Barefoot running and foot dynamics

In a review article published in the journal Nature in 2004 (vol 432, pp 345-352) Bramble and Lieberman from Harvard University compared the efficiency of human running with the running of other mammals.  They concluded that it was likely that the human musculo-skeletal system had evolved in a manner that allowed our forebears to run for long distances in pursuit of game, in the distant past before the development of spears or guns.  Unlike the lioness which relies on a sprint of a few hundred metres to capture her prey, it appears that our forebears relied on endurance rather than speed; perhaps they pursued some ungulate for hour after hour until their quarry was exhausted.  Clearly one of the major adaptations is the ability to get airborne and to land in way that captures gravitational energy as elastic potential energy.  The quads and calf muscles are well suited to this purpose.  But what about the feet?   In an era long before the engineers from companies such as Adidas and Nike developed the heavily padded running shoe that makes heel-striking feasible – humans ran barefoot. A more recent study of habitual barefoot runners by Lieberman demonstrates clearly what many had expected: habitual barefoot runners land on the fore-foot or mid-foot, and that this style of landing does not produce the sharp initial rise in ground reaction force that is seen when the runner lands on the heel (Nature, 463: 531-565; 2010).

The design of the human foot

It is scarcely surprising that barefoot runners land on forefoot or mid-foot because the design of the human foot is well suited to such a landing.  The two most striking features are the longitudinal arch which extends from the metatarsal heads at the front of the foot to the heel, and subtalar joint that allows the movements known as pronation and supination. 

The longitudinal arch

The longitudinal arch is most pronounced along the medial (inside) border of the foot.  It is a spring-like structure well suited to absorbing impact, storing the energy of impact as elastic energy distributed within the connective tissue of the foot, the Achilles tendon and the calf muscles, and then releasing this energy as the body is propelled upwards by the recoil.

The subtalar joint

The subtalar joint is a compound joint involving the head of the talus (the bone within the ankle that links the long bones of the lower leg with the rest of the foot); the calcaneus (the heel bone located beneath the body of the talus) and the navicular bone (one of the major tarsal bones that acts as the keystone of the medial longitudinal arch).  The anatomy of the sub-talar joint is rather complex, but the crucial point is that it allows the foot to rotate onto its inner edge, locking the medial arch so that it can support the weight of the body.  This movement is known as pronation.  Rotation in the opposite direction around the subtalar joint rolls the foot onto its outer edge and takes the weight off the longitudinal arch

Foot dynamics when running

Because a runner’s swinging leg must move towards the midline of the body as it descends  towards footfall, the natural tendency for a barefoot runner is to land on the outer edge of the foot.  As the weight is taken up, the subtalar joint allows pronation which transfers the load to the longitudinal arch.  As the Achilles tendon and calf muscles take up the strain the heel drops towards the ground.  Provided the calf has not been maintained in an artificially tense state, by midstance the heel will be on the ground and the arch will be spring-loaded, exerting pressure of the ground.  The ground reaction force drives the body upwards and as recoil occurs, the foot tends to supinate transferring the load towards the outer edge of the foot so that it is distributed across the metatarsal heads immediately before lift-off.

Thus, within a period of duration in the range 80-200 milliseconds, the foot has undergone a complex series of movements.  Not only is there the hinge-like transition from slight plantar flexion of the ankle at footfall to marked dorsiflexion at mid-stance and plantar flexion at lift off, but also the inward roll (pronation) and the final outward supination.    A moderate degree of pronation (typically up to 15 degrees) is a crucial feature of distributing the load onto the longitudinal arch. 

Torque at hip and knee

The way the foot lands has a substantial influence on the twisting forces at hip and knee.  In a comparison of the torques occurring at the various joints of the leg during running during barefoot running compared with running in typical trainers with a padded heel, Kerrigan and colleagues demonstrated greater torques at hip and knee during shod running (PM &R: The Journal of Injury, Function and Rehabilitation, Vol. 1, pp 1058-1063, 2009).  The magnitude of the torques during running in shoes were in fact greater than those observed during walking with high heels.   The barefoot runners were observed to adopt a shorter stride length, which would be expected to result in lower ground reaction forces, but only a small proportion of the relative reduction torque when barefoot could be accounted for the reduced stride length.  It is likely that the altered distribution of load on the springy longitudinal arch was a major factor in the lower torques when barefoot.   

Is it the shoes or running style?

