Editor’s Note: In July 2025, TrainingPeaks officially launched Fueling Insights, a feature that allows coaches to offer more personalized, effective fueling plans based on world-tour proven metabolic research from Dr. Iñigo San Millán. In this article, Dr. San Millán addresses the methodology behind Fueling Insights and explores the limitations of applying a fixed Gross Efficiency model for accurately reflecting the biological complexity of energy expenditure during exercise.
The calculation of calorie expenditure during exercise is a complex area of exercise physiology.
While methods such as Gross Efficiency (GE) have historically been used, recent advancements in our understanding of metabolism highlight the need for more nuanced approaches.
Gross Efficiency (GE) Is Not a Universal Constant
Gross Efficiency — calculated as Work Rate in Watts divided by Energy Expenditure in J/s multiplied by 100 — appears straightforward. However, this formula overlooks the intricate biological processes of metabolism.
Human exercise performance converts chemical energy into mechanical energy, with watts being the result, not the source. Assuming all metabolic pathways are uniform across individuals, as the GE formula implicitly does, is inaccurate.
Energy expenditure is highly dependent on substrate utilization. For instance, a glycolytic athlete (with relatively high carbohydrate oxidation) expends more kilocalories than an aerobically trained one to generate the same wattage. GE, being “blind to metabolism,” cannot account for these differences.
A fixed GE of 21% is a research average and not a universal constant. GE can vary between 17% and 25%, influenced by:
- Substrate use (fat vs. carbohydrate)
- Training status
- Exercise intensity
- Muscle fiber recruitment
- Mitochondrial health and efficiency
Applying a universal 21% GE is an oversimplification, akin to claiming all cars achieve the same fuel efficiency regardless of their characteristics or operating conditions.
Energy Cost: Derived from Substrate Use, Not Solely Mechanical Work
The cost of ATP production varies significantly depending on the fuel source: approximately 9.75 kcal/L O₂ for fat and 5.05 kcal/L O₂ for carbohydrates. This means identical wattage can result in vastly different kilocalorie expenditures based on an individual’s metabolic state. GE provides little insight into substrate use and therefore cannot inform carbohydrate needs, fueling strategies, or recovery protocols with a high degree of accuracy.
Fueling Insights models these dynamics by considering:
- Carbohydrate and fat oxidation (g/min)
- RER shifts based on W/kg
- Athlete category (training status)
This model is grounded in decades of indirect calorimetry data.
Illustrative Case Study: Energy Expenditure at Varying Body Weights for a Fixed Power Output
Consider a 317-watt, 60-minute cycling session across two different body weights:
- Case A: 119 kg (2.66 W/kg)
- Carbohydrate oxidation: 2.34 g/min → 140 g/hr → 561 kcal
- Fat oxidation: 0.42 g/min → 25 g/hr → 227 kcal
- Total: 788 kcal/hour
- Case B: 80 kg (3.96 W/kg)
- Carbohydrate oxidation: 4.95 g/min → 297 g/hr → 1188 kcal
- Fat oxidation: 0.04 g/min → 2.4 g/hr → 21.6 kcal
- Total: 1209 kcal/hour
Note that Fueling Insights predicts the shift in substrate use due to higher relative intensity. GE would produce the same kilocalorie output at both weights, failing to account for this physiological reality.
The Physiological Impact of Body Mass Changes
Changes in body mass are not metabolically neutral when considering fixed power output. For example, a significant loss of body fat would result in an increase in relative intensity, higher CHO oxidation, and higher kCal demand.
But if the loss includes muscle, power output may decline, mitochondrial density may drop, and overall energy efficiency can decrease. Body mass changes inherently impact the metabolic engine.
All this to say: two cyclists maintaining the same wattage may be generating mechanical work from vastly different fuel sources. One might be highly glycolytic and the other more aerobic (greater fat reliance).
Fueling Insights Adds Nuance
Fueling Insights addresses this physiological heterogeneity, offering customization based on real-world data from hundreds of athletes across various categories (Tour de France, Competitive, Masters, and Recreational).
In contrast, GE assumes a fixed energy source, disregarding the influence of factors such as W/kg, VO₂, RER, or metabolic status, and considers substrate choice irrelevant. These factors are critical when considering accurate fueling strategies. Fueling Insights is a physiological model, whereas GE is a mechanical approximation.
Final Conclusion and Summary
Fueling Insights captures the inverted U-curve of fat oxidation, the exponential rise in glycolysis, the shift in substrate use with relative intensity, and athlete-specific metabolic efficiency. Unlike GE, which assumes a “one size fits all” approach, Fueling Insights reflects the complexities of real athletic physiology.
While Gross Efficiency was a valuable tool in early exercise physiology, it is no longer sufficient for a comprehensive understanding of energy expenditure. It overlooks substrate selection, the metabolic component of exercise, and fails to capture relative effort, thus limiting its ability to inform fueling strategies.
Fueling Insights offers a more accurate and biologically relevant model, as demonstrated by the case study above. Even with a fixed power output, substrate utilization shifts with relative intensity — a better reflection of how human physiology operates.
