Carbon plate shoes work. Just not by 4% for most runners who buy them. The 4% headline came from a study of 18 elite men averaging 64 kg, running at sub-4:20 per km. If you weigh 85+ kg and race at 5:30–6:30 per km, the honest number is closer to 1–2%, and at the slowest end, the benefit isn’t statistically significant at all (Joubert 2023).
That’s the short answer. Read on to understand why, and what it means for your next race.
What the 4% Study Actually Tested
The landmark paper was published by Hoogkamer and colleagues in Sports Medicine in 2018 (Hoogkamer 2018). It showed that Nike’s ZoomX Vaporfly reduced the running economy, how much oxygen your body uses to hold a given pace (lower is better), of 18 competitive male runners by an average of 4%.
Here’s what most people skip over when citing that number.
The 18 runners had a mean body mass of 64.3 ± 4.7 kg, roughly 142 lb. The heaviest runner in the study was likely around 75 kg (165 lb). The three test speeds were 14, 16, and 18 km/h. That’s 4:17 to 3:20 per km. Sub-6:52 per mile. These are the paces you’d expect near the front of a major marathon, not the middle of the pack.
The 4% study tested runners who weigh less than most recreational athletes and run faster than most recreational athletes. Both at the same time.
Individual responses in the same study ranged from –1.6% to –6.3%. So even within that elite, lightweight, fast group, some runners got almost nothing.
The Pace Problem: What Happens Below 12 km/h
Joubert and colleagues (2023) asked a simpler question: what happens to the benefit at slower speeds?
They tested the Nike ZoomX Vaporfly Next% 2 at 10 km/h and 12 km/h, paces closer to how most recreational runners actually train and race.
At 12 km/h (5:00/km, roughly a 3:30 marathon pace), the improvement was 1.4%. That’s real and statistically significant.
At 10 km/h (6:00/km, roughly a 4:13 marathon pace), the improvement was 0.9%. That result was not statistically significant (p = 0.065). Meaning: at that pace, we can’t confidently say the shoe did anything at all beyond random variation.
The shoe-speed interaction was clear: p = 0.021 for the interaction effect. The shoe works. It just works a lot less at slower speeds.
At 6:00 per kilometer, the pace many heavier runners actually race, the benefit drops to under 1% and that result isn’t even statistically significant.
Carbon Plate Shoes for Heavy Runners: The Body Weight Gap
The 2025 systematic review by Bolliger and colleagues analyzed 14 studies covering 271 runners. Their pooled estimate: 2.75% improvement across all subjects (Bolliger 2025). Not 4%.
They also flagged something important. The authors couldn’t stratify results by body mass because individual participant data wasn’t available across studies. Their recommendation: future research must compare three groups: under 60 kg, 60–75 kg, and over 75 kg. We simply don’t know what the benefit looks like at higher body weights.
The heaviest published sample in a well-designed shoe study is the Van Hooren 2025 work: mean body mass 77.3 ± 8.8 kg. That’s heavier than most prior studies, but still well short of the 85–100 kg range where plenty of recreational marathoners live.
No published randomized trial has tested runners above 85 kg as the primary sample. That’s a real gap.
Why Mass Matters: The Foam Physics
The performance benefit of carbon plate shoes comes mainly from the foam, not the carbon plate. Beck and colleagues (2020) showed that the plate alone accounts for only 0.3 ± 2.2% of the benefit (Beck 2020). The soleus muscle was completely unaffected by plate stiffness (p = 0.538).
Think of PEBA foam like a high-quality spring. It compresses under load and snaps back, returning about 87% of the energy you put in. Compare that to standard EVA foam, which returns around 60–65%. That difference is where the benefit lives.
Here’s where body mass enters the picture. Running generates a ground reaction force, the push the ground sends back up through your foot with each step, of roughly 2.2–3.0 times your body weight.
| Body Mass | Est. Peak Force (2.5× BW) | Est. Peak Force (3.0× BW) | Context |
|---|---|---|---|
| 63 kg (139 lb) | 1,575 N | 1,890 N | Mean of Hoogkamer 2018 subjects |
| 75 kg (165 lb) | 1,838 N | 2,205 N | Upper edge of most research samples |
| 85 kg (187 lb) | 2,081 N | 2,498 N | Common recreational runner; near standard lab test ceiling |
| 95 kg (209 lb) | 2,325 N | 2,790 N | Force exceeds standard lab test load (~2,000 N) |
The standard mechanical test force for elite racing shoes is around 2,000 N. A 95 kg runner can generate 2,300–2,800 N per step. Whether that over-compresses the foam or simply operates it in a different part of its performance curve, nobody has published the answer yet (Hoogkamer 2018, mechanical data).
A 90 kg runner also has different leg spring stiffness than a 63 kg runner. Stiffness scales with body mass to the two-thirds power, giving a 90 kg runner about 27% more absolute stiffness. The shoe’s carbon plate has fixed bending stiffness designed around elite runners (McLeod 2022). It may be slightly under-loaded for a heavier runner, though the foam does most of the work anyway.
The honest answer is: the physics suggests heavier runners might see less benefit, but we don’t have the data to confirm it.
Individual Variability: The Bigger Problem
Even within the narrow slice of runners who have been studied, the response to carbon plate shoes varies wildly.
Knopp and colleagues (2023) compared Kenyan world-class runners (mean 59.9 kg) and European amateur runners (mean 72.1 kg). In the Kenyan group, individual responses ranged from –11.3% to +11.4%. That’s right: some elite runners ran worse in the shoes. European amateurs ranged from –1.1% to +9.7% (Knopp 2023).
In the Van Hooren 2025 study of heavier recreational runners (mean 77.3 kg), the individual shoe ranking agreement across participants was just 0.16 on a 0–1 scale. Near zero agreement. The shoe that ranked first for one runner ranked last for another.
Your response isn’t the average. No one knows your response until you measure it.
Marcus: A Real-World Estimate for a 92 kg Marathoner
Marcus is 41, 92 kg (203 lb), targeting a 4:15 marathon finish. He’s been running for three years, logs about 45 miles per week, and his current training shoes are a standard EVA trainer.
He buys the Nike ZoomX Vaporfly Next% 3. The marketing math looks tempting: 4:15 × 0.96 = 4:04:48. Ten minutes off. Easy.
The problem: Marcus’s target marathon pace is 6:03 per km (9:43 per mile). That’s 9.9 km/h. Right at the speed where Joubert 2023 found 0.9% improvement with a p-value of 0.065, below the threshold for statistical confidence.
A more realistic estimate applies 0.9–1.5% improvement: 4:15 × 0.985 = roughly 4:11 to 4:12. Two to four minutes saved, not ten.
Marcus still might benefit. The Vaporfly’s cushioning could reduce muscle fatigue in miles 18–26 even if the metabolic gain is small. But his pacing plan should account for a 2–3 minute shoe benefit, not 10 minutes. Building his race prediction around an inflated number is a recipe for going out too fast and dying in the last six miles.
That’s the risk of applying group averages to individual race plans.
Individual Variability Chart
Notice how the pessimistic line at 10 km/h crosses zero. That represents the real-world possibility: some runners at slow paces see no benefit at all from carbon plate shoes.
What the Aerobic Base Has to Do With It
Here’s a connection most shoe reviews miss. Runners with strong aerobic decoupling, who can hold pace without their heart rate drifting, tend to have more consistent ground contact mechanics late in a race. A runner whose form deteriorates in miles 18–26 engages the foam differently than in mile 3.
This matters because the shoe studies all measure running economy in controlled, fresh conditions. Your ability to stay on top of the shoe’s mechanics late in a race depends on your aerobic base. A deep Zone 2 base built over months affects whether you can actually harvest the shoe’s benefit when it counts most.
Shoe performance isn’t just about the shoe.
How AthleteOS Models Shoe Benefit Without the Hype
AthleteOS factors body mass and target pace into running-economy modeling when generating marathon finish time predictions. A 92 kg runner targeting 6:00/km doesn’t receive the same shoe multiplier as a 63 kg runner targeting 4:20/km. The model applies a pace-adjusted estimate based on the published data, closer to 1% at slow recreational paces rather than 4%.
This matters because a 6-minute overpromise on finish time leads to a real pacing error on race day.
You can also track your training load and fitness score in AthleteOS across the months leading into a goal marathon. The fitness score (CTL) combined with drift ratio data tells you whether your base is ready to extract the shoe’s benefit at all.
Get a personalized marathon prediction that accounts for your body mass and shoe class at AthleteOS.
Should You Still Buy Carbon Plate Shoes?
Probably yes, with realistic expectations.
Even a 1–2% improvement translates to 2–5 minutes in a 4-hour marathon. That’s real. The cushioning benefits late in a race may add value beyond what the metabolic studies capture.
What you shouldn’t do: assume the 4% number applies to you, build your race pacing plan around it, and blow up in the second half because you went out 8 seconds per km too fast.
Measure your own response. Run two 10-mile efforts at target marathon pace. Use your trainer for one and the racing shoe for the other. Then compare average heart rate at identical paces. That’s your personal data point. It’s worth more than any group average.
Sources
-
Hoogkamer W, Kipp S, Frank JH, et al. A Comparison of the Energetic Cost of Running in Marathon Racing Shoes. Sports Medicine. 2018;48(4):1009–1019. PMC
-
Bolliger A, Spengler CM, Beltrami FG. Metabolic effects of carbon-plated running shoes: a systematic review and meta-analysis. Frontiers in Sports and Active Living. 2025. PMC
-
Joubert DP, Dominy TA, Burns GT. Effects of Highly Cushioned and Resilient Racing Shoes on Running Economy at Slower Running Speeds. Int J Sports Physiol Perform. 2023;18(2):164–170. PubMed
-
Knopp M, Muñiz-Pardos B, Wackerhage H, et al. Variability in Running Economy of Kenyan World-Class and European Amateur Male Runners with Advanced Footwear Running Technology. Sports Medicine. 2023;53(3):605–616. PubMed
-
Van Hooren B, Copier R, Pedersen S, Balamouti Z, Meijer K. The Mediating Effect of Running Biomechanics, Anthropometrics, Muscle Architecture, and Comfort on Running Economy Across Different Shoes. J Sports Sciences. 2025. PMC
-
Beck ON, Golyski PR, Sawicki GS. Adding carbon fiber to shoe soles may not improve running economy: a muscle-level explanation. Scientific Reports. 2020. PMC
-
McLeod AR et al. Leg stiffness during running in highly cushioned shoes with a carbon-fiber plate and traditional shoes. Gait & Posture. 2022. ScienceDirect
-
Bolliger A, Spengler CM, Beltrami FG. Impact of Advanced Footwear Technology on Running Economy at Slower Running Speeds. Sports Medicine Open. 2026. PMC
-
Seglina I, Torniainen K, Kvarforth L, et al. Sex and Isolated Anthropometric Measures Do Not Explain Individual Differences in Responsiveness to Advanced Footwear Technology. 2026. PMC