The main focus in the studies of Lieberman and of Kerrigan is on the differences of between barefoot running and shod running, but the issue is confounded by the differences in foot dynamics.  As about 20-30% of shod runners land on forefoot or mid-foot, it is important to know whether the apparent advantages of barefoot running are primarily a consequence of the being unshod or, instead, a consequence of the style of landing.  

One of the striking differences between the barefoot runner landing on the forefoot and shod heel-striker is the sharp rise in ground reaction force shortly after impact.  Although no study has demonstrated that this sharp spike is responsible for increased risk of injury, it is a prime suspect, as it would be expected to initiate a jarring impulse that would be transmitted up the leg to the knee, hips and spine.   It is easy to envisage how landing on the heel would result in an abrupt braking action that would generate the sharp spike of ground reaction force (GRF).  The role of the running shoe in causing or ameliorating this sharp spike is less clear. 

In fact, force plate data recorded during heel striking with and without shoes, which Lieberman and colleagues present on their website, demonstrates that shoes reduce the initial spike, but do not abolish it (  However, comparing barefoot heel striking with barefoot forefoot demonstrates that forefoot landing produces a far greater reduction in the initial spike.  Thus there is little doubt that the style of landing plays the major role in reducing the initial spike in GRF.  As expected, shoes are potentially helpful rather than harmful for the heel-striker.

However, the cardinal comparison of interest is that between barefoot and shod forefoot landing.  In fact the force plate data indicates that for forefoot runners, the initial sharp spike is virtually entirely abolished for both barefoot and shod running. Furthermore, there is very little difference between fore-foot striking in racing flats compared with conventional running shoes.  Thus, at least as far as the potentially harmful initial spike of GRF is concerned, there is little evidence of difference between barefoot and shod running with a fore-foot landing.

Non-conscious adaptation

One feature of particular note about the data presented by both Lieberman and colleagues and by Kerrigan and colleagues is that barefoot runners automatically adjust their running style to adapt to the lack of protection.  Not only do they tend to change from heel-strike to forefoot or mid-foot strike, but they also shorten their stride.  This has several implications.

First of all it demonstrates that the neuromuscular system acts automatically to adjust running action so as to minimize risk of injury.  This might be advanced as an argument in favour of running barefoot as it suggests that the bare foot is more responsive to the running surface whereas shoes dull this sensibility.  The automatic shortening of stride might also be regarded as an advantage in at least some circumstances.  The impact forces will be less, and in addition, running with a shorter stride and a higher cadence is more efficient, simply because less work is done against gravity in a larger number of shorter strides that cover the same distance.  However, shortening stride improves efficiency only up to a certain point beyond which it becomes less efficient (as discussed in my post on 31st Dec 2009.

In his book ‘Programmed to Run’ , Tom Miller reports data recorded by Buckalew in the US Olympic women’s marathon trial is 1984.  They recorded cadence and stride length in 40 women early in the race and again late in the race.  They compared the 10 women who achieved the fastest finishing times with the 10 who recorded the slowest finishing times.   Early in the race there was no difference between the two groups of women, but late in the race, the slower women exhibited a shorter stride despite maintaining the same cadence as the faster woman.  This loss of stride length despite maintenance of cadence is similar to that observed as runners grow older.  Thus, the decrease of stride when running barefoot might potentially impair performance unless an effort is made to ensure that cadence is maintained. 

Is it worth abandoning your shoes?

Despite the evidence from Lieberman’s studies indicating that the initial sharp spike of GRF is similar for a fore-foot runner whether running barefoot, in racing flats, or in conventional running shoes, a thick padded heel actually makes it more difficult to land on the fore-foot.  Furthermore, rigid motion control that prevented adequate pronation would hinder the transfer of load to the medial longitudinal arch, and built-in arch support might hinder the natural flattening of the arch as it takes up the load, thereby impeding the storage of elastic energy.  These considerations provide reason to consider abandoning shoes with heavily padded heel, arch support and rigid control of pronation. However, even before making a change such as this, it should be noted that the connective tissues supporting longitudinal arch, the Achilles tendon and calf muscle will be at risk of injury unless these structures have been adequately prepared for the task of capturing and storing the energy of impact, and the required neuromuscular coordination has been developed.

If a lifetime of wearing shoes has caused atrophy of tissues and loss of the required strength and sensitive neuromuscular coordination, then any transition to a more minimal shoe should be done gradually.

 The question of whether it is worth taking the further step of abandoning a minimal running shoe once one has acquired the strength and coordination to run comfortably and safely with fore-foot or mid-foot landing, is less easy to answer.  I am not aware of any evidence that indicates that this extra step is worthwhile. On the other hand, it would carry some risk: not only the risk of laceration or bruising of the foot on sharp objects, but also the possibility that stride would automatically decrease as an unconscious protective measure. This might be especially troublesome in the later stages of a long distance race.

There is one additional issue to consider.  Several studies have demonstrated that running barefoot uses about 5% less energy than shod running ( e.g. Squadrone and Gallozzi, J Sports Med Phys Fitness. 49(1):6-13; 2009.).  However it is not clear whether that is mainly accounted for by the weight of the shoe or the mechanics of running.  A study by Divert and colleagues (Int J Sports Med. 29(6):512-8, 2008) concluded that the soft padding of the shoe might result in a decrease of the storage and restitution of elastic energy capacity which could explain the lower net efficiency reported in shod running. 

For some people the feeling of freedom and the satisfaction of running naturally might be adequate justification for running barefoot.  However, I am inclined to regard shoes as the product of a natural phenomenon: the inventiveness of mankind.  I am as happy to celebrate human inventiveness as I am to celebrate primitive naturalism.  As I sit here typing, I am very pleased that human inventiveness had produced reading glasses. 

While there might be some satisfaction in running as our remote ancestors did, I suspect that if I had been born on the African savanna a million years ago, I would not be spending the few weeks before my 64th birthday contemplating how to maximize the chances that I will still be enjoying running in 10 years time.  It is more likely I would be worrying about how to avoid becoming a gristly morsel for some predatory feline within the next 10 days, if that fate had not already befallen me.  So I am perfectly happy to live in the modern era and to celebrate the inventiveness of my forebears.  However I hope I am wise enough to recognize when inventiveness has become counter-productive. I think the modern running shoe might be an example of counter-productive invention



In my youth I initially ran in sandshoes, but in the late 1960’s I bought an expensive pair of Adidas shoes – I do not remember the model – they had arch support and heavy padding especially under the heel.  In the shop they felt luxurious, so I eagerly bought them, took them home and set out on a trial run.  I could scarcely believe how uncomfortable they were.  Because I couldn’t believe it, I did not take them back, but persevered with them, but they did not get much better.  My next shoe purchase was a pair of Onisuka tigers; very lightweight racing flats which I wore for many years. I ran all my best marathons in my tigers.  Even now I regret that during one of the many house moves over the past four decades I jettisoned these dilapidated old friends.   The Onitsuka Company eventually became Asics, and I recently bought a pair of Asics hyperspeeds.   The hyperspeeds are lightweight and have a fairly minimal sole, though not quite as minimal as my old tigers.  I am enjoying running in them – but they have one irritating feature. Unlike the slightly rippled undersurface of the tigers, the hyperspeeds have a criss-cross pattern that is ideal for trapping small stones.   So I will be true to my age and murmur nostalgically that running shoes were far better when I was a lad.


14 Responses to “Barefoot running and foot dynamics”

  1. Rick Says:

    Yes Canute my Nike Free’s are always picking up stones between the tread :[
    Have you read this on why kenyan runners are fater

  2. Rick Says:

    ‘FASTER’ :]

  3. canute1 Says:

    Thanks for that link. The authors acknowledge that it was a small study without a comparison group and therefore they are cautious about drawing conclusions. The features characteristic of elite Kenyans that are consistent with my prior expectation are a history of running to and from school; slender legs including slender calf, and short time on stance.
    The features that I would be been less confident in predicting were the lower leg strength relative to European elite runners and the high ratio of hamstring eccentric strength to quadriceps concentric strength (H:Q ratio). The high H:Q ratio should help protect against hamstring injury. I would have been interested to see data for quadriceps eccentric strength as I believe that is important for achieving short time on stance. The authors speculate that short time on stance is due to stiffness of the legs but they do not present any evidence to support this speculation.

    I was also interested to note that time on stance was less for the right leg at all speeds tested – this was a small effect that is unlikely to have an observable effect on running efficiency, but interested me because there was no strength difference between right and left leg, so the difference is likely to be due to better neuromuscular coordination on the right. It confirms my belief that it is potentially worthwhile to work on exercises designed to improve neuromuscular coordination.

  4. Ewen Says:

    That’s particularly interesting about Non-conscious adaptation of running ‘style’ to barefoot running.

    Having done lots of barefoot running in the ‘old days’, I’m a big fan of the Frees (although they don’t look that minimal with the heel, they feel minimal and like running barefoot). As Rick says, small stones are a problem, but you don’t notice them until after, and don’t get stones on grass.

    I think ‘minimal’ shoes like your old Tigers are the go — they offer just enough protection so there’s less non-conscious adaptation (protective running), yet allow plenty of foot movement because of their flexibility.

  5. Ewen Says:

    Canute, there’s a good blog post covering studies on footstrike by Steve Magness:

    It was particularly interesting about how footstrike can change during the course of a race (even in a 1500m race) as fatigue sets in — also how ground contact time increases with fatigue. For runners interested in speed, it would seem that forefoot or midfoot landing is more efficient (more elastic energy is stored and used with that technique).

  6. canute1 Says:

    Thanks for that link. I am in complete agreement with Steve Magness regarding the importance of minimizing ground contact time, and about the need to maintain a short ground contact time throughout the race. This is partly a matter of developing strength endurance – and again I agree that repetitions of relatively long hill help this. It is also partly a matter of developing a style of foot dynamics that maximizes the capture of elastic energy – and I think that fore-foot or mid-foot striking is more efficient than heel striking in this regard.

    While I agree with all the main points made by Steve Magness, there is an additional point which he does not address, that I am still pondering. Is it best to land with a degree of planar flexion of the foot or alternatively, to land on forefoot with slight dorsiflexion. If the foot is plantar flexed at footfall, the calf can stretch further and might perhaps capture more of the gravitational energy, but the stretch will take longer, and this might result in longer time on stance. If the foot is already slightly dorsiflexed a foot-strike, the calf will already be under some tension (generated by tibialis anterior). It will develop it maximum tension more quickly and the peak tension might be greater. So I think that landing with slight plantar flexion will assist a quicker lift off, and is therefore preferable when sprinting or maybe also for middle distance races. However I suspect that the pre-tensioning of the calf by action of tibialis anterior is likely to be more exhausting than allowing the gravitational energy of footfall do more of the stretching of the calf. Hence, for a long distance runner I think that a greater degree of planar flexion and a more relaxed ankle might be better. However the more strength endurance training the athlete has done, the longer he/she will be able to maintain that slight degree of plantar flexion that promotes a rapid lift off. I also think that developing good neuromuscular coordination is important.

  7. Ewen Says:

    Canute, that’s an interesting point. It would be hard to evaluate. A high-speed video camera like Peter Larson uses would be handy –
    – then you could possibly calculate time on stance trying both planar flexion and dorsiflexion of the foot. High speed video of a group of elite athletes with similar PBs might be more useful to see if there’s a preference for either technique.

  8. canute1 Says:

    I agree that a high speed camera would be useful. It would be good to compare time on stance with different degrees of plantar (or dorsi) flexion at footfall. However, because it is likely that neuromuscular coordination is also crucial, it would be necessary to spend a substantial period achieving good coordination for each degree of plantar flexion.
    Comparing different elites would also be interesting though differences in strength and differences in degree of tension in quads and other muscles would also be expected to affect time on stance, so on balance I think the most convincing evidence would be obtained using the same runners for all degrees of plantar flexion, with adequate practice of each style.

  9. Aaron Says:

    I just discovered this blog, and I thought I’d contribute some anecdotal observations from my own barefoot training.

    Any workouts on rubberized track is risky business for barefoot running, but after spending all of last summer adapting to long distances on various terrain I could finally do it for a full set of intervals in the fall without blistering my feet. Generally I would run a modest 3 minute 800m with ~190 steps/min cadence. The foot strike that allowed me the least amount of contact time, rotation, and impact was a relaxed midfoot planting with some dorsiflexion. More hip rotation played a bit part in it too, probably by reducing the amount of horizontal push off and keeping my knees in line with my center of gravity.

  10. Aaron Says:

    Correction: I was thinking about horizontal vs vertical movement when I wrote “horizontal push off”; that should read “vertical push-off”.

  11. canute1 Says:

    Thanks for that comment. I am interested to hear that you found mid-foot landing with slight dorsiflexion worked well.

  12. Oliver Says:

    Ah how lovely to read about the Tiger Marathons. Were they the Cubs that were made in canvas with strips of leather at the toes? They had the identical soles to the leather and later the nylon versions. Nothing like ’em.

  13. running/how to run faster/quickness training/speed training/quickness speed training Says:

    running/how to run faster/quickness training/speed training/quickness speed training…

    […]Barefoot running and foot dynamics « Canute’s Efficient Running Site[…]…

  14. Natural running « Canute’s Efficient Running Site Says:

    […] 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 […]

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: