Journal of Science and Cycling

Wear analysis of bicycle chain joints

Martin Rank-Isepp

The rapid electrification of bicycles has fundamentally transformed drivetrain dynamics, leading to significantly higher loads on mechanical components such as chains. Unlike in traditional bicycles, where wear and efficiency losses primarily concerned competitive riders, the widespread adoption of e-bikes and other high-performance bicycles has made these issues critical for everyday cycling. Increased torque and sustained power assistance place unprecedented mechanical and tribological demands on drivetrains, making durability, friction reduction, and efficiency vital considerations for both manufacturers and users. As a result, research and testing methods that were once used mainly for industrial chain applications are increasingly relevant for the cycling industry. This study addresses these emerging challenges by adapting proven methodologies from mechanical and automotive engineering. Through controlled laboratory testing, bicycle chains and lubricants can be evaluated under standardized conditions that replicate real-world usage scenarios. Such testing does not only enable precise comparisons between different chain and lubricant systems but also provides insight into the fundamental wear mechanisms and efficiency losses occurring under varying loads, speeds, and environmental conditions. The approach supports the tailored design of drivetrain components for diverse cycling applications, including urban e-bikes, high-speed track bicycles, and rugged mountain bikes. By bridging disciplines, this work paves the way for innovations that enhance performance, reliability, and sustainability in modern cycling. Ultimately, it provides a framework for drivetrain optimization that meets the evolving demands of electrified and specialized bicycles, ensuring both increasing efficiency and resilience across applications.

BMX Cycling, Warm-up, Fatigue A self-reported analysis of BMX rider’s warm-up behaviours pre-race

Lee Paul Rylands

Coaches and athletes have long noted the importance of including a warm-up as part of their pre-competition routine. However, no research has looked at the pre-race warm-up routines in BMX cycling. Therefore, the aims of this current study were to gain an understanding of current BMX warm-up strategies. The study had two main themes: the first was to ascertain if riders performed a pre-race warm-up routine and if so, what those strategies entailed. The second was to gain an understanding of the respondents’ opinions of warm-up strategies pre-race. The study found that 56 % of riders did not perform one. Respondents were asked about their perception of the impact of a warm-up on performance and gave a range of responses from strongly agree to strongly disagree. 53 % of the respondents who did perform a warm-up noted their routine contained high intensity intervals. The results suggest riders are unsure if a warm-up will aid performance and this may be a rationale as to why they do not perform one.

The Protective Bicycle Handlebar

Boris Sobočan

The difference in the magnitude of forces in this direction during cycling and during dangerous impacts enables the design of a Protective bicycle handlebar which provides safe cycling and efficiently and effectively reduces the risk of dangerous injuries to cyclists caused by impacts with the bicycle handlebar.

The Protective Bicycle Handlebar

Boris Sobočan

While cycling, the forces on the handlebars in the direction of the longitudinal axe of the bicycle handlebar are substantially smaller than cyclist’s bodyweight.

Reliability of Cyclotronics Smart Trainer Power Output in Trained Cyclists

Vitor P Costa

The aim of this study was to assess the test–retest reliability of the Cyclotronics Smart Trainer during a graded exercise test (GXT) and a power profile test (PPT) in trained cyclists. For the GXT reliability analysis, 22 cyclists performed three GXTs, each separated by seven days. For the PPT reliability assessment, a smaller group of 10 trained participants (2 females, including 1 triathlete and 7 cyclists) completed one GXT followed by three PPTs, also separated by seven days. The first test in each protocol was used for familiarization and excluded from the reliability analysis. During the GXT, the coefficient of variation (CV) ranged from 5.5% at the lowest workload (100 W) to 1.6% at the highest workload (300 W), demonstrating reduced variability at higher intensities. A very low CV was also observed for peak power output (PPO), indicating high consistency in maximal performance measures. Regarding the PPT, acceptable test–retest reliability was observed, particularly for efforts lasting 120 seconds or more. As expected, shorter efforts showed slightly greater variability, which is common in high-intensity performance tests with a smaller sample size. In conclusion, the Cyclotronics Smart Trainer demonstrated reliable measurements during both GXT and PPT protocols. These findings support its use in repeated testing scenarios where measurement consistency is essential

