REF-719 · Hoogkamer, Kipp, Frank, Farina, Luo & Kram, 2018 · Sports Medicine, 2018 Apr;48 · Paper · PubMed
The founding "4%" study: with shoe mass matched, prototype plated-foam shoes lowered the energetic cost of running in all 18 athletes tested
The effect is the midsole and plate, not the weight
Protocol
n=18 high-caliber athletes ran six 5-min trials (three shoes × two replicates) at 14, 16 and 18 km/h across three sessions.
Shoes: a Nike prototype with a highly compliant, resilient midsole plus an embedded stiff plate (NP), versus two established marathon racing shoes, NS and AB (identified in the full text as Nike Zoom Streak 6 and adidas adizero Adios Boost 2; full text, checked 2026-09-27).
Submaximal oxygen uptake and CO₂ production were measured during minutes 3-5 and energetic cost (W/kg) averaged over the two trials per shoe.
Shoe mass was matched across conditions — the decisive design choice, since footwear mass alone changes energetic cost.
Key Findings
Energetic cost in the prototype was 4.16% and 4.01% lower than in NS and AB respectively, with mass matched (16.45 ± 0.89 vs 17.16 ± 0.92 and 17.14 ± 0.97 W/kg; both p < 0.001).
The reduction appeared in all 18 subjects tested. No non-responders in this sample.
The percentage change was independent of running velocity across 14-18 km/h.
The authors' conclusion is a prediction, not a measurement: with these shoes top athletes "could run substantially faster" and achieve the first sub-2-hour marathon.
Practical Application
In 18 high-caliber runners, mass-matched prototype plated-foam shoes lowered the energetic cost of running by about 4% (4.16% and 4.01% vs two established racing shoes) in every one of the 18, independent of speed across 14-18 km/h. This is where "4%" comes from: a laboratory oxygen-cost number, not a race-time promise.Not from this paper: in real elite marathon racing the same technology was associated with 1.0% faster race speed (REF-721) — a different outcome, so not a strict like-for-like comparison.
The mass-matched design is why this study matters. Footwear mass alone changes energetic cost; this shows the gain survives when weight is held constant, so the mechanism is the midsole-and-plate system rather than "less shoe".
Treat a personal economy gain as plausible but unmeasured for any individual athlete — pooled evidence puts the typical effect lower (REF-720, plated vs non-plated shoes).
Limitations
n=18, single laboratory, high-caliber runners (all male per the full text, checked 2026-09-27). No recreational runners in this study (REF-723 covers slower runners).
Treadmill oxygen cost at fixed speeds, not racing. Energetic cost is an input to performance, not performance.
Inference, not tested: the "established" comparison shoes have since been superseded, so the margin over a modern conventional racing shoe is not necessarily this number.
The sub-2-hour claim is explicitly an extrapolation by the authors from laboratory data.
Industry-relevant prototype tested against competitor products; funding and equipment provenance deserve the reader's attention when weighing the size of the effect.
How this entry was checked and changed
Content correction 2026-09-27
Lead rewritten as the study's finding (4.16%/4.01%, all 18 runners, 14-18 km/h); "roughly a quarter of it in racing" replaced by the REF-721 figure with a note that it is a different outcome.
Shoe model names and "male only" labelled as full text; minimalist-ideology point and the superseded-baseline point labelled as inference.
Lines labelled "full text; not checked against the abstract" were checked against the open-access full text (PMC5856879): every number in them appears there; relabelled "full text, checked 2026-09-27".