Why Limit Sole Thickness in Running Shoes?
A scientist who argued for stack height limitations explains his reasoning and current thoughts one year after the World Athletics ruling.
Heading out the door? Read this article on the new Outside+ app available now on iOS devices for members! Download the app.
One year after World Athletics delivered their ruling on fairness in shoes, the debate continues whether it was justified or appropriate. When thinking about regulating shoes, there are two important, valid questions that often get packaged together in a single debate: 1) Should we regulate footwear? and 2) If so, how?
Criteria for Regulation
Setting aside the first question for now, a good answer to the second question — how to regulate shoes — must meet several important requirements.
- First, we need an operational definition of a “fair” shoe, as the old regulations had no clear definition of what constituted an “unfair assistance or advantage.”
- Second, this definition obviously needs to effectively control what you’re seeking to control. In this case, it’s the extent to which a shoe can augment an individual’s performance.
- Third, this definition needs to be logistically feasible to enforce. A law’s ultimate utility (from a Hobbesian perspective) is conditional on its level of enforceability. The more complex the definition of a fair shoe becomes, the more contentious each criteria will become, and correspondingly, the more burdensome the policing of that definition would become for World Athletics and race organizers.
- Fourth, we would ideally have a regulation that is somewhat “future proof” or forward looking, rather than something reactive — e.g., simply banning new things that come out. This is important for the appeal and generalizability of the sport. It’s also critical for the companies, as reactive regulations open the door for corruption and bias within WA and stifles their ability or desire to develop new products, always having the fear that WA might react against their new product if it’s “too good”.
Midsole Thickness as a Simple Solution
Limiting the stack height — as I and Nicolas Tam suggested in the fall of 2019 in a British Journal of Sports Medicine editorial — seemed like an elegant, straightforward solution that ticked those boxes and also afforded a compromise to allow for innovation. It’s essentially an engineering design constraint — a geometric black box that defines what a shoe is on an athlete (literally, the physical space that a shoe can occupy on an athlete), and encourages optimization within that space. If you think of the shoe as a nebulous addition to bottom of a person’s foot, it simply regulates the size or extent of that addition.
To choose a handful of mechanical parameters on which to regulate (e.g., resiliency — commonly called “energy return,” stiffness, material composition, etc.) would be an enormous burden for World Athletics to enforce, and it would also make the sport somewhat intractable, as it would lose the universal appeal. This is the route swimming took with tech suits in 2009. They set limits on the buoyancy, permeability, and composition of the suits as well as a defined, allowed space for coverage on the athlete’s body. New suits now have to be approved by a FINA lab.
My thought was that if World Athletics were to adopt a similar strategy with an elaborate, technical definition of an allowed shoe, it would create far more problems than it would solve. Choosing which mechanical characteristics to regulate, what the actual limits or allowed design space for each are, then writing the test methods to explicitly police them would be contentious at every step. It would likely be effective at limiting the effect of the shoes (the second criteria above), but I thought it would be a convoluted band-aid.
This has already played out with World Athletics’ recent regulations and the current crop of “approved” super shoes. Rule 5.13.1, which added to the height limitation, stipulates that shoes “…must not contain more than one rigid plate or blade made from carbon fibre or another material with similar properties or producing similar effects” and Note (ii) adds: “The one rigid plate or blade referred to in Rule 5.13.1 may be in more than one part but those parts must be located sequentially, in one plane, not in parallel (i.e., not stacked above each other), and must not overlap.” That is all ridiculous.
In the Nike AlphaFly, are the more rigid materials above and below the air bags not sufficiently stiff to be counted against the curved plate in the shoe? How “rigid” would they have to be to “produce similar effects?” What about the energy rods in the Adidas Adios Pro? They’re not in the same plane as the plate in the heel. The plates in most of those shoes aren’t even in one “plane” anyways, as a plane is, by definition, “a flat two-dimensional surface.” Wouldn’t every super shoe fail this criteria?
I’m not calling for those shoes to be banned, but simply want to highlight the convoluted contradictions that are inevitably going to consistently pop up the more you try and craft explicit definitions of architectural features. I am advocating that the rules be simplified (and made more robust), as every unnecessary complexity added to the ecosystem of running is a step further in abstracting athletic performances from the physicality of the athletes.
The stack height limit alone was operationally appealing, as the eyeball test grossly keeps it in check, and you essentially just need a pair of calipers or something similar to measure or enforce it. A good test method or device is still lacking from the WA rules to make that exact measurement explicit and make the rule length-dependent — what “marginal increase in sole thickness” is allowed for larger sizes of the shoes beyond an EU 42? — but hopefully that’s being worked on.
As a geometric constraint, limiting stack height serves several regulatory functions. It limits the complexity of the shoe architecture and the extent to which that shoe can afford a performance advantage. With less “space” — literally and figuratively — within which to work, the extent to which materials and architectures can be arranged to create something substantially beneficial is limited. For example, with the plates, a more and more constrained midsole would limit the variation and shape of those plates. At some point, if it was thin enough, it may not even be beneficial anymore to have a plate there — particularly as we learn more about the importance of the plate’s curve and its specific location within the three-dimensional midsole.
On the flip side, the thicker and thicker the shoe becomes, the more complex the architecture within it would likely become to stabilize it and realize the advantage. The AlphaFly gave us a taste of this — what would that shoe, or it’s next version, look like without a 40mm limit? Moreover, it effectively limits the amount of mechanical energy that can be stored and released for given stiffnesses and compliances, i.e., a 40 mm spring stores twice as much strain energy as a 20mm spring.
However, importantly, this limitation does allow for optimization within that space. Personally, I’m not at all opposed to shoe enhancements — EVA foam is quite mediocre. If we could have foams that approached 100% energy return (i.e. perfect springs), that would be awesome in my book. There may be a lot to be explored about what potential health benefits may be afforded by better foams. I would simply advocate for drawing a line so that we can eventually (and preferably soon) reach an asymptote in the effect of the shoes on performance. Optimize within our definition of a shoe, then get back to enjoying the comparative performances of athletes (ourselves included).
One critical point to make clear when thinking about stack height and midsole thickness in relation to performance is this: it is NOT a sufficient condition for an advantageous shoe, but it likely is a necessary condition. That is, a thick shoe is not necessarily going to be beneficial — you’re probably not going to be winning any World Marathon Majors in the Hoka Bondi. But, an advantageous shoe is likely going to be necessarily thick — you’re probably not going to get 4+% economy improvements with a 25mm racing flat.
How Thick is Too Thick?
As to what that stack height limit should be, that’s an open question whose answer will inevitably be arbitrary. However, that’s true of most rules in sport — they’re arbitrary definitions that provide a framework for us to understand the competition and performances. Why are soccer pitches 90–120 meters long? Why not 125 meters? Those were the rules that FIFA adopted in 1904 because it best suited what they understood as a reasonable pitch that suited the game in the form they wanted to propagate. Why is the pitching mound 10 inches high in the major leagues? Because that is the height MLB set it in 1969 — lowering it from 15 inches — to reduce the pitcher’s advantage over the batter and increase run scoring.
In suggesting a height restriction, I personally felt less strongly about what exactly the limit should be, but rather that it simply be defined so we can start working and optimizing within it. However, for the reasons described above, I would have leaned towards a more restrictive limit, like 30 mm, as it would have brought down the proportion of that performance that is afforded by the equipment.
Another important consideration with respect to “how much” shoe is allowed is this: the more and more space that a shoe can be, the more and more intellectual property becomes an issue. So, the greater the space in which to develop patentable architecture, the more and more brands have control over their effects. We see this with Nike’s patent on the shape of the plate in the VaporFly. As that space grows, so too do the possibilities to control advantageous designs within it, further taking the performances out of the hands (rather, feet) of the athletes.
The more and more complex the shoes become (afforded by a greater and greater design space), the more and more they also become fragile to the context in which they’re used. Things like road surface, turn radius and frequency, weather conditions, and even the speed at which you’re running will start to have distinctly different optimal configurations that would afford different equipment substantially distinct advantages. This exists grossly now (road flats aren’t going to be great on a cross-country course), but you’re not going to wear different flats because of certain types of pavement or the tightness of turns.
In cycling, World Tour riders usually have at least 3 different bikes for a grand tour — a normal bike, at least one TT bike, and a climbing bike. And that’s with tight restrictions on bike geometry. Do we want that level of equipment selection to determine outcomes or to be part of our race-day decision tree? Do we want race-day discussions to be centered on whether the runner choose the right shoes for the conditions, like wax is for Nordic skiing? Some may, but I’d prefer to limit that.
In the new shoes, I love running in the AlphaFly on a track. I hate running in it on roads. Anything that’s a bit uneven in the surface feels like it throws me off for a few strides, so I’m seemingly fighting the shoe, whereas on the track I feel like I slip into great, undisturbed rhythm. It’s fun when it works, but I resent that variability a bit (and would hate to assess a race on whether the surface is sufficiently smooth for it).
Moreover, at what point do these architectures become something that aren’t advantageous to begin with, but can be trained with to better harness their advantage? That’s interesting scientifically, but I’d really lean away from having that as an additional variable with which we’d have to tango in training. And that’s not even considering the more uncomfortable hypothetical with respect to trainability or selective optimization with more complex footwear architectures: grappling with the development options and trajectories for youth runners from both biomechanical and availability standpoints. Perhaps that’s another can of resilient, carbon-fiber-plated worms, but the point is: complexity in begs complexity out, and in a sport whose value proposition is brutal simplicity, how much in do we want to invite?
Finally, I also hoped that with a reasonable stack height/thickness limit, you could get rid of the rules around plates (limit the number/geometry of them) and the rules on prototypes. My personal feeling is that if we’re going to allow shoes to be 40mm, we might as well develop as best we can within it (and not worry about adding more inevitably contentious arbitrary determinations). Moreover, if an athlete wants custom shoes, they and their company should be allowed to design, adapt, and create something unique within that agreed-upon space. Again, just a single standard to freely optimize within.
At some point, we can always loosen the restriction in the future if there’s an appetite for it. If something is worth protecting, I’d say it’s more prudent to regulate first to protect it and relax accordingly in the future, rather than letting it play out unregulated and risk losing or permanently damaging it.
Why Regulate at All?
As for the first question of whether we should regulate footwear at all, that certainly is a valid debate. There is one camp that would say we should either be barefoot or have a free-for-all. I actually think from a strictly biomechanical perspective, this is theoretically sound (and intellectually interesting). To the former contention, we’re a sport — and a society — that has largely known footwear to be a component, so barefoot would be unreasonable for most. To the latter, stack height is certainly not something that adheres to the adage: “If some is good, more is better.” Running on stilts (even optimized carbon-fiber ones!) is likely not going to be the best way to cover variable ground on foot quickly (though it may be the case that that becomes one configuration that is optimal to one hyper-specific person or context). There is an optimality.
At some theoretical point, we might find the ideal height and architecture for each person and each context, and the whole thing would limit itself. However, as described above, the greater and greater the extent of that addition to the athlete becomes, the more and more that it becomes fragile to specific contexts. And how long until we even get there? Two years? Five years? Ten years? Never? That optimality may get rearranged every time we have an advancement in materials (i.e., foams with new properties). Without a framework or agreed-upon definition of a shoe, we would leave the question open forever, “Is that performance the athlete or the equipment?” That’s interesting and cool for some, but I’d suspect it would ultimately erode the appeal of running for most.
For the lion’s share of the modern history of our sport, when you crossed the finish line of a race or finished your favorite loop, there was a number on the clock. Some of that number, maybe it was minutes, maybe it was only seconds, are because of the shoes you were wearing — as opposed to being barefoot. So some fraction of that percentage of your performance was always due to engineering. However, with the VaporFly and the advent of super shoes, that percentage increased substantially. Now we have to face the question — how much are we okay with? What fraction of that pie that is our time on the clock do we want to give to the engineers and what fraction do we want to preserve for ourselves? Speaking as both an excited engineer and a bleeding romantic in the sport, my sense is to minimize it within reason. The more and more that we allow that percentage to grow, the more and more baggage we bring with that in terms of the competitive structure, our emotional relationship with the sport, and how it exists sociologically in terms of comparative performances.
Running the middle ground — limiting the extent of the shoes — will inevitably beg a definition that’s arbitrary, but that definition will provide a framework with which to bring our efforts against our opponents, the clock, and ourselves into relief.
The author would like to thank the research teams at the University of Colorado, Brigham Young University, and Grand Valley State University for their exceptional work helping us understand the physiology and mechanics of the new generation of racing shoes, as well as Dr. Rodger Kram, as his tweets never fail to stimulate ideas and keep the shoe conversation honest.