BMX race–start dynamics: coupling on–track measurements and physics–based modeling for performance optimisation

christophe clanet

A fast and stable launch is decisive in Bicycle Moto-Cross (BMX) racing, yet the underlying interplay between rider power production, bicycle gearing and ramp geometry has received little attention in the engineering literature. Using high-frequency field measurements gathered on the Olympic track of SaintQuentin-en-Yvelines (France), we develop and validate a physics-based model that predicts the rider’s velocity profile over the start mound. The formulation combines (i) a Hill-type linear torque–cadence relationship fitted to instrumented-crank data, (ii) an energy balance that accounts for aerodynamic and rolling losses, and (iii) the measured track slope. Model predictions agree with experiments for three gear ratios. We then explore two practical optimisation criteria—minimum time to the first bump and maximum mean acceleration on the mound—to identify gearing strategies tailored to individual athletes. The framework illustrates how on-track sensing can be translated into actionable engineering guidelines for coaches and riders.

Braid Complexity and Peloton Phase Dynamics

Hugh Trenchard

A peloton is a group of cyclists that exhibits complex positional and energetic dynamics. This study introduces a method to distinguish between peloton phases by examining cyclists’ positional changes and crossings—referred to as braid complexity—combined with the peloton convergence ratio, a normalized energy coupling metric. Previous studies of peloton dynamics have quantified stretched and compact phases according to energetic thresholds, in addition to qualitative descriptions of other subphases. The method proposed enables higher fidelity distinctions between phases and subphases previously quantified or described qualitatively. This methodology may have application in other collective systems involving agents with identifiable energetic thresholds which continuously shift their relative positions within the system. Similarly, phase analysis may be practically applied in competitive pelotons to guide coaches and managers with greater precision to optimally position team members within the peloton.

titulo de prueba

Julio Alberto Ramos Martinez

To TT or not TT: Considerations for Children’s Cycling in Triathlon

Stuart Evans — 2023-07-02

Triathlon and Time Trial (TT) bicycles are engineered for performance by optimizing aerodynamics for maximum speed and race efficiency. As going fast on a bicycle essentially involves the cyclist overcoming air resistance, the objective is to devise ways to lower the influence of drag. For adults participating in triathlon, cycling at high speed in what is known as an aerodynamic position can be physically demanding where every athlete is constantly looking for the smallest improvements to enhance their performance. In road cycling there has been recent conjecture that training on TT bicycles is dangerous, with some professional cyclists calling for the use of TT bicycles in training to be banned due to the safety considerations. Despite this, a recent trend in junior (children) triathlon events has seen the use of aerodynamic bars and TT bicycles appear in age groups starting from ages 12 and upwards for both sexes. The desire for children to win and compete with the latest bicycle that provides a mechanical advantage has created debate with some jurisdictions banning the use of TT bicycles in junior triathlon events. Scientists involved in long-term athletic development (LTAD), biomechanics and exercise physiology that concern children have discussed that a balanced and long-term approach is needed when introducing children to triathlon and the use of TT bicycles. The purpose of this communication is to enable coaches to recognise the risks and limitations involved when working with junior triathletes and the considerations of using TT bicycles.

