The big debates of the past decade: 2) shoe design

For almost a decade many runners have been captivated by the issue of running shoe design – a preoccupation fuelled two opposing factors.  On the one hand padding is expected to provide protection and in particular, provides shock absorption attenuating the impact of foot-strike.  On the other hand, there is the allure of the idealistic notion of barefoot running – based at least partially on the rational argument that if our distant ancestors survived by persistence hunting, the human frame must be well adapted to barefooted running.   These opposing influences have led to fluctuating enthusiasm for fashions ranging from barefoot (or minimalist shoes such as the Vibram Five Fingers) to the heavily padded Hoka one-one.

In addition to these two opposing influences there is the issue of the effects, either helpful or harmful, that shoes might have on the twisting movements that occur at the joints of foot and leg. Most notable are the compound motion of pronation occurring at the forefoot and ankle that allows the foot to roll inwards transferring weight onto the medial longitudinal arch as the leg is loaded during stance; and the inwards bend of the leg below the knee (varus deformation) that places pressure on the vulnerable medial aspect of the loaded knee joint while also dragging the ilio-tibial band towards the lateral femoral condyle.  Although pronation is a natural movement, shoe companies have placed strong emphasis on the potential dangers of over-pronation.  To prevent this, they have marketed motion control shoes with a medial post, a structure embedded in the medial side of the shoe that arrests the inward roll.  This affects not only the impact absorbing capacity of the foot, but also modifying the varus torque acting at the knee.


The question of high technology shoe design also brings with it the issue of the ethics of unfair technical enhancement of natural ability.   While this ethical issue can only be dismissed entirely by adopting barefoot running, it might be argued that in the modern man-made environment, denying at least a modest degree of protection would be unreasonable.  In principle there is a difference between basic protection and the overt assistance provided by embedded springs such as in the Spira.  However, once any layer of fabric is interposed between foot and ground, there is a continuum of assistance provided depending on the elastic properties of the material.   Nonetheless, most runners accept that the assistance provided by the bulk properties of a compressible material primarily designed for protection against either shock or penetrating injury is reasonable.

Cadence and foot-strike

The protective effect of shoes is clearly demonstrated by two automatic responses seen in most habitually shod runners when they change to barefoot running.  Self-selected cadence increases leading to decreased length of airborne time during each gait cycle, thereby decreasing the magnitude of the vertical force required to get airborne.   Furthermore, as discussed in my recent post on style, foot-strike tends to change.  The investigation led by Daniel Lieberman of Harvard University indicates that barefoot runners are more likely to adopt a mid-foot or forefoot strike rather than the rear foot strike typically seen in about 75% of shod runners.  This change in foot strike abolished the potentially harmful sharp rise in vertical ground reaction force that is generated by heel striking.  Nonetheless, it is noteworthy that avoidance of rear foot strike is not necessarily the case in habitual barefoot runners.  For example the study of north Kenyan habitual barefoot runners by Hatala found that 72% were heel strikers at their self-selected endurance pace, though the majority landed on mid or forefoot when sprinting, when vertical forces are greater.

As discussed in the post on style, there is little evidence that fore-foot strike is more metabolically efficient whereas several studies actually show rear-foot strike is most efficient at low speeds.  The situation with regard to risk of injury is mixed, with greater risk to knee with heel strike and greater risk to structures around the ankle with forefoot strike.  The balance of risks favours a mid-foot strike. Therefore, a shoe style that allows this is preferable.

Joint torques

Kerrigan and colleagues reported that the torques acting at ankle, knee and hip occurring in runners wearing Brooks Adrenaline shoes were increased in comparison with barefoot running.  The Adrenaline is described as a neutral shoe, meaning that it is not designed to strongly inhibit pronation, and has midsole thickness ranging for 24 mm at the heel to 12 mm at the front.  The increases in torque when shod were especially marked for knee varus (increased by 38%); knee flexion (36%) and hip internal rotation (54%) around mid-stance.   Only a minor portion of these increases in torque could be accounted for by the lower cadence of the shod runners.

Knee varus places stress on medial aspect of knee joint at a site especially prone to osteo-arthritis.  It also drags the ilio-tibial band towards the lateral femoral condyle increasing the risk of iliotibal band syndrome.  Knee flexion torque flexion places stress on patella-femoral  joint and increases load on patella tendon and quads.  It should be noted that tension in the patella tendon at mid-stance  is not necessarily bad, as it would be expected to increase the eccentric loading of the quads and facilitate the upward drive of the body that occurs after mid-stance.   Similarly a moderate degree of internal rotation of the hip is required as the pelvis rotates around the hip joint during stance, so a torque promoting internal rotation torque is not necessarily bad, though it is noteworthy that some runners do develop osteo arthritis of the hip.

The heel-toe drop

Elevation of the heel relative to the toe is the most likely explanation for the additional knee flexion torque revealed in Kerrigan’s study of joint torque at the joints. Furthermore despite providing padding, the presence of a bulky heel makes it difficult to avoid localized impact at the heel, and thereby make a substantial contribution in the rapidly rising spike of vertical ground reaction force observed in heel strikers.  As shown by Zadpoor, a rapid of rise of vertical force increases risk of injures such as tibial stress fracture.  Thus, it would appear that shoes with minimal or no drop from heel to toe that allows initial contact further forwards might  be safer, and will tend to be lighter.

An interesting alternative is the Healus, a shoe without a heel.  A slanted sole ensures that the runner avoids heel contact but instead makes contact via a well-padded mid foot.   Force plate data demonstrates that it abolishes the initial spike in vertical ground reaction force.  The padding under midfoot provides maximum protection when vertical ground reaction force is at its peak.  However, despite an endorsement by former European 5,000m record holder Dave Moorcroft , it does not appear to have achieved much popularity, possibly because it is produced by a small company.

Ankle and forefoot motion control

The inward rolling of the foot that occurs with excessive pronation has several potentially adverse consequences.  The ankle tends to be displaced towards the midline thereby increasing varus deformation at the knee, enhancing risk of iliotibial band syndrome and perhaps also osteo arthritis of the medial aspect of the knee joint.   The medial longitudinal arch of the foot is flattened increasing tension in the plantar fascia increasing risk of plantar fasciitis.  Thus, in runners with excessive pronation, a shoe with a medial post that limits pronation might be beneficial.  However it should be noted that Kerrigan observed increased knee varus torque in shod runners relative to barefoot.  The Brooks Adrenaline is a neutral shoe but nonetheless has a modest medial post and hence it might appear surprising that there was increased knee varus torque.  However the shoe had not been matched to the specific needs of individuals.  It is plausible that the one consequence of being shod was that individuals lacked the sensation and freedom of movement within the shoe required to produce optimal adjustment of the motion at the ankle according to their individual needs.   It might be argued that at least for non-injured runners, that light weight shoes or bare feet providing the freedom to adapt the ankle and foot motion according to individual needs and changing surface conditions, would be preferable.

Nonetheless, there is evidence that customised orthotics designed specifically to control ankle motion for each individual can reduce pain in runners with an established problem.  For example, Maclean and colleagues studied the effects of 6 weeks use of customised orthotics in a group of female recreational runners (15 to 40 km per week) who had a history of overuse running knee injury in the 6 months leading up to the study. The intervention decreased pain significantly and led to significant decreases in maxima for ankle inversion moment and angular impulse during the loading phase, impact peak, and vertical loading rate, though the effects at the knee were complex


Because the shoe is at the far end of the swinging leg, its mass makes a relatively large contribution to energy cost of repositioning the leg during the swing phase.   However, there is growing evidence that at least a small amount of padding brings a benefit that compensates for the additional weight.  Franz and colleagues from Roger Kram’s lab in Colorado compared oxygen consumption during running barefoot with that when wearing lightweight cushioned shoes (approximately 150 gm per shoe) in 12 runners with substantial barefoot  experience, running  with midfoot strike on a treadmill.  In additional trials to determine the effect added weight, they attached small lead strips to each foot/shoe (150, 300, and 450 g).   They found that in the absence of added weight there was no significant difference between shod and unshod running.  Adding weight led to an increased metabolic cost of 1% for each 100 gm of added weight.  When adjusting to equalise mass in shod and unshod condition, shod running had ∼3%-4% lower metabolic cost.

In a further experiment for the same lab, Tung and colleagues measured the metabolic costs of  barefoot running  on an unpadded treadmill and after adding strips of padding of either 10mm or 20 mm thickness to the surface of the treadmill.  They also measured the costs of running shod in lightweight shoes on the unpadded treadmill.  They found that when running barefoot, 10 mm of foam cushioning (approximately the thickness of the forefoot shoe midsole) afforded a benefit of 1.91%.   There was no significant difference between metabolic costs of shod and unshod running on the unpadded treadmill, indicating that the positive effect of shoe cushioning counteracts the negative effects of added mass.

Thus, running barefoot offers no metabolic advantage over running in lightweight, cushioned shoes. The explanation for this remains speculative.  One possible explanation is that when running barefoot, a runner maintains a lesser degree of stiffness in the legs, resulting in less efficient capture of impact energy as elastic energy, in the same manner as a floppy spring produced a less efficient recoil that a stiff spring.

While light weight shoes might offer adequate protection in short and medium distance events, it is necessary to consider the possibility that in a marathon or ultra-marathon, the cumulative damage from repeated eccentric contraction will result in a significant loss of power.  A little more padding might protect against this loss of power.    Similar issues apply during periods of high volume training.  Last summer, while training for a half-marathon, I built up my total training load to a substantially higher volume than during any recent year and found that I suffered a gradual accumulation of aches in my legs.  Hence, at least for an elderly person, light weight shoes should be employed sparingly, but nonetheless, frequently enough to produce the adaptive changes required if they are to be used for racing.


My overall conclusion is that for racing distance up to half a marathon, light weight shoes with near zero drop from heel to toe are preferable, as these give the optimum combination of efficiency and protection.  Unless the legs have been very well conditioned to the rigours of long races, for the marathon and ultra-marathons it might be preferable to use a little more padding.  Similarly, during periods of very high volume training, a modest amount of additional padding might provide helpful additional protection.  Motion control is only sensible if there is a clear need

11 Responses to “The big debates of the past decade: 2) shoe design”

  1. Robert Osfield Says:

    Hi Canute,

    Are you aware of how the elastic efficiency of mid-sole of modern shoes compares the efficiency of elastic recoil of runner?

    My expectation is that is should be relatively straight forward to produce a mid-sole that is more efficient at returning energy than human muscles/tendons. The study with the barefoot runners on padded treadmill suggest that was achieved, but only for a modest amount thickness.

    If man made mid-sole is better elastically than human tissue then with a simplistic view would suggest that more cushioning the better, but the padded treadmill study suggested the gain in efficiency rapidly drops off. I wonder if too much cushioning effects running mechanics enough that one would need time to readjust to get the best out of it. In terms of time, I’d expect this adaptation period to be weeks, rather than minutes as were likely the case in the studies.

    From my own preferences I find around 5 to 10mm works well for me, less and I have to adjust my gait to a more barefoot form and slow myself down. Last month I bought a pair of Nike Kiger which have easily twice as much cushioning as the F-lite’s I wear for most of my training. I bought them to allow me to rotate between shoes of quite different characteristics so that I can avoid over stressing any one part of my foot as up the mileage.

    Initially my runs felt horrible in the Kiger’s, as it was so squishy and unstable, but as the weeks have past I have tuned into them and am now far less bothered by the amount I sink into the shoe on each step. I suspect my efficiency has also recovered from the initial shock too.

    The weight of the Kiger is pretty similar to the F-Lite so it’s differences in efficiency will be down to effect of mid-sole efficiency and the effect it has on my gait. I still feel the F-Lite’s have the edge on efficiency, but it’s pretty close now.

    • canute1 Says:


      Thanks for your comment. I do not know of direct comparison of the elastic properties of human tissue compared with the foam used in running shoes. However, it is generally accepted that the elastic recoil of the steel springs in Spira shoes offers an unfair advantage, and hence they were banned in the US.

      However, it is possible that the way in which the runner adjusts tension in the muscles that maintain stiffness at the knee and ankle in response to the compliance of the surface will have a greater effect on the amount of energy recovered via elastic recoil and hence a greater effect on efficiency than the elastic energy stored in the shoe itself. The runner is likely to increase stiffness at the joints on a softer surface. This will improve elastic recoil provided the surface is not too soft. However when the surface is very soft, the capture of elastic energy will be less efficient despite the increased stiffness of the joints.

      As an illustration of the way in which joint stiffness is adjusted automatically, Alison Gruber and colleagues have shown that stiffness of the knee is less during heel striking compared with forefoot striking whereas the stiffness of the ankle is greater for heel striking. Furthermore, when choice is not constrained by the investigator, the choice of heel strike v forefoot strike is influenced by the compliance of the surface. Heel strike is favoured on softer surfaces. (This might be relevant to the differences between the findings of Lieberman and Hatala that I have mentioned in several posts)

      • Robert Osfield Says:

        There are curious differences between shod and barefoot runners with mid-foot/forefoot strike.

        Barefoot runners look to have a relatively high ground contact time vs time in air, while shod mid-foot/forefoot runners have relatively low ground contact time vs time in air. Heel strikers also have higher ground contact times vs time in air compared to shod mid-foot/forefoot runners. In this respect Barefoot runners may actual have greater similarities to heel strikers w.r.t ground contact time/time in air.

        As for heel striking barefoot runners, I suspect the Hatala study over emphasises this due to how the study was carried out. The participants had to land on a pressure map which was quite short relatively to the stride length so the participants are likely to have adjusted their stride to hit the pad, this means in some cases they will be over-striding and heel striking. The Figure 1 they include with the paper clearly shows a case of heel strike where the lower leg is abnormally reaching out :

        This abnormal stride is very likely to be down to the participant trying to land on the mat. Without seeing video of all the tests there is no way to know how many test runs were affected in this way, but I suspect it’s a significant number.

    • canute1 Says:

      Thanks for your comments. .
      It is true that barefoot runners and shod heel strikers both tend to spend longer on stance relative to airborne time, compared with shod mid-foot/forefoot strikers. I think the reasons are similar, but differ in detail. In both instances there is likely to a bias against incurring high vGRF. All runners face a trade-off between the cost of getting airborne (which entails high vertical ground reaction force) and braking. In the case of barefoot runners, there is likely to be a bias towards accepting a higher braking cost in order to avoid a high vGRF acting through unprotected feet. In the case of shod heel strikers, the rate of rise of vGRF is greater than for shod forefoot strikers. However accepting a greater braking penalty and shorter airborne time results in a less vertical motion and less impact at foot-strike, thereby at least partially compensating for the rapid rise in vGRF. These automatic adjustments are probably achieved by adjustment of tension in the muscles that regulate stiffness of knee and ankle joints.

      I agree that the short mat used in the Hatala study might have contributed to the observed tendency towards heel striking, but perhaps this would be a more serious problem at high speed, where stride length is longer, rather than at low speed. The highest proportion of heel striking was observed at low speeds, suggesting that speed does influence foot strike for barefoot runners, just as it does for shod runners. I am inclined to think the softness of the surface is likely to have played a greater role in the observed high proportion of heel strikers.

  2. Ewen Says:

    Thanks Canute. In your conclusion you mention ‘near zero drop’ shoes as being ideal in terms of efficiency and protection. Interestingly when running in the zero drop Saucony Virratas I always tend to pull up sore the following day. Also, for a supposedly well cushioned shoe, I find they produce a hard landing. On the other hand, the ‘small drop’ (4mm) Saucony Kinvara don’t leave me feeling sore and I don’t feel like I’m landing hard on the ground with each stride. Perhaps just an individual preference of a natural heel striker?

    • canute1 Says:


      Thanks for your comment. While I think that a large heel to toe drop contributes to excessive flexion torque at the knee as illustrated by Kerrigan’s studies, a complete absence of drop is likely to place greater strain on the Achilles. Furthermore since it is generally best if at least some of the load is borne on the heel at slow and medium paces, a modest amount of padding at the heel is probably helpful. Those are the reasons why I used the phrase ‘ near zero drop’. I suspect individual differences in strength and elasticity of various muscles and tendons determine just how much drop suits each individual.

    • Robert Osfield Says:

      Hi Ewen,

      It’ll take time to adapt between different shoes with different levels of cushioning and heel drop. The adaptations required will likely be combination of physical changes in the muscle/tendon lengths and the neuromuscular side. From my experience to fully adapt it takes many months.

      Over the last three years I have migrated from traditional built up shoes like the Nike Pegasus to minimal cushioned and zero drop shoes. I for sure wouldn’t go back, but can still enjoy a run in a shoe with a modest heel drop – 3mm drop feels fine so have a couple of shoes that are 0 drop and couple of 3mm drop shoes in my shoe rotation.

      However, 6mm drop now feels bit awkward, while tradition 12mm drop feels horrible and clunky, so I’ve either chucked these shoes out or relegated them to the back of the cupboard.

      As for needing a but of cushioning under your heel for slow paces like Canute suggest, I don’t personally find this. At slow paces the forces are pretty gentle and don’t tax my feet and can go zero drop zero cushioning without issue. Running faster, especially when charging downhill I really find that forces are significant enough that I prefer to have some cushioning, if only to ensure the forces are better distributed across my foot. I find that on descents I have to hold my pace back if I’m wearing really minimal shoes.

      My advice would be to be patient, keep your zero drop runs slow and short initially, and use a shoe rotation that loads your body in slightly different ways so you don’t over stress on part. I would also recommend running with relaxed lower legs, let the ankle and calf relax as you recover the leg – relaxing the wrist and forearm is a good cue for helping you relax the lower leg.

      Best of luck.

    • canute1 Says:

      The story about decreased life expectancy with a high protein diet in middle-age (especially if from animal sources) based on the human data was on the front pages of the newspapers in the UK today as well. However, for older individuals, protein (at least from plant sources) appeared to increase life expectancy. That is plausible, but I will look into the data a bit more closely before jumping to any conclusions.

    • canute1 Says:

      The reason why padding beneath the heel for a heel striker might be helpful even at slow and moderate paces, remains speculative. There are three issues to consider. First, if the padding has elastic properties, it will capture energy at foot strike and might return a useful proportion of this energy later in the stance phase. Secondly, if padding reduces jarring, this might reduce transmission of a sudden damaging impulse though the musculoskeletal system thereby reducing risk of microscopic damage to muscles in the short term and repetitive strain injury in the long term. Studies such as that by Del Coso, which I discussed in my post on 5th Jan, reveal appreciable muscle damage in a substantial proportion of individuals, even at the medium paces typical of non-elite marathon runners. Thirdly, a modest degree of cushioning might allow viscoelastic tendons and muscles to store and release elastic energy on the optimum time scale. If a viscoelastic structure is loaded very rapidly, it recoils very rapidly. However, the optimum leg stiffness is difficult to estimate, so I would not wish to argue too strongly for this third mechanism.

      There is abundant anecdotal evidence that individuals can adapt to running with less padding, but very little evidence that minimal padding improves efficiency, race performance or safety. I think that the studies by Franz and colleagues from Roger Kram’s lab indicate that the weight penalty of a few extra mm of padding at the heel is likely to be small compared with penalty from microscopic muscle damage suffered by almost half the runners studied by Del Coso. If the few extra mm of padding help reduce this damage, it worthwhile. However I accept that this is speculation.

  3. eth us Says:

    eth us

    The big debates of the past decade: 2) shoe design | Canute’s Efficient Running Site

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: