all statements need proof, but if you are looking for definitive proof of anything, I think we have to wait a long time. Romanov in his latest book provides various empirical observations and calculations, and I myself provided a pretty simple empirical experiment in the post above. Can you provide an experiment or observation where you push yourself forwards, and rules out any gravitational torque contribution? It is not enough to push, which is easy to do, but the experiment must also rule out any contribution of gravitational torque and a balanced fall, as describe above.

    • canute1 Says:


      I do disagree with the stick model.

      You have provided a thoughtful attempt to describe your own experience of employing a fall to promote forward motion. I have two comments about this:

      First if you ‘lean forward into a fall’ as you say, and the height of your COG actually decreases you would have to push against the ground to recover the height you had lost. When moving at a steady average speed on level ground, the only vertical force that can lift the body is vGRF.

      Secondly, If you look a the diagram I drew in my post on ’Problems with Pose’ you will see that a realistic force diagram (based on data from a force plate) demonstrates that the resultant force acting on the COM after mid-stance is forwards and up, not forwards and down as depicted in Dr Romanov’s article. Observation of video recordings of runners demonstrates that runners do not fall after midstance. The COG rises in accord with my diagram.

      However, one of the issues in reaching a definitive agreement about the physics of Pose is that most of the language is not precise enough to allow direct confirmation or refutation. Although Dr Romanov says that the pull is the only important action in running, I do not think he provides a description of the pull that makes sense in the language of Newtonian mechanics. I think that the most productive way to understand the Pose pull is to regard it as a mental image rather than a mechanical description.

      If we address the question of what mental images are helpful (or harmful) I am quite happy to accept that the image created by the word push is likely to be harmful unless the runner has exquisite proprioception and muscular control. Newtonian physics clearly demonstrates that a push must occur but for many of us, thinking of pushing is likely to result in pushing for too long and furthermore is likely to lead to wasteful contraction of other muscles.

      Many of my posts have dealt with the Newtonian mechanics of running, becasue I think this it is foolish (at least for a coach) to be ignorant of the actual forces that occur. However, for the runner, the major question is what mental images will help him/her to run efficiently with only a small risk of injury. I do not have a definite answer. I am inclined to think that focussing on a arm swing with an economical range of movement is helpful – but that is only a speculation based on subjective experience.

      I am grateful for the links to articles addressing running from a Feldenkrais perspective.

      • Hans Holter Solhjell Says:

        I agree with you that there is upwards force and an absolute need for vertical movement, and that vertical lift is provided by elasticity and muscle action, regardless of if we call this a push or not. But I do not see how this denies the role of gravitational torque and the stick model. Let’s expand the stick man somewhat, and say we have a stick where the bottom part consist of a compressed spring. If we tilt the stick slightly forward and releases the spring the stick will have both the fall to propel the top forward, as well as push upwards. Obviously the stick does not know how to move with balance, but a human does, and can also let the other leg come foreward to create a new support for the next balanced fall as well as the next vertical lift.

        I do not agree, as written elsewhere on your blog and in various comments, that the fall has to be canceled completely, if one by this means bringing the stick back into vertical, or stopped or canceled completely somehow by an opposite force. The fall only has to be balanced. We do this by letting the next leg fall to the ground to provide support, and the nervous system initiates various muscle action to provide both stability, dynamic balance and vertical lift, while momentum carries us over the sticky spot and into a new balanced fall. So this is more of a finely tuned continues balanced fall, rather than a complete start/stop/start/stop fall, or a push, push, push action.

        Balance might be something I myself feel is underdeveloped in POSE, even if Romanov’s talks about this in his latest book and includes various balance oriented exercises. Personally I also prefer the term dynamic balance rather than Pose, and release weight rather than lean, as more useful mental images. Pull I think is good.

        So my formula is more “dynamic balance, release weight, pull” but obviously inspired by pose, lean, pull. Including “light vertical lift” could be an option, but I prefer to keep that as a secondary concern, background image rather than foreground image, useful for plyometrics and so on.

        I also agree with you that not all article’s and everything Romanov says or has written is complete or accurate in each individual case, and it seems to me that he has also developed his ideas over time, and the later writings are more accurate than the earlier, although I have not done a thorough historical study but read what you say he has written previously. In most cases I prefer to look at the latest edition, and latest book of a thinker, as this most often, but not always, is the most mature and precise expression of a set of ideas. I often see you referring to older articles, and his first book, rather than his latest book on triathlon technique, where he goes into more detail on the physics part of his theory. I wonder if you have read this book, and what your views are on it. Since you spend a lot of time writing about your views pros and cons of POSE, which I appreciate and learn from, I wonder why you prefer to comment these older versions, rather than commenting the latest and most in dept version. Where he also clearly states that vertical lift is needed and occur, and in his view is provided for, and one should train for, muscle elasticity.

        This seem to be a major point of several critics of POSE, and I am not really sure why, as POSE certainly includes vertical movement, provided by elastic recoil. The other aspect of your criticism, that push is underplayed and muscle action is needed and should be more focused more in various types of training I find easier to understand and is more a matter of degree and emphasis, and very different from saying that Romanov’s model does not include vertical lift.

      • Hans Holter Solhjell Says:

        Do I understand you correctly if I summarize your view on foreword movement in running that it is created by first, muscle action push to accelerate forward and to get airborne, then less intense muscle action but still forward push, plus vertical push, muscle action to maintain speed and counteract braking effects and wind resistance? Including momentum to maintain speed once at speed? Further acceleration while at speed needs more forward push, not increased lean?

    • canute1 Says:


      Thanks for your comments.

      If you read through the very long debate I had with Robert Osfield in December 2011 and January 2012 in the comments section associated with my calculations page, you will see that I reach a conclusion that has some similarity with the conclusion that you draw from your spring and stick model. I concluded that gravitational torque does produce an increase in head forward and down angular momentum in late stance. It is perhaps interesting that the calculations show angular velocity of the line joining point of support to COG actually decreases, but there is an increase in angular momentum because the moment of inertia increases. However this small increase in angular momentum does not support the Pose model because a decrease in head forward angular momentum occurs before mid-stance. In your spring and stick model, in early stance, when the point of support is in front of the COG, the stick is leaning back and gravitational torque tends to produces a head-back and down rotation. Even if we accept the concept that gravitational torque has appreciable effects during stance, it does not provide forward propulsion – though perhaps it can contribute to overcoming wind resistance.

      I criticise the Pose theory of gravitational torque because the erroneous theory is still presented on the Pose Tech website, and furthermore Dr Romanov’s more recent writings still contain contradictory and/or misleading statements. Even in the presentation on accelerated running presented at the 22nd ISBS conference in Limerick, Ireland, in 2009, in which Dr Romanov and his colleagues acknowledged that ‘gravity completes no net work during stance in constant speed running,’ there were errors and contradictions, which I discussed briefly in my post on Problems with Pose (14th Feb 2010) . I consider that Dr Romanov’s vague statement to Tim Huntley when Tim recently sought his opinion about Usain Bolt’s running style (see my post March 11th 2012) continues to sustain confusion caused by the original Pose theory of gravitational torque. Dr Romanov has not presented a clear description of the errors in his earlier theory, but continues to make statements that appear to endorse the notion that gravitational torque provides forward propulsion,

      As Klas pointed out in his response to your previous comment, Pose method does not need the concept of gravitational torque. I frequently point out the positive features of Pose theory and practice (especially for recreational runners) but I also believe that it is important to be clear about the aspects of Pose theory that are wrong, especially as the Pose Tech site still presents those errors and some Pose disciples still believe the original theory.

      With regard to your summary of my account of running mechanics, I would word it a little differently. At constant running speed, momentum is the main factor that maintains that speed. A vertical push is required to get airborne in order to minimise braking, while a moderate horizontal push is need to compensate for the braking that cannot be entirely eliminated, and to compensate for wind resistance. I have never said that one should not lean to increase speed. In fact I think leaning does help increase speed, though the propulsion comes from muscular push. The unbalancing produced by a lean helps elicit a stronger push, just as a lean promotes a strong push as a sprinter pushes off from the blocks.

      • jhuff Says:


        Could you elaborate on the idea of a “horizontal push”?

      • canute1 Says:


        After mid-stance the leg is angled down and backwards. The push against the ground has a vertical and horizontal component. The backward directed horizontal component of the push will produce an equal and opposite forward horizontal ground reaction force. This horizontal ground reaction compensates for the braking effect that occurred before mid-stance when the leg was angled down and forwards.

        Around mid-stance, gluteus maximus and hamstrings cease to fire. The mechanical effect occurs about 50 milliseconds after the neural firing, so the contribution from these muscles will cease shortly after mid-stance. Gastrocnemius continues to fire a little after the cessation of the contraction of gluteus maximus and the hamstrings so I suspect that gastrocnemius is mainly responsible for generating the horizontal component of push late in stance.

      • jhuff Says:


        Thx for explaining. It fascinates me that you believe that. I am so confused as to what technique you actually use for your running.

      • canute1 Says:


        Sorry if I appeared to be making too big an issue of material from ‘running mechanics 101’. Accounting for the horizontal push does require some careful consideration in light of the so-called extensor paradox.

      • jhuff Says:


        No worries, I am just thankful that I view the “push” as vertical. It appears we found another area to disagree about.

  8. jhuff Says:


    Found this article on technique: , is this more or less how you practice and think about running?

    • canute1 Says:


      Thanks for that link.

      I think Lauren runs with a good style. Her legs and feet function well. The range of forward and back motion of her arms is greater than I would recommend, while as far as one can see, the rotation of shoulders is less than ideal. However I think the relaxation of her upper body is good. Perhaps her head is very slightly too far forward but it appears well balanced.

      I disagree with many features of Dr Yessis’ description. One the one hand he criticises her for landing too far in front of her torso yet he also criticises her for too much upward motion. He does not appear to appreciate that you can only maintain a very short time on stance if you exert a strong upward push. He seriously underestimates the importance of the upward push. By making a stronger upward push the runner can spend less time on the ground and thereby minimise braking (and also minimise the need for horizontal push). I think that for 5000m pace, Lauren has got the swing action about right, though in fact I consider that the swing is not as important as the push from stance. I think Lauren is a better at running than Dr Yessis is at biomechanics.

  9. Hans Holter Solhjell Says:

    Thank you as well Canute. You have gone trough this several times I see so thank for your patience in giving answers. Regarding the discussion you are referring to, I could not find one with those dates and name. I wonder if it could be this one, ?

    • canute1 Says:


      Sorry, I gave you the wrong directions to the discussion with Robert Osfield about torque and angular mometum. It is in the comment section on the page entitled ‘Running: a dance with the devil’ which is accessible via the right side bar. I am afraid it is an extremely long discussion. The most relevant entry is my comment on 11th Jan 2012.

      In that comment you will note that I refer to preparing a full report on my computations. The first and second installments of that report are my blog posts on 16th Jan (on the effects of increasing vGRF on braking and elevation costs) and 7th Feb (on the effect of increasing cadence on braking and elevation costs). I have yet to do the detailed report on the computation of changes in angular momentum whilst on stance. I got side-tracked by the issue of the energy costs of limb repositioning, because I realised that when discussing the effects of increasing cadence, it is essential to take repositioning costs into account. I have just published a post on cadence taking account of repositoning costs, today. I plan to address the question of the effect of wind resistance before returning to the topic of torque and angular momentum. I think that the most relevant effect of ‘gravitational torque’ during running is the role it plays in compensating for the torque associated with wind resistance.

      Once I have completed reprots on all of the sub-sections of this huge topic, I will prepare a summary of the entire set of computations in a single document that contains links to documents giving details of the individual computations. The entire enterprise is likley to take me at least another 6 months, I am afraid. Meanwhile, the vigorous comments from readers are providing me with a good indication of many issues that are of interest to readers. I hope this wil make the completed ‘opus’ more informative.

  10. Hans Holter Solhjell Says:

    Hi again Canute.

    Again I am very impressed with the intellectual effort and stamina you are putting into this, although I am still not convinced that gravitational torque can be ruled out as a mayor or main contributor to acceleration and forward movement in running. The main point I got from your discussion with Robert is that it is extremely difficult to prove or disprove this by calculations and limited experimental data alone, even though you in the end ended up with a conclusion I will describe as surprisingly close to the POSE theory, even though clearly refuting it.

    On january 14th you wrote,

    “However, in the case of a runner, whose COG is forward of the point of support in late stance phase, the decrease in angular velocity is less than would be expected due to the change in body geometry resulting from extension of the leg, confirming that gravity does exert a turning effect that increases angular momentum in a head-forward and downwards direction, as proposed in Pose theory. However as I have pointed out many times, this increase in angular momentum does not provide forward propulsion. It is cancelled by an oppositely directed change in angular momentum in the first half of stance”.

    The first half of stance is what I in my little model above (not highly original) called the sticky point, and that momentum carries us trough this point, and I would guess would counteract the “backwards” rotational component you are describing. This speed, momentum, also carries into an accelerated start for the fall in the second half of stance. Also the backwards angle is newer as steep as the forward angle.

    I have no idea about how to calculate this my self, and guess you might have done so previously, or otherwise can tell me why this is wrong.

    I also think practical, observable and repeatable experiments might be a better way than highly complex calculations, with incomplete input and choices of framework, of determining some of these questions, and there are some very easy experiments everyone can do to get some concrete experiences with the questions in play. Also, calculations can only be deemed as correct in relation to the real world, not only mathematically, if they fit with empirical findings and experiment.

    Above I demonstrated one such experiment which can be repeated by anyone who are interested. This might seem like a very simple experiment and little high tech, but I think this is a hugh advantage, as the high tech analysis, like pressure plate data, while interesting, only shows you a small piece of a puzzle and can be interpreted in several ways. So, I present the first experiment again, and three variations on it.

    Stand on one foot, release your weight forward, let the airborne foot come to the ground, and see where you are. In most cases you will end up forward, meaning you have both accelerated and moved forward, simply by gravitational torque (or whatever word for a certain interplay of forces one would prefer too choose) and without push. There certainly is upwards push, grf, as the next foot lands (There might even be a minimal forward component, at a steep upwards angle, as I, most likely not very skilled, initiates the pull of the first leg of the ground, somewhere in there but at least in my perception it is highly unlikely to contribute significantly compared to the rotational component.This is not the lift by pull you have referred to, but an actual push I seem to be making to initiate the pull)

    In the first part of the experiment I myself by muscular effort stopped the forward motion. The second part of this experiment is to let the speed, momentum, created by the first fall, carry me over the support leg, releasing my weight into a new fall, and so on.

    In the third part of this experiment I stand on one foot on one step in a stairway. I release my weigh slightly forward and let the airborne foot land on the next downwards step. Even though there is vertical push from the second foot as it lands on the second, lower step, I challenge anyone to push forward from the first step.

    In the fourth part of the experiment I turn around in the stairway, stand on one foot and release my weight forward and let the airborne foot land on the next higher step (this can be done with a hamstring pull up to a certain height of the next step, coordinated with the release forward of body weight). I then release my weight even more forward over the upper foot and leg until all weight now is over this leg. If I get my body weight in the proper balance over the upper leg at some point it becomes very easy to create a vertical push. This push feels, almost reflexive, like something my body simply does as a consequence of finding the right balance and coordination (you might know a better word for it).

    I am now standing on the next step, having moved upwards and forwards, no horizontal pushing involved. Only following the formula from experiment one. I can continue experiment two also in the downwards and upwards condition (which I do every time I walk in stairs or hills. I particularly like the London (on holidays, as I live in Oslo, Norway) subway stations where one can take the stairs rather than the elevators. Perfect for practicing these principals both in walking and running) but I will leave that for now.

    I myeself at least can easily both create initial acceleration and continues movement this way. I am still not running, but by increasing the lean, and allowing for vertical push and the proper coordination, I will move forward, running both on level ground, downwards and upwards. I can also put up a video of this for further inspection and clarifying.

    There might of course be something I am missing here, and I might be doing something, like horizontal pushing, that I am not aware of, so if you can explain it to me, I am grateful. At least this is a framework for empirical analysis, and I guess there are equipment both for analysis of external forces as well as which muscles and so on are doing what work when.

    To counter these experiments, and the role of gravitational torque, helped by momentum, it should be possible, or necessary, for anyone who wants too rule out gravitational torque, and propose that acceleration is primarily a horizontal push action to create an experiment where forward movement is created primarily by push, continued movement is caused by either push, or momentum or both (depending on ones view on this) and gravitational torque is taken out of the picture. I think it would be far easier to create an experiment where you both have gravitational torque and push, than to rule out gravitational torque completely.

    I invite you to try out these experiments, as well as to design experiments as suggested in the latest paragraph, and/or explain what is problematic about the performed experiments in relation to the question of whether gravitational torque (as the term is commonly used and understood on this blog) is a contributor to acceleration and steady pace running.

    • canute1 Says:


      You express concern about the validity of the calculations I have presented, and point out that Robert and I have disagreed about some issues. There are some aspects of the calculations that are based on clearly defined approximations (eg the time course of vGRF) and some aspects where the approximations are less clearly definable (eg the computation of repositioning costs; and the estimate of the proportion of the work that can be recovered from elastic recoil). However there are major aspects of the computations that are incontrovertible if one accepts Newtonian mechanics. A vast body of evidence supports the extremely high accuracy of Newtonian mechanics in accounting for the movement of human sized objects moving at running speeds.

      Although Robert and I have our points of disagreement about some aspects of the problem, I think you will find we are in agreement regarding the account presented in my posts of Jan16th, Feb 6th and Feb 27th. You might note his comment on my post of 27th February: ‘Hi Canute, Thanks for another excellent essay on efficiency.’

      We also achieved near complete agreement (after much wrangling) in our long discussion in Dec and Jan in the comment section of ‘The Dance with the Devil’. I think our only substantial disagreement is Robert’s rejection of the term ‘gravitational torque’ to describe the cause of the changes in angular momentum about a pivot point on the ground.

      Your suggestion that linear momentum might somehow cancel out a change in angular momentum is counter to Newtonian mechanics. However, I accept that any mathematical account that is valid must account for experimental evidence provided the experiment is performed and interpreted correctly so I am happy to comment on the experiments you describe.

      When you movement forward as a result of slight unbalancing as you drop your foot after standing on one leg, you will move forwards. The height of your COM has decreased so it is indeed reasonable to say that gravity contributed to you forward movement, just as it accounts for the lateral movement of a felled tree. However to recover the height of your COM you must press against the ground. You acknowledge this, though Dr Romanov denies it in some of his articles that are still on Pose Tech (eg 000229). More relevant with regard to running, there is no evidence that the COM falls after mid-stance during running. The force plate data demonstrates that the net force on the body is forward and up. According to Newton’s second law, if a body is subjected to a net force that is forward and up, it accelerates in a forward and upwards direction. I think that in your second experiment you failed to record exactly what happened. In your fourth experiment, you describe the push as ‘almost reflexive’. I find this credible. I explicitly recommend that non-elite runners should avoid a deliberate push unless they have good proprioception and muscular control. However I do advise them to be aware the occurrence of the push so that they can prepare themselves to provide this ‘reflexive’ push effectively and safely.

      With regard to your invitation that I design an experiment that test theories of accelerated running, so far I have not attempted to propose a detailed account of the acceleration phase of running. For a distance runner, the acceleration phase is of very minor importance. Even in my discussion of Usain Bolt, I was describing the steady state (as was Dr R in Pose Tech article 000781). However, I believe that in acceleration from the blocks, the horizontal component of the push is the main force that produces acceleration, though gravity does play a crucial role. If there was no gravity, the initial forward and upward acceleration would launch the runner into flight. If he/she could no longer press against the ground he/she would not be able to accelerate any more. Running is a dance with the devil on the surface of the earth. It differs from intergalactic space travel.

      Because running brings us into forceful contact with the ground, it has risks. For both elite and recreational runners it is sensible to minimise the risks, though the preferable strategies are likely to be different in different circumstances. I think both Mark Cucuzella and Jae Gruenke have developed sensible approaches. Both erroneously recommend landing under the COG, but at least aiming to land nearly under the COG is better than reaching forward with the foot. Mark places a bit more emphasis on pushing; Jae focuses in a wholistic mental attitude. We all need to find what mental image works best for us. For many recreational runners, focus on posture, balance, and relaxation rather than muscular pushing is likely to be beneficial. Dr Romanov deserves credit for drawing attention to the potential benefits of less focus on a muscular push, but, as Klas says, Pose doesn’t need gravitational torque.

  11. Hans Holter Solhjell Says:

    Thank for your in depth reply again Canute.

    I noticed the high degree of agreement, and the long discussion about the term gravitational torque. And the conclusion at the end. I can not follow the math myself, but understand the conclusions and can relate them to real world experiences, experiments. But obviously might put things into words in a way that is less precise or even wrong from a certain perspective. And I appreciate you pointing it out, as I learn from that. As I see it the question of gravitational torque vs push has quite big consequences for choices in technique as well as mental images and pedagogy, so I appreciate your effort in this regard.

    I am not sure if cancel is the same as counteract in the terminology of newtonian mechanics. I wrote counteract and by that I mean that yes, there will be a backwards rotational component before mid stance, but real world experience tell me that if a stick, leaning back, is pushed with enough force, it will travel forward, past mid stance, and into a new fall. In my example in experiment two the acceleration in the first fall produces enough speed, and following momentum, to bring my body past mid stance and into a new fall. The acceleration from the first fall seems to produce enough speed to bring me past mid stance not only just barely, but with enough speed so the next fall is given an accelerated, flying start and is potentially faster than the first fall if not counteracted by me balancing it out.

    Possibly this is all an misunderstanding on my part, or it might not relate to running and my experiment number two at all, but on the face of it it certainly seems so.

    This is of course only correct if I do not perform an unnoticed (and still not well explained) horizontal push of some magnitude. I have actually tried the experiment while also pushing actively of, using the gastrocnemius, as suggested by you to play a role in horizontal push, and by others (not you) to perform a more active flick backwards of the foot, and can produce an fairly strong active push this way, but only at a fairly steep angle, not directly on the horizontal. And it also feels less coordinated, and more like a bouncy semi jump style running. But even if I do a lesser, more coordinated horizontal push that I am not aware of (as seem to be your suggestion, although still not explained exactly how this push occurs, as far as I have read. If you have written this out, please point me to the relevant pages), this light push does not rule out gravitational torque.

    First of all, momentum carries the body over the first half of stance, and over mid stance. No muscular push happens, as far as I know, before mid stance, at least not forward, so momentum is not helped in this part of stance, as a stick might be helped with a little flick if momentum was not enough to get over mid stance? I guess grf kicks in, muscular stabilization happens as well as loading of the elastic tissues. As I have understood it, grf is not the same as muscular push, and grf are present during the whole stance in various degree. So even if there is some hGRF, a fairly small amount compared to vGRF if I have understood it correctly, this does not mean the same as muscular horizontal push?

    But, I agree, there might be a light well coordinated push with an horizontal component in the second half of stance, but to me it seems more likely that this adds to the speed initiated by torque, and is not alone responsible for creating enough speed to pass over the first part of stance over mid stance. More like the flick of the stick mentioned over, not the main push. But I am not convinced that this is the case.

    Do I understand you correctly when you write that since GOM does not fall but rises after mid stance this rules out gravitational torque? But I can not see how vertical push, a rising GOM and gravitational torque can not work together at the same time, plus various other things that also happens, creating the dance with the devil. On a stick of unevenly distributed weight, the higher the GOM, the more unstable it is, and the more likely it is to fall. A rising GOM might maybe influence the rate of fall, but not cancel it?

    Also, a possible horizontal push seems to be at a pretty steep upwards angel, not anything like what a sprinter have in his starting blocks.

    Regarding your comment that running does not need gravitational torque, I would like to clarify one thing, as I am not sure of the correct terms here. As I understand it, gravitational torque happens as my GOM moves forward of my base of support, my standing foot. It can not not happen. Is this correct, or wrong? If it is wrong, how, or with what term would you describe what is happing when GOM goes forward of the support foot?

    I just tried a new experiment. Lets call it experiment number five.
    I stand on one leg, and put my airborne foot forward, but eliminate gravitational torque by adjusting my body weight so GOM stays in place over the support foot. What happens? Not much, I am stuck and even if I push as hard as I can I only go up. I can not move forward without releasing my body weight forward, allowing gravitational torque to happen. No gravitational torque, no forward movement. Not even push with a horizontal component. But I might misunderstand the terms here. At least, without releasing my GOM forward of my base of support, no forward movement. I guess this is the case both in the acceleration phase, and to maintain speed in the steady speed phase.
    Can you explain this to me?

    You also did not comment on experiment three, where GOM actually drops, going down one step. There almost certainly are no horizontal pushing in this experiment, but still forwards movement. And as we know gravity without torque can only be responsible for the downwards movement, not forwards.
    What do you suggest is going on here, and especially if we expand the experiment and goes forwards down several steps at a steady speed?

    Maybe the math does work out inside the framework and inputs used, but as it still does not match my experience and the suggested experiments, and the horisontal push is still more suggested than explained, especially in light of the extensor paradox, I find it hard to see a full “picture” or “story” that explaines forward motion better than a model that also includes gravitational torque.

  12. Hans Holter Solhjell Says:

    GOM, should be COG of course….

  13. canute1 Says:

    Hans, Thanks for your continuing comments.

    With regard to experiment three, there is a small horizontal push. When a tree is felled, it ends up with its COM displaced down and sideways. In fact as the tree rotates it exerts a push against the ground that has horizontal and vertical components. The horizontal component pushes the COM sideways. In your experiment three on the step, the sideways force was imperceptible. If the upper step was covered in ice, you might have been aware of a tendency for the foot on the upper step to slide backwards. The energy comes from gravitational potential energy, but there is no violation of the laws of physics because a fall occurs. However no fall occurs after mid stance during running, so such an effect does not occur when running.

    With regard to experiment five, if while maintaining your COM over your point of support, you flex the hip and knee of the stance leg and then extend them so that you exert a push against the ground that exceeds your weight, you will accelerate upwards. Simply pulling your foot towards your torso will not get you airborne as that would be pulling yourself up by your own bootlaces. If you were already moving horizontally at cruising speed (assuming you were sliding on ice so that you were not retarded by friction while your foot was on the ground), you will of course continue to maintain that horizontal speed. The main reason we continue to move horizontally when running is simply that a moving body will continue to move at a constant velocity unless acted upon by a force. That is Newton’s first law. The horizontal push when running is only required to compensate for braking, and to overcome wind resistance.

    The horizontal push it is far smaller than the vertical push, but it occurs. It can be seen in force plate data such as that produced by Cavagna and Lafortune. The peak magnitude is typically about 25 % of body weight. I provide a reasonably accurate sketch of this in my Dance with the Devil page. Provided one spends a large proportion of the time in the air, braking cost is relatively small. That is why the resultant force is much more upwards than forwards when running at constant speed.

    Gravity actually opposes the upwards acceleration, so the push needs to be larger than it would be on the moon. Once we have managed to get up to our cruising speed on the moon, it would cost less energy to maintain that speed than on earth. However, as discussed earlier, when accelerating we need to be pulled back to the ground so we can push hard against the ground.
    In summary, the force plate data provides irrefutable evidence that a push occurs during running. The data reveals that both vertical and horizontal ground reaction forces occur. The rate if initial rise of vertical force is greater for heel strikers than for forefoot or midfoot strikers, but the average forces are similar. These ground reaction forces are a reaction to a push by the foot against the ground (Newtons third law). Dr R claims that the push can be provided by elastic recoil, but direct observational evidence, consistent with our knowledge about how muscles and tendons function, indicates that elastic recoil can account for only about 35-50% of the energy required. So a muscular push must be applied. In the case of the vertical component of push, this is large. The peak push typically 3 times body weight at moderate speed, but can be much higher for elite sprinters.
    I think the more challenging issue is the question of the best mental image to evoke the appropriate muscle action. In general, for complex motor actions it is best to let the non-conscious motor control system in our brain regulate the recruitment of the muscles. The conscious brain need to provide the mental image that engages the non-conscious control system. I think we an elite sprinter, the best image is invoked by words like ‘drive’ but for recreational distance runners, an image of a quick light pull might actually work better, despite the fact that the required action is a push.

  14. Hans Holter Solhjell Says:

    Good morning Canute. At least it is morning here.

    I am not sure wether to use COM or COG, so since you seem to use COM, I will use this as well.

    I agree with you that there is both vGRF and hGRF and therefore a definite push.

    But this was not my question. I asked if vGRF and hGRF can occur without an active push movement, however well or poorly coordinated, made by my body, and I guess the answer seem to be yes, as the tree can not make a push movement of any kind, but can provide the stiffness and weight required for both v and hGRF, as well as gravitational torque to occur.

    Horizontal push movement made by the body, in addition to the GRF that also applies to a tree, or not is the main problem I am interested in clarifying, not GRF, which I fully accept plays an important role. (Romanov also talks about the importance of GRF in his 2008 book, even though you seem so have been able to find articles where he has managed to formulate himself in a way that can be interpreted otherwise)

    Neither does it seem like hGRF can occur without the forward angle, meaning the COM is forward of support, which also implies that gravitational torque is happening at the same time, simultaneously. You did not answer my question regarding this either, and I am still interested in reading your answer to this.

    My uneducated guess, although I have good non mathematical understanding of systems theory and the interplay of multiple interacting factors, and as pointed out by others in comments on this blog, is that hGRF and gravitational torque can not happen without the other occurring on the same time, and neither is the direct cause of the other, and they are both dependent on several other factors as well, like gravity, stiffness, and in my experiment, my intent to release my weight forward, and discontinuing of efforts to maintain vertical balance.

    So in experiment three there is v and hGRF, but no perceptible horizontal body made push movement, in addition to the stiffness and release of weight required for hGRF and torque.

    Secondly you did not answer my question of whether upwards movement of COM denies the role of gravitational torque. Let’s look at what happens if we put a very heavy object in the falling three, like King Kong, and let him climb upwards at high speed, at the same time as the tree falls. If King Kong climbs fast enough, and up to a certain point in the forward fall of the tree, he will contribute to the COM of the tree rising, while the top of the tree continues falling to the ground. So the rising COM does not cancel or in some way deny the gravitational torque. I am not sure wether it influences the rate of fall or not, but I am pretty sure, the higher the COM, the more unstable the system is.

    I have also had a new look at experiment one, where I simply stand on one leg and release my weight forward. It is both clear that a push is perceptible, and that there is vertical lift happening in addition to forward movement. The vertical lift was confirmed using a mirror as well. But what is the source of the vertical lift in this experiment, as I am making my best effort not to make neither a upwards or horizontal push movement? The mechanics of the foot seem to be the answer, interacting with gravitational torque.

    As I stand with my weight evenly distributed on the forefoot and heel, and release my weight forward and gravitational torque kicks in, my weight moves forward to my forefoot as my heel lifts. The stiffness of my foot, working as a lever, now forces my body upwards, a mechanical push powered by gravitational torque, without any need for a push movement (which does not rule out that imperceptible push movement does not happen, and slightly contributes, but there are high tech equipment that can tell us this more precisely. I am also sure others have described this lever effect, I just discovered it for my self here trough my practical experimentation. You might also have talked about it before, and made relevant calculations. Again, if so, please point me to the relevant pages).

    So, it seems gravitational torque and other interacting forces can both provide forward motion as well as contribute to vertical rise of the COM, without push movement occurring. We are not airborne yet though, we simply have a rising COM while the foot is in contact with the ground.

    Which brings me to experiment six.
    I stand on one foot, releasing my weight forward. As my weight is released forward my COM rises as a consequence of mechanical push created by the lever of the foot, powered by gravitational torque (v and hGRF is present as well but no additional body made push movement, jump). At the peak of the rise of COM, with my weight on forefoot and the heel elevated, and my COM also moving forwards, I now pull the foot of the ground with a hamstring pull.

    For a short time I am now airborne as well as moving forward. I have recreated the basic properties of running, without perceptible body movement horizontal push. To completely rule out muscle push we need more advance equipment, but we already have the data showing the extensor paradox, and the data you referred to for the gastrocnemius can probably be interpreted in various ways, not only as providing the main horizontal impulse.

    To me it seems likely that this lever effect contributes to vertical rise of COM and getting airborne together with muscle elasticity, further reducing the need for muscular push movement in both vertical and horizontal movement of COM.

    With regard to experiment five you did not answer wether forward movement can initiate without my COM being released forward of the support point. You point out that it is possible to create vertical push and get airborne vertically, as I myself did, but the question was forward movement, and wether gravitational torque is needed in running or not. If it is need to accelerate, it is needed, and can not be ignored.

    Also, if one sees a horizontal push as need to maintain speed, and/or to accelerate, you need a COM leaning forward over support, if not you are only pushing upwards, so gravitational torque is happing at the same time. As my experiment 1 tells me I can accelerate and move forward simply by gravitational torque and GRF, without extra body made push movement, and experiment 6 also shows me I can get airborne this way, and also knowing that acceleration is harder than maintaining speed, I can not help but conclude that gravitational torque can be a contributor, together with momentum, also in maintaining even running speed, and that the role of body made push movement can be very small.

    I have not yet been able to create an experiment that demonstrates forward movement by horizontal push, without also releasing my weight forwards of support, whereby I also introduce gravitational torque. So forward movement very much seem to rely on the COM forwards of support, which again can not help but introduce gravitational torque. Therefore, no gravitational torque, no forward movement, and no running.

    I still invite anyone to create a real world experiment, performable using ones own body, whereby forward motion is created and maintained solely by push, ruling out any contribution from gravitational torque.

    I would also suggest that gravitational torque also is needed in braking. What do we do to stop running fast? Adjust our weight backwards, which both create higher hGRF in the opposite direction (speculation, as I have no force plate data, but it seems likely) and backwards gravitational torque. As we slow down, we use remaining momentum to adjust the backwards lean into a more vertical position, and finally into completely vertical as we stop completely.

  15. canute1 Says:


    I usually use the term COG, but I started using COM in my recent discussion with Klas because he uses COM. I do not think it matters much, but I will try be consistent in using the term COG

    Yes, an object can generate vGRF due to its own weight (ie the effect of gravity) but this is of no use in running. Greater weight increases the energy cost of running. An object can also generate hGRF if it is unbalanced, and this can propel the COG sideways, backwards or forwards provided a fall occurs. I also think this effect is generally of no use in running on a level surface in the absence of wind resistance. I am still not completely sure about the implication of this effect when there is wind resistance, though I am confident that muscular work is required to compensate for the effect of wind resistance .

    A theory that denies that an active push is required when running conflicts with the laws of physics and physiology. When we run on the surface of the earth, gravity plays a large role, but as Dr Romanov and his colleagues acknowledged in their internally inconstant presentation at the 22nd ISBS conference in Limerick, Ireland, in 2009: ‘Gravity completes no net work during stance in constant speed running,’ There is some truth in many of the observations you have reported from your various experiments. Your observations do not alter that conclusion. Do you agree or disagree with that conclusion?

    The practice of Pose does not need the theory of gravitational torque.. I think that Robert Osfield’s term ‘smoke and mirrors’ is not too harsh a description to apply to many of Dr Romanov’s statements, including his recent misleading response to Tim Huntley. Despite the thoughtfulness of your approach, I think that you are in danger of being mislead by the smoke.

  16. Hans Holter Solhjell Says:

    I guess this is the full quote from the article you are referring to:

    “Gravity completes no net work during stance in constant speed running, but achieves angular work via a gravitational torque accelerating the COM in both constant speed and accelerated running.”.

    This fits very well with the performed experiments, and must therefore be seen to have a certain support in observable events in the physical world. I therefore can not see how it can be counter to basic physics. It is also easy to demonstrate and explain the string of events that makes forward motion happen if one includes gravitational torque, and very hard to do so if one tries do exclude it, as demonstrated in experiment five.

    I asked several very concrete questions relating to the described experiments in my two last post, as well as to your understanding of the physics involved, and you have so far chosen not to answer any of them, but refer to previous conclusions that my questions and experiments are designed to challenge.

    So I did not get that much from your last reply as previous ones. I fully admit that I can be wrong, but this has not been demonstrated trough argument, only previous conclusions are presented.

    I also admit that an active push can be involved in running, but I have also demonstrated in experiment six that forward acceleration as well a rise of COM and even getting airborn can be achieved without active push. And also that horizontal pushing is not possible without the COM forward of support, which demonstrate that gravitational torque at least interacts with a possible push.

    Your statement that a theory and practice of running does not need gravitational torque needs to counter this experiment, and provide some empirical observable sequence of events that demonstrates how forward movement can be created and sustained without gravitational torque being involved and therefore playing a role.

    So far I have enjoyed this conversation a lot and it has brought me several new insight as well as having been stimulating, so if I am proven wrong, by experiment that is, I still have learned a lot and am happy with the result.

    As for the smoke and mirror’s, I am trying to cut trough the smoke by relating my argument’s, or rather, drawing my arguments and conclusions from observable, easily repeatable, real world, whole body, whole movement experiments, and I think I have at least succeeded in raising several, still unanswered, questions and raised a couple of challenges to previous conclusions.

    I hope to be countered by arguments informed by other experiments with similar qualities, or be explained why my experiments are not relevant or otherwise problematic. This can only further my understanding and refining of the experiments.

    • canute1 Says:


      As I said, that article by Dr Romanov and colleagues is inconsistent. Perhaps for completeness, I should have quoted the inaccurate parts along with the accurate part. The second part of the sentence that begins with a correct acknowledgment the gravity does no net work in the stance phase, adds the claim that gravity achieves angular work via a gravitational torque accelerating the COM in constant speed running. The computation which I presented in my discussion with Robert in December and January shows that the angular velocity (in a head forward direction) of the line joining point of support to COM actually decreases during the second half of stance. Robert also demonstrated this using a somewhat different mathematical procedure. I believe there was an inaccuracy in the computations performed by Dr Romanov and colleagues.

      I have done my best to respond to the descriptions of your experiments, and have agreed that some of what you describe in these observations is correct, though as Simon points out (below), it is very difficult to exclude non-conscious reflexive muscle actions that provide a push against the ground. The reason why I do not think your experiments are relevant is that they do not address my two main objections to Pose theory:

      1) the statement made by Dr Romanov in Pose Tech article 00229 that no push is required during running.

      2) the implication of the classic Pose prescription: ‘pose, fall, pull’, there is a fall after the pose at mid-stance.

      If you agree that these statements are incorrect, I do not think we have any significant disagreement about Pose theory. If you do believe that they are correct, I am happy to do my best to respond to any argument that you offer supporting these statements.

      With regard to Pose practice, I have few objections except the way in which some Pose coaches underestimate the forces and the risks that are involved.

      With regard to the marketing of Pose, I disapprove of what I consider to be vague and misleading statements such as Dr Romanov’s analysis of Usain Bolt’s sprinting in Post Tech article 000791. I disapprove of the unjustified claim that Pose is a good way to achieve-elite level performance. I disapprove of the attempts by the Pose organization to encourage young runners who have realistic hopes of becoming elite athletes to adopt the Pose technique. However for older runners whose highest priority is minimising stress on their knees, I think Pose has some merit.

  17. Hans Holter Solhjell Says:

    Canute, can you point me to the relevant laws of physics and physiology that you write is defied if one denies and active push in running? I guess you here are referring to the active movement of pushing, and not the push from GRF, or mechanical push upwards from the lever of the foot.

    I am, as pointed out above, not denying that a push movement can play a role together with other factors, but would like to know more about your argument, and learn a thing or two about physics and physiology on the way.

    I am also a bit confused, as you several times refer to h and vGRF which led me to believe you put more meaning to it, but in your last post you seem to say that you think the practical role of these two are negligible. But at least I guess they must be taken into account, as they play a part in the whole composition of forward and upward movement.

    • canute1 Says:


      The relevant laws of physics are:
      1) The law of conservation of energy. If a body has the same gravitational potential energy at the end of each gait cycle, gravity has not provided ‘free energy’.
      2) The law of conservation of momentum. This law is equivalent to Newton’s first law of motion: ‘a body continues in a state of uniform motion in straight line unless acted upon by a force.’ If a body has the same forward momentum at the end of each gait cycle, then any loss of momentum during the gait cycle due to braking must be compensated for by a horizontal push. There is also a similar law of conservation of angular momentum.
      3) Newton’s second law of motion: Force = mass times acceleration. If the net force acting after mid-stance is forward and upward (as is shown by force plate data) the body accelerates forward and upward.
      4) Newton’s third law: ‘Action and action are equal and opposite’. If the force plate reveals a backward and down force exerted by the foot on the plate after mid-stance, the plate (or ground) exerts an equal and opposite forward and upward force.

      The relevant muscle physiology is that a muscle and tendon can only capture impact energy in the form of elastic energy if the muscle actively develops a force while it is being stretched. Thus the capture of elastic energy requires active expenditure of energy. Experimental observation (reviewed by Alexander (J.Exp.Biol.160,55–69,1991) demonstrates that at most 50% of the energy required for subsequent contraction can be provided by storage of impact energy as elastic energy.

      I certainly did not mean to imply that vGRF or hGRF are negligible during running. In general vGRF is much larger than hGRF. When time on stance is short so that the braking effect is relatively small, the work done generating vGRF (which is required for elevating the body) is greater than that required to generate hGRF (which is required to compensate for braking). Conversely, when time on stance is long, the work done overcoming braking is greater than that required to elevate the body. Even in this circumstance peak hGRF is smaller than peak vGRF but it acts over a greater distance so more work is done to compensate for braking.

  18. Simon Says:


    Like you I have tried to make sense of how a ‘fall’ could work in running. I have argued the case both for and against many times in a few places to gain greater understanding from both the proponents and detractors of the idea.

    The simple inescapable fact is that for gravity to do any kind of net work, the COG needs to decrease in height. As the COG of a runner has a mean height that is unchanged on a level running surface, it is clear that gravity can do no net work and so must have zero net effect on propulsion.

    If you can accept that fact (and this is the only way I have found to make sense out of complex scenarios), you will find you can figure out why gravity may be ‘propulsive’ under some circumstances, but the propulsion must be paid for with an opposite torque and corresponding opposite linear acceleration at another point. So even if there were a propulsive phase which is mainly gravity based (the heavy acceleration at the start of a sprint race may have the right conditions for this), you will find that there is an equal and opposite effect that would completely neutralise the propulsion if you look hard enough.

    Whilst most people who are sympathetic to the falling stick model can grasp gravity being neutral when resolving the vertical position and forces, it just takes the same logic to see that it is the same with horizontal and rotational forces too.

    So when you look at real world experiments, like a single step ‘fall’ you need to be careful of what is happening. A single step experiment actually proves the neutrality of gravity very nicely if you are careful not to add any muscular effort i.e. make sure you do not extend your leg at all as you fall forwards – fall with a straight leg.
    You will then find that when you fall forwards, COM height decreases, you fall like a stick and then you plant your other foot. At this point you have to be very careful not to use your leg muscles – a straight leg is the only way to not accidentally ‘cheat’. As the foot is ahead of support, there is complete counteraction of forward motion as the COM tries to rise against gravity, stalls a little before getting to vertical and starts to fall backwards. The net effect is that you cannot quite complete a single step unless you help it along somewhere using your muscles.

    To then see how the action of running relates to falling, instead of doing the above on a straight leg, start with a flexed leg and extend it as you ‘fall’. If done in balance with the movement, you can make it so that the COM does not decrease in height. You have now switched from driving the movement with gravity to driving the movement with your leg muscles; that leg extension provides the small amount of force necessary to shift your mass horizontally. You can provide a new support point with the COM at the same height and perpetuate this movement, never changing the COM height and so never working with or against gravity.

    Real running is obviously different and more complex, but the fundamental principles can be applied if you explore it with an open mind.

    Physics aside, I think the fall is a nice mental image for relaxed running and have seen it used to good effect. I often use it when I find myself stiffening up or overly muscling it.

  19. Hans Holter Solhjell Says:

    Thank you Canute and Simon.

    I agree that it is difficult to exclude imperceptible reflex muscle push action in the experiments. This was also my own main objection to these experiments, but at the moment I can not see how this rules out all of the points and questions I made with the experiments as an exit point.

    Just to clarify my own understanding, I ask you to give exact answers or comments to these questions, that are based on observations in the experiments, rather than general comments about laws of physics. Of course the laws of physics and physiology must be obeyed and any explanation must be in accord with these laws, so what I am interested in is knowing exactly how these laws relates to the observations from the experiments.

    1. First of all, in experiment one, I can see that my COG first rises and moves forward, not falls, as I release my COG forward of support, and this is mainly due to mechanical push due to the lever effect of the foot and gravitational torque, even if a slight push movement might occur as well. Correct or not?

    After the initial, perceived, rise of COG, there is perceived fall, drop, that is sharper than the initial rise. How does this lever effect rise and following drop factor in in the overall equation? I would guess that the calculations are somewhat different than for a regular stick, and the pressure plate data would be somewhat different as well. Or does the lever effect not exist, and is actually a push movement?

    More precisely what I observe is my hip and everything above first move upwards relative to a line on a mirror I am looking at. My COG might actually fall even though the visual impression is of the body moving up relative to this fixed line, as well as giving a distinct up feeling, and a following sharp drop, but I am not able to see how this can occur myself. I can not see or feel a significant push movement being made, but can of course be wrong here as well.

    2. In my observation of my self in the mirror while doing these experiments, it looks like my foreword angle is quite a lot steeper than my backward angle, and my support foot stays in contact with the ground for a longer distance after mid stance than before mid stance. My initial guess is that for the backward torque to even out the forward motion created by the forward torque it needs to be of a certain length, not to much smaller than the forward traveled distance. Correct or not?

    3. Also, in relation to experiment one, I ask that if a push movement is to give me an horizontal acceleration, my COG needs to be in front of my base of support. At this moment, forward gravitational torque is working as well, and if it was not, any push could not carry me forward and initial acceleration would not happen. Correct or not? I guess one could argue here that the angle is what is import ant, not torque. But can the angle be created without torque in the first place? One can lean forward one foot in front in a static position, or supported on the hands, as a sprinter I guess, but in most situations forward movement is initiated from the vertical position.

    4. In experiment 3, going down the stairs, is the contribution of gravity, via torque, larger here in the horizontal direction since COG now obviously moves down, and the need for horizontal push less?

    5. In experiment 4, going up a step, I am pretty sure I can perform this without horizontal push movement, and only weight shifting by releasing my weight forward from the first standing leg fully over the second one, one step up, and then performing a vertical push upwards to stand on one leg on the second step. Meaning I have forward and up movement, quite slow of course, but still without horizontal push movement. Comments?

  20. canute1 Says:

    My first comment is until I know what you are trying to demonstrate by these experiments, any comments I make have only a limited chance of being relevant to your train of thought. However I will offer comments on what I consider are the key features of each of your experiments, attempting to answer your multiple specific questions as I proceed.

    *Experiment 1: ‘Stand on one foot, release your weight forward, let the airborne foot come to the ground, and see where you are. In most cases you will end up forward, meaning you have both accelerated and moved forward, simply by gravitational torque.’
    Although you almost certainly made a small push to unbalance yourself, gravity provided the energy for the fall and the forwards acceleration. Since the COG has fallen, this interpretation does not violate any laws of physics. Without any force to elevate the cog back to its starting height, this experiment does not in itself does not address the validity of Pose theory.

    *Experiment 2: ‘let the speed, momentum, created by the first fall, carry me over the support leg, releasing my weight into a new fall, and so on.’
    If your COG returned to starting height before the second and subsequent fall, you must exert a push against the ground. To propose that the COG can be returned to its starting height (and therefore to recover it gravitational potential energy) without application of an upward force would violate the law of conservation of energy.

    *Experiment 3: ‘I stand on one foot on one step in a stairway. I release my weigh slightly forward and let the airborne foot land on the next downwards step.’
    You probably made a small push to unbalance yourself, but gravity did most of the work. The price you pay for this movement is a large loss of gravitational potential energy.

    *Experiment 4: ‘I turn around in the stairway, stand on one foot and release my weight forward and let the airborne foot land on the next higher step (this can be done with a hamstring pull up to a certain height of the next step, coordinated with the release forward of body weight). I then release my weight even more forward over the upper foot and leg until all weight now is over this leg. If I get my body weight in the proper balance over the upper leg at some point it becomes very easy to create a vertical push.’
    As you say, it is easy to generate the upward push. Your muscles (mainly the hip extensors assisted by the knee extensors) do the work of elevating your COG. You have also moved forwards. This requires very little energy compared to that for lifting, but your foot on the lower step probably exerted a slight horizontal push; alternatively, as your upper foot landed, the hip extension pulled the torso forward over the leg. As the amount of work required for the horizontal movement was a very small fraction of that required for the elevation, it would be virtually impossible to perceive this small additional load consciously.

    *Experiment 5: ‘I stand on one leg, and put my airborne foot forward, but eliminate gravitational torque by adjusting my body weight so GOM stays in place over the support foot. What happens? Not much, I am stuck and even if I push as hard as I can I only go up. I can not move forward without releasing my body weight forward, allowing gravitational torque to happen. No gravitational torque, no forward movement. Not even push with a horizontal component. But I might misunderstand the terms here. At least, without releasing my GOM forward of my base of support, no forward movement. I guess this is the case both in the acceleration phase, and to maintain speed in the steady speed phase.’
    As you say moving your foot forward while leaning back does not produce any unbalancing. The only GRF is that generated by your weight. This balances the downward force of gravity. You remain in equilibrium and do not move. To move you need to unbalance and allow yourself to fall as in exp 1 or you employ your hip and knee flexors to push yourself forward and up as Newton’s laws tell us you must do in exp 6. If you were already moving horizontally at constant speed (for example, skating friction-free on ice) this posture would allow you to continue to move horizontally. However when running we also need to perform elevation and overcome braking during each step.

    *Experiment 6: ‘I stand on one foot, releasing my weight forward. As my weight is released forward my COM rises as a consequence of mechanical push created by the lever of the foot, powered by gravitational torque (v and hGRF is present as well but no additional body made push movement, jump). At the peak of the rise of COM, with my weight on forefoot and the heel elevated, and my COM also moving forwards, I now pull the foot of the ground with a hamstring pull. For a short time I am now airborne as well as moving forward. I have recreated the basic properties of running, without perceptible body movement horizontal push.’

    The initial elevation of the COM was not powered by gravitational torque. According to Newton’s laws, the COM an only be elevated by an upwards force. If your COM rose, you must have pushed down. Perhaps the hip and knee of the stance leg were slightly flexed at the start, allowing extension to produce elevation. If you COM moved forwards as well as upwards, some active horizontal push must also have occurred. I suspect that your stance leg was angled slightly as you elevated you COM and the backward angle allowed the hip and knee extension to generate a hGRF. You have created some of the elements of running. You might not have perceived the forces you exerted because you did not move far or fast, but Newton’s Laws tell us that these forces must have been produced.

    In summary, in experiments 1 and 3, the COM ended up below its starting height, and gravity did most of the work. In exp 5 you remained in equilibrium in accord with Newton’s laws. In 2, 4 and 6 Newton’s laws tell us you must have pushed against the ground, even though in most instances, the push might have been imperceptible. The fact that you found it easy to elevate your COM in exp 4 despite the fact that you raised the COM by a relatively large amount (step height is more than the typical elevation from mid-stance to mid-flight when running) demonstrates that it is fairly easy to exert the forces required for these movements. Perhaps the conclusion should be that running slowly is fairly easy (at least for the few thousand steps) and recreational runners need to be encouraged to relax.

  21. canute1 Says:


    I have been trying to get inside your mind to understand what question you are trying the answer with your experiments. Maybe your question is: If gravity provides energy for horizontal acceleration when the COG falls as an object moves forwards, why do we not say it provides energy for horizontal acceleration when the COG rises as it moves forward. If that is your question, consider three further experiments. Because you have lived on earth under the rule of Newtonian physics for many years, I think you will be able to imagine the results without actually doing these experiments.

    1) Place a billiard ball on a very smooth flat surface and strike it with a cue. The ball rolls forwards with almost no slowing. (If the surfaces of ball and table were perfectly smooth and rigid, it would not get slower.)
    2) Jack up one side of the table; as the ball is placed on the table, strike it with the cue in the downhill direction. It will accelerate as it rolls. Gravity provides the energy for this.
    3) As the ball is placed on the sloping table surface, strike it with the cue in the uphill direction. It will get slower as it rolls.

    In these experiments, the cue corresponds to the runner’s initial push against the starting blocks. In exp 1, no further push was required to maintain speed. You could say the ball continues under its own momentum in accord with Newton’s first law. Gravity does no work
    In exp 2, gravity does net work as the COG falls.
    In exp 3, gravity does net work in slowing the ball. This situation resembles the situation in late stance when the body is being propelled forward and upwards by a resultant force that is directed forward and upward, just as the cue strike and table reaction generate a resultant forward and upward force that propels the billiard ball forward and upwards. It is true that gravity contributes to the resultant force, but its effect is to slow the forward and upward acceleration. It would be misleading to say that gravity provided ‘free energy’ to promote the forwards propulsion.

    These experiments are not running but they illustrate relevant Newtonian principles in action

  22. Hans Holter Solhjell Says:

    Thank you again Canute, for your generous answers.

    I will reply a bit more when I have more time. But since you asked what question I am trying to answer I have given it a bit more thought. One question I at least is trying to answer is, how little push is needed. I am also interested in relating this to push and drive models of running, vs a pull oriented style. I will expand more on this later. Several of your comments touched in on this as well.

    • canute1 Says:


      I look forward to your further comments. My own thought on how much horizontal push is required is that it is the amount required to overcome braking. If we could spend zero time on stance, there would be no braking and no need for a horizontal push to maintain the forward momentum. However to completely avoid braking we would need to exert an infinite upward push.

      The crucial point about braking is that it occurs when the point of support is in front of the COG. You have previously noted that the duration with support in front of the COG is shorter than the duration with support behind the COG. So far I have not addressed this in my writing. I will do a blog post on this fact soon.

      You might also note that I have added an additional sentence to the discussion of experiment 3 with the billiard ball to emphasize the similarity with late stance.

  23. Hans Holter Solhjell Says:

    A new short question.

    Let’s say we take a half circle, with the bowl shape up, and place’s steel ball on the top left end, holding it with a finger. If I let go of the ball, it will role down and a fair distance up the other end, but not over the top. If I give it a hard enough push it will go over the top.

    If we cut of half of the right part the ball will fly over the top and forwards even without a push, but obviously not to the same height as on the left side. If we use a magnet too lift the ball up to the same height, simulating the vertical lift that happens in running, and place a new cut of half circle here, the ball will role down and over the top again. The energy input comes from the magnet to perform the vertical lift, and the forward movement is powered by gravity, as well as the gradual slowing of the ball going up the right end, and over the top.

    I guess this is somewhat similar to what would happen when the forward torque in running works longer than the backwards torque? The forward movement is

    In your reply to experiment one you noted that a push is required to move me out of equilibrium. As I have understood it this is the case in an object with a stable equilibrium, but not in the case of an object with an unstable equilibrium, like the human body. All that is required for me to fall forward is to stop balancing by reducing the stiffness of my ankle joints, while maintaining enough rigidity in my body to not collapse straight down. This seems like an important point, but I might have understood it wrong. But if it is right it is worth considering in the overall picture, and in construction of experiments. It would be similar to your experiment nr 2, with the downwards rolling ball, but no need to strike the ball, if it’s starting point is in the slope. It also affects the interpretation of all my experiments, whether there is push input in the start or not.

    • canute1 Says:


      I agree with you that the ball will roll over the lip of the bowl. I disagree that the horizontal propulsion can be attributed to gravity. I note from your subsequent comment that you also have recognised that in fact the energy is provided by the magnetic fleid, but I will nonetheless provide the response that I had prepared before I saw you later comment, as I think this is a thought provoking experiment that justifies careful analysis.

      As the ball descends the left hand side, the horizontal force that produces the acceleration is not gravity, it is the bowl reaction force. This bowl reaction force is generated by the pressure of the ball against the bowl. However it is true that the energy required for this acceleration is provided by gravity. A fixed passive object, the bowl, has redirected the kinetic energy imparted by the loss of gravitational potential energy.

      On the ascent, in the presence of an upward magnetic field, the bowl reaction force is decreased. By the time the ball reaches its initial height, the net right to left horizontal impulse delivered by the bowl reaction force is less than the left to right impulse delivered during the descent, so the ball retains a net horizontal velocity. Where has the energy come from? Because the ball has returned to the initial height, it has the same gravitational potential energy at the beginning. So gravity has done no net work. In fact the magnetic field has provided the ball with kinetic energy. In summary, by appropriate adjustment of vertical forces (gravity and magnetic field) that cause the ball to exert pressure on the bowl, we have created a situation in which the bowl imparts a net horizontal impulse to the ball. The kinetic energy that remains once the ball has returned to its initial height has been provided by the magnetic field. Although this is rather thought-provoking experiment, in fact it is a quite common occurrence for the reaction force generated by pressure on a fixed passive object to redirect kinetic energy. The crucial issue here is identifying the source of the kinetic energy.

      Note also that if the bowl were resting on a frictionless surface, the bowl would have moved to the left during the descent and to the right during the ascent, so that there was no horizontal displacement of the combined COG of ball and bowl. When the ball reached the right side lip, it would have been moving vertically as a result of the increased kinetic energy imparted by the magnetic filed.

      With regard to the fact that the period for foot strike to mid-stance is less than for mid-stance to lift off, note that in the absence of wind resistance, the law of conservation of angular momentum dictates that any angular momentum generated before mid-stance must equal that generated after mid-stance or the runner would rotate onto their face (or back). In estimating the consequence of differences in the duration before and after mid-stance it is necessary to take account of the fact that the average value of vGRF is greater between foot strike and mid-stance. Also be aware that the change in angular momentum is not due only to gravity. In fact the net force acting on the COG is upwards during most of the period on stance. The head forward rotation rate increases before mid-stance and decreases after mid-stance. This is the opposite to what Pose theory suggests on the basis of an under-estimation of the upwards vGRF leading to a mistaken assumption regarding the direction of the net force acting after mid-stance

      • Hans Holter Solhjell Says:

        You put the various physics factors into words far better than I am able to. Though for the question I was trying to find an answer too I don’t think it matters much if the horizontal movement is caused by gravity, or by the bowl’s reaction force. Even so, It is a more precise and clarifying formulation.

        What I think is relevant here, is that the horizontal movement happens without the need to simulate a horizontal push movement performed by the ball. The magnet is a stand in for the body’s capabilities to create vertical lift.

        I am not entirely sure from your answer if you are saying, that even so, due to other factors, this does not apply to running?

        Also, you write,
        “the law of conservation of angular momentum dictates that any angular momentum generated before mid-stance must equal that generated after mid-stance or the runner would rotate onto their face (or back)”.
        I guess this is definitely true for an inanimate object that has no ability to balance. Does it also hold for a living organism that can balance itself in relationship to the various forces acting on it?

        You have mentioned this before,
        “The head forward rotation rate increases before mid-stance and decreases after mid-stance”.
        Can you explain why and how this happens?

      • Simon Says:

        Hi Canute,

        Couple of things you wrote there that I don’t think are quite right, not a big deal but probably worth clarifying.

        Firstly, you say that the magnetic field supplied the ball with kinetic energy which is not really true. Gravitational potential was used to created the kinetic energy and the magnetic field restored the gravitational potential energy so the final state was a ball with kinetic energy and the same amount of gravitational potential energy it started with.

        Secondly, “the law of conservation of angular momentum dictates that any angular momentum generated before mid-stance must equal that generated after mid-stance or the runner would rotate onto their face (or back)”. This is not quite true – you are making assumptions about the kind of torque that can be produced in running or assuming that there is no net acceleration. A push in late stance produces a head back torque, and as we see in the special case of sprinters leaving the blocks, angular momentum can be conserved completely with no landing ahead of the COG. A similar situation would probably exist if running into a strong enough headwind.

      • canute1 Says:


        Thanks for the clarification

        I agree that it is more precise to say that gravity provided the initial kinetic energy and the magnet restored the gravitational potential energy.

        My statement about the equal rotational effects before and after mid-stance refers to running on a level surface at constant speed in the absence of wind resistance. In general, that is the situation I have been describing in my posts on the mechanism of running, and occasionally I fail to include that modifying clause in my statements.

    • canute1 Says:


      I think the situation is different during running. The ground is flat, so the ground reaction cannot generate a horizontal impulse in the same way in the ball and bowl experiment. Because the line from COG to point of support is not vertical except at mid-stance, a push acting along this line generates a horizontal reaction force. This has a braking effect before mid-stance and provides forward acceleration after mid-stance.

      You do not need to think about generating a separate hGRF and vGRF. You simply need a push along the line from COG to point of support. Near mid-stance, when this line is near vertical, the push can be provided by gluteus maximus and hamstrings acting in concert with the quads. (This push is partly achieved by elastic recoil.) Late in stance when the line from COG to support is further from vertical, I think that gastrocnemius (and recoil of Achilles and plantar fascia) play a bigger role.

      With regard to conservation of angular momentum this applies even for an animate body. Internal re-arrangements of body parts do not change angular momentum of the body around a pivot point.

      The reason that the head-forwards angular velocity of the line from support to COG, around the pivot point at the point of support, decreases after mid-stance is that the vertical component of resultant force acting on the body is directed upwards, rather than downwards, during most of this part of the gait cycle.

      • Hans Holter Solhjell Says:


        I look forward to your blog post on the relationship between forwards and backwards torque, as I think I read you said you will write. In the mean time I have some thoughts on this my self.

        As you say the ground is flat, so the ground can not generate the impulse to move. But as I noted some posts ago, the standing human body is in a state of unstable equilibrium, and the forward impulse, initiating the torque, can be generated by the stored kinetic energy of the body, If the human decides to reduce the effort to balance.

        Also, what I meant to demonstrate with the steel ball experiment is that since the left side fall of the ball, directed in the horizontal by the GRF of the circular shape, is twice as long as the right side rise of the ball, the ball will jump over the right edge, and all that is required for it to repeat this is for us to provide vertical lift.

        This I would say is similar to what happens in the relationship between the longer torque after mid stance, and the shorter torque before mid stance. Therefore I look forward to reading your thoughts on this.

        In my thinking at this point, this means in my experiment one, I can initiate the forward movement without push. All that is needed to produce one full step is for me to provide vertical lift. Therefore restoring my COG to its previous height, coming into a new state of unstable equilibrium.

        What seems problematic here is that the highest point of COG is not at mid stance, unlike the steel ball which is at it’s highest point at the start of the half circle. But the ball is not suspended over its base of support, and on a flat surface it would be in a state of stable equilibrium. Since the forward torque of the human body is not affected by the rise of COG, and also is not an inanimate object without the capability to balance and create a new BOS, the COG simply needs to be in front or back of BOS for torque to work, and to be provided the balance and vertical lift not to fall to the ground.

        If we now draw an half circle, which I guess is not really a full half circle, but more likely a drawn out shape, with a longer part before the highest point than after, following the head of the runner, the fall of the head, and COG, actually happens in the second part of the upside down half circle. This is also where the first half of stance happens, and the backwards torque.

        I am not sure how the rise and fall of COG affects the speed of the torque, but I guess this can be both measured and calculated, if there is an effect at all.

        The main point here seem to be that there need to be a lift of COG, but that it does not have to happen during mid stance. Energy input is needed, like the magnet, to provide the up movement of COG, but it does not have to happen during mid stance, It can be provided anywhere during the cycle.

        Also, since the vertical lift happens not during mid stance where COG is at its low point, but while forward torque is working during the second part of stance while the body now is at an angle, the mechanisms of the body that produces vertical lift now also must have an horizontal component. This does not seem to deny the point I was making with regard to forward torque being longer than backwards torque, and that the upright body is in unstable equilibrium and can move forward without needing to make a push movement. So as a horizontal push seem to need to happen, due to the angle at which horizontal push happens, it at most makes a relative contribution, not all. Also the angle is quite steep, and hardly optimal for a powerful horizontal impulse.

        Also the horizontal push component is not needed for initial acceleration, as the body is in unstable equilibrium. I also fail to see how it is possible from the vertical position to create an horizontal push. The angle has to happen first, and can only be created by COG moving in front of BOS, which again implies forward torque working before the push.

        I also think it is safe to say that the horizontal push component is a side effect of the body’s need to create vertical lift and the mechanisms by which it does so.

        As my main question is how much muscular push movement has to be made to create vertical lift, and the side effect of the horizontal component, I at least think it is safe to exclude push from the initial acceleration.

        Further, relating to my experiment 2, I think my steel ball experiment shows that continued movement, the next step, is possible, only by providing for the vertical lift. Even though horizontal component seems to happen by necessity, due too the angle created by torque, it is not really a necessary component.

        This further affects also the interpretation of the rest of my experiments.

    • canute1 Says:

      First, a minor point of terminology, since we are likely to end up mis-communicating if we use terms differently. In physics, the term kinetic energy is used to described the energy of a moving body. The energy that might be used to make an initially stationary body move (or to accelerate a body that is initially moving at constant speed) is called potential energy. There are many forms of potential energy. If a body is in a position that would allow it to fall to a lower position if has gravitational potential energy. If a body has stored chemical energy, for example, in the gasoline that can be used to create motion in an internal combustion engine, it has chemical potential energy. In the cases of living systems, the chemical energy is often called metabolic energy. The molecule ATP, which can be created from glucose in body tissues, can provide fuel that produces movement in muscle. The energy contained in ATP is a form of chemical energy that is usually called metabolic energy.

      This brings us to an even more important issue. Newton’s laws apply to animate bodies just as much as to inanimate bodies. However, because of its store of metabolic potential energy a human body can do things a billiard ball cannot do. As even more awe-inspiring feature of the human body is the way in which very subtle effects in the brain can send messages via the nervous system to initiate the conversion of metabolic energy to mechanical energy in the muscles. Provided the muscle is not fixed so it cannot contract, the conversion of metabolic energy to mechanical energy creates movement. Some people question whether or not the mind is bound by the law of conservation of energy. I consider that there are good reasons to propose that even the process by which mind/brain initiates movement is subject to Newton’s laws, but that is a debate that is too complex for this comment section. Whatever we consider to be the mechanism by which the person decides to move, once the decision has been made, we are dealing with interactions between physical objects. In the case of the human body, the intricate neuromuscular system that can mobilise metabolic potential energy to provide propulsion, is bound by the laws of conservation of energy, momentum and angular momentum.

      The standing human body is unstable and any extremely small movement would cause it to collapse if it were not for the fact that we continually use energy in the postural muscles to hold ourselves upright. (As a side issue, the art of using the minimum tension required to sustain posture is one of the arts cultivated by Alexander technique). When we decide to move forwards, an extremely small movement, coupled with the release of tension in some postural muscles will allow us to begin to fall. Gravitational potential energy can be converted to kinetic energy. However we would simply fall to the ground unless we activated other muscles to propel us forwards.
      Because our body begins to lean as we start to topple over, a push along the line from the COG to the point of support has a vertical and a horizontal component. The vertical component helps restores our lost gravitational potential energy and even gets us airborne if we push hard enough. The horizontal component accelerates us forwards. A sprinter must push hard to produce rapid acceleration. A recreational endurance runner can afford to accelerate quite slowly.

      Therefore there is a sense in which the horizontal push is a consequence of the vertical push, though I think that it makes more sense to say that the vertical push is an intrinsic part of the mechanism by which the body generates a horizontal push.

      • Hans Holter Solhjell Says:


        Thank you for helping out with the correct terminology. I am not familiar enough with the terminology to get it right, but hope to learn along the way.

        Could you go into more detail on the extremely small push that you are referring too? Obviously a lot of various movements happens regardless of style of running, and there is need for metabolic input.

        We can also describe ceasing an action, like balancing or releasing the stiffness of the ancles, as an action requiring energy in it self. Inhibition is a concrete function of the nervous system and neural activity does not just stop by it self (unprecise description I know), but I have no idea of the metabolic costs of stopping or inhibiting an action.

        For my purpose it is more relevant to understand if the extremely small movement you describe as necessary for the fall to initiate can be understood as a push within a push and drive sense of running or not. If the extremely small movement more reasonably can be described as a weight shift I would not put it in this category, even if there is a miniscule push against the ground somehow.

        Also, it is important to note that the COG of the standing human body spends most of the time not over BOS, but is in a state of constant movement around it, back and forth. If it is this movement you are referring to, I will agree, but this is not a problem for me, and does not qualify as a push in the sense I am thinking about, even if it might be described as a push from the perspective you are using.

        As long as the COG is outside of BOS, I fail to see how a movement that reasonably can be described as pushing is needed. I hope you can clarify this?

      • canute1 Says:


        First I should emphasize that my account of running mechanism has focussed almost entirely on the question of the most efficient way to maintain a cruising speed, I have not hitherto devoted much thought to the most effective way to initiate running action. However, here is what I do to get started.

        The initial small push is plantar flexion of one ankle (usually the right) produced by gentle contraction of gastrocnemius. My COG moves up and forwards but the elevation is minimal because I allow the knee of that leg (the initial stance leg) to flex. Almost simultaneously I flex the other hip. The consequence of these first three movements (ankle plantar flexion, flexion of the stance knee and flexion of the swing hip) is that my COG moves in front of my BOS. After a scarcely perceptible initial rise, my COG falls –at this point the action is Pose- like because there was no strong hip extension around mid-stance to propel the COG upwards. As my COG falls the swing hip flexion is arrested by the hip extensors (gluteus max and hams) and the swing foot drops to the ground, typically 30-40 cm forward of its starting point. Around this time, contraction of the hams of the initial stance leg flexes the knee, breaking ground contact, and hip flexion (iliopsoas) initiates the forward swing. In that first step, I scarcely become airborne at all because there had been no powerful push in midstance, As the initially swinging foot lands, the hip extensors (G max and hams contract in concert with the quads. This action captures impact energy as elastic energy; immediately after this eccentric contraction, a concentric contraction amplifies the push that will get me airborne on the next step. At this stage I am running, though for the first few steps, the vertical impulse is less than it will be a cruising speed, so the initial steps are short and fairly quick with a short airborne time.

        I should emphasize that I do not exert conscious control over the sequence of actions. In the first few steps I usually place greater mental focus on making sure that my shoulders are relaxed and my arm action is economical.

        I think that there are other ways in which one could produce the initial de-stabilization, but it seems to me this is a sensible way to do it as it initiates the sequence of actions that will occur at cruising speed. However this start would be very unsuitable for a sprinter.

        Your final statement assumes that the COG is already outside the BOS. Provided the COG was forward of the BOS at the instant one wants to start running, the initial ankle plantar flexion would not be essential, though I think that for me, it happens automatically. It has the advantage of starting with a small forward and upward motion. If you watch other runners, you will see that for some the first movement is a forward collapse without the initial slight ‘forward and upward’ motion due to calf muscle contraction. What I am describing as a calf muscle contraction is what you describe as a lever action of the foot. Usually this will cause the COG to rise at least a little before the fall.

    • Simon Says:


      There are systems that can utilise gravitational potential to create forward motion as you have illustrated via redirecting the downwards motion created by the force of gravity. Running has some specific issues that means it cannot.

      We could probably make a machine that was a big ball with a small heavy ball inside it. By displacing the small ball internally, the big ball would roll forwards. When the small ball was inline with the big balls COG it could be rapidly shunted above the large balls COG and so would continually add forwards motion and roll along.
      Because we cannot roll along as bipeds (or at least that is not how we chose to run), we have to balance any angular motion that is created which means we cannot make use of a fall directly to produce forwards motion. And if we try, we either face plant or are forced to overstride to catch ourselves.

      • Hans Holter Solhjell Says:

        Hi Simon.

        Thank you for commenting on my experiment and attempt to relate it to running.

        You write,
        “we cannot make use of a fall directly to produce forwards motion”

        In my experiment one, I am not suggesting that we are using the fall directly, but as I fall forward, there is also a lift created by the lever of the foot, as long as I make sure to keep it stiff. This could maybe be described as a kind of a redirecting of the downwards motion of the fall
        As far as my current knowledge goes, even though I am discussing this with Canute at the moment, this happens without a push movement to initiate the fall.

        I also need to keep my balance all along the way, but I guess this is not the kind of balance you are referring too.

        I could be mistaken here, and no lift due to the lever of the foot is occurring, even though I clearly can see my body rising up relative to a line in the mirror. If this is the case, or I am making other mistakes here, please help me to understand this.

      • Simon Says:

        Hi Hans,

        I think you can lift yourself up over your foot as you describe, but to do so is working against gravity (if the COG is raised) and so will require some external force to make it possible.

        When running, we have momentum. The action of pivoting over a foot like you describe could be used and would decrease some of our forwards momentum, converting it to upwards and forwards momentum at the cost of some of the initial forwards momentum.

        The problem with that is that you would slow down with each step, rather than maintain pace. To pivot upwards over your foot makes you work against gravity, so unless it was accompanied by some pushing, you would be loosing momentum and slowing down.

        However, I am having trouble relating this mechanism to your experiment 1. In experiment 1 as I understood it, you simply toppled forwards on one leg and caught yourself on the other. That should mean that you finished with your COG lower than when you started if there was no other external force such as pusging.
        If you are now saying that you pivot UP and over your foot to ‘fall’ forwards, then it is not a gravity based fall as the force of gravity can never be redirected upwards.

        Perhaps I have misunderstood what you are getting at?

      • canute1 Says:


        There is no doubt that it is possible to create forward motion by redirecting the energy of a fall. You can easily demonstrate this by falling forward from a stationary position. You can even use such a fall to initiate running. This experience has tempted many individuals to believe that a fall might provide free energy when running at a steady speed, but as we have been discussing in recent days, you cannot continue to convert gravitational potential energy into kinetic energy without progressive loss of height of your COG unless you use some energy consuming mechanism to recover the height of the COG – and for a runner the only mechanism available to elevate the COG is a push against the ground.

        It is possible to devise a gait in which the torso (and COG) falls after mid-stance – this happens during walking. However, above a speed of around 8 Km/hour (2.2 m/sec) this becomes very inefficient because of the large costs of swinging the leg and overcoming braking. The only practical answer is to use a vertical push to get airborne. This reduces braking costs. It also reduces swing costs because the range of motion of the foot relative to the torso is greatly reduced. The price we pay for these savings is the cost of getting airborne, but it can easily be shown (as in my post on Jan 15th) that the saving from reduced braking costs is greater than the increased energy expenditure required to elevate the COG. The solution of Newton’s equations (when using a reasonable estimate of actual force plate data) demonstrates that the COG rises after mid-stance. This is confirmed by video recordings of runners, which provide clear evidence that the torso does not fall between after mid-stance until after mid-flight. Therefore, a fall does not occur during late stance.

        You proposed a spring and stick model which led you to believe that it might be possible to extract useful energy via gravitational torque as the body is propelled upwards. In fact the upward acceleration that is an essential feature of running actually causes the line from the point of support to COG to rotate in the opposite direction than would result from gravitational torque. Gravity does have an effect – but this effect is to increase the energy cost by opposing the effect of the push. If you want to work out how to run safely and efficiently, there is little point is trying to replace the push by a fall after mid-stance. It might be more profitable to focus on how one can capture impact energy from the fall that precedes stance.

      • Hans Holter Solhjell Says:


        Regarding your comment on the 9. april. I have made comments about this further down, but would like to add my opinion here, for completeness, and as you have not commented further.

        You say that an external force is needed to lift COG by means of the lever. I don’t think this is the case. As the human body is in a state of unstable equilibrium, and has potential gravitational energy, I can simply release my weight forward, and the lever will lift COG, powered by kinetic energy.

        To illustrate this we can make an addition to the stick model, and put a 20 cm stick on the bottom of a 180 cm stick, with a hinge that we can adjust a little bit of the time to make the vertical stick lean gradually more forward over the bottom part. As we adjust the lean forward, at one point the top part will start to fall forward, and the “heel” end of the bottom stick will raise up, and lift the COG of the stick a few centimeters, before it falls to the ground.

        For the human mover, these first few centimeters is all that is needed, and he will not fall to the ground as he can both move COG forward, and balance.

        There are other aspects involved, and I have commented on those further down, so this comment is only about needing an external force to lift COG.

    • canute1 Says:


      You continue to provide thought provoking comments. Although this comment was addressed to Simon, I will offer my thoughts on you model in which there is a 20cm foot attached to the bottom of an 180cm stick.

      If the COG of any object is initially stationary, Newton’s laws require that it cannot move upwards unless an upward force acts on it. You have not specified the initial orientation of the 20 cm ‘foot;’ but I assume you intend that it is horizontal. If the hinge allows free rotation, I see no reason why the heel should rise as the upper part of the upper stick falls. If the hinge is stiff but nonetheless sufficiently mobile to allow the upper stick to fall, the heel might rise. However, you have provided no evidence that the COG of the combined unit rises. I believe that the COG of the foot will rise but the COG of the upper stick will continue to fall, and the COG of the combined unit will also continue to fall.

      Nonetheless, it is easy to devise a system in which the energy derived from a fall can cause elevation of the COG, though never to a height greater than the initial height. Consider a ball encountering a hump as it rolls down a sloping path. Provided the top of the hump is below the starting height, the ball will rise over the hump. This is simlar to your experiment with a ball in a bowl. The upwards propulsion is achieved via a ground reaction force that elevates the ball. Similarly, I suspect you could design a compound stick with hinges and springs that would allow the COG of the entire system to rise part of the way back to its starting height using energy derived from an initial fall, but it would be necessary for such a device to create a ground reaction force that redirected the kinetic energy derived from the fall into kinetic energy associated with upward motion.

      • Hans Holter Solhjell Says:

        Thank you Canute.

        As I am still learning here I appreciate your comment, applying the relevant laws to my experiment.

        As this is fairly easy to confirm or not by an empirical experiment, and have to conform to the relevant laws, I guess you are correct, which means an additional vertical impulse is needed to raise COG. And as this vertical impulse happens at an angle it also will have an horizontal element.

        As I have agreed previously that there is an absolute need to do work in the vertical direction during running, this would be in harmony with that. and the magnet in the half circle experiment.

        So what is left for me is to come to a final conclusion regarding the horizontal element that I both understand, and find that rules out the role of torque, or find a role for it. I am not all there yet but will look more into your latest comment. There is not as much time for this now as during easter, but I will get to it.

        I think all this back and forth with you have been a great educational process, both in terms of learning the relevant physics, terms and basic understanding, but also to learn by challenging the established view and trying out other perspectives.

        And I am also quite pleased with developing my understanding of the push vs springy movement, which works just as well regardless of whether there is a absolute need for horizontal push or not.

        As a curiosty i can report that I attended a two hour running class two days ago organisec by the morst prominent running store chain here, and could not help thinking of our discussion as running with gravity was mentioned as a explanation for the technique they thought.

  24. Hans Holter Solhjell Says:

    I think I formulated myself not quite precisely regarding the work gravity is doing. Would it be better to say that the extra energy input required, in addition to the work gravity is doing, is supplied by the magnet doing the last part of the vertical lift?

  25. Hans Holter Solhjell Says:

    Regarding your comment to my experiment two, you are right, I forgot to include the necessary vertical lift, but have previously several times said that this is necessary and of course requires energy input. I should have been more precise.

  26. Hans Holter Solhjell Says:


    As you asked for my motivation and what I want to question I want to answer, I will give you a little background on my general motivation and interest in this. My goal is simply to understand this better, not to prove or disprove any theory. So if it is as you say that both vertical and horizontal push must happen due to the laws of nature I have no problem accepting that. In regards to the experiments I am simply checking my perceptions in relation to this, and as you say, the most important conclusion may be that effort is required, but actually not that much. But I still have more questions to ask, and if you have the time and motivation to answer I am happy for your help. I understand that you might have gone trough this line of reasoning with several others before me and answered most of my questions before, so I hope this is not too tiring for you.

    What I am looking for is not really a theory of running, but how to run. But in reading about POSE I also got in contact with the physics of movement, and I find this interesting for several reasons, regardless of if Romanon’s theory is wrong or not. In this way POSE theory has been valuable for me, both as a new way of improving my running, which it has a lot, and as a stepping stone for improving my understanding of the physics involved in movement, for example as in discussions like this. If I have come forward as a strong defender of POSE theory, that has not been my point. I am in general not a believer in holding too hard onto ideas of most kinds, but find it useful to at times use a strong stance to explore the possibilities and limitations of that stance, and learn from what happens holding the position. I might look not to smart doing so at times, but that is not my main concern.

    Another reason for my interest in this is a long standing interest in the body and movement as a part of, or the base for, the psychological self and self experience, as well as body/mind oriented methods of learning, self development and therapy. As a Feldenkrais practitioner I am obviously interested in movement, and running is a fundamental movement pattern, and can be very enjoyable when performed right. I have come to prefer methods of learning that I can describe as organic, or body/mind, learning as experiments, perception oriented, and not only mechanical, repeat, repeat, drill oriented.

    As you are familiar with both the Feldenkrais method and Alexander techique and maybe other similar methods I guess you can appreciate this difference. As I see it POSE pedagogy has a way to go here, and for instance for swimming I far prefer the Total Immersion method which I see as one of the few athleticly oriented systems who gets this point, not only in theory, but in practical implementation, in teaching materials, dvd’s and so on.

    I have agreed previously, based on the data you have referred to several times, that there must be a vertical push movement being made. You prefer to describe it as strong, while I prefer the word light. We also seem to agree that it is better off being well time, coordinated, unconscious, at least partly reflexive and not actively performed. I am still not clear what happens in the horizontal push, but most likely the same qualities are needed.

    This brings me to my next motivation, and this is also based on my experience of changing from whatever style I had growing up, to more of a heel landing push and drive style, to a pull oriented style, which we agree must include some kind of push, both vertically, and horizontally (Though I will continue to challenge the horizontal push as new questions pops up, and the amount of vertical muscle push needed, and if my questions turns out to be based on a faulty understanding, I at least know how to counter similar ideas when I come in contact with them).

    Even though the push and drive style was informed by the effort reduction ideas of the Feldenkrais method, I got into problems both using the previously growing up undefined style, and the push and drive style. These where mainly hip and knee problems. Especially the right hip flexors kept me from running more than very short distances, 15 to 20 minutes, and 30 minutes at best, and I had pain just walking for a couple of days afterwards. Obviously this might not be due to the push and drive style as such, but my implementation of it, and/or my body’s structures ability to tolerate this style.

    As I switched from a push and drive style to a pull oriented style several things happened that I as a Feldenkrais practitioner, and a person who have worked at lot over more than a decade, to reduce effort and tension in action, as well as chronic tension levels, found very interesting, but also somewhat frustrating, as the push and drive style was something I had learned in the Feldenkrais community.

    First, at the point in time where I finally managed to reduce or eliminate the active movement of driving the knee forward the mayor part of the hip flexor problem went away. This actually happened more or less in an instant, while running with a focus on letting the knee drop while the foot is pulled. More or less exactly the same moment I got this right, I could feel the tension and pain in the hip melt away. To the degree the problem has returned later (never as much as previously) I have always been able to make it go away by the focus on technique, and various exercises designed to improve the function of pulling, while minimizing or removing the active movement of driving. Continued work on this has also shown me how I have been holding a lot of unescessary tension in both my leg and hip due to the drive movement of the leg.

    Secondly, as many others I got calf problems trying to learn POSE, and this problem got solved learning to push off less. I had worked on pushing less as I had read about that as one likely cause of these problem, and thought I had come quite close, but Jhuff clearly showed me that I was pushing off a lot more than I thought, so I worked a lot more on this. During the first run where I got this close enough to right, I could clearly feel how several layers of really deep tension let go in my back. Clearly my movement pattern inspired by push and drive had led me to hold a lot more tension than actually needed.

    Continued work in this direction has shown me how I can run a lot more relaxed at all speeds, with the pull oriented style, compared to the push and drive style, and that the chosen movement pattern affect tension patterns and levels throughout the whole body. Running also feels a lot better, lighter, flow like at best, and I am working to make the most of this style of running. The less active movement the better it feels. When I get it most right, it feels more like movement is happening to my body, rather than something I do actively.

    I can use weight shift and a small exercise as an example. If you sit on the edge of a chair with a fairly straight back, and both feet on the floor, and then shift all your weight to the left, so your upper body is now supported on the left part of your pelvis, and the right part in the air (your spine is in a curve, with the head more or less over the left side of your pelvis, not tilted to the left on a straight spine) the weight of your right pelvis can now move you back to neutral again and somewhat over, and only very light muscle action is needed to create the same potential movement on the other side.

    This might take some repetitions for most people to work out, and to notice unnecessary muscle action and blocking tension, but the result can be a very light, easy flow like movement, where it feels more like my body is beeing moved by its own weight, rather than actively muscleing side to side. This same feeling I can get in running, at best, and I am interested in finding the minimum necessary muscle input needed. Mostly in real running, but also understanding the relevant theory behind this.

    I also now from time to time do experiments with a push and drive style (also in walking) to see what happens, If I can run faster, what happens to tension levels and the overall experience of running. Mostly I find it not so good, with higher tension, more wear and tear on the body, and no benefit in speed and endurance. Obviously also a subjective experience, and subject to various properties of my own body.

    As a note I should say that I do not credit POSE alone with all the release of tension and reduced effort, as I have been working also with various other methods for this, but chancing the movement patterns of walking and running in the direction of what is prescribed in POSE has helped a lot.

    • canute1 Says:


      It has been a delight discussing running mechanics with you and I hope we will continue to do so. Although it appears that you have not previously used the mathematical form of Newton’s laws to analyse motion, you clearly have a very good appreciation of many aspects of movement. I agree that s is very likely that in the past, the push and drive approach led you to hold a lot more tension than was actually needed. I find it very easy to believe that Pose method helped you reduce the unnecessary tension. However I also anticipate that as a Fedenkrais practitioner you would prefer an understanding that is mechanically accurate while allowing you to exert the optimum amount of force in the right direction and at the right time.

      I think the image of a pull rather than a push can be helpful. I suspect that the image of the pull contributes usefully in two ways: first, I think that it prompt the non-conscious motor control system in the brain to create a light, rapid upward movement of the leg. I believe that provided we do not allow our conscious control mechanism to interfere too much, the non-conscious motor control system is capable of selecting previously rehearsed motor patterns to achieve the goal. I think that for an elite sprinter, the conscious goal should include an element of forcefulness, but for a recreational endurance runner that is often quite counter-productive. Secondly, I think that insofar as an intention to pull does influence the actual motor program that is engaged, the intention to pull produces engagement of hip flexors very early in the swing phase. I think this promotes an efficient swing. However at this stage I do not have a strong opinion about how the intention to pull might be beneficial and I will value any thoughts you have about this.

      • Hans Holter Solhjell Says:

        Thank you Canute, for your generous reply.

        It is quite easy to not look to smart entering into a discussion on the physics of running, not having opened a text book on physics since high school and not knowing the math, so I am happy you find the discussion delightful and that you find that I have a good appreciation of many aspects of movement.

        And thank you for your patience in answering my questions.

  27. Hans Holter Solhjell Says:


    I would like to ask you for a comment on the use of mathematical form of Newton’s laws to analyze the motion of living organisms.

    As I understand it Newton’s laws and the related math is designed to analyze the movement of inanimate objects. And while living objects also must obey the laws of nature, they relate to these laws in a very different way than an inanimate object, for instance by being able to balance, move by their own volition, not only being moved, and so on.

    This partly makes the math a lot more complex, and the input not only more complex, but also changing as the organism changes what it does. Now some types of movement, like walking and running, are fairly predictable and on some level simple, but still extremely a lot more complex than a falling stick.

    For instance, the falling stick has no choice but to fall to the ground. But a human can employ several different strategies to remain balanced, even while in movement, and with a changing angle back and forth.

    Also, the stick, standing on a completely flat surface and itself being completely flat on the bottom end, is in a stable equilibrium and therefore needs a push to initiate the forward torque, while the human body due to its construction is in an unstable equilibrium and has an inherent kinetic energy. What I need to do to release this energy is to reduce my effort to balance, not increase effort like in the case of pushing the stick.

    Further, the human body only need to be lifted to conserve it’s energy. This difference of the stick and the human body exist’s throughout the whole gait cycle. As the human body is lifted by it’s own various lift creating systems, it restores the kinetic energy needed to start the next gait cycle. Again very much unlike the stick.

    • canute1 Says:

      Hans ,
      As I described above, the human body is subject to the laws of conservation of energy, momentum and angular momentum.

      I agree that the human body is quite different from a stick insofar as it can mobilise metabolic energy to exert a push that has both a vertical and horizontal component. The vertical component recovers the lost gravitational potential energy, while the horizontal component provides initial acceleration when we start to run, and overcomes braking while we are running at constant speed. For a recreational runner, the initial acceleration can be slow, so the horizontal force can be small. For any runner, once at cruising speed, the need to overcome braking can be minimized by being airborne for a substantial part of the gait cycle. So it is more efficient at cruising speed to provide a greater vertical impulse than horizontal impulse. But the horizontal impulse is an essential feature of running.

  28. Hans Holter Solhjell Says:


    In my previous post I use the term lift creating systems.

    When I refer to lift creating systems, I do so not to confuse the vertical lift with muscular movement push. My main concern is to find out how much muscular movement push is needed, and a fair part is beeing done by muscle elasticity, but clearly not all. I am not certain that a push movement has to make all the rest of the vertical lift. I have previously mentioned the foot as a lever that creates mechanical lift. You have not commented on this yet, but I guess this also plays a role in creating the vertical lift.

    Could you comment on the role of the foot as lever in creating vertical lift?

    • canute1 Says:


      When I previously referred to gastrocnemius and Achilles providing a large portion of the push in late stance, I was referring to the fact that the calf muscles and Achilles tendon employ the foot as a lever. This produces a mechanically efficient push. The direction of the push is mainly along the line from point of support to COG and therefore the resulting acceleration of the COG has both vertical and horizontal components. The vertical component contributes to lift.

      Earlier in stance I believe that the push provided by the more powerful gluteus maximus and hamstrings acting on concert with the quads makes a much greater contribution to the push. In the case of the push provided by gastrocnemius and also that provided by gluteus maximus, hams and quads, a substantial proportion of the energy can be obtained via elastic recoil, but much evidence (reviewed by Alexander) shows that elastic recoil provides at most 50% of the energy.

  29. Hans Holter Solhjell Says:


    You have mentioned how the head forwards direction increase and degreases during the gait cycle. Could you refer me to more information on this?

    I have aslo trie to locate the article by Alexander regarding elastic recoil, but have not been able to find it. Do you know if it is available on the open internet? I my searches I also found a couple of books where his work is referred to. These books seem to say that there are more sources for elasticity than those calculated by Alexander, but does not give exact estimates.–N4gSN54XwBw&redir_esc=y#v=onepage&q=Alexander%20elastic%20recoil&f=false–N4gSN54XwBw&redir_esc=y#v=onepage&q=Alexander%20elastic%20recoil&f=false

    • canute1 Says:


      In a hypothetical situation in which the COG moves over the point of support at constant speed the line from point of support to COG rotates in a head forwards-direction. If the horizontal velocity is increasing, the speed of rotation in the head forward direction will increase. If the height of the COG increases after it moves past the point where the line is vertical, while the horizontal velocity remains constant, an increase in height makes the angle of the line nearer to vertical so the line doesn’t move away from vertical as rapidly, and the speed of rotation decreases.

      The question is: what actually happens during running? After mid-stance, the horizontal velocity increases because of the horizontal component of the push but at the same time the height increases because of the vertical component of the push. The vertical component of the push is much greater than the horizontal component (because the line from point of support to COG is never very far from vertical while on stance. Acceleration is proportional to force and in the direction of the force (Newton’s second law). Therefore the upward acceleration is greater than the horizontal acceleration and the rate of head forward rotation slows. This can also be demontrated numerically from the equations I presented in my post of Jan 16th

      Thanks for posting the link to the book chapter by Alexander. His JEB review can be downloaded free from

      If we can work out a strategy for maximising recovery of elastic energy that would be useful, though the mechanism by which a muscle captures elastic energy means that we can never achieve 100%. One of the reasons I have been interested in Pose is that I think that maybe the mental mechanism of Pose does achieve effective recovery of elastic energy, despite the fact that the basic theory of Pose is based on a poor understanding of physics.

  30. Hans Holter Solhjell Says:

    I see know that the chapter in the book I was looking at was written by Alexander himself:) And that I linked to the same book twice. I looked trough several google hits and thought I had seen two different books.

    On page 27 and 28 of the book he uses the example of a runner of 70 kg. He says that the energy requiered for vertical lift is 100 j. He says about half of this is stored as strain energy and returned in an elastic recoil. And estimates 17 j to come from the ligaments of the arch of the foot, and 35 j from the achilles tendon. This is a total of 52 j. This suggest that 48 j has to be provided by muscle work.

    However, he also notes that even less work might be required by muscle work to make up the remaining 48 j, as they did not estimate the storage of strain energy that must occur in the tendons of the quadriceps muscles.

    There is no mention of other muscles and tendons also involved. For instance the gluteus maximus and it’s ligaments are not mentioned. I would guess these also should be calculated into the overall picture.

    The foot as a lever, and the potential of the lever effect to provide vertical lift is also not mentioned. I will suggest that the foot both work as a spring, and a lever (simply by beeing kept fairly stiff, which also is useful for elastic recoil) and therefore provides more than 17% of the vertical lift capacity that would be the result, if we only considered the foots elastic recoil capacity.

    I am not trying to exclude muscle work and a push movement as a factor in creating vertical lift, but I will suggest that the total picture here is a bit more complex than what is included in a conclusion that 50% of vertical lift must be provided by muscle work by performing a push movement.

    • canute1 Says:


      There is no doubt that the estimate of 50% recovery via elastic recoil is approximate. Published figures range from 35% to 50%. The proportion might even be higher than 50% but you should not assume that this is likely simply on the basis of the approximate estimates provided by Alexander in that chapter. One crucial fact is that it must be substantially less than 100%.

      However even more important, from the point of view of injury prevention, it is probably the amount of push that matters. Whether or not this push is produced by elastic recoil or additional energy consumption in the muscle might not greatly affect the injury risk. One of my biggest criticisms of some Pose coaches is that they underestimate the forces that are involved. For many runners, especially recreational runners, injury risk is actually a more important issue than energy efficiency.

      However, the main points of my blog posts on running mechanics is not to point out the problems of Pose coaches. I am much more concerned to present a realistic estimate of the magnitude of the push. I readily accept that the uncertainly about the proportion of elastic recovery makes the estimates of metabolic cost approximate.

  31. Hans Holter Solhjell Says:


    I aggre that it is important to be informed of the significant forces and risks involved, and that a substantial lift must be accomplished to raise the COG repeatedly during running, and that injury risk is more important than absolute efficiency. And also that there are risks involved with POSE and other forms of forefoot running that are often not made clear enough. I think you are doing job on your blog to raise awareness of this.

    But then again I also think that having the wrong idea of how much push is needed, or the mental image of a push that is very easily created, also involves it’s own risks. Also as shown by my own history of injuries.

    So I think it makes a difference whether we create an image of the leg as a pushing device, or as a quite springy device. The resulting movements can look and feel quite different. Obviously not without risks for any of them.

    If you have further references in addition to Alexander that tries to calculate the relative contribution of muscle work vs elastic recoil in vertical lift during running I am very interested in having a look.

    • Hans Holter Solhjell Says:

      That should be,

      doing a great job on your blog to raise awareness of this.

      Also, I am not trying to say that 100% is provided by elasticity. As I have understood it this is Romanov’s position, but I have not seen him provide the actual empirical scientific support for this. I have seen the 30 – 50 % quote in several articles on the internet, but apart from the work of Alexander, I have not seen any other scientific article or literature that is based on experiment and attempts to calculate this exactly. But, I this was my first look into this, so there might be a lot more information out there that I am not aware of.

      Anyway, based on the work and statements of Alexander, I think it is good reason to not be too pessimistic in the search for a higher percentage of elastic recoil vs muscle work.

      • Simon Says:


        Bare in mind that elastic response in soft tissue generally follows the pattern of a stretch-shortening cycle that has both elastic and concentric phases both of which require muscle work to stabilise the tendons/ligaments. Whilst it appears to be very efficient, it is not free energy and seems very unlikely to me to be more than 50% efficient and indeed probably much less in real world conditions.

        Another thing to consider when looking at how much elasticity we may expect returned in running is how inanimate objects behave.
        A very good springy steel ball dropped onto concrete may bounce up to 90% of its original height, so is 90% efficient.
        Take the same ball and drop it on a sandy beach – 0% efficient!

        I think this illustrates that the human ability for storing and releasing elastic energy is impressive but it obviously requires some work to make it happen and more work under non-ideal conditions.
        The work is more than just positioning and stabilising, there is active eccentric muscle work as the tendon/ligament stretches and active muscle work as the tendon/ligament shortens.

        The great thing about it is we don’t need to think about it – it just happens. But the fact that it just happens in the background does not mean it has no cost.

      • Hans Holter Solhjell Says:

        Thank you Simon.

        I should maybe clarify my position a bit, and want to note that I am not looking for free energy, in the sense that some has intepreted POSE theory as saying is possible, by using gravity to propel us forward. Neither am I saying no muscle work is required to perform the various movements, stabilization and so on that is required for running. I am trying to learn more about the role, function and amount of muscle movement push, both in the horizontal and vertical, as well as learn about more about the physics, biomechanics and physiology of running as well. Just to clarify my position, and not to be confused with other positions of free energy and so on. For instance, in discussions with Canute I have discussed the merit of using the terms strong vs light, quick push. And I want to explore the differences off a push and drive style of running, and a pull oriented style.

        For example, it could be that the leg works in a way, or makes a movement, that better can be described and experienced as springy, rather than a push movement, but that muscle work is involved in creating the springy effect, not only to stabilize but also to create vertical lift. This would happen as you say without conscious effort. This would also be congruent with a pull style of running, but not with a push and drive style, where the leg is used more actively, with a conscious image, to create a push movement backwards.

        You make a very important point regarding the quality of the surface and its effects on the elastic resonse, as well as providing important input on the details of the functioning of the elastic response of the soft tissues.

        Can I ask you what your thoughts are on running style?

      • Simon Says:


        As far as the physics of running are concerned, I think Canute describes it quite well.

        In terms of the language to describe running in a way that leads a runner to a better style, I think that is a very broad discussion. My own background is Pose, Alexander technique plus some knowledge of the various other running techniques.
        My style has been influenced by the knowledge I have gained but fundamentally I had to find my own imagery to overcome injury issues and gain more speed.

        I think that focusing on a push in running is a terrible idea at any level. Being aware that you are pushing is fine though – being aware of the factors that change the push magnitude are important, probably more for coach than runner but often these are one and the same person.

        Equally, focusing on a pull is not always a good idea. It depends on the runner in question – focusing on pulling has never worked for me.

        The way I feel I run when running well is springy and light. I am sure that I am exerting large forces on the ground but they are reactive to balance (to avoid face planting) and are necessitated by my short time on stance. So rather than focus on ‘push’ to move or achieve short stance, I focus on springy brief quck footfalls and a stable moving trunk that tracks closely to the gradient I am running on.

        So I agree entirely that in the language of what helps us run, ‘push’ is useless if physically accurate.

        I think it is generally easier to keep the conversations of language for movement and the cold hard physics seperate as they do not go well together. Perhaps Canute will have success with bringing them together one day.

      • Hans Holter Solhjell Says:


        Good point about the pull not always being a good mental focus. Even though I have worked a lot to understand the pull, and separate it from the vertical lift, previously more done by more of push movement, and the swing, previously done more by a drive, I am not thinking pull, pull, pull as I run. I have read some giving this advice, and maybe as a practice, and for learning it might work, if one pays attention, awareness to overall organization as well, but as a main focus during running I don’t do it either.

        Your focus of springy and light I can refer to. I am not sure a by standing onlooker would agree though:)

        I also focus on not holding unnecessary tension, as well as noticing if I can reduce effort, relative to the speed I am running at.

    • canute1 Says:


      I think we are in complete agreement about the desirability of avoiding an over-estimate of how much push is required. It is one of the reasons why I considerer it is best for many recreational runners to avoid a deliberate push

  32. Hans Holter Solhjell Says:


    In my previous reply to Simon I came up with a new formulation, that might be a new way at looking at the elasticity vs muscle issue, and that I think would with well with a pull oriented style of running.
    I would like to ask for your thoughts on this formulation.

    “For example, it could be that the leg works in a way, or makes a movement, that better can be described and experienced as springy, rather than a push movement, but that muscle work is involved in creating the springy effect, not only to stabilize but also to create vertical lift. This would happen as you say without conscious effort. This would also be congruent with a pull style of running, but not with a push and drive style, where the leg is used more actively, with a conscious image, to create a push movement backwards”.

    • Hans Holter Solhjell Says:

      This would also mean that I would change my image of a light quick push, to an image of “allowing a springy movement to happen”.

      This springy movement then also has both vertical and horizontal components.

      I think that makes a difference in terms of ideas of practical implementation, and what movement we make with, or let the leg do.

  33. Hans Holter Solhjell Says:

    I make to many spelling and mistakes in formulations. Is there anyway to correct my comments here? I can not find one.

  34. canute1 Says:

    Hans and Simon

    I doubt that there is a single best mental strategy for all individuals and all circumstances. However, I do believe that understanding the physics and physiology helps. In my opinion, Glen Mills and Usain Bolt have married the theory with the practice very successfully. I think a serious sprinter would be wise to examine Bolt’s physical and mental action closely. As described in my post of 11 March, his mental strategy includes driving during the acceleration phase and pushing while at top speed.

    The situation is less clear with regard to endurance running.

    Simon, I was interested to hear that you include a focus on ‘springy, brief, quick footfalls’. I strongly agree with the adjectives ‘brief quick’ and am happy to add the adjectives springy and/or light. When trying to go fairly fast I often focus on a brief, quick descent of the contralateral arm with emphasis on the hand moving rapidly from upper chest height towards the waist, with lightly opposed forefinger and thumb, while avoiding elevation of my shoulders. If I think about the legs at all, I am aware of a rapid hip flexion early in swing. Maybe this is similar to the Pose pull, though I think Pose experts would regard this as heresy. I am usually fairly aware of the pattern of pressure of the soles of my feet. As speed increases the load moves more to the forefoot, but at slower speeds and on down-slopes I am very careful to keep the weight reasonable far back. As a kid I ran on my forefoot. About twenty-five years ago, after a long period of not running, I decided one day to go for a run. I tore my calf muscle after about 250 metres. It took ages to recover and ever since have been very cautious about putting too much stress on the calf.

    In summary, my mental approach is based partly on physics; partly on (very incomplete) knowledge of how the brain works, partly on a wish to avoid unnecessary muscle tension and partly on past experiences – good and bad. It suits my temperament to try to match the images reasonably closely to what I understand is actually happening.

    I practice a version of the Change of Stance drill to consolidate in my repertoire of non-conscious motor programmes a programme that links the descent of the leg to the brief quick descent of the contralateral arm, and the Swing Drill to ensure that the hip flexion that propels the forward swing as arrested smoothly by hip extension in late swing.

    Hans, I am afraid that I cannot find an option in WordPress that would allow you to edit your own comments once they have been posted.

  35. Hans Holter Solhjell Says:

    Regarding the topic of which words to choose and mental image to make, I do not think it is only a matter of what works best for us on an individual level, but also a matter of what words we use for describing various types of movements, and different qualities of movement. This happens both on an individual level and also collectively. If our language is not rich enough describe important differences regarding to a topic discussed, we run into serious problems both of knowledge, fundamental understanding and communication.

    I will try to make clearer what I mean by a difference of a push movement and a springy movement.

    If I stand on my both feet and lowers my COG slowly, stop and then pushes hard enough up I will lift off the ground and then fall down again to a full stop. I will use the term push for this movement. The is a very similar movement to the one I perform if I just lower my COG, and slowly rise until I stand on my toes, without a lift. Also it is similar to what happens while performing a squat movement, either standing, but also sitting in a squat machine in a gym.

    If I stand on my both feet and perform a hopping, skipping movement, similar to what I do when using a jump rope. This movement has some similarities to a push, especially at the first hop, but after that the important differences appear. My legs now works more as springs, contributing to what is better described as oscillatory motion with the springy movement of my legs on the one end creating lift, and the pull of gravity on my body moving me down, repeatedly.

    Some google searches helped me learn that this kind of movement is called simple harmonic motion. For the oscillation to continue one needs to add energy to the system, if not, the oscillation gradually gets smaller and stops. Therefore, muscle work drawing on metabolic energy also has to be done to create the springy movement of my legs. Elasticity alone is not enough, just as two springs strapped up in each end of a box, connected with an object in the middle, will make the object oscillate back and forth somewhat until the oscillation stops unless we add some energy into the system. If no energy is added, leading the motion to stop, it is called a damped simple harmonic motion.

    This also supports your point that the contribution of muscle elasticity can never be 100%, even if I prefer to use the word springy. This does not imply, or need, muscle elasticity to do all the work. Whether it is 10 or 90% is not important from this perspective, although it might be important for other reasons, like metabolic costs. The movement the leg makes for creating vertical lift just have to have this springy quality, regardless of what the proportion of elasticity vs muscle work is. And not be a simple push movement. The springy movement can also have a horizontal component, and create hGRF.

    So in this sense, a Pose, fall, pull style of running does not need a theory of free energy from gravity, as long as the vertical lift movement the support leg makes is springy, not pushy. Neither does it need a 100% contribution from elasticity to create the vertical lift.

    We can also train to create very powerful springy movements. Looking at some great runners, I feel that their running has more of this springy, not to be confused with bouncy, quality, rather than a push, push quality, even though some of them, like Usain Bolt, use this terminology.

    My main point is that the kind of movement the leg is performing while running is not well described as a push, even though it has some of the qualities of a push. And that this use of terminology, even though it is very common, might obscure our view of what is going on, both in a real sense, in our communication of what is going on, and in terms of mental images and connected real movements we create.

    Some great talents can use the image of a push to create powerful springy movements, but a lot of us can not and has to work a lot to learn it, so it becomes a quite important distinction also for real running.

    As a curiosity I will mention that my native Norwegian language has several words for running. One of them is “springe”.

    • canute1 Says:


      I agree that the comparison with simple harmonic motion is interesting. If you look at Figure 4 in my post of Feb 6th (the effect of cadence on efficiency), you will see that the vertical displacement of the COG does have an approximately sinusoidal variation with time. Pure simple harmonic motion would produce a perfect sine wave. Pure simple harmonic motion would occur of the force that restores the COG to its average height was proportional to the displacement from the average vertical position. That is not quite true for running. It would be true for the case of a spring where the restoring force is proportional to the displacement form the rest position. Nonetheless, I think that the image of a spring action might be useful in imaging the vertical motion.

      It is not as relevant to the variation in forward velocity during running, but as we have discussed previously, the vertical forces are much greater than the horizontal and it is perhaps best to let the horizontal force arise without any conscious awareness.

      My own approach, which I have described in some detail in a new post, places greater emphasis on a holistic appreciation of the sensations and movements of running, but as I explain in that post, I think it is generally more effective to be consciously aware of the movement of the hands than the feet.

      • Hans Holter Solhjell Says:

        Thank you for comment Canute.

        I see you mentioned our discussion in your blog post of today and am glad that you have found my comments worthy of a mention.

        This part of your reply I did not entirely understand.
        “….of the force that restores the COG to its average height was proportional to the displacement from the average vertical position”.

        “That is not quite true for running”.
        What is different?

        I am also preparing a last attempt to see if I can make an account of vertical movement in running without horizontal push, and look forward to your counterargument.

    • canute1 Says:


      Consider first a spring: Hooke’s law states that amount of displacement of the end of the spring (y) is proportional to the force (F): F=c*y where c is a constant.
      The acceleration of the end of the spring is proportional to the force acting on it. Therefore the acceleration is proportional to the displacement. A path of an object for which acceleration is proportional to displacement is a sine wave. (This can be proved using calculus. If you are familiar with calculus you will know that the first derivative a sine wave is a cosine wave, while the first derivative of a cosine wave is sine wave so the second derivative of a sine wave is a sine wave).

      However for the runner, the vertical force (per unit mass) during the airborne phase is –g (where g is acceleration due to gravity), and the vertical force when on stance is (vGRF-g). This force is not proportional to the displacement from the average position, so the body does not move in the same way as a spring. The motion cannot be described as simple harmonic motion.

      However that does not prove that elasticity does not matter. If time on stance is much shorter than airborne time (as in the case of a kangaroo or a human on a pogo stick), the trajectory is a series of parabolic arches separated by a rapid reversal of direction of movement from downwards to upwards during the impact. So it is still be possible for elasticity to play an important part even if the motion is not simple harmonic motion.

      I think we agree that elasticity is very important. I think we also agree that only a fraction of the energy can be provided by elasticity. We also agree that it is efficient to run in a way that maximises the conversion of impact energy to elastic energy, followed by recovery of that elastic energy to propel the body upwards. I think you are more optimistic about the chance of recovering more than 50% of the energy. However I do not think this difference of opinion is a serious difference.
      The main point that I would make is that I believe it is possible to develop a holistic sense of what is happening to the body, even though our attention is not focussed consciously on every aspect. There is a very informative picture in today’s Guardian newspaper, showing Prince Harry and Usain Bolt being silly for the sake of a photo-opportunity. They are imitating a well known advertisement for Richard Branson’s company, Virgin. In the advertisement, Branson’s face is superimposed on Bolt’s body, as he mimes shooting an arrow from a bow. In this Guardian photo of Harry and Bolt, note how the index finger of Bolt’s right hand is aligned perfectly with the index finger of his left hand. I suspect he wasn’t consciously thinking about this as he posed for the photo. Simply, his brain has an extremely good sense of where the ends of his limbs are at all times. I think that is one of the reasons Bolt is the world’s fastest sprinter. I think we can improve our running by improving our bodily awareness. Furthermore, awareness of the end of the index finger can probably associated with subliminal awareness of the location of the foot.

      • Hans Holter Solhjell Says:

        Thanks Canute.

        I think even though the movement of running cannot be described as simple harmonic motion, what I learned from reading about simple harmonic motion, and the conclusions I drew in the post mentioning it, regarding push vs springy, and that for the leg to perform a springy movement where it repeatedly lifts COG to the same height there also needs to be some kind of energy input, meaning muscle work, seems to hold. Even if the leg was 100% made of elastic tissues, it would not be able to repeatedly lift COG, because there would be no energy input. You seem to agree on this?

        Pretty easy to see in the picture who is most comfortable being out of the vertical position. Prince Harry looks a lot more tense, and has to twist his neck to keep his head in the vertical to balance and/or orient, while Bolt easily maintains relaxed balance and orientation even with his head in line with his spine.

      • canute1 Says:

        A perfectly elastic spring could lift the COG repeatedly without requiring additional energy input, but this is not really relevant to running because the leg is not perfectly elastic, and also because the we require energy to overcome the braking force. Some of the braking impact can also be recovered via recoil, but not all. Furthermore we require energy to reposition the legs.

        Again, to summarise my view, we should try to recover as much energy as possible to help meet the needs of all three main energy consuming processes of running: getting airborne overcoming braking and limb repositioning. During the gait cycle there are several opportunities for storing elastic energy in addition to the impact at foot fall. These include the hip extension in late stance (which preloads the hip flexors) to help propel the swing, and hip flexion in late swing which preloads the hip flexors) to help exert the push against the ground. These various energy saving mechanisms improve efficiency, but they have a cost, since reloading a muscle puts is at greater risk of microscopic damage. This is probably why runners are much more likely to experience DOMS than cyclists.

      • Hans Holter Solhjell Says:


        Do you have an example of a perfectly elastic spring? Or a simple harmonic motion system that does not require energy input. For instance, the spring and load systems showed on this page,
        will they just continue forever? Are these real, or a theoretical construction?

        Also, could you give me some input on breaking forces in running? Or point me to where you have written about this previously. I guess there is backwards torque, as we have discussed, and friction at ground contact. Anythings else, or should it be worded differently?

      • canute1 Says:

        The oscillators shown on that website are not damped so once started they will continue to oscillate forever. They are a theoretical construction.
        You might have noted that the solution for the equation describing the time course of the load on the spring is given as a cosine wave, whereas I described it as a sine wave in my previous message,. This is not a contradiction because a cosine wave is simply a sine wave moved by a quarter of an oscillation along the time axis. If we consider time is measured from the point where the load on the spring at the middle point, the time course is a sine wave. If we consider that time is measured from the extreme point of the oscillation, the time course is a cosine wave.

        The braking force is shown in figure 1 of my post on Jan 16th – hGRF before mid-stance pushes the body backwards and so has a braking effect; hGRF after mid-stance accelerates the body forward to compensate for the braking effect.

      • Hans Holter Solhjell Says:

        Thank you for explaining.

      • Hans Holter Solhjell Says:


        I googled for the term imperfect harmonic notion and less showed up than for simple harmonic motion. This page dicusses the term.
        As you say running can not be described as simple harmonic motion, but can it be described as an imperfect harmonic motion?

      • Hans Holter Solhjell Says:

        Maybe stretching it a bit, as the web page belongs to a science fiction writer, but I like to open minded, at least in terms of where inspiration comes from, as inspiration and facts are not the same thing:)

    • canute1 Says:

      The type of imperfect harmonic motion described in that article is not really like running. That article describes the motion of a complex oscillation arising when the force that pulls the object back towards the mid-point of its trajectory is a symmetrical parabolic function – in which force increases as the square of the distance from the mid-point, equally on both sides of the mid-point.

      During running, the force acting on the COG is asymmetrical about the mid-height of its trajectory. For most of the period during which the COG is below the mid-height, the force that propels it vertically towards its mid-height is vGRF, while the downward force that acts when the COG is above mid-height is gravity (The force per unit mass is g). The upwards force, vGRF is represented fairly accurately in fig 1 of my post on Jan 16th, and again in fig 3 in the post of Feb 6th while the trajectory of the COG predicted by Newton’s equations is given in fig 4 on Feb 6th. If you compare the trajectory of the runners COG shown in fig 4 with the trajectory of the black spring shown in Daniel Russel’s website describing simple harmonic motion
      you will see that there are strong similarities, but there are noticeable differences.

      The upwards turn of the trajectory of the runner’s COG at mid-stance is a little tighter than the down-turn at midflight. This is because the upward force at mid-stance (peak vGRF) is stronger than the downward force at mid-flight (g) so the upward acceleration immediately after mid-stance is greater than the downward acceleration immediately after mid-flight. The similarity emphasises the fact that provided we can capture the kinetic energy of the downward fall under gravity as elastic energy, we can use this to help provide the upward acceleration – but as we have discussed previously, the process of elastic recoil only captures fraction of the energy required.

  36. Hans Holter Solhjell Says:

    Simon and Canute.

    As a result of the discussions and comments from you in the comments in this blogpost by Canute, as well as my own investigations using google, my previous knowledge, and the performed and suggested experiments, I have learned a little more about physics relating to running than I did before, and also about the weaknesses of my, and others, previous attempts to formulate an account of how continuous forward movement in running can happen without the need for a horizontal push movement. Obviously any movement must abide by the laws of physics and the account must be formulated in such a way that it conforms to the usual vocabulary of physics, as well as the related biomechanics, physiology and other empirically observable facts about running.

    Using what I have learned so far I will now make a new attempt at formulating an account of how horizontal movement in running can happen without the need for a horizontal push movement. You both might have heard and countered this account before, or something similar, but I have not myself, and if I am wrong I at least know the counterarguments. As previously stated I am here not trying to defend a particular position, but trying to learn by experiment, construction of models and discussion. Also I find the topic challenging and fun as well.

    I will for this account expand my formulation of my experiment one, and relate it to the steel ball, half circle and magnet experiment. Note that in this experiment the right half of the half circle is cut of and shorter than the left side, so the steel ball will role down the left side and jump over the right edge. Also, as I have noted in my previous comments, while standing, we are in a state of unstable equilibrium, there is constant postural sway, and our COG is outside of BOS most of the time. This means my body has gravitational potential energy and can fall forwards without a push being applied externally or internally. All I have to do is to cease balancing. Due to the construction of the human body the most likely direction it will fall is forward. I can also adjust the direction of the fall. The foot also works as a lever, and can redirect the forward and down motion of the fall to a forward and up movement of my COG, for a distance of a few centimeters upward. Also, vertical lift is provided by, after the first gait cycle, by a spring movement requiring both elastic recoil and muscle work, that has a metabolic cost.

    As I stand on my right foot, in a state of unstable equilibrium, there is postural sway, and my body has gravitational potential energy. I can cease balancing and release my weight forward, and the potential energy of my body turns into kinetic energy as I fall forward (I am here trying to use the terminology as I understood it from Canute’s explanation above. Please correct me if I am using it incorrectly, while at the same time relating to the underlying argument).

    After the fall forward has started and kinetic energy and gravitational torque is moving my COG forwards, I stiffen the foot stiff so it can work as a lever. This will rise my COG, but also decelerate it. But the speed and momentum created by the kinetic energy is large enough to carry me over the top point of the trajectory.

    During the trajectory the left leg also moves forward, due to it’s own weight, being hinged in the hip and the changing angle of my body, relative to the rest of my body so it can land slightly in front of COG. I am not sure if this changes the position of my COG forward relative to BOS during the trajectory, helping out somehow, but have to look a bit more into that.

    After the top point of the trajectory my COG now goes into an accelerated fall down and I land on the left foot.

    As I land on the left foot, there is an opposite torque working during the first part of stance. But since this first part of stance is shorter than the second part (as in the cut of half circle and steel ball experiment) the momentum created by the speed that remains after the climb of COG up the lever plus the fall after the max height reached by COG I will now at least reach a new vertical position, a new state of unstable equilibrium, potential energy is restored, ready for a new cycle.

    If there now is excess speed available, so the body is not only restored to the vertical but continues over the mid point. I will now not fall on the face as a stick would, since I knows how to balance and can move my BOS forward.

    I am now still on the ground all the time, so it can not be called running, but is more similar to walking. But if I at the top of the trajectory of COG pull the foot of the ground, I am now airborn, similar to running. As I land on my left foot and leg, the elastic tissues of my leg are loaded, and together with muscle work are ready to perform a spring movement, adding to the vertical lift of the lever of the foot in the next gait cycle. This vertical lift again raises my COG, and the same cycle is repeated. I am now running.

    There are many similarities here to the steel ball experiment. The body, as the steel ball, has potential energy. As we let go of the ball so it can role down the half circle, and as I cease balancing and release my weight forward, potential energy becomes kinetic energy.

    The lever of the foot working together with the ground, GRF, does the job of the half circle to redirect the fall of COG upwards, just as the half circle and its GRF redirects the fall of the ball.

    Powered by the kinetic energy in the first part of the fall before the lever kicks in, and GRF, the lever lifts the GOG, playing the role of the magnet as well. This rise in COG comes at an initial cost of deceleration and loss of kinetic energy as the COG is lifted to the top point of its trajectory. During running the spring movement will add to the vertical lift.

    During the forward and up movement of COG there is also h and vGRF, but the hGRF is caused by the kinetic energy moving my body into an angle. I am not doing a push movement in this experimental condition, and at this point I can not see that it is necessary, even if possible. Just as a falling tree also has both h and vGRF. But the tree does not have something that resembles the lever of the foot to lift the COG of the tree.

    At the top point of the trajectory the COG of my body now has new potential energy, created by my body’s initial kinetic energy, potentiated by the lever.

    I now pull the right foot, and becomes airborne.

    This potential energy now turns into kinectic energy and new acceleration is created as the GOG falls down in the second part of the trajectory, and creates enough momentum to overcome the backwards torgue and friction in the first part of stance, much like the momentum of the speed of the steel ball created by the fall down the left side slope can trow the ball over the edge of the right side if this side is somewhat shorter than the left side.

    The lever of the foot seems to be a key here, as it both redirects the fall of COG and potentiates the kinetic energy of the body so it can create a new store of potential energy and subsequent kinetic energy as the COG tips over the top of the the trajectory, and accelerates down on the other side. And also carries me over the mid point of the next stance.

    According to this account, we have potential energy at two points in the gait cycle, first during midstance, and then at the top of the trajectory of COG. We also have kinetic energy in both halfs of the cycle. No push is needed to turn potential energy into kinetic energy as the body at mid stance is in a state of unstable equilibrium, and on the top of the trajectory the COG simply falls down, but is carried forward by momentum as it falls.

    Also, a vertical spring movement will also have a horizontal element due to the angle at which the lift has to happen and this will show up on a pressure plate as hGRF, but since the angle that is necessary for this to occur is due to the torque, and my body is, according to this suggested account, moving forwards anyway, the horizontal push component and following hGRF is a side effect of the vertical lift and torque, and not what is moving my body forwards.

    Running then partly becomes an act of balancing and adjusting the forwards and backwards angle of the torque, stabilizing and balancing the body, creating vertical lift, and repositioning the limbs to move the BOS forward, all of which requires muscle work and has metabolic cost. I can accelerate by adjusting the torque forward, maintain even speed by balancing the torque, reduce speed by adjusting it towards the vertical, or stop by adjusting it backwards before moving into the vertical.

    I can not find the holes in the argument my self, or where it does not fit with a normal understanding of physics, observable facts about running or other factors, so please help out.

    I also think it should be possible to calculate this, as the parameters are pretty clear and the various formulas should exist. Unless I have left something out of course and this does not make sense. I have no idea of how to do the math myself.

    • canute1 Says:

      As you acknowledge, whenever a vertical force is applied, there will be at least a small horizontal force due to the fact that the direction from COG to support is oblique, unless the COG is stationary above the point of support for the duration of application of the force. This duration must be greater than zero. Otherwise an infinite force would be required to achieve elevation. If the time is non-zero, the COG moves over the point of support during this time and the line from COG to support is oblique for some of the time

      A force acting along a line that is oblique to the ground exerts either a backward or forwards component of push against the ground. It is not correct to say this horizontal force is due to torque. Even if the only force was along the line from COG to support (in the hypothetical situation of zero gravity) there would be a horizontal component when the line is oblique, even though in this situation there would be no torque.

      According to Newton’s first law, a body continues in a state of uniform motion unless acted upon by a force. Thus when the runner is already at cruising speed, the only need for any horizontal force is to overcome the braking due to obliquity (and to overcome any wind resistance). I think that you might not fully appreciate that if the vertical force is applied for a very short period, the body only suffers a small amount of braking and therefore only needs a small horizontal push. This is why I emphasise the increase in efficiency obtained by short time on stance, but short time on stance demands a strong vertical force.

      Since obliquity when the support is in front of the COG produces the braking, I am not sure why you are unwilling to accept that obliquity when COG is in front of the support provides the compensation for this.

      An additional point to note is that we cannot lift the COG by a pull. When running we initiate the upward acceleration of the body with a push. In late stance, a pull might be used to flex the knee and thereby contribute to breaking contact with the ground, but a pull in the absence of a push would bring simply bring the hips nearer to the foot.

      • Hans Holter Solhjell Says:


        Just a short comment. It is late here, so I will read your comment more closely tomorrow.

        Did you interpret my comment as saying that COG is lifted by a pull? That was not my intention. The pull does not lift the COG. The pull breaks contact with the ground. The lever of the foot powered by gravitational torque, or kinetic energy, or whatever term might be the most precise, is doing the vertical lift in the first step, and the pull then breaks the contact with the ground. In the second step there also is additional vertical lift from elastic recoil and muscle work. The pull is never involved in the vertical lift.

      • canute1 Says:

        We agree that the pull does not provide vertical lift. I mentioned it to try to establish what level of agreement we have about the pull. I do not think it is essential for a pull to be employed to break contact with the ground, but I accept that many runners exhibit a pronounced knee flexion in early swing.
        I think it is probably best if this is mainly initiated by gastrocnemius, which crosses the knee joint and is active in late stance, when it contributes to the lever action which you describe. I remain unsure whether a shortening of hamstrings is beneficial at this time as it is might interfere with the initiation of the forward swing. However, as the hip flexor contract in early swing, if the hamstring retains a modest isometric contraction, the knee will be further flexed because the hamstrings cross both hip and knee. Overall, I think the hip flexors (especially iliopsoas) are probably the muscles that do the major work in propelling the swing. However the mechanism of the pull is not our main concern at present
        Whatever the mechanism of the pull, the key issue is that any vertical force must act for a duration greater than zero if it is to exert an upward impulse. While the upward force is being applied, the COG moves over the support and there must be a horizontal component. In fact I think you accept this, but for a reason that is not clear to me you consider that this does no work.

      • jhuff Says:


        The foot tapping drill clearly demonstrates pulling. You have seen a video of that exercise correct? In my experience it is necessary that the hamstring be used to accomplish foot and leg recovery as prescribed by pose method. Indeed one can run and recover the leg without using a hamstring initiated pull.

    • canute1 Says:


      I have seen many video recording of foot tapping, including one in which you demonstrate it very neatly. I believe that the foot tapping drill is useful insofar as it promotes the initiation of the swing with an action that draws the foot towards the buttocks, thereby minimising a tendency to allow the foot to follow a long arching path behind the torso.

      I believe the hamstring plays a role in foot taping, but I do not think that contraction of the hamstring is the principal muscle action, though a largely isometric hamstring contraction does play a part. Foot tapping involves flexion of the hip and knee, together with a passive plantar flexion of the ankle. The major muscles involved in flexing the hip are the hip flexors, such as iliopsoas, though because this is located deep, it is not easy to perceive its contraction. The hamstring also shortens a little, though on account of the fact that the major hamstrings (apart from the short head of biceps femoris) cross both the hip and knee joint, only a minor degree of shortening of the muscle is required in order to achieve knee flexion. The fact that one of the major roles of the hamstrings is hip extension yet foot tapping involves hip flexion indicates that concentric contraction of the hamstrings is not the principal muscle action involved.

    • canute1 Says:

      Yes. There is no hip flexion in this exercise, so flexing the knee requires contracting the hams

      • jhuff Says:


        Ok…so why does adding freedom of movement of the hip/knee/thigh cause the principal muscle use to ne changed?

    • canute1 Says:

      First, some muscle other than the hamstring must produce the pronounced hip flexion (in the tapping drill and also in the swing phase of running) Second, when the hip is flexed, only an isometric contraction that maintains the length of the hams is required to produce knee flexion.

      • jhuff Says:


        Why must it? Experience tells me that the principal muscle used must ne the hamstring or it will not be precise.

    • canute1 Says:


      I agree that the question of which muscle plays the key role is important because I think it shapes s the trajectory of the foot during swing. It is clear that while running and during the foot tapping drill that both hip and knee flexion occur at lift off. From my perspective the question of which muscle initiates the swing is of importance because it is probably more efficient if the foot lifts towards the buttocks rather than rising too far behind the buttocks. Nonetheless I accept that in fact the foot does rise behind the buttocks and this helps create a short lever during the swing. However, the foot should not be allowed to lag too far behind and I believe the swing drill encourages a lift towards the buttocks.

      If the lead role is played by the hip flexors combined with an approximately isometric contraction of the hamstring, the knee flexion will follow the hip flexion. If shortening of the hamstring occurs before hip flexion, the foot will swing further behind the buttocks.
      In practice we let the non-conscious brain select the order of muscle firing. However I believe that by practicing the foot tapping drill, which involves a hip flexion together with relatively small amount of shortening of the hamstrings, when we run the non-conscious brain is more likely to lead with the hip flexion and an associated approximately isometric tensioning of the hamstring, which will minimise the lag of the foot.

    • canute1 Says:


      I am sorry that I have not yet made a video of the swing drill. In that drill I do allow the foot to rise a little behind the buttocks. I am not sure whether or not I should place more emphasis on lifting the foot towards the buttocks, as appears to be recommended in both Pose and BK.
      I am intrigued by the fact that elite sprinters (eg Usain Bolt) and 10,000m runners (eg Haile G in his prime) tend to bring the foot up very high behind the buttocks. In a comment of my article on ‘Mechanism of efficient Running’, Matt McGinn stated that after Bolt contracts his hamstring his knee drive is straight forward with his hamstring still fully contracted. There is little doubt that having the knee fully flexed during swing will decrease the energy required for the swing, but I am very dubious of using a hip extensor such as the hamstrings to initiate a movement which aims to move the foot forwards rapidly. Therefore I am inclined to think that drills such as foot tapping, that encourage simultaneous hip flexion and knee flexion (thereby minimizing the amount that the foot lags behind the buttocks) are useful.

      • jhuff Says:


        In the slow motion video I have of usain bolt he does not bring his foot up behind…but under? Do you have picks of what you are refering to?

    • canute1 Says:


      I was referring to the comment by Matt McGinn on my article on ‘Mechanics of Efficient Running’ (See the side bar of this blog)
      My current view is that the hamstring tensioning that contributes to knee flexion should be simultaneous with the hip flexion. I think that foot tapping encourages this simultaneous flexion of hip and tensioning of hams.

      We do not differ in our view of the value of foot tapping. We appear to differ insofar as you appear to consider that hamstring contraction initiates the movement, whereas I think that hip flexion plays a leading role. I do acknowledge that there should be simultaneous tensioning of the hams. During the swing drill, it I put my hand on the hams I can feel the tensioning, though the fact that the hip is beginning to flex at that time demonstrates that the hip flexors are active. I suspect that iliopsoas is the major contributor, but it is not possible to place a hand on iliopsoas.

      • jhuff Says:


        It would appear that I present a stronger argument that the hamstring is indeed the principal mover during the tapping exercise 🙂 As usual you are free do disagree at your own risk 🙂 I do look forward to the day I see a video of the swing drill.

    • canute1 Says:

      What evidence is there that the hamstring is the principal mover in the tapping drill?

      • jhuff Says:


        Unfortunately I don’t think there is any concrete scientific evidence that proves one way or the other. Thankfully there doesn’t have to be. It forces us to think and be responsible and accountable for our choices. My evidence comes from feeling primarily, based on simple exercises that show me that it must be the hamstring. I love to see you perform the exercise and actually try to use another muscle as the principal mover. My guess is that you would not be able to bring the heel to touch under your buttocks. Given your procrastination to becoming video capable I guess I will never see what I think would happen……

    • canute1 Says:


      Do you know what a hip flexor does? The reason I am arguing that hip flexion occurs at the beginning of the tapping drill is that the thigh begins to swing forwards as the foot leaves the ground. I suggest you examine a video recording of the drill.
      The hamstrings are hip extensors. If the hamstrings fire without hip flexion occurring, there would hip extension together with knee flexion and the foot would move backward as it leaves the ground.

      If you read what I have been saying on this topic you will see that I was quoting the observation of Usain Bolt as possible support for your argument that the hamstrings might act before the hip flexors at the beginning of the swing. In particular I referred you to Matt McGinn’s comment that appears to support your claim that hamstring contraction plays the leading role at the beginning of swing. However you disagree that Bolt brings his foot up behind his buttocks, so you even reject my attempt to present some evidence in favour of your argument. While I do not mind if you reject my attempts to help your argument, it does make me wonder if you know what hip flexors and extensors do.

      • jhuff Says:


        So…now you agree that hamstrings are the principal muscle? That is great I must have been confused by all your wordy comments. I will go back and re-read them to see where I got lost. Glad to see we agree though:-)

    • canute1 Says:


      I do not agree that hamstrings are the leading muscle in either the tapping drill nor the swing phase of running. Please read the preceding comments.

      The crucial issue in the tapping drill is that the foot does not swing back as it would if the hamstrings (which area hip extensors) were acting before the hip flexors. Therefore, I believe that the hip flexors must be active right at the beginning of the movement. In addition, the hip flexion is accompanied by approximately isometric contraction in the hamstrings thereby producing the knee flexion that accompanies the hip flexion.

      • Jeremy Says:

        Canute, I am following and I do understand the role of hip flexors and extensors. In the foot tapping and in running I think the role of the flexors activity is to be minimized to stabilize posture of the body so the hamstring can bring the foot to the position under the body. They assist/accompany the work done from the hamstrings. Since the flexors role is to pull the knee upward they are not the prinicipal muscle advised to be used during the foot tap or in running. Yes they still have some activity as we are a unit. No argument there. Do we agree that the hip flexors raise the knee upward around the joint and toward the torso? They don’t directly raise the ankle/foot in the direction of the buttocks…..correct?

    • canute1 Says:


      It appears that it comes down to what we regard as the principal action, as we agree that both the hip flexors and the hamstring are involved. My opinion is the hip flexor does more of the work, both during the tapping drill and during the early part of the swing phase of running. In practise, it is best to avoid consciously selecting which muscle to use. Furthermore, I appreciate the danger that conscious focus on hip flexor contraction during swing might increase the risk of over-striding. However, I do not think we should under-estimate the work that hip flexors do.

      So maybe we will have to agree to differ in the emphasis we place in the role of hamstrings and hip flexors

      However, if we are going to strengthen the appropriate muscles for running, it is helpful to know what muscles do the work at the various stages of the gait cycle. There is no doubt that hamstrings play a major role during several phases, such as in arresting the swing after mid-swing and also in assisting gluteus maximus in creating the ground reaction force that is essential to get us airborne. So it is clear that the hamstrings need to strong. However, I think it is also important not to under-estimate the role of hip flexors. I believe that ‘high knees’ is useful. In recent times, Chris McDougall had popularised Walter George’s 100-up, which in my opinion is intermediate between Change of Stance and high knees. I personally prefer to do CoS to improve precision of timing and ‘high knees’ to increase hip flexor power. Marching is also good for developing endurance of the hip flexors.

  37. Hans Holter Solhjell Says:


    I think that in my previous comment I showed that there are two horizontal components in running.

    One is provided by the combination of the unstable equilibrium of the body and it’s potential gravitational energy, kinetic energy, gravitational torque, momentum and the foot that works as lever as well as the spring movement of the leg that redirects the fall of COG from forward and down to forward and up before a subsequent fall, from the top of the trajectory of COG, happens, as well as the the body’s ability to balance the torque, so it does not fall to the ground as a stick. You did not comment on this account and so far it has not been show not to hold.

    The second is provided by the horizontal push that has to happen at the same time that vertical lift is created, as the vertical lift happens at an angle.

    So this is somewhat puzzling, and since there is a horizontal push it is very intuitive to assume that this push is responsible for acceleration and maintaining steady pace. Especially if one already has concluded that a push is necessary and has ruled out gravitational torque. I myself can not rule out gravitational torque yet, as my account above has not been countered, and I can not see that it violates the laws of nature. I therefore has to explain the role of the horizontal push within that account better, and understand it better myself.

    I therefore tried to construct several thought experiments to illuminate the situation, and the relative, possible or necessary contribution of each in relation to horizontal propulsion. The solution seems to be quite simple, but somewhat more complex than just saying that acceleration and steady pace happens either by torque, or push.

    I will use the half circle, where the right side is shorter than the left, and the steel ball again. As we release the steel ball and it rolls down the left side, somewhere just after the bottom we can now give it a short extra push, slightly increasing it’s speed. But as the ball would have jumped over the right side anyway, this push is possible, but not necessary for the ball to jump over the right side. The kinetic energy of the ball and it’s momentum would have carried it over the right side anyway. And if we add enough energy into the system to lift it to the same height as its starting point, as with the magnet, and there is a new half circle at this point, the ball will have an accelerated start for the next go, reaching a higher speed on this half circle. The push add a minimal, but not necessary, acceleration. It’s relative contribution is very small. On the other hand, if we applied the push at the beginning, it would still not be necessary, but it’s relative contribution is larger.

    For running, this means that I can initiate running both with, and without a horizontal push. For explosive acceleration, as for a sprinter, or in a sport where there is a lot of start/stop movement, like tennis, more of a push is necessary. For steady pace running, the horizontal push is there, but is not really necessary, and is not what is the main mover. Momentum and torque is. To accelerate while running, I can now either adjust the lean, or I can push harder.

    As the horizontal push can only happen if the body is at an angle, torque has to be working as well. The only time this is not true is at the first step. It is possible, but not necessary, to place yourself at an angle while COG is still over BOS and push of from there. This is what sprinters do, and also what you, Canute, exemplified as your prefered starting move. This is also similar to pushing the ball at the beginning of the half circle. After this first push, torque is working, and any extra push happens together with torque and has to be coordinated with it. For the sprinter, who I think spends some steps with the COG in front of BOS, has to push hard enough to raise the “body stick” high enough so it is possible to balance the torque, and now use torque for further acceleration.

    • canute1 Says:


      I agree that when leg is oriented obliquely while on stance there is both horizontal ground reaction and a torque due to gravity.

      There are two crucial points to make:

      1) On account of Newton’s second law, a horizontal ground reaction force produces braking before mid-stance and forward acceleration after mid-stance.

      2) Torque produces rotation about the pivot point, not forwards propulsion. In fact the rate of rotation after mid-stance increases in a head backward direction (due to the simultaneous upward acceleration, as described in several of my previous comments), but the angular momentum in the head forward direction increases due to the increasing moment of inertia. However, any rotation of the body must be cancelled at some time in the gait cycle. In the absence of wind resistance, the increase in angular momentum after mid-stance is balanced by an opposite decrease before mid-stance.

  38. Hans Holter Solhjell Says:

    Also, since the parameters of the framework I am suggesting is quite clear, I think it should be possible to calculate mathematically what the optimal amount of torque vs push is at any give speed, and for a given acceleration.

    • canute1 Says:

      I have presented calculations demonstrating the energy costs associated with braking in posts on Jan 16th and Feb 7th. As torque contributes to rotation but not propulsion, and furthermore, its affects on rotation before and after mid-stance cancel, it is not directly relevant to the energy cost, at least in the absence of wind resistance. I plan to address the effect of wind resistance in a later post

  39. Hans Holter Solhjell Says:


    Waking up today, having slept on our exchange yesterday, a new question popped up.

    In a situation where gravity does no net work in the horizontal direction via torque, and there is a definite need for a horizontal push, we can still say that this push needs to be smaller or larger, depending on how well we utilize kinetic energy, torque and so on?

    In the example of the half circle and steel ball, the less friction there is and the easier the ball can role, the further it will role up the other side, and the less of an extra push is needed to push it all the way to the top, and over the edge.

    And for practical purposes, during running, we then have to take torque, kinetic energy, fall and lean into account, to maximize the contribution of these factors, even though they are not enough to sustain forward motion without a horizontal push element.

    I am not concluding here on my own behalf regarding the role of torque, but am trying to make sense of the various technique elements, as well as my own experience of running from the perspective of gravity doing no net work, and that includes a horizontal push element. Since having worked on technique Pose style, I experience my running as fairly push free, and I can both accelerate and sustain max speed without any movement i can describe as active pushing. As discussed earlier, my experience is more of a springy movement, that also has an horizontal component. I also feel that i can adjust my speed by adjust my lean, and run faster for longer by leaning better rather than pushing more. It could be that this feeling of more flow, leaning into speed experience is then based on a better utilization of my kinetic energy, torque, fall (please help out with a better description here, if mine is incorrect) reducing but not eliminating the need for a horizontal push component at any given speed, as well as in acceleration.

    This can then again support the subjective feeling of moving more freely, with less effort, and more flow like, even though a small amount of horizontal effort is still made.

    Your thoughts on this as well as help to formulate the physics aspect is appreciated.

    • canute1 Says:


      First, it is important to note the role of gravity is accelerating, either at the start of a run or when changing speed. Leaning does allow ius to use gravity to generate kinetic energy which can then be re-directed to provide horizontal acceleration by mean of ground reaction force. This involves a push but this push is mainly a reflex action to stop falling on ones face. So there is not doubt that gravity helps acceleration. For a sprinter, he/she consciously pushes against the ground or the blocks but an endurance runner rarely perceives the push.

      When running at a steady speed, Newton’s first law, which states that a body continues in a state of uniform motion unless acted upon by a force, tells us that we can minimise the need for any push by minimising braking. We minimise braking by spending small time on stance. There are two feature of Pose that minimise braking. First, high cadence results in shorter gait cycles, including shorter time on stance and shorter airborne time. The shorter airborne helps by virtue of the fact that the impulse required to get airborne can be delivered within a shorter stance time (for a specified value of average vGRF). The other relevant feature of Pose is the mental focus on a rapid pull. I fact it is an illusion that pulling gets us airborne. It is fact it is a push that gets us airborne. though a substantial portion of the energy is provided via elastic recoil. Furthermore, it is misleading to claim that the hamstring does the pulling because a larger amount of work is done by the hip flexors such a psoas which most runners are scarcely aware of, (unless they injure their psoas). Thus despite being based on erroneous physics and erroneous biomechanics, Pose does encourage the runner to engage muscles that achieve efficient running in an apparently less effortful manner.

      If we are to run well we need to avoid unnecessary or mistimed pushing. In particular we need to avoid wasting kinetic energy by necessary braking and we need to learn how to capture impact energy via elastic recoil.

      I therefore think that for recreational endurance runners Pose is better than a running style that is based on the mistaken belief that strong conscious pushing is required. Elite sprinters do need to push consciously, but that is not our present topic.

      However, while Pose has advantages for the recreational runner, there are two types of problem with Pose. First, it creates the illusion that large forces are not required and this illusion does predispose to some injuries. Secondly, for a recreational runner who wishes to achieve his/her best possible performance, there is a risk of discouraging the need to do the type of training that is required. For example, Change of Stance and High Knees involved similar movements. However, CoS promotes precise timing while High Knees develops powerful hip flexors. Pose places a disproportional emphasis on CoS at the expense of High Knees.

      • Hans Holter Solhjell Says:

        Thank you Canute.

        I will not go into the biomechanics aspect of this and what muscles do what in which movement for now, apart from noting that I find the pulling movement and the high knee movement to be two quite different movements, even though they might look somewhat similar to most people. Personally I do not see the relevance of the high knee drill/movement, at least not compared too the pulling drill/movement. If one thinks of the high knee drill as a muscle strengthening exercise I can understand, but I see many using it as a “the way to move your legs while running” drill and can not see the relevance. I also have not seen a single one of the persons I have seen perform this drill actually move their legs this way while running. Neither do their posture and angle of the body resemble anything they actually do while running. The pull drills mostly loosk much more similiar to what people of all abilities actually do while running somewhat skillfully, regardless of what muscles one understand as beeing involved or not.

        Personally my experience has been that for me it has been beneficial to reduce hip flexor involvement, and increase hamstring involvement. I have also been somewhat suprised by the depth of change in muscular patterns involved in my own process of learning this and the “fine” motor control aspect involved in these gross motor movements. Especially in the process of doing more and more refined small aspect drills, getting into the small details of the movement, current and habitual activation and deactivation of motorpatterns and changing these. In my last couple of coaching sessions with Jeremy over the last couple of weeks, one issue he observed in my drilling was an element of active driving down of my leg, especially the left leg, after the pull, rather than just releasing it and letting it fall by it’s own weight to the ground. This took some quite intensive looking into the details of what I was actually doing, breaking and slowing the movement down into it’s smaller parts to solve better. While walking I also became aware of moving my lower leg more forward than necessary, and therefore activating and keeping more muscle tension in my lower legs than needed. Working with this has even further reduced my lower leg issues.

        Over the last eight weeks I have had some problems with shin splints after overdoing plyometrics training doing one leg jumps in a stairway and have therefore not been able to run much. After the sessions with Jeremy and my own experimentation I could feel real progress happening and was on saturday able to participate in a 10 k run without lower leg issues. I could at one point, around 7 k, notice my old hip flexor problem flaring up. From experience I know this is a result of letting the leg trail to much, having a late pull and too much hip flexor involvement, and was able to adjust my technique, performing a better timed pull and reduce hip flexor activity, even though somewhat fatiqued at this point, and had no mayor problems after that, and no pain the day after, apart from regular muscle soreness.

        To sum up my current understanding, based on the discussion of the last few weeks, I now find POSE method of running to not really be in need of a horizontal push not occurring, and gravity and a redirected fall to be solely responsible for horizontal acceleration and steady pace running. It works just as well even if a horizontal push component is definitely needed, and this occurs as a part of the springy movement that provides vertical lift. Apart from this, no extra horizontal push is needed, apart from in cases of explosive acceleration.

        Still, this does not rule out the usefulness of an understanding of the role of physics, gravity, the body’s unstable equilibrium, kinetic energy, gravitational torque and so on. Even in a situation where these factors can not provide all the horizontal acceleration and forward impuls needed to create and sustain forward motion during running, the potential gravitational energy and kinetic energy of the body seems to play a very large role in the overall picture of horizontal movement, and working to maximize it’s potential should be a key focus of technique work in all kinds of running, even sprinting. This is, as far as my impression goes, largely underemphasised or not emphasised at all in most discussions on running technique, outside of POSE, your blog, and maybe a few others that I am less knowledgeable about.

        I think this is pretty close to what you have been saying, and I also agree with you that the forces involved in running should not be underemphasised and should be described as strong, compared to the structure of the body, even though we strive for lightness and ease in movement and experience of runnning, and than one should train properly to avoid injury.

        Even so, as I have no problem accepting the view that there is a definite need for a horizontal push component (and that there definitly is a horizontal push component present, it showes up in the pressure plate experiments, and can not be removed even in theory) I still am not completely satisfied with all of the arguments on the side of ruling out the possibility for horizontal movement happening without a definite need for horizontal push, and that the horizontal push is present as a side effect of the horizontal push and not an essential component of horizontal propulsion. I understand that from a physics perspective GRF is a push and whether hGRF is due to a redirected fall and the vertical lift capacity, springy movement, or a needed horizontal push movement is not really important, but from a technique perspective and in terms of how we visualize and talk about running it most definitly think this is a key difference.

        So I look forward to your writing on the role of torque in running and the fact that the torque after mid stance is longer than before mid stance. If you could include some thoughts on the difference between a stick, and a human being able to redirect the fall as well as balance this torque just so to maintain forward acceleration, and thereafter steady pace, and also not falling, rotating to ones face, that would be interesting too.

    • canute1 Says:


      Thanks for your comment. I agree that high knees is not a drill whose primary purpose is developing good form. I consider it is a drill for developing power, especially power in hip flexors such as psoas, though it might improve form indirectly as powerful muscles can help improve form.

      With regard to your description of your hip flexor problem, if your swing leg was dragging, the problem was almost certainly a late contraction of iliopsoas. During early swing if you only contract hamstrings (which extend the hip) and delay the contraction of hip flexors, your swing leg will drag.

      A subsequent late hip flexion is likely to lead to over-striding.

      I think that the CoS drill is very helpful for a person who has a delayed hip flexion because the main muscle action in CoS is hip flexion. The hamstring contraction flexes the knee which is a secondary effect, but the primary action is flexing the hip. As I have remarked frequently Pose can be quite effective for recreational runners, even though the theory is based on erroneous physics and biomechanics.

      I personally think that Pose could be a very good technique if only the underlying theory was abandoned, as this would make it possible Pose coaches to deal with risk of injury in a realistic manner and also allow them to acknowledge fully that a runner who wants to achieve peak performance needs to develop power as well as precise timing.

      • Hans Holter Solhjell Says:


        I think if we are to come to some kind of agreement on what muscles do what when we need to strap ourself up with some electrodes and measuring equipment and see what they tell us. But I am not sure if that will help us in our everyday running as we then will have to rely on our senses primarily. I think we agree that a well timed pull is a good thing, and that my pull was late, contributing to the problem, and I could fix it myself while running, by focusing on a better timed pull as well as a few other factors that influence the pull. Someone posted an article in a comment on one of your recent blogs, advising to let the leg trail until recoil, swinging foreword simply because it is attached to the upper body that moves forward, brings it forward. This most definitely does not work well for me, both causing over striding and hip flexor problems.

        Your pointing out of the iliopsoas is interesting, as it is an important but often forgotten muscle deep in the core of the body. If you can expand on your understanding of its role in running that would be nice to read.

        As for high knees as a strengthening drill I don’t think POSE’ers would have a problem with that, as long as one is aware of the difference compared to a technique drill. Most people are not. Even the instructor, an aspiring athlete, in the workshop I participated in recently had no idea of the purpose of this drill, and thought as a technique drill, even though doing nothing similar to it while running himself. Jeremy has also thought me several exercises that have more or less the same strengthening effect as the high knees drill, and there are are several strengthening exercises one is advised to do in the POSE literature and articles, although I have not studied this in depth myself.

    • canute1 Says:


      Interestingly, EMG is not a very useful way to determine the amount of work done by a muscle because the electrode is placed in the vicinity of only a fraction the junctions between nerve fibres and muscle fibres, and furthermore EMG does not distinguish isometric from concentric contraction, so the size of the electrical signals is not a good indicator of the amount of fibre contraction. EMG does indicate timing (though one needs to allow for approximately 50 ms between electrical activity and mechanical effect). Magnetic resonance imaging (MRI)can be used to determine how much activity (relative to the maximum possible activity level for that muscle) occurred during running. However it cannot determine at what stage in the gait cycle the muscle was active.

      Therefore I consider that the best way to estimate which muscles contribute most to an action during a particular part of the gait cycle is an estimate of the effects of contraction based on the known attachments of the muscle. The reason I believe iliopsoas plays the major part in early swing is simply that observation shows that flexion of hip and knee occurs at this time. Hamstring contraction that was stronger than the contraction of the hip flexors would cause flexion of knee and extension of the hip which would delay the swing. Contraction of quads would cause flexion of hip but prevent flexion of knee. Hence, when considering the effects of these there muscles, the combination that will produce a rapid flexion of both hip and knee is a strong iliopsoas contraction accompanied by a moderate hamstring contraction.

      The EMG evidence confirmed that iliopsoas is active in early swing. However MRI shows that during running it is only activated at approximately 60% of its maximum possible activity level. I think this is consistent with the fact that iliopsoas is a strong postural muscle and therefore has considerable reserve strength. Because its strength is largely based on postural activity (which requires mainly type 1 fibres) I suspect the development of type 2A fibres is not likely to be great, and therefore I believe high knees is useful to develop power (ie ability to generate a rapid strong pull.)
      With regard to hamstrings, they are maximally active in late swing when they act to arrest the swing. MRI evidence indicates that they are strongly activated (typically within the range 75-85 % of their maximum). The hamstrings are not as ‘naturally’ strong as iliopsoas. I therefore think that exercises to strengthen hamstrings even more important that exercises to develop power of the hip flexors, as the hamstrings are at risk of damage when running at speed.

      However as you imply the most important drill for a person who has a delayed initiation of the swing is a drill that improves accuracy of timing of the flexion of iliopsoas and hamstring . In my opinion change of stance is the best drill for this though there are several other Pose drills that are likely to be helpful

  40. Hans Holter Solhjell Says:

    Hi Canute.

    It has been a long time since our latest exchange, and I hope you are doing well.

    I have just read trough our discussion again and given the topic some thought over the last weeks. I still fell there are several unanswered questions and unclear points and would like to ask you too clarify a few points in relation too torque.

    First of all, how do you explain how backward torque cancels forward torque when the torque works longer after mid stance than before mid stance.

    Secondly, If the forward torque is not canceled by the backwards torque, due to it being shorter than the forward one, I can not see how this is a problem for a human runner. A stick would fall to the ground, but a human have the ability to balance, to adjust the lean, and to move our BOS forward. So even if a stick would fall to the ground in this condition, a human does not. I can not see that you are taking this into consideration. Neither have I found any writing where you or anyone else provide the data that shows that the torque is actually canceled in the way you suggest.

    Thirdly, as I previously mentioned in this thread, the human body is in a state of unstable equilibrium, and I can accelerate forward in the first step only by ceasing to balance and allow myself to rotate forward, while at the same time providing vertical lift so my COG moves upwards. No horizontal push is needed for this initial acceleration, even though a horizontal hGRF occurs due to the angle of the pressure against the ground. This happens in the second half of stance. Let’s say that we for sake of this argument conclude that the torque from the second part of stance is canceled by the torque in the first part of stance, and I return to a full stop. I have now returned to a standing position, and my body is again in a state of unstable equilibrium, with its potential kinetic energy restored, ready to again rotate forward. So even in a situation where the torque is perfectly balanced in a purely mechanical way, restoring my body to its vertical position only, I still will have the potential to move forward, simply due to the unstable construction of my body and its potential kinetic energy.

    • canute1 Says:


      It is good to hear from you again. Sorry that I have been slow responding. I simply haven’t had much time for blogging recently.

      With regard to your first question about the cancellation of torque, I agree that it appears that there is more torque applied in the second half of stance but the human body, like all other physical objects larger than sub-atomic particular and moving at the speeds much less than the speed of light, obeys the laws of Newtonian mechanics. Therefore, angular momentum must be conserved. Adopting a less theoretical stance, if head forward rotation was not cancelled the runner would fall flat on his/her face after a few strides. There are some amusing U-tube clips that demonstrate this. You argue that the body does not fall because of adjustments at the joints. I agree that the body is not a rigid stick and it is probable that adjustments of shape due to flexion at the joints (together with the effect of upward propulsion that lengthens the body and increases the moment of inertia about the pivot at the point of ground contact) explains the apparent paradox of a longer time on stance after mid-stance than before. But the fact is that the head forward and down rotation that Pose theory proposes is a source of forward propulsion must be cancelled at some other part of the gait cycle. The crucial issue is that the body is propelled forward and up during late stance, and the figure that Dr Romanov has presented in his scientific paper on the topic is simply wrong – as discussed in my post on problems with Pose. As the COG moves forward and upwards, the line from point of support to COG actually rotates in a head-backward direction – the direct opposite of what is claimed in Pose theory.

      With regard to the initiation of movement by falling forward, I agree that the COG moves forward as the body rotates in a head forward direction. However, as you acknowledge, the COG falls to a lower level. You propose correcting this by applying an upward vertical force to the COG. However you cannot apply a purely vertical force at this stage, because the leg is oblique. You must either push forwards and upwards from the current point of support, or you move your foot forwards to a position in front of the COG. In fact both of these things usually happen when a runner starts from a stationary position.

      The combination of gravitational torque and purely vertical push cannot provide horizontal propulsion. To accelerate the body forward there must be a net backward push against the ground. At constant speed (in the absence of wind resistance) there is no need for a net backward push. Momentum carries the body forward.

  41. Hans Holter Solhjell Says:

    Hi Canute, and thank you for your reply.

    I think that we previously have agreed that there are various problems with the way Romanov formulates his theory and that they are partly incorrect in the use of the language of physics, and at times he is imprecise in describing what actually goes on a more descriptive level. So my questions was not really related to to Romanov and his ideas, but to my own attempt understand and describe what goes on during running.

    I agree with you that the direction of movement is forward and up, and that it is not correct to describe this as a fall. A fall actually happens after peak elevation has occured, but this is not what Romanov is describing. Still, in various places, it seems that Romanov agrees that the movement is forward and up, but still choose to describe this as a fall, which creates some confusion, at least for people who care about precise language and description.

    It seems from your answer that you agree that since the human body is not an inanimate object but has the abiltity to balance, adjust it’s angle and move support forward, we do not have to cancel out the torque in the same way a stick does. But we still have to take the balance of forward and backward torque into consideration. As the natural tendency of the human body is to rotate forward over its pivot point, as it is an ustable system with a forward tendency, and the forward torque works longer than the backwards torque we actually have to balance it out in a different way than a stick. You describe this as adjustments of the joints, which I agree with. I described this as balancing, but this is more of choice of words and images, not a fundamental difference, as a part of balancing involves adjustments of the joints.

    In my opinion this means that we actually can control and use torque actively, both to accelerate, keep a steady speed, and decelerate. You mention the fact that the vertical lift involves a push, and since this push happens at an angle, there will also be a horizontal push at the same time. I tried to formulate some thoughts about that in an earlier post, and it is obviously somewhat puzzling to account for, if one tries to formulate a theory of how gravitational torque is the main mover in the horizontal direction.

    I am not sure what you mean when you say that “As the COG moves forward and upwards, the line from point of support to COG actually rotates in a head-backward direction”. What does this mean? We actually do move forward, so there is some terminology or concept I am not familiar with here and does not understand. Can you elaborate?

    • canute1 Says:


      Thanks for your continuing comments. We agree about many of the important points though sometimes we use different words to describe the situation.

      With regard to my statement about the rotation of the line from point of support to COG, I was a little imprecise. To be more precise, I should have said that the rate of head forward and downward rotation slows in late stance. The explanation is complicated and probably not really very important . However, if you want to follow my argument it is probably easiest to visualise if make a sketch of the line from point of support to COG, taking account of the following facts. After the COG moves over the point of support, the line from point of support to COG is inclined obliquely forwards and upward from the point of support. In late stance, the COG is rising as it moves forwards. The forward component of the motion of the COG tends to make the line from support to COG more near to horizontal while the vertical component of motion tends to make the line from support to COG more vertical, so the actual rate of change in angle of during late stance depends on which of these two opposing effects is greater. The calculations reported in several of my posts in the early months of this year indicated that in late stance the effect of the upward motion of the COG has a strong enough effect to decrease the rate at which forward motion cause the line to become horizontal. In other words, this rate of head forward and down rotation of this line actually slows rather than increase, as appears to be predicted in Pose Theory.
      However, we need to consider not only the speed of rotation but also the effect of the lengthening of the leg as the knee straightens in late stance. For a rotating body, the angular momentum is the product of angular velocity times the moment of inertia. An enlongated object has a greater moment of inertia. As the runners leg straightens the moment of inertia increases. It might be helpful to consider the for example an ice-skater who stretches his/her arms outward. With arms out-stretched, the moment of moment is greater. In the absence of an external torque the amount of angular momentum must remain the same, so the skater’s angular velocity decreases as the arms stretch out. In the case of the runner in late stance, as the leg extends, the moment of inertia around the pivot point on the ground increases. Unlike the situation of an ice-skater, the interaction of gravity and ground reaction force generates a torque, acting in a forwards-and-down direction. Even though the rotation slows, the increase in moment of inertia more than compensates for this, so in fact the angular momentum in the head forward and down direction does increase, as Pose theory predicts. However the angular momentum only increases by an amount that compensates for braking in the first half of stance (when running at a steady speed in the absence of wind resistance). If it exceeded the effect of braking, the runner would get faster.

      But I do not think any of this rather complex mechanical description really matters very much. I quoted it merely to illustrate that the simplistic view that the body falls like a stick, is very misleading. But we already agree on this. From the practical point of view, the important fact is that there is a substantial push against the ground, and as a result, the ground reaction force pushes the body forward and up. The upward push gets the body airborne. The forward push compensates for the braking that occurred in the first half of stance. In the absence of wind resistance, the impulse produced by forward push must be exactly equal to the backward impulse due to braking, when running a constant forward speed. If the forward push generates a greater impulse than the backward impulse due to braking, the runner will accelerate.

  42. Hans Holter Solhjell Says:

    Hi Canute,

    Thank you for your in depth explanation. Although quite theoretical and complex for a non physicist, I had to read it several times, it seems to me that it also shed some light on some aspects of technique.

    For instance, it is generally accepted that pushing to much, creating excessive lift, is not a good idea. Good runners typically run with less vertical displacement than less technical accomplished runners, and with less knee extension. One effect of this for the proficient runner should then be that the vertical impulse is smaller, COG rises less in late stance, and then the effect you describe, that the moment of inertia increases for an elongated object, is smaller for the proficient runner with less vertical displacement, than for the not so good runner. The proficient runner in this way utilizes the torque in the second phase of stance better. Correct?

    • canute1 Says:

      Like many things in life, one should aim for the Goldilocks option when running: not too much vertical displacement nor too little.

      Too much displacement wastes energy and increases the impact forces. However running is an efficient form of locomotion at speeds in the range 3 m/sec to 12 m/sec largely because it involves getting airborne. Getting airborne requires the upwards push that makes a longer stride possible. Unfortunately, it results in substantial impact forces. We can find the Goldilocks solution by adjusting cadence and time on stance. At higher cadence we do not need as long a stride length to achieve a given speed, but at very high cadence the energy cost of repositioning the limbs becomes very high. Short time on stance minimises braking, but requires a large vertical force.

      So, it is all a matter of balancing various different things to achieve the optimum combination of vertical elevation, time on stance, stride-length and cadence. I do not think that torque itself is a very important part of the calculation. In practice I think it is best to concentrate on a fairly high cadence and a fairly short time on stance at moderate and high speeds, though at low speeds cadence should be a little lower and time on stance a little longer.

      • Hans Holter Solhjell Says:

        Hi Canute.

        I think we agree on most aspects of running technique in a general sense, but I am trying to sort out and understand better the various physics aspects of running, and also increase my understanding of the physics of movement, so I continue to probe the details here.

        I agree that there are many elements that has to be work together and balanced. But we can still break things down and look at each element by itself, while keeping the overall picture in the background.

        You did not answer my question directly, and I would like to as for a more precise comment relating my my proposal, question. Do you agree or not that a better runner, who has less vertical displacement will have a smaller increase in the moment of inertia than a runner who have a larger vertical displacement, and therefore also utilize the torque in the second part of stance better?

    • canute1 Says:

      Since torque is proportional to the distance from point of support to COG, a smaller amount of lengthening of this distance will result in a very slightly smaller torque when there is less vertical displacement of the COG while on stance. So the runner with less lengthening of the leg will experience a slightly smaller increase in rotational momentum about the pivot point during the second half of stance. However, the lesser amount of rotational energy is trivial compared with the energy costs of elevating the body. Furthermore, considering the energy costs of only part of the gait cycle in isolation gives a misleading impression. When the entire gait cycle is taken into account, the balance of energy cost and energy gain from rotational acceleration and deceleration produced by torque makes no net contribution to running efficiency.

      • Hans Holter Solhjell Says:

        Hi Canute,

        I understand. But does that also apply in the same way to an object where the COG is moving in one way relative to the support point?

        When the iceskater pulls the arms inwards, he moves faster, and the further the arms extends the slower he moves, so I thought maybe the same would apply in some way, but I guess there also is a difference in how this works in the horizontal and vertical plane, or these are two seperate phenomena?

        Regarding your statement that torque makes no net contribution, I guess you are now still talking about steady pace running? I think I can agree with this. Obviously, you also have the calculations showing this. And it is a fairly simple observation that backwards and forwards torque must be balanced out over time, or the runner will fall to the ground at some point.

        At first, I thought that you meant that this balancing out happened in a purly mechanical fashion, in such a way that the backwards and forwards torque cancels each other out by beeing of equal length, but we seem to agree that this is not the case, and that forward torque works for longer than backwards torque, and that the runner actually have to perform the act of balancing out this imbalance, making sure to create a situation where torque produces no net contribution to steady pace running. Otherwise, he would at some point fall to the ground.

        One should think that this should be a clear win for a push and drive style understandung of running, and that gravity based models would have to yield at this point. But I still have some more questions and points to make.

    • canute1 Says:


      The difference between the situation of the runner and the ice skater is that no external force exerts a torque around the relevant axis (the vertical line from point of support to crown of head of the ice-skater, so there is no change in angular momentum. Therefore when the moment of inertia changes due to internal re-arrangement of limbs, the speed of angular rotation must change to conserve angular momentum. In the case of the runner, gravity is an external force that exerts a torque around the horizontal axis that passes through the pivot point at point of support, so there is a change in angular momentum about this axis while the runner is on stance.

      However, as we agree, the change in angular momentum due to gravity makes no contribution to forward propulsion at constant speed. As you point out, we also agree that the period on stance after the COG passes over the point of support is longer than the period before, and this can be accounted for by re-arrangement of the angles of the various joints of the body.

      I will of course be interested in the further points you wish to raise regarding ‘push and drive’ compared with a ‘gravity’ style of running. However, until someone produces force plate data showing that a Pose runner exerts a lesser impulse (produce of force times duration) against at the ground at a given speed and cadence, I will be inclined to believe that pushing against the ground is a much more important feature of running than gravitational torque. If we are to improve efficiency we need to maximise the proportion of the impact energy that can be conserved as elastic energy and re-used at lift-off, and also ensure that we push at the most effective time. If we are to minimise injury, we need to arrange the limbs so that the forces are sustained by the parts of the body best able to withstand them, in addition to strengthening these body parts.

      With regard to Pose, I am aware of no credible evidence that it improves efficiency directly, though if it can reduce injury this is likely to result in better performance. I believe the evidence indicates that that provided the appropriate strengthening exercises are performed, Pose can reduce some types of injury, and for this reason, it is a beneficial style for some runners. I think this is because some of the features of Pose (eg fairly high cadence; encouraging the capture of elastic energy in the arch of the foot; avoiding reaching out with the leading foot at footfall etc) are features that are supported by a realistic analysis of running mechanics. However my own view is that it is even more beneficial to develop a style based on a realistic understanding of running mechanics.

  43. Hans Holter Solhjell Says:

    Hi Canute,

    I am preparing some more thoughts on the matter, but would like to ask your for some clarification about the concept of torque providing no net input. Could you elaborate on how you conclude this, and also what in your view the practical implication of this is?

    At the moment, my conclusion is at least that we are balancing forward and backwards torque in such a way that we are not falling to the ground, but maintain an erect position during the process of moving forward. I am not familiar enough with the concept of no net input to make precise use of it in my thinking at the moment or to challenge the concept. But regarding the balancing of backwards and forwards torque and it’s role or not in producing steady pace forward movement I have several more points to make/explore. I will find time in the next couple of days to write out my train of thought on the matter.

    • canute1 Says:

      By ‘no net input’ I meant that torque does not provide a net forward propulsive force. Torque produces rotation, which does not propel the body forwards. During the acceleration phase, when we lean forward as we start from rest, the rotation does encourage us to exert a force against the ground and this can help propel us forwards. As we have discussed on several occasions previously, when running at constant speed, not only does the torque only produce rotation but this rotation cancels to zero when averaged over the entire gait cycle.

      • Hans Holter Solhjell Says:

        How do you separate between the amount of rotation that occurs in running, and forward movement? If GOC rotated 360 degrees around a fixed point I can see your point, no forward movement is produced, but as GOC in running moves forward both during forward and backwards torque, rotation, and BOS also moves forward, I am not sure what you mean. Can you clarify?

    • canute1 Says:

      Rotation around the point of support is in itself quite unimportant. We could also describe rotation about any arbitrary point that is not on the path of the COG, but this would merely be a description of what a stationary observer sees, and would not have any influence on what actually happens to the body, so can be ignored entirely.
      The reason that Pose theory pays attention to the rotation due to gravity is that this is associated with a change in angular momentum of the body due to a torque and might does influence the movement of the body. However as we have discussed, it has no net propulsive effect during constant pace running. Although the roatation due to gravity does not produce a complete 360 degrees, it is to and fro (alternating clockwise and anti-clockwise) and therefore has no net effect.

      • Hans Holter Solhjell Says:

        Hi again Canute,

        I would like to have a clarification of the point you are making here, on september 30th at 6:32. You write that the rotation is to and fro and therefore has no net effect. I am not sure that the rotation of COG over the point of support that occurs during running can be described as to and fro in the way I would normally understand this term. For the rotation to be described as to and fro I would normally expect that this is in relation to one, unchanging pivot point, or point of support. In this case, no net forward movement is produced, regardless of if the energy input for the initial movement comes from an external push, an internal impulse, for simply from the weight of the COG.

        But this not the case during running as when our COG rotates over BOS and reaches the end of the second part of stance, there is no backwards rotation over the same BOS, but rather we move BOS forward, rebalances our system and make sure that GOG starts its forward rotation from approximatly the same relative angle to BOS as in the last step.

        So, it seems to be that in this sense there is no to and fro, only to, and no fro. I agree that there is a backwards torque working during the first part of stance, but the rotation during this part of stance is still to, and not fro.

        So, the rotation over BOS, regardless of what we see as the energy source to produce this rotational movement, that happens during stance is vitality important for forward movement in running. Without this rotation, or if a full counterrotation occured, we would not move forward at all.

        Looking forward to your comment.

        I also enjoy your strength training posts a lot, and needs to do a lot more strength training myself. Especially at the moment as I have been sick for several weeks, and even developed a blood clot in my right lower leg due to lying in bed for 5 days and getting dehydrated due to food poisoning. I have also negleted strenght training for many years, mostly due to discomfort from muscle soreness, as well as time constraints due to family and work. Anyway, it’s time to take this part of training more seriously, and especially now before starting building up after no or little activity for the last weeks.

    • canute1 Says:

      Hans ,

      Thanks for your continuing comments.

      During stance the line from point of support to COG rotates in a head forward direction at approximately constant rational speed as the COG moves forward over the point of support. However the speed of this rotational motion is not constant. In the first half of stance, when the line from COG to support slopes down and forwards, gravitational torque imparts a small amount of rotational momentum in a head-back direction and thereby tends to slow the rotation. In the second half of stance, the gravitational torque imparts a small amount of angular momentum in the head forward direction and thereby tends to reverse the effect produced in the first part of stance. This is the sense in which I meant that the effect of gravitational torque is to and fro. This to and fro effect is superimposed on a steady head forward rotation sustained by the forward momentum of the body. I agree with you that the net rotation is of course always in a head forward direction. The major factor that keeps the runner moving forward is the forward momentum that was generated when the runner accelerated from rest (or from a slower speed). Effects sush as braking in the first half of stance, the forward propulsion in the second half of stance required to compensate for this braking, and gravitaional torque produce only a small to and fro fluctuations in the approximately constant forward progress of the runner.
      As an analogy, imagine a pendulum swinging in a moving railway carriage. The net movement of the pendulum bob is always forward. Relative to an observer in the carriage, the pendulum moves to and fro. The to and fro motion does not affect the average forward movement. Similarly, gravitational torque does not add to the average head-forward motion of the runner.
      It should be noted that due to the changing length of the leg, the actual movement of the COG is a little more complex than described above, but the effects due to change in leg length do not impart angular momentum so are not relevant to this issue.

      • Hans Holter Solhjell Says:

        Hi Canute,

        Thank you for elaborating on your point.

        It seems to me that the forward rotation and the mechanism whereby this happens is more important that what it is my impression that you give it. We have previously discussed the topic of redirection of energy, like the ball that rolls down the slope and is redirected in the forward direction.

        Whatever we think is the energy source, a horizontal MDPM, or primarily a vertical one, pluss redirecting of the pull of gravity by torque, it still seems that redirection by rotation over support plays an important role in the forward movement. As we fall down from the high point of the trajectory, and the only force is straight down, the pull of gravity, which is first redirected forward by momentum, it also seems that from the moment the foot touches the ground, and support is established, in combination with the stiffness of the body, that this further redirects the energy provided by momentum in the forward direction. If not for this support, and the stiffness of the body, COG will rapidly move more down that forward.

        So redirection of energy by rotation, whatever the source of the energy, seems a vital part of running.

        Also, I can not quite agree that the situation of the pendulum inside the forward moving train is the same as the rotation over support that happens in running. The pendulum in the train is fixed to a stable unmoving point, which is a vital difference to the change of BOS that happens in running. If we somehow made a device that made a well timed change of the fixed point of the pendulum at the time when it reached its outer point of its swing, the pendulum would move forward relative to the train.

    • canute1 Says:

      I do not agree that the downward pull of gravity is deflected forwards by the forward momentum. As the body falls towards earth in late flight phase, the momentum imparted by gravity is downwards, and the ongoing motion due to horizontal momentum contunes to be horizontal. The resultant motion of the COG is forward and down. The horizontal momentum can only be changed by an external force – namely the horizontal component of ground reaction force after foot fall.

      The situation of a ball rolling down a slope where the horizontal component of the reaction by the surface imparts a horizontal motion, which we discussed previously, illustrates the fact that a vertical motion can be deflected to horizontal motion by a surface reaction force, I do not think it is a good illustration of what happens during running on a level surface.

      As the COG continues to descend in the first part of stance, the horizontal component of ground reaction force acts backwards and exerts a braking effect that slows the forward progression of the body. In the second half of stance, horizontal component of GRF is forwards. The net effect of hGRF is a ‘to and fro’ fluctuation in the forward progression of the body.

      I did not intend the pendulum in the railway carriage to be an exact description of running. It was merely an analogy to illustrate what I meant by to and fro motion superimposed on a steady forward motion.

  44. Hans Holter Solhjell Says:

    Hi Canute.

    Thank you for your patience in answering my questions and comments so far.

    I will here try to work a bit more with the topic of torque making no net contribution, and see again if it is possible to make sense of it within a “flow with gravity, no horizontal push” perspective on running. For the most part, we have been trough all the major points several times, and my understanding of the topics has increased a lot. I have also read up on physics, although not diving into the maths and formulas, but relating to it from an experiential perspective.

    There are obvious weaknesses in the original account of Romanov, and I can see several places where he is inacurate and incomplete, and where he has not been able to answer the most important counterarguments. At least as far as my reading of his work goes, and have read his major books, website and googled the relevant topics.

    Anyway, I here go trough the whole thing again, and present some new ways of looking at what happens when we run at at steady pace.

    As previously established, the standing human body is in a state of unstable equilibrium and has potential energy. A part of this potential energy is always in play as kinetic energy, as the standing human body is continuosly moving, with it’s COG outside of BOS.

    A primary task for the standing human body is to balance, create stability by managing this kinetic energy and keep it, and COG, within certain limits, as well as balancing out all the various forces that might influence it, like wind, a push and so on.

    Also, the standing human body is constructed in such a way that COG has a tendency to move forward, rather than sideways of backwards, due to more of it’s weight being in front of BOS. This also means that no push is needed to turn more of the potential energy into kinetic energy and initiate movement of COG forward in the first step. I only has to reduce my effort to stabilize, and increase the unstability of my system, or readjust, my balance, and allow my COG to release forward, more outside of BOS.

    To stand still, what we must do is to hold back, balance and restrain our potential energy, and provide stability to the system, to not move forward. We are in fact very much like the steel ball in the experiment previously described, having to be held in place to not move, with a large store of gravitational potential energy.

    As I release my COG, and it moves, in the first part of this trajectory, the major componen of the movementt is forward, and only minimally downward, but the loss in height and potential energy needs to be restored.

    So to run I also need to provide a vertical push to get airborn, and regain the height of COG, and restore it’s potential energy. At this point, in the second half of stance, we have forward torque working, as well as the vertical push, and also a horizontal component as the vertical push happens at an angle.

    In this first part of the first step (second part of stance for later steps) the relative contribution from the horizontal component of the push can be larger or smaller, and can play an important role in explosive acceleration. It also possible, as you described, to put oneself at an angle an push off from there. What is important here is that at a slow to medium first step acceleration the horizontal component is a present component, but from an energy perspective not necessary to create forward movement of COG. There is enough potential energy stored in my body to move my COG forwards. Pushing can add energy, to increase acceleration, but this is not necessary to accelerate. As I take further steps I can further accelerate simply by letting forward torque work for longer than backwards torque.

    Forward torque at this point, as long as I accelerate by allowing forward torque to work for longer than backward torque, and as far as I have come to understand this trough our discussion, provides net energy.

    Obviosly this is only sustainable for a short period, or I will continue to rotate forward and fall to the ground. At some point I will have to stop accelerating, rebalance and increase the stability of my system. Our question here is what happens after initial acceleration and when I continue at a steady pace.

    As we have agreed, when we run at steady pace, the backwards and forwards torque has to be balanced. But I would like us to look more closely at the situation at this point, and I suggest that there are more factors involved.

    As described, the standing human body is in a state of unstable equilibrium and has potential energy. A part of this potential is always in play as kinetic energy, and a main task for the standing human body is to create stability, balance and create stability.

    The same is true for the running human body, which is also in a state of even more, and more dynamic, unstable equilibrium, with the COG most of the time even more outside of BOS than while standing, and still with a construction that has a strong forward bias. In addition to this, we know that forward torque in running will work for longer than backwards torque if we don’t restrain it by balancing. Also, in addition to this, comes the small horizontal component of the push against the ground.

    So, we have several forward components all working at the same time. These factors actually provide a large surplus of potential forward movement in the system. Working in the opposite direction is ground friction and wind resistance, but the three forward components provide a lot more forward potential than is needed to overcome these two factors.

    It seems very likely that there is more than enough gravitational potential energy in the system, redirected by torque in the forward direction, to overcome wind and ground friction, so that the extra horizontal push component is not needed from an energy perspective to overcome these opposing factors, even if it is still there.

    In fact, the potential energy and kinetic energy provided is so large that we have to restrain it, just as in standing, by balancing to maintain a dynamic, unstable equilibrium. If I adjust down my effort to restrain it, I will first accelerate over several steps, and then in the end fall to the ground, or I can choose to create a new equilibrium at a higher speed. If I completely stop to restrain it in one moment, I will fall straight down.

    I suggest that the job we are performing in steady speed running, after initial acceleration, is not pushing forward to overcome friction and wind, but balancing, actually restraining ourself, to keep all the forces that influence us, the forward factors described, including the horizontal push, and the opposite forces of friction and wind in a balance, making sure that the sum of all the forces are zero.

    And when the sum of all forces influencing a system is zero, there is no change neither in the forwards or backwards direction. The system continues to do what it does, which in the case of the runner, after initial acceleration, is to continue forward at a steady pace. This state, where the sum of all forces influencing a system is zero is called mechanical equilibrium.

    I am suggesting that the sum of all the various forces influencing the runner at steady speed over the whole gait cycle is zero, and that we are in a state of mechanical equilibrium. And this is the case because we make it so, by balancing all the forces that influence us, including the kinetic energy of our bodies and the torques.

    I suggest that the same is the case while running at steady pace. We have to hold back, balance the forward and backward torque, so as not to let the forward torque dominate, and not fall to the ground. But we can balance this in such a way that we can overcome friction and wind resistance and maintain a steady pace. And we can use torque for the initial acceleration, and the final deceleration, as well as for acceleration and deceleration while running.

    So in this way, I suggest that gravitational potential energy provides the horizontal movement component during the whole run, even at steady pace. The movement push, hGRF, that show up in the pressure plate data, is not what brings us forward, but is as previously suggested a present but not nescesery horisontal impulse. Apart from in cases of explosive acceleration from a stand still or low speed where we can push of from a much sharper angle than at steady pace running, it is a side effect of the nescesary effort to create vertical lift, which happens slightly off from the vertical position. Also, the small extra horizontal energy provided by the push only adds to the much larger surplus of potential forward movement that is already built into the human body by it’s construction, and therefore also has to be balanced out by our effort to balance the various energy factors influencing our system, to keep the sum of the factors at zero. There is no need for net horizontal energy from the push.

  45. Hans Holter Solhjell Says:

    I look forward to your comment’s and critique.

    • canute1 Says:


      You describe your perspective as “flow with gravity, no horizontal push” I too favour minimization of horizontal push but this must be balanced against the cost of minimising horizontal push. At constant speed, in the absence of air resistance, horizontal push is only required to overcome braking. Braking can minimised by short time on stance. Short time on stance is achieved by the combination of large vertical push and relatively high cadence. So in general, small horizontal push demands strong vertical push – and a compromise must be reached. My main critique of the way that Pose misrepresents the push is that it underestimates the magnitude of vertical push (as discussed in my post of 10 Feb 2010).

      With regard to your description of the acceleration phase, I would describe things a little differently, but I do not disagree with concept that de-stabilization of the body by leaning forward provides the stimulus to acceleration. I would emphasise that the net work that is required to achieve acceleration is done by the muscles pushing against the ground.

      With regard to your description of running at a steady speed, we agree that there is no need for much horizontal propulsion. I disagree that gravity provides any energy at all for horizontal propulsion. Producing rotation does not add to horizontal propulsion. Furthermore as we agree the rotation must be cancelled to avoid a fall.

      As stated in my first paragraph, the main task at constant speed is to exert a strong enough vertical push to get airborne. We can reduce the cost of this by ensuring that we capture kinetic energy as elastic energy at foot fall, but regardless of whether the energy is provided from stored elastic energy or muscle contraction, the vertical push is large and we must make our bodies strong enough to withstand this.

      The second main task is moving the swing leg forward during the swing phase. This does not involve a push against the ground.

      The third task is to minimise braking by spending as short a time on stance as is practical, but because this time cannot be reduced to zero we do need to exert a small horizontal push in the second half of stance to balance the loss of energy due to braking. Gravity cannot help with this because it cannot provide net horizontal propulsion.

      We agree that at steady speed, horizontal forces should be minimal. I disagree that ‘flow with gravity’ contributes usefully. Just as a vortex in a stream of water does not contribute to downstream flow, the small rotational effect of gravity does not contribute to forward flow.

  46. Hans Holter Solhjell Says:

    Hi Canute,

    Since we already have agreed several times that the there needs to be a vertical push, and that this involves muscular work and metabolic cost, and that there is no “free” energy and so on involved, I am more than happy to agree that the net energy in running comes from the push. But since we have discussed several aspects of the qualities of the push, I am really interested in the horizontal aspect and the details of that, as well has how to explain the forward movement from a physics perspective. Both in the acceleration phase and at steady speed. Do we push forward, or do we release our weight forward by adjusting our balancing effort?
    Also, is there at steady pace a need for net energy input from a movement push? Or is there more than enough gravitational energy, restored for each step by the vertical lift, that can be redirected into forward movement, making the need for net energy from movement push superfluous?

    I find that it has a large impact on the choice of technique and discussions of various aspects of it whether one believes there is a need for a horizontal push movement, that actually pushes you forward and provides net energy that is needed to move forward, or if there is no such need for net horizontal push movement.

    For instance, a runner running at steady pace, but who believes he has to push forward and as a technique based on this belief, rather than releasing his weight forward at the same time he pushes to create a vertical lift, will actually, if my account above is correct, waste energy both in trying to push forward, which most likely only creates a larger vertical lift, but he also has to spend more energy balancing and stabilizing, making sure not that the sum of all forces influencing his body is zero. He is pushing only against his own effort to balance and stabilize, as well as up.

    You seem to agree that acceleration can happen the way I described. This also implies that we have the ability to adjust the torques, and that there is surplus gravitational energy in the system that we have to actively balance.

    You also say that there is a need for a movement push to overcome braking forces, even if these can be minimized, as well as wind resistance. This is where we seem to disagree at this point.

    You also say that torque creates rotation, not forward movement. I am not really sure what you mean by this, and I agree, torgue on it self, left unchecked, will rotate the body forward to the ground. But if we look at what happens and how the various components work together the direction is clearly forward. In the second part of stance, we have forward torque, as well as vertical push, which together creates upwards rotation, in the forward direction. After ultimate height is reached, momentum carries the body forward, and gravity pulls it down, in a forward rotation, leading into the first part of stance and forward torque, and to the low point of COG. So the line of movement of COG is clearly forward, but in a partly circular fashion, created by torque in interaction with the vertical push. If push did not have the torque to interact with, the movement would just be up and down, not forward.

    I previously agreed that the torques has to be balanced out and did not provide net energy. My new position is that at steady speed all the forces influencing the body, it’s own potential and kinetic energy, the torques, the horizontal push, the braking and wind resistance needs to be balanced out so that the sum of all forces is zero, to make sure that we stay at at steady pace. Any addition or reducing of net energy will lead to acceleration or deceleration.

    Also, my position is that there is a large surplus of energy in the system, due to the body’s unstable construction and inherent potential energy, which trough torque redirects in the forward direction, and is restored trough, as well as augmented trough the vertical lift. This energy needs to be restrained, by balancing. At steady speed, there is no need for net energy from a movement push, and the smaller the actually occurring push is, the better.

    But I will moderate myself, and say that as there are several components working in the forwards direction, it is not really possible to single out one and say it is not necessary or is not the main mover. The important aspect is that the sum of all forces at steady pace is zero. But it also implies that we can make the horizontal push component as small as possible, and probably the smaller the better.

    I still agree that the forces involved are large, especially as we land and stabilize in the first part of stance, and as we create vertical lift in the second part of stance, and that there is significant stress on the body, and that one should prepare properly to deal with the potential risks involved.

    Also, the need for net energy from the push will vary depending on intention. A larger initial acceleration for instance, needs a net input from movement push. Especially a sprinter in the starting blocks, who actually has to raise is COG quite a lot in the initial steps, and create backwards torque while at the same time accelerating explosively forward.

    I would like your comment more specifically on the two new points I made in the account above,

    1. The point that the sum of all forces when we run at steady speed, after initial acceleration, should be zero.

    2. The point that there is a surplus of gravitational energy (that we previously have agreed can be redirected in the forward direction) and forward movement potential in the system that needs to restrained by balancing, and that we can adjust this balance in such a way as to overcome braking and wind resistance, and to keep the sum of all forces in the forward and backwards direction at zero.

    • canute1 Says:

      I agree that the sum of all impulses averaged over the gait cycle is zero when running at constant speed in the absence of wind resistance. (Impulse = force x time) This is a direct consequence of the law of conservation of momentum

      I do not agree that a surplus of gravitational energy can be used in such a way as to overcome braking and wind resistance. Braking acts backwards. Gravity acts down and therefore these two cannot balance each other.

      The only sense in which gravity can contribute to forward motion is by encouraging us to move the swing leg forward and press against the ground to stop us falling flat. Shortly after footfall,, this pressure against the ground produces a braking effect, but after mid-stance, the pressure against the ground exerts a forward propulsive effect that compensates for the braking effect.

      • Hans Holter Solhjell Says:

        Why only in the absence of wind resistance? Also in the presence of wind resistance this is the case, or not? I would say that we can adjust COG, releasing it, forward, to overcome wind resistance as well.

        Gravity works down, yes, but you have previously agreed that the direction of an object that is pulled down by gravity can be redirected by GRF, like in the steel ball experiment, where the steel ball is moved in the forward direction due to the bowl. And in the case of the human body, COG, as we reduce our effort to balance, will move forward, due to the remaining rigidity of our body, and ground friction. It will not move straight down, unless I release all rigidity. If I unbalance and release my weight, COG, forward, as well as push, I will now move forward and up. Torque, powered by the weight of my weight that is outside of BOS, moves me forward, while the energy provided by my push moves me up, resulting in an forward, upward, circular trajectory. Do you disagree with this, or would you prefer to word it differently, with more precise physics terms?

      • Hans Holter Solhjell Says:

        You mention the law of conservation of momentum. I am not sure if this is exactly the same situation or not so maybe you can elaborate. Momentum as I understand it will obviously play a large role once we start moving, or even while standing, but then in relation to the small movements involved as COG moves relative to BOS.

        But in the presence of friction, to keep a steady pace pace there need to be some energy input to maintain the steady pace, so this is not the exact same as momentum? Momentum will only maintain steady pace in a completely frictionless condition, correct?

    • canute1 Says:

      In the presence of wind resistance an external force acts on the body and tends to slow it. Therefore a net forward impulse must be applied to prevent a loss of momentum due to the wind. This forward impulse is provided by the push against the ground.

      We agree that an object falling down can be redirected by a horizontal push (as in a ball rolling down a incline) but the forward propulsion is produced by the push. Furthermore we have agreed that if there is no net loss of height, any gain in kinetic energy due to a fall must be compensated for by an upwards push that return the object to its original height. Therefore I do not agree with your proposal that a surplus of gravitational energy can be used in such a way as to overcome braking and wind resistance.

      Unless the runner’s foot slips on the ground, friction does not contribute to horizontal work. In fact strictly speaking, when there is no sliding we should describe the force that holds the shoe in place n the ground as stiction, not friction. Friction only occurs when one surface slides over another surface.

      The law of conservation of momentum states that unless an external force acts on a body to produce acceleration, the body does not experience a change momentum. For a body with constant mass, this means there is no change in velocity. Conversely, when a body maintains a constant horizontal velocity (averaged over the gait cycle) the net horizontal impulse (force x time) averaged over the gait cycle must be zero

      • Hans Holter Solhjell Says:

        Thank you for elaborating Canute.

        Maybe we should also challenge the push model a bit. A push model of horizontal movement in running should also fit with various observable facts.

        For instance, better runners seem to run faster with less vertical displacement than not so good runners. A larger push should according too this model make one run faster, but due to angle of pushing, one should also have more vertical lift. This does not seem to happen in the better runner.

        Also the push seem to happen at a fairly steep angle, how do this angle bring about the kind of trajectory of COG that can be observed?

        And what about the extensor paradox?

        How do the horizontal push interact with the torques? One should be able to calculate the optimal time and amount to push trough a model like this, and then practice timing and so on, but I have not seen anything like this presented.

      • Hans Holter Solhjell Says:

        You write,

        “We agree that an object falling down can be redirected by a horizontal push (as in a ball rolling down a incline) but the forward propulsion is produced by the push”.

        You are here talking about the push created by the incline, a passive object which produces GRF. If this is the push you are talking about, I agree with you, the redirection happens by this kind of push. But this is not the type of push created by the human body, a movement requiring metabolic energy. It might be that our language is not precise enough to have a clear and precise dialogue so not to confuse these two important differences. You are also coming from a physics background, where you are very familiar with the terms GRF and relate to this as the upwards push of the ground, while I come from a experiential movement background and use the term pushing in relation to human movement.

        To your other point, we also have to remember that the vertical push, push movement, in running not only raises COG to its previous standing height but several centimeters above it, and that the point of maximal potential gravitational energy and instability is at the top of the trajectory of COG. At this point the kinetic energy of COG is directed forward by momentum only.

        I am not sure if you would like a different formulation better, that all the energy for both the vertical and horizontal movement during steady pace comes from the vertical push. This vertical push both gets the body airborne, and repeatedly elevates COG, so as to be able to pulled down by gravity and redirected forward repeatedly by GRF, ground passive push, as well as momentum during the flight fase. Gravity plays a role here, by providing the downwards pull that can be redirected forward via torque, and due to the body’s unstable equilibrium, but can only do so because a repeated effort is performed to raise COG above it’s starting height, reposition the limbs, as well as the effort to stabilize and balance the body and all the forces influencing it, keeping the net horizontal forces at zero.

        This fits with most of what I wrote earlier, and there is no net input from a horizontal push movement

      • Hans Holter Solhjell Says:


        I think it would be easier to be precise if we used different terms for the different types of push. For instance GRFpush for the push created by the ground by any object that pushes down on it, regardless of the qualities of the object. And Metabolic Demanding Push Movement for the push push created by the body by pressing the leg actively down or backwards, MDPM, metabolically demanding push movement. This obviously will also imply GRFpush, but so will weight alone, without a MDPM.

        So, in the post above, what I am suggesting is that all the energy for both the vertical as well as the horizontal movement in running is supplied by the vertical MDPM, and that this energy is redirected in the forward direction by the mechanisms described. There is more than enough energy provided by the vertical MDPM to redirect the movement forward and to overcome the various types of resistance, as well as to accelerate while at speed.

        There is, if I am correct in this, no need for a net metabolic energy contribution from a horizontal MDPM, but there will of course be both a large vGRF push as well as a smaller hGRF push that can be measured.

        So gravity does no net work in this case (or at most the contribution required to overcome wind and ground resistance and to keep the overall sum in the horizontal direction at zero) but it provides an essential part of the mechanism whereby the vertical MDPM is redirected into forward movement.

        The body, as such, only contains the potential gravitational energy for one fall straight down, and any redirection and restoration of this energy requires metabolic energy, so there in any case is a lot of work for the body to do, and no free energy, just more or less efficient use of it.

      • canute1 Says:

        With regard to your post at 10:39, the most important issue regarding what you call the push model is that time on stance should be as short as practical to minimise braking, and thereby minimise the need for horizontal push. I believe that the reason that Pose help to reduce injury in runners who previously produced too much horizontal push in a misguide attempt to increase stride length, is that Pose encourages high cadence and short time on stance. Both of these factors both decrease the need for horizontal push and hence minimise some types of injury (though a very short time on stance creates different risks). These benefits of Pose have nothing to do with gravitational torque.

        I do not know of good evidence that better runner shave less vertical displacement of the COG. They often have less vertical displacement of the head because they lift the swinging leg higher, and this contributes to elevation of the COG so the torso rises less. However Weyand’s data demonstrates that less good runners spend less time airborne Shorter airborne period must result in lesser elevation of the COG. Overall the evidence indicates that less good runners produce slightly less elevation of the COG.

        The extensor paradox is only a paradox if you ignore the glutes. The glutes are active during stance and have the potential make a substantial contribution to capturing elastic energy required for both the vertical and horizontal push. The glutes continue to be active in late stance, at which stage they assist the horizontal push. I think that some people consider there is an extensor paradox because the knee extensors (quads ) are not strongly active prior to footfall, but quads also flex the hip, so quad activation during late swing would be counter-productive. At this stage the hams are active. However, immediately after footfall, the quads become active. I do not think there is any paradox.

        With regard to calculating optimal time and amount of push, there is some debate about what should happen in late stance. Some people advocate allowing the hip to extend as much as possible before ground contact is broken in order to maximise the storage of elastic energy to drive the swing. I am not convinced by this. I certainly think it is risky to deliberately try to keep the foot on the ground in late stance. I agree with the Pose approach of aiming to get off stance quickly.

      • canute1 Says:


        With regard to your comment at 11:05, even an animated body obeys the laws of Newtonian mechanics.

        I agree that at the peak of the airborne phase the kinetic energy of the COG is directed forwards and is maintained by momentum. I also agree that gravity pulls the body down, but as explained several times, it makes no sense to say that the horizontal push redirects the kinetic energy generated by gravity in a forward direction. The energy generated by the fall must be repaid by a vertical push. In early stance, the horizontal GRF actually slows the body but in late stance the direction of hGRF is reversed, and it generates the small amount of forward impulse required to compensate for braking.

    • canute1 Says:


      With regard to your comment at 12:34, I do not like the term MDPM because it does not make clear that a portion of the push is generated by elastic recoil. However that is only a matter of terminology. More importantly, the push by the leg acts along the line from COG to point of support. It has a large vertical component and a small horizontal component. As discussed several times, at constant speed in the absence of wind resistance, the net horizontal impulse over the whole gait cycle is zero. However, in the first half of stance, there is a backward directed impulse, and in the second half, there is a forward directed impulse. The relevant force that produces this impulse is the inevitable horizontal component generated when the leg pushes obliquely against the ground. In the presence of wind resistance, a net forward impulse is required and the leg must generate a stronger forward impulse. As discussed many times, there is no way that gravity can do this.

      • Hans Holter Solhjell Says:

        Good morning Canute.

        Thank you for your replies.

        I would like to ask more about the reason for way you mean that the torques at steady speed has to be balanced, so forward torque does not work for longer than backwards torque, and therefore can not provide net energy during steady pace.

        It seems we agree that for some steps at least, forward torque can work for longer than backwards torque, and therefore for these steps provide net energy. This surplus energy will lead to acceleration of the body in the forward direction, as long as we also provide the vertical lift.

        Can we agree on this, even if you might choose to use different words to describe this process?

        My understanding of your position is that you further say that forward torque can not work for longer than backwards torque indefinitely, because over more steps, COG will end up on the ground due to the excess forward and down rotation provided by the forward torque.

        Is my understanding of your position correct, would you prefer to choose different words to describe this, or did I misunderstand what you meant?

        If so, I at least one challenge to this position to present.

    • canute1 Says:


      I agree that under many circumstances forward torque works for longer than backwards torque. However, this does not mean that there is net work done by the torques, or that net energy is acquired by the body. The net change in angular momentum produced by a torque depends on the magnitude of the torque and on the duration of action. The magnitude of the torque depends on the force which in this case is gravity, and the length of the relevant lever arm. The length of the lever arm is the horizontal distance from the pivot point (point of support) to the COG. This changes during stance in a manner that depends on the trajectory of the COG. The exact trajectory of the COG is a little difficult to compute. In my slightly simplified model, the calculation showed that there is no net change in angular momentum. If I employed accurate data for vGRF and hGRF I could compute the exact path of the COG, but I have not yet done this. I believe that in even if I had done the calculation with perfect accuracy, the net increase in head-forward angular momentum due to gravity would equal the head-backward angular momentum imparted by the wind. In the absence of wind resistance, this would be zero. However, since I have not done the precise calculation, I am prepared to accept during the present discussion that it remains to be demonstrated whether or not there is a net increase in rotational momentum (and in rotational energy) during the stance phase.

      Nonetheless, even if there were to be an increase in rotational momentum that would not move the body forwards. As we have discussed, it is necessary to apply some horizontal force to redirect movement in a forward direction. Furthermore, if there was any net gain in energy due to the action of gravity while the body was on stance, this could only have occurred if there was lowering of the COG. (Gravitational potential energy can only be converted to kinetic energy by a lowering of the height of the COG. This is the law of conservation of energy that must be obeyed by all matter, whether animate or inanimate). So if there had been any gain in kinetic energy due to lowering of the COG, the lowering would have to be reversed by an upward push. The energy required to get the body back to its initial height would exactly cancel the energy gained from the fall. Thus gravity cannot provide surplus energy for horizontal propulsion. The energy required to compensate for the loss of horizontal kinetic energy due to braking in the first part of stance can only be provided by the horizontal push.

      So in summary, I do not agree that gravitational torque can provide surplus energy that leads to acceleration of the body in the forward direction.

      • Hans Holter Solhjell Says:

        Hi Canute.

        You have given me a lot to think about here, and I need to reread your post a few more times when I have a bit more time. But i do get that if there is no net fall of COG, no net energy from gravity.

        One thing I would like to question in the mean time is the idea that if forward torque works for longer than backwards torque over several steps, in the end COG will end up on the ground. I find it hard to understand the mechanics of this.

        For this to be the case it seems to me that the extra length forward torque works longer than backwards torque for each step would have to be added to each other. I can not see how this can be the case, but that rather, each step should be considered a separate event in this regard.

        Lets say we start at mid stance at 90 degrees and let forward torque work for 10 degrees to a 100 degrees. We also push off and get airborne, and in the air reposition our limbs and adjust our body so that we land behind BOS at an 85 degrees angle. We again rotate forward to a 100 degrees angle while pushing off, and repeat. For each step we start at an 85 degrees angle, due to our repositioning and adjustments of our body in the air, so even if forward torque works for 5 degrees longer for each and every step, it never ads up, and forwards torque always ends at a 100 degree angle where we push of.

        I am not able to see the mechanical reason for why the torques need to be balanced in the way suggested to avoid ending up on the ground. I might have misunderstood the meaning of balancing the torques though.

      • Hans Holter Solhjell Says:

        Hi Canute.

        You describe the factors that influence the magnitude of the torque. It seems to me that since the lever arm lengthens during forward torque, and shortens during backwards torque, this implies that even for an equal working length of the torques, forwards torque have a (here I am not sure what is the precise language, is of a, have a, or produces a larger force) larger magnitude than backwards torque. Does it not?

        I agree that there is a need for a horizontal force to move the body forward, and as we know, there is the hGRF involved. I am also ready to be convinced that there also is a definite need for a hMDPM (lets just for the sake of making this easy say that this also includes elastic energy) but I still have question in this regard. Obviously, and as I previously have agreed and discussed, there also is a hMDPM component, but I am trying to understand both the exact mechanics, the physics, and also relate this to my movement experiments, and my experience of running, and the various changes I have made in my technique over the last years. Also, I find it easier to create an image of all the mechanical aspect of what happens in the gait cycle, and what I see in good runners

        You have been very patient in answering my questions, which at times might be a bit naive in relation to the physics, so let me know if your patience or time runs out. For myself I find the discussion very stimulating, even though time consuming.

        Regarding the need for a net loss of height for COG for gravity to do work I agree that this is the case. And as we have agreed many times, there is indeed a MDPM as well as other actions that has to occur to counter this drop or we will for sure end up on the ground. And as we know this push happens during forwards torque.

        But as we also know, it is possible for torque to work, the pull of gravity redirected by stiffness and stiction, and as a consequence COG to move forward and rise at the same time for a certain part of the trajectory. This demands some other energy input than gravity, and this in the runner is provided by the MDPM. So we can in a way say that the runner pays for the subsequent drop in COG in advance by lifting COG while forwards torque work. But from mid stance to mid stance and during one whole gait cycle there is no overall drop in COG even if COG moves up and down during the cycle. I assume this is what you are referring to. I have to give some more thought to the significance of that, and I have some questions in this regard to various elements of the gait cycle.

        But my first, and maybe naive thought, is that it seems to me that the only thing required for COG to move up is a vertical push, and that the subsequent fall can be redirected forward. How much forward, and against what amount of resistance, friction might then be a relevant question, if this is not all wrong to start with.

    • canute1 Says:

      If you accept that gravity contributes no net energy, I do not understand the reason for your claim (at 9:36 on 7th Oct) that the surplus energy generated by gravitational torque will lead to acceleration of the body in the forward direction. If you agree that any kinetic energy created by the fall must be repaid by an upward force, what is the source of the surplus energy?

      I have never intentionally stated that if forward torque works for longer than backwards torque over several steps, in the end COG will end up on the ground. The amount of angular momentum produced by a torque depends not only on the time for which it acts, but also on the magnitude of the torque. This is not constant during the stance phase as discussed in my recent response. So the amount of rotation produced does not depend only on time of action of the torque

      I have previously stated that if there is a rotation in a head forward direction on each step that is not balanced by a rotation in the opposite direction, you will end up on your face, simply because in each successive step the rotation will bring the body nearer to a horizontal position. There are several u-tube videos showing this effect. One shows a man holding a baby as he reaches the bottom of a slide. He stands up an because of his forward momentum he keeps running. However because the combined COG of man and baby is in front of his torso, he lands a little behind the combined COG on each step. He continue to rotate at an increasing speed. After about 5 steps both man and baby crash to the ground. (As far as I know neither were hurt). Unfortunately I cannot find that U-tube video now.

      However, the crucial issue is the fact that gravitational torque provides no surplus energy. The debate about rotation is largely irrelevant. In fact, In the second half of stance, the net vertical force on the body is up, not down as shown in the Pose pictures. Because of the effect of the upward displacement of the COG, the head-forward rotation of the line from point of support to the COG (due to the fact that the COG is moving forwards) actually slows a little rather than accelerating during the second half of stance . I described this effect in some detail a few weeks ago

      • canute1 Says:

        Our responses have got out of order. This is a response to you comment at 9:54 pm on 9th Oct.
        Despite saying you accept that gravity cannot contribute energy to horizontal propulsion you appear to continue to argue that gravitational torque might somehow provide surplus energy forward propulsion. The law of conservation of energy makes it clear that if there is no net change in height of the COG, gravitational torque can provide no surplus energy.

        With regard to your specific question about the changing lever arm in the first and second half of stance, the question of whether it is increasing or decreasing is less important that the average value during the relevant time period. As I said before, when I did the calculation using my approximate model, the net torque was zero. I have not done the calculation using precise information about the trajectory of the COG but I am fairly confident that the net torque will be very near zero. Maybe one day I will do the precise calculation, but it will almost certainly add nothing useful to the understanding of the energy costs of running, firstly because the law of conservation of energy makes it clear that gravity can provide no surplus energy for horizontal propulsion, and secondly because the energy associated with rotational motion is trivial compared with other energy costs.

        The energy associated with torque is certainly trivial in magnitude compared the energy required to lift the body from its lowest point at mid-stance to its high point at mid-flight. It is gravity that makes this such an effortful task, though of course the gravitation potential energy gained is returned as kinetic energy as the body falls and a fraction of this energy is retained as elastic energy. If you are concerned with the role that gravity plays in the energy costs of running, the issue of lifting the body vertically far outweighs any rotational effect due to torque. And furthermore, as we have discussed many times, torque produces rotation, not forward propulsion. Any deflection forwards is achieved by a horizontal push. The push is a consequence of the obliquity of the leg and it balances the effect of the oppositely directed obliquity of the leg in the first half of stance.
        If you want to take account of the actual energy costs of running you must consider first the cost of elevating the body, secondly the horizontal push and thirdly the cost of repositioning the limbs. Everything else is trivial.

      • Hans Holter Solhjell Says:

        Hi Canute.

        It might be more precise to suggest that gravitational torque does not provide net energy, but that the vertical push, MDPM, provides the net energy by lifting COG, and that this energy, by torque is then redirected in the forward direction. The human body, as described, seem to have some important characteristics that enables such a redirection, and fine tuning and balancing of it.

        So torque does not produce net energy, but might redirect the energy provided by the body to lift COG. So the source of the surplus energy is the MDPM, and we have several times agreed that COG has to be lifted and that the body produces the lift, but I still find the further mechanism of horizontal redirection, or horizontal push not to be beyond questioning, and with important consequences.

        My suggestion might not be factually correct, but anyway it might be a better wording of my suggestion. You speak the language of physics as a native, while I am more of a foreigner who might know what I want to say but might choose the wrong words to say it.

        Regarding COG ending up on the ground, It seems I have misinterpreted your formulation for forward and backward rotation as implying torque, and also that the forward and backward rotation would have to be of equal length. This does not seem to be the case, and we seem to have agree that the human body has the capability to balance the torques, as well as forward and backward rotation without them having to be of equal length.

        I reread your previous explanation of the effects of forward rotation and rise of COG but will have to read it again at a later point when I have more time.

      • canute1 Says:

        Even if there is unbalanced torque (which I consider unlikely) I see no evidence at all that torque redirects vertical push into the horizontal direction. It is the obliquity of the leg that ensures that the push has a horizontal component. The obliquity arises from rotation, but the major cause of the rotation is the forward movement the body sustained by momentum.

        To summarise:

        Gravity contributes no net energy to horizontal motion.

        The amount of energy associated with rotation due to gravitation torque is very small compared with the energy associated with other movements that occur when running.

        All the available evidence suggest that the change angular momentum attributed to torque is balanced, but even if you disagree with this, the first two points make this an issue of trivial importance.

  47. Hans Holter Solhjell Says:

    I see that at the top I ask whether there is a need for net energy from the push. I then mean a horizontal push, not the vertical, where agree there is a need for energy input. I made a couple of more mistakes throughout, so to bad there is noe editing function here. I refer to forward torque at one point where it should be backward torque. I trust that you know what I mean from the context.

    • canute1 Says:

      That is OK. I understood what you meant.

      At constant speed in the absence of wind, there is no net horizontal work done. The work done by braking force pushing backwards is balanced by the work done by the leg pushing against the ground after mid-stance. However, unless we can capture the braking energy as elastic energy, then we do face a net energy cost providing forward propulsion in the absence of wind.

  48. unique endurance challenge Says:

    With havin so much content do you ever run into any issues of plagorism or copyright infringement?
    My blog has a lot of unique content I’ve either written myself or
    outsourced but it seems a lot of it is popping it up all
    over the internet without my agreement. Do you know any ways to help reduce content from
    being ripped off? I’d really appreciate it.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: