From Carbon to Track

ParaAthleteTrackRunning
Written By Cat Hoh

Blade design and running performance. Why the evidence is thinner than it looks.

How stiff should a running blade be? It is the first question any sprinter, coach or prosthetist asks, and thirty years of laboratory work has not produced an answer anyone can lean on. Everyone assumes they know. A blade is a spring, it hands energy back, and a stiffer one hands back more. The first two are roughly right. The third has been assumed since the nineties, tested a dozen times over, and never actually established. Which is an odd thing to discover about a piece of equipment that costs thousands of dollars and gets chosen on exactly that basis.

It's a spring, and that's the whole trick

Start with what a blade does, because it is less than most people imagine. It stores energy when the ground reaction force bends it during stance, and gives that energy back as it unloads. It generates nothing. Every joule returning to the athlete went in off their own leg, minus whatever the carbon ate on the way through.

Czerniecki measured this back when the old prosthetic feet and the new ones still competed for the same customers. The numbers are the clearest thing in this entire literature.

Figure 01

Energy return efficiency: how much of what goes in comes back

31%

SACH foot

Solid ankle, cushion heel. The old standard.

52%

Seattle Foot

Mid-1980s. The first move toward dynamic response.

84%

Flex Foot

Grandfather of every blade on a start line today.

Values from Czerniecki et al., as reported in Rahnama, Soulis & Geil (2024). The leap that created modern para-sprinting was a leap in waste reduction. Nothing was ever added.

Because it's a spring, the blade's properties feed into the stiffness of the whole leg, and leg stiffness governs a great deal of what happens in running. That's the chain the whole field is built on. Material determines structure, structure determines leg stiffness, leg stiffness determines performance. It's a perfectly sensible chain and it starts coming apart at about the second link.

Four dials that turn out to be one

Blades are made of carbon fibre, almost universally, and a fair slice of the engineering literature exists to ask whether they ought to be. Custom-built comparisons put glass fibre and carbon fibre a factor of two apart in stiffness, somewhere around 28 N/mm against 57. Hybrid lay-ups get investigated for the reason hybrids always get investigated in composites: glass fails gently and carbon fails all at once.

But none of this settles much, because the material is only one of the inputs and probably not the dominant one. A blade is a long cantilever beam, and its geometry does at least as much work as its chemistry. The commercial models split into C-shapes and J-shapes. Shape governs how the thing mounts, which governs its height, which feeds back into its stiffness. It very likely sets the natural frequency of the structure too, though almost nobody has looked. And a shorter beam may simply need to be stiffer to survive, because the stress has to concentrate somewhere.

So you can't move one of these without moving the others. Geometry, height, stiffness and material aren't four dials on a console. They're one structural problem wearing four different names, and the evidence for that is more awkward than any theory. When Beck's group put fifty-five commercial blades through a testing machine, they found stiffness ran 10 to 39 per cent lower at the mounting angles typical of actual running (10 to 25 degrees) than at neutral. The number on the box is measured in a posture no athlete has ever occupied. They also found that stiffness fell as height rose in J-shaped blades, and didn't in C-shaped ones. Same variable, different behaviour, depending on the geometry it happens to live inside.

Then you attach a person

And the engineering account starts losing arguments.

Groothuis and Houdijk changed blade alignment angle and watched stiffness fall away in a clean straight line, exactly as a mechanical engineer would predict and hope: 17 to 19 per cent softer across fifteen degrees. Their runners were less cooperative. They compensated, stiffening the residual leg by landing with a straighter knee, and held total leg stiffness very nearly constant across the whole range.

Figure 02 / Interactive

Change the blade. Watch the runner undo it.

Alignment angle / tap to change

0° 5° 10° 15°
0° 5° 10° 15° STIFFNESS
Blade stiffness Total leg stiffness

The engineer changed the spring. The runner changed themselves. The output barely moved. Schematic. Shows the direction and shape of the relationship reported by Groothuis & Houdijk (2019) at the four alignment angles they tested, not their measured values. Blade stiffness fell 17% (imposed step frequency) to 19% (free) across the range; total leg stiffness showed no significant change in the free condition.

One caveat that matters more than it sounds. Groothuis and Houdijk did not test amputee athletes. They tested ten able-bodied runners wearing prosthetic simulators, and said plainly that whether the same compensations are available to amputee athletes needs further investigation. An earlier study they cite found that transtibial athletes did not appear able to regulate residual leg stiffness the way these able-bodied runners did, possibly limited by residual muscle strength or socket fit. So read the figure as a demonstration of the mechanism, not a measurement of your athlete.

Barnett's group found the mirror image in actual prosthesis users. Alter blade stiffness acutely and unilateral users reorganise their spring-mass behaviour on the spot to accommodate it.

There's a hard practical lesson buried in that, and it's the one coaches most often miss. If you judge a blade off a single session you are measuring adaptation rather than equipment. The athlete's residual limb, and for a unilateral athlete their intact limb as well, are active participants in setting leg stiffness. They are not mounting brackets.

And this is very likely what explains the field's strangest split. In athletes with bilateral transtibial amputations, softer blades lowered the metabolic cost of running, while the same research group found no significant effect in unilateral athletes. Read it the obvious way. A unilateral athlete has a whole biological leg available to paper over whatever the blade gets wrong. A bilateral athlete doesn't. The equipment matters more when there's less human left to compensate for it.

Same blade, two answers

Beck's group tried to reproduce a stiffness figure that Brüggemann had published for a particular blade, matching the applied force, the foot orientation, the loading mode, the loading rate. Then they tested the same foot again, with a higher applied force, force-controlled loading, and a plantarflexed orientation.

Figure 03 / Interactive

One piece of carbon. Tap a test protocol.

A. Replicating Brüggemann B. Higher force, plantarflexed
34.2kN/m

Matched conditions. Same applied force, foot orientation, displacement-controlled loading and loading rate as the original published test.

29.2kN/m

Changed conditions. Higher applied force, force-controlled loading mode, plantarflexed orientation angle. Same blade, 15% softer on paper.

Beck, Taboga & Grabowski (2016), via Rahnama et al. (2024). A 15% swing, from the same blade, on the same day.

Their reading of this reframes everything that comes before it. These tests measure structural properties and not material properties, and the distinction isn't pedantry. A genuine material property holds steady across loading magnitudes within the elastic region. A blade is several materials arranged into a shape, and what you measure depends on how hard you push, how fast, and from what angle. Every stiffness figure in this literature is a fact about a blade in a particular rig on a particular afternoon. None of them is a fact about carbon fibre.

And the rigs are all over the place. Loading rates in the published work run from 50 to 1,000 mm/min. Peak loads from 1,500 N to 3,500. Some studies report the load-deflection relationship as linear and others as polynomial, and there is currently no way to tell whether that reflects genuinely different feet or just different laboratories. Mounting technique matters. Friction matters. Ambient temperature and humidity matter. There is no standard, which means there is no meaningful comparison, which means the accumulated numerical record of this field is closer to an anthology than a dataset.

Hysteresis, the energy the blade eats instead of returning, has barely been studied at all. Where it's reported it sits low, around 4.3% for J-shaped blades and 5.1% for C-shaped in one dataset, 3.7 to 8.1 across a larger sample. Exactly one study has connected it to running speed. Natural frequency, viscoelastic behaviour, vertical displacement: essentially unexamined.

The claim that fell over

Which brings this back to where it started.

In 2020 a well-known study concluded that prosthetic shape, but not stiffness or height, determined maximum sprint speed in athletes with bilateral transtibial amputations. PLOS ONE retracted it in December 2022, and the notice repays reading. A statistical reviewer found the conclusions unsupported and the statistics misread. Specifically, that non-significance in a small sample is not evidence of no effect. The paper's own appendix tables, the reviewer pointed out, contained evidence that stiffness and height both substantially affected running speed, which is the precise opposite of what the title said.

A change in running speed of 0.3% would give a top runner at least one extra medal in every ten races. PLOS ONE statistical reviewer, retraction notice, December 2022

In fairness to the authors, they disputed this throughout and all three disagree with the retraction. The sample was five athletes, which the academic editor noted is a normal constraint in prosthetics research, since very few people meet the criteria. That is the real bind. The population is tiny, so the studies are small, so the statistics can't carry the weight the conclusions put on them.

The companion paper on unilateral athletes was never retracted and still stands in the literature. It also shares the design, the team, and the shape of the claim. Three of the studies concluding that stiffness isn't the main driver of sprint speed came out of the same research group working from largely the same dataset, while at least one independent study found the reverse: that stiffer blades, in combination with better shape, produced higher sprint velocity.

So the defensible position isn't that stiffness doesn't matter. It's that stiffness hasn't been shown to matter, by a small and heavily interconnected body of work that may never have had the statistical power to detect effects big enough to decide a race. Those two statements sound similar and they are not remotely the same, and only one of them has any business influencing what goes on an athlete's leg.

So what do you actually do

Two systematic reviews and a literature review, coming at this from different angles, land in roughly the same place. The relationship between blade properties and running performance can't currently be pinned down to any single conclusive finding. Manufacturers don't publish standardised structural specifications. A clinician who wants an objective comparison has to go to a literature that disagrees with itself. And so, as the review evidence states without much enthusiasm, blade prescription mostly comes down to practitioner experience and athlete comfort.

That reads as an indictment of the practitioners, but it shouldn't. Absent a defensible general rule, comfort and patient systematic trialling aren't a poor substitute for evidence. They're the best method anyone currently has, and case-by-case optimisation is what the evidence genuinely supports rather than a fallback from it.

Which frees you up to do a few sensible things. Read manufacturer stiffness categories as rough labels rather than measurements, remembering they're generated at angles nobody runs in. Give any change enough weeks to let adaptation settle before you decide it's working. And treat alignment as a performance variable rather than a fitting detail, because it demonstrably shifts blade stiffness even when the runner's own compensation hides the effect from you.

One last thing, which nobody in this literature seems especially comfortable about. Every study reviewed here examined adult prostheses. Blades are manufactured for children. On junior athletes, the entire evidence base is silent.

References

Cat Hoh
Next

Mental Adjustment After In-Race Crisis Events