On foam and plates
Getting a surprise
I'd been geeking out over lightweight running shoes, finding and buying racing flats like "New Balance Hanzo s V2" (170 g) and "Saucony type a8" (167 g) on eBay, as they are no longer available from shops. These are both very minimal shoes, with firm cushioning, and c. 10 mm of foam and rubber between you and the ground, these feel incredibly light, and thus fast on your feet (though both are desperately narrow for my feet).
After doing few short runs in these, feeling i put in quite the effort, my partner bought a pair of Puma deviate nitro. That shoe has a nylon plate, and Puma's "super foam". I thought I'd borrow them for a run. They felt super weird, all squishy and I could feel the plate in the shoe, but at the end of the run I'd run really fast...
What to do? I set up a spreadsheet and used meters per heartbeat as a proxy (inspired by fellrnr) for running efficiency. Guess what, my partner's shoes where better than my racing flats - better than any of my shoes...!
Method
I decided on a local loop of 4,5 km with c. 30 m of elevation and ran that route almost religiously every second day until I had a minimum of 5 runs per shoe. The shoe choice was random but biased towards shoes with fewer runs, ensuring I got even number of runs in each shoe, but that the specific order shouldn't be a bias.
The intensity was high for all runs, near threshold, with an extra set of 4 runs at a set heart rate to at least have a single set at equal heart rate. I only used heart rate data from after the first 5 min, to only have steady state in the data.
I further adjusted for my fitness changing (improving) over time, by fitting a 16 day smoothing to the m/beat curve (16 days chosen as this minimised the residuals of the model).
The "score" of the shoe on a given run was the excess m/beat compared to the expected, thus, at least partially, correcting for changes in fitness over time.
As a side note I used my "heart rate reserve" (activity heart rate - resting heart rate) as the heart rate for the activity. This further served to compensate for gains in fitness as my resting heart rate dropped somewhat during the "experiment".
Results
The mean "fitness adjusted m/beat" for the shoes were:
Saucony type a8: 1.98
Scott speed rc: 1.98
Altra escalante 3: 2.01
Puma deviate nitro 2: 2.04
or in charts:
As relative efficiency:
And trying my hand at some simple stats in google sheets:
 | |
| Probability two shoes are equal in in efficiency, two-tailed test, unequal variance. Green indicates <5 % probability. |
Shoe Summry
Saucony type a8
Light classic racing flat with tight fit and firm foam (EVA type, possible with added bubbles). No stability features or rigid structures. The shoe is practically zero drop. This shoe was new during these runs. c. 167 g per shoe.
Scott speed rc
Scotts full supershoe, with high stack height, carbon plate and Scott's fanciest foam (olifin EVA mix) which is very firm. The shoe has slight drop but a very pronounced "rocker" that literally makes you fall/lean forward and want to speed up. This shoe was new. c. 240 g per shoe.
Altra escalante
My normal running shoe. Mid stack height, but zero drop, with roomy foot, that actually is shaped like a normal foot. The foam is TPU based, and is much springier than EVA foam, even after the >500 km I have put into them. c. 260 g per shoe.
Puma deviate Nitro 2
A speedy shoe for short to mid distance, with low drop. The midsole is Puma's "nitro" a TPE material which is very squishy. The shoe has a nylon plate. the shoe was new when I started this. c. 216 g per shoe.
What did i learn?
- Meters per heart beat is quite constant over a range of paces.
- The presence of a carbon/nylon plate makes no difference to my running efficiency.
- More and softer foam is better for running economy (i.e. not EVA).
- A small weight gain (e.g. from 170 g to 260 g) is not significant.
- One can measure small differences, here 3 %, with a simple setup.
A 3 % efficiency might mean:
4:30/km becomes 4:22/km
4:00/km becomes 3:53/km
3:45/km becomes 3:38/km
3:30/km becomes 3:24/km
It's just that racing flats are much more fun...!
