Shoe Bending Stiffness Increased Estimated Achilles Load in 31 Runners

Shoe bending stiffness increased both peak Achilles tendon force and total tendon impulse in a modelling study of 31 male rearfoot strikers. Runner experience changed the shape of the loading curve, but it did not interact with bending stiffness.

Bending stiffness is the property that makes a carbon-plated racing shoe resist being flexed at the forefoot, and it has become one of the most-marketed and least-explained numbers in running footwear. A study published on November 7, 2026 in Scientific Reports put three bending stiffness levels on the same 31 runners and estimated what each one did to the Achilles tendon.

The work comes from a group spanning three Chinese institutions: Shixin Lin, Jiaying Shao and Ye Ma of the Research Academy of Grand Health in the Faculty of Sports Sciences at Ningbo University; Hao Yuan and Mianfang Ruan of the College of Physical Education and Health at Wenzhou University; and Dongwei Liu of the School of Information Technology and Artificial Intelligence at Zhejiang University of Finance and Economics.

Thirty-one male rearfoot strikers took part, split into 18 experienced and 13 novice runners. All of them ran at 4.17 metres per second, roughly a 4-minute-per-kilometre pace, in carbon-fibre shoes built to three bending stiffness levels: 10.9, 17.5 and 32.9 newton metres per radian, labelled T1, T2 and T3.

Achilles tendon force was not measured directly. It was estimated in a subject-specific musculoskeletal model using static optimization. Discrete outcomes were compared with mixed two-way analysis of variance and a Holm correction, and the full loading waveforms were analysed with one-dimensional statistical parametric mapping.

How the Bending Stiffness Conditions Were Built

The design choice that gives this paper its value is simple: the same runners ran in all three shoes. That makes each person their own comparison and removes the problem that plagues footwear research, where differences between people swamp differences between shoes.

The three bending stiffness levels were not arbitrary. T1 at 10.9 newton metres per radian sits near a conventional trainer. T3 at 32.9 is about three times stiffer, in the territory of aggressive racing plates. T2 at 17.5 fills the middle. Spanning that range matters, because a plate too stiff for a given runner is a real possibility rather than a theoretical one.

Every runner was a rearfoot striker, and all ran at one controlled speed. Both decisions tighten the comparison and narrow who the results describe, a trade-off that runs through most biomechanics work of this kind.

Stiffer Plates Raised Both Peak Force and Total Impulse

Two separate things can happen to a tendon when footwear changes. The single hardest moment of loading can get harder, or the total amount of loading accumulated across the step can grow. The study found both, and found the second more convincingly than the first.

Peak Achilles tendon force rose with bending stiffness, with the stiffest condition exceeding the least stiff one at a Holm-corrected p of 0.013 and a partial eta squared of 0.159. Tendon impulse, which integrates force over time, showed a stronger effect at p = 0.003 and a partial eta squared of 0.264, with both T2 and T3 above T1.

That impulse result is arguably the more interesting of the two. It suggests the middle bending stiffness was already enough to increase cumulative tendon loading, even where it did not clearly raise the single peak. For a tissue whose injuries are usually described as accumulated rather than acute, cumulative load is the more relevant currency.

Why the Ankle Explains Part of It

The mechanism the authors point to for the bending stiffness effect is at the ankle. Peak ankle dorsiflexion angle changed with bending stiffness, and the effect there was the largest in the paper, with a partial eta squared of 0.472 and a p value below 0.001. A stiffer forefoot resists bending, the joint travels differently, and the tendon crossing it sees a different demand.

There is a caveat the authors flag themselves: the dorsiflexion effect appeared in the discrete peak values without corresponding significant effects in the waveform analysis. The two methods disagreeing is a signal to treat the mechanism as suggested rather than demonstrated.

Those percentages of stance where differences did appear are narrow windows. Loading differences showed up across roughly 12.7 to 18.6 percent, 23.3 to 63.4 percent and 79.7 to 88.9 percent of the stance phase, with a brief reversal early on at 3.6 to 5.7 percent.

Experience Changed the Curve, Not the Peak

The comparison between experienced and novice runners produced a subtler picture than the shoe comparison. Group effects on peak Achilles tendon force and on tendon impulse were not significant. Experience did not simply mean more load.

What did differ was how quickly load arrived. Achilles tendon loading rate was higher in the experienced runners, at a Holm-corrected p of 0.030 and a partial eta squared of 0.208, and the waveform analysis showed higher estimated force across several stretches of stance with a short period of lower force in early contact.

Reading that as better or worse would be premature. A faster loading rate can reflect a more efficient, springier stride, or it can reflect a tendon absorbing demand more abruptly. A model of one running speed cannot distinguish an adaptation from a risk.

The Missing Interaction Is the Useful Result

No runner level by bending stiffness interaction was detected for the tendon loading outcomes. Statistically that is a null finding. Practically it is the most portable thing in the paper.

An interaction would have meant that bending stiffness affected experienced and novice runners differently, and that shoe advice would have to be split by ability. Its absence suggests the two influences operated through what the authors describe as distinct and independent pathways: bending stiffness did its work, experience did its own, and neither amplified the other.

A null result in 31 people is weak evidence, though. Interactions need larger samples to detect than main effects do, so the honest reading is that this study did not find one, not that none exists.

What Bending Stiffness Findings Mean for Choosing Shoes

Nothing here says carbon plates are harmful, and nothing here says they make anyone faster. This study did not measure performance, injuries or running economy. It estimated tendon load in a laboratory, over single steps, in one group of men.

The bending stiffness result does add weight to a caution that has been circulating among coaches for several years. If a stiffer plate increases cumulative Achilles loading, then a sudden move into very stiff racing shoes is a change in training stress, not merely a change in equipment. Treating it as the former is the sensible default.

Some general, non-individualized points follow reasonably from that:

• Introduce a much stiffer shoe gradually rather than debuting it on a long or hard session.
• Rotating footwear means the tendon is not loaded the same way every run, which is a reasonable hedge given the uncertainty.
• The middle bending stiffness condition already raised tendon impulse, so stiffer is not automatically a bigger dose of the same thing.
• Persistent Achilles pain is a reason to see a clinician, not a reason to change shoes and hope.

Footwear research has arrived at a similar place to gait research. Work on retraining runners away from a rearfoot strike found that changing one part of the chain shifts load somewhere else rather than removing it, and reviews of proximal exercise for flexible flatfoot reach the same conclusion from the opposite direction.

That is the frame this paper fits into. Load moves; it rarely disappears. Related laboratory work on foot function screening and running performance and on lung function in marathon runners shows how narrow each individual measurement in this field tends to be, and how much the useful picture depends on stacking several of them together.