Tour de Physiology: The Exceptional Power Outputs and V̇O₂ of climbing in the Tour de France

Ole Kristian Berg

The Tour de France stands as perhaps the most demanding endurance competition in the world, requiring athletes to sustain exceptional physiological performance over three weeks of racing. Central to success in this event is a remarkably high maximal oxygen uptake (V̇O₂max). The 2024 Tour de France showcased unprecedented climbing performances, with new records set on iconic ascents. This brief report analyzes the physiological demands of these performances by estimating the V̇O₂ and power output required during three decisive climbs (Plateau de Beille, Isola 2000, Col de la Couillole) by the race winner, Tadej Pogačar. Using publicly available climb data, rider anthropometrics, and validated mechanical models of cycling power output, results indicate estimated mean power outputs of 442 ± 13 W and corresponding mean oxygen consumptions of 80 ± 2 mL·kg⁻¹·min⁻¹ sustained over ~38.5 min. Extrapolating from these efforts and known relationships between critical power and V̇O₂max suggests that Pogačar’s V̇O₂max during the race likely exceeded 90 mL·kg⁻¹·min⁻¹. These findings underscore the extraordinary aerobic capacity required to achieve record-breaking performances in Grand Tour cycling. They also highlight how ongoing improvements in training, equipment, and rider physiology continue to push the limits of human endurance performance to the enjoyment of the spectators.

Is "Big Gear" Effective for Road Cyclists? A Coach’s Standpoint to Guide Decision-Making.

Jose Luiz Dantas

Backgroung: "Big Gear" training, a common low-cadence strategy among road cyclists, involves pedalling at 40–60 rpm under high torque, typically at or just below Functional Threshold Power. Despite its widespread use and perceived similarity to resistance training, Big Gear training effectiveness remains debated. Purpose: This viewpoint critically examines the evidence surrounding Big Gear training role in enhancing muscular strength, endurance, and overall performance in road cyclists. Key Findings: Contrary to popular belief, Big Gear training does not meet the intensity threshold required to induce strength adaptations comparable to traditional resistance training. Evidence shows that Big Gear training produces lower neuromuscular strain and operates at <50% of Maximal Dynamic Force—well below the >60% threshold needed for strength gains. Furthermore, studies indicate that resistance training leads to greater improvements in maximal force production than Big Gear training. Similarly, Big Gear training shows no consistent advantage in promoting endurance adaptations or performance outcomes. Notably, studies that suggest benefits of low-cadence training often employ cadences and intensities higher than those typical of Big Gear training protocols. Some evidence even raises concerns about potential negative effects of Big Gear training on key performance determinants. Practical Application and Conclusions: The available evidence suggests that Big Gear training is either ineffective or, at best, questionable, being insufficient to trigger the positive adaptations associated with low-cadence training. It is recommended that coaches consider the efficacy of traditional Big Gear training in comparison to protocols that have been demonstrated to have more robust evidence-based outcomes.

Inter-rater variability in 2D kinematic cycling analysis using Kinovea®: a cross-sectional study with 53 bike fitting professionals in Brazil

Andrei Pereira Pernambuco — 2025-09-10

Cycling kinematic analysis plays a central role in the bike fitting process, directly influencing decisions related to performance, comfort, and injury prevention. Consistency of measurements across evaluators is therefore essential for ensuring reliable outcomes in both clinical and performance contexts. This cross-sectional inter-observer agreement study evaluated inter-examiner variability in two-dimensional (2D) kinematic measurements obtained with Kinovea® software. A sample of 53 professional bike fitters from different regions of Brazil analyzed the same 40-second video of a cyclist pedaling on a stationary mountain bike. Each participant independently selected frames and measured seven predefined joint and positional angles. Statistical analyses included descriptive measures, Shapiro–Wilk normality testing, bootstrap confidence intervals, one-sided chi-square variance tests with Holm corrections, bias and empirical limits of agreement, Brown–Forsythe tests of dispersion, and Fleiss’ κ for categorical KOPS classification. The results showed notable inter-examiner variability, particularly for knee extension (CV = 6.2%), trunk flexion (CV = 5.8%), and plantar flexion (CV = 4.8%), which exceeded predefined tolerance thresholds of 2–4% of the mean. By contrast, hip flexion, knee flexion, and armpit angle showed greater consistency. Subgroup analyses revealed no significant effect of professional experience or software used on measurement variability. These findings highlight that, even under identical testing conditions, methodological differences among raters can substantially influence kinematic measurements. The study underscores the need for standardized protocols and structured training in 2D motion analysis to improve reliability in bike fitting practice and ensure safer, more effective adjustments for cyclists.

The Influence of Crank Length on Metabolic Output and Lower Limb Muscle Activity in Well-Trained Cyclists

Hannah Payton

Cycle cranks are an industry standard length of between 165 mm and 175 mm. Crank lengths that are shorter than industry standard have been shown to have benefits in maximal cycling performance, however less is known about the effect on sub-maximal cycling. The aim of this study was to determine the effect of three different crank lengths (130 mm, 150 mm and 170 mm) on metabolic output and muscle activity during sub-maximal cycling in well-trained cyclists. Participants (n=9, 4 males, 5 females, aged 31.1 ± 9.7 years, 173.1 ± 11.4cm, 71.5 ± 12.9kg) performed a VO2 Max test on a cycle ergometer with 170 mm cranks. On a separate day, three 13-minute sub-maximal tests were then performed with each of the different crank lengths, with five minutes at an intensity of 80% of VO2 max power. Metabolic output was recorded as well as surface electromyographic data from nine muscles. Results were analysed using RM-ANOVA. There was no significant main effect of crank length on metabolic output (p=0.154). There was a significant main effect of crank length on intensity (p=0.028), but no significant interaction between crank length and intensity on metabolic output (p=0.493). There was no significant main effect (p≥0.144) or interaction effect (p≥0.107) of crank length on muscle activity for any of the nine muscles. There was a significant main effect of crank length on intensity for gluteus maximus (p=0.028), vastus medialis (p=0.022), rectus femoris (p=0.015), tibialis anterior (p=0.015), medial gastrocnemius (p=0.013) and soleus (p=0.009). It was concluded that shorter cranks are not detrimental to metabolic cost or muscle activity, yet could offer a number of other potential benefits, including reducing injury risk.

The impact of tyre width, pressure and surface condition on rolling resistance and vibration transmission of bicycle tyres

Jens Buder

Rolling resistance and vibration transmission are critical parameters influencing the mechanical efficiency and comfort in competitive cycling. Optimising a tyre setup requires the consideration various metrics such as rider-system mass, tyre width, tyre filling pressure, and surface conditions in parallel.
In this study, the combined effects of tyre width and inflation pressure on rolling resistance and frame-transmitted vibrations were evaluated under laboratory conditions. Five tyre widths of one tyre model were tested across four rider-system weights, two surface condition (tarmac and cobblestones), and multiple filling pressures. The experiments showed significant non-linear relationships between tyre width, filling pressure, and rolling resistance, with surface conditions contributing a secondary effect.
A Vibration analysis revealed that wider tyres substantially reduce surface-induced accelerations, whereas tyre pressure showed a minor but detectable influence. These findings support the development of rider-specific optimisation indices that balance rolling resistance and vibration sensitivity. Further research should integrate physiological responses to vibratory stimuli to refine setup recommendations for individual athletes.

The complete performance package – A case study

Loes Stessens

Improving cycling performance isn’t about just one thing—it’s a combination of factors, from biomechanics and aerodynamics to how the body responds to different environments and training loads. At the elite level, even the smallest changes can lead to those all-important marginal gains. This case study takes a deep dive into the full performance evaluation of a cyclist at a high-performance center equipped with everything from a bike fitting studio and climate chamber to a physiological lab and a wind tunnel designed specifically for cycling.

The process starts with a detailed bike fitting session using advanced tools like motion capture and pressure mapping. The goal here is to fine-tune the rider’s position for maximum efficiency, comfort, and power output, while also helping to prevent injury. Once the position is dialed in, the cyclist moves into the climate chamber for a series of tests under different temperature and humidity conditions. These sessions help us understand how the athlete’s body handles heat, loses fluids, and uses energy—key insights for racing in diverse environments.

From there, the rider undergoes a battery of physiological tests in the lab, including incremental efforts to exhaustion on a cycling ergometer. These tests are designed to measure things like VO2max, lactate threshold, heart rate variability, and muscle fatigue. By combining metabolic data, blood analysis, and real-time heart monitoring, we get a clear picture of the rider’s endurance capacity and physical limits.

Finally, it’s time to optimize aerodynamics in the wind tunnel. Using tools like computational fluid dynamics and force balance sensors, the cyclist’s position and equipment are adjusted to reduce drag as much as possible. The beauty of this approach is that each phase informs the next—every bit of data helps guide tweaks in posture, training plans, and race strategies.

By looking at performance through multiple lenses—mechanical, environmental, physiological, and aerodynamic—we can build a complete picture of what’s holding an athlete back and where gains can be made. It's a continuous cycle of testing, learning, and refining to unlock every bit of potential.

Effect of Heat on Psycho-Physiological Responses During a Self-Paced Cycling Time Trial

Thibaud PIRLOT

Introduction: As global temperatures rise and competitions increasingly take place in hot environments, understanding the impact of heat on performance is critical (Hunter et al., 2002). Thermal stress induces cardiovascular strain, alters brain activity, and modifies perceptual and motivational responses, affecting pacing and endurance (Périard et al., 2015). While studies have reported changes in heart rate (Girard & Racinais, 2014), muscle oxygenation, and cognitive function under heat (Nybo & Nielsen, 2001), few have focused on self-paced exercise. This study examines how heat affects performance, perceptual responses, motivation, and neural efficiency in trained cyclists during a simulated self-paced time trial. Methods: Twelve trained male elite cyclists were recruited for this study. Each participant performed two 26.2 km time trials on home trainer in random order: one in temperate and one in hot conditions. Core body temperature, skin temperature, core-to-skin temperature gradient, heart rate and respiratory frequency, power, cadence and electroencephalography activity were continuously monitored. Data were averaged over 30-s windows at each 20% segment of time trial completion for analysis. Perceptual variables were assessed at baseline and every 20% of time trial distance. These included: perceived exertion, pain, perceived pleasure, affective load, and motivational level. A significant condition effect was reported for the core to skin gradient temperature, with higher gradients in the TEMP condition. Results: A significant main effect of condition for power output was found, with higher values in TEMP than HOT. No significant difference was found between condition for RPE, pain but for pleasure and for affective load. No significant difference was found for heart rate, respiratory rate, α/β ratio on prefrontal, motor, and parietal cortex. Discussion: This study reveals that performance during a self-paced TT is negatively impacted by heat, primarily due to reduced core to skin temperature gradient and altered perceptual responses. Core temperature alone did not differentiate performance, suggesting that skin temperature and heat dissipation capacity (core-to-skin gradient) are more relevant. Despite reduced power output, cyclists maintained cognitive functions linked to effort management, likely due to anticipatory pacing strategies. Perceived pleasure, seem to be a key factor in self-regulated exercise, decreased in the heat. Neural efficiency remained stable, the self-regulated nature of the exercise likely enabled participants to limit cognitive strain by adjusting effort in real time. Conclusion: Exercising in hot conditions impairs cycling performance primarily through the decrease of core to skin temperature gradient and, reduced perceived pleasure. Despite this, RPE, pain, motivation, and neural efficiency remain intact, indicating that self-pacing allows athletes to regulate their effort to avoid a catastrophic failure.

Performance analysis in high-level virtual cycling: data from the 2023 Zwift® UCI Cycling E-Sports World Championships

Giorgio Varesco — 2025-09-09

This study aimed to characterize the performance at the 2023 UCI Cycling E-Sports World Championships in 18 women and 27 men competing in this three-race knockout series with a direct elimination finale. We evaluated their power profile, critical power (women: 249 ± 29 W; men: 362 ± 26 W) and W′ (women: 14.4 ± 1.5 kJ; men: 23.5 ± 4.2 kJ) over the 12 months preceding the competition and during the event. Energy depletion during races was analyzed using Bartram’s W′ balance models. Women and men advancing to the next race developed respectively ~15% and ~16% more power (W/Kg) compared to those eliminated in Race 1, for durations of ≤1 min (p < 0.001). In women, similar W′ depletion was observed in Race 1 and Race 2 (p = 0.35). In men, greater depletion was observed at the end of Race 1 compared to Race 2 (8 ± 8% difference, p = 0.007). The direct-elimination format of Race 3, involving repeated sprints, led to lower W′ depletion for both sexes. The results suggest that power developed for efforts up to 60 seconds and the ability to recover across races are pivotal for performance at the 2023 Cycling E-Sports World Championships.

Not Small Men: The lack, the need and the implementation of female input in sport engineering, with an emphasis on the cycling industry

Mary O Donnell

‘If Manny is for a male mannequin, then is there a Wo-Manny for females?’ A rather whimsical, yet, relatively insightful question that arose on completion and presentation of my Master’s thesis ‘Optimising Bicycle Fitting for Aerodynamic Performance’ . The question forms the basis of this paper, which highlights the absence of female physical and computational models within sports engineering with an emphasis on cycling performance. This paper discusses the implication of male assumption on female athletes and how inclusion can improve the future of research and development. The research was performed by surveying prominent sports engineering researchers, particularly those involved with creation of 3D models for CFD analysis, as well as physiologists and female athletes. Current literature and personal insight has been included to further highlight the research. A thematic analysis showed that there is a need to challenge, not just the cycling industry, but the sports engineering industry as a whole when it comes to the inclusion of female input and inclusion. This paper serves as a call to action for the cause.

The Effects of Different Indoor and Outdoor Cycling Environments on Lower Limb Muscle Activation

Chia-Hsiang CHEN

Abstract: This study investigates the effects of different indoor and outdoor cycling environments on lower limb muscle activation. Eighteen female participants performed both indoor and outdoor cycling while a wireless electromyography (EMG) system recorded muscle activation in the rectus femoris, biceps femoris, tibialis anterior, and gastrocnemius of the dominant leg. The outdoor route included downhill, flat, and uphill sections, while the indoor setup matched the outdoor gradients and resistance levels. A paired-sample t-test compared muscle activation between conditions. Results showed that rectus femoris and gastrocnemius activation was significantly higher during indoor cycling. These findings suggest that cycling environment influences muscle activation, particularly in the rectus femoris and gastrocnemius, which may have implications for training strategies and performance optimization.

Multidisciplinary Optimization of an Aerobar: Aerodynamics, Biomechanics, Mass and Structure

Sebastien Terol

The optimization of cycling position involves integrating two distinct analyses: individualized aerodynamic analysis and position-change power analysis. Aerodynamic analysis, conducted via Computational Fluid Dynamics (CFD) simulation or wind tunnel testing, provides the aerodynamic efficiency of each position. Power analysis quantifies the maximum power output of the cyclist for each position. By comparing these analyses, a balance between aerodynamic efficiency and power output can be achieved, estimating optimal positions for individual cyclists and given conditions. Indeed, the optimal position is specific to a particular cyclist, and depends as much on the race length and profile as on UCI or triathlon regulations. Based on the methodology of Fintelman et al., 2015, the torso angle as well as the position of the extenders allows us to precisely define a position on which to confront aerodynamic efficiency and power production. A digital twin of a professional athlete on his time trial bike was created using a 3D scan, with automatic generation of extensions based on forearm positions. This digital replica facilitates the automated design of extensions tailored to each position, enabling a comprehensive and customized approach to performance enhancement. The study aims to compare numerical study results with wind tunnel data to evaluate the algorithm's effectiveness. Subsequently, the automated aerobar design can be refined by optimizing its shape, structure, and mass, leading to a multidisciplinary optimization of the aerobars according to the cyclist's position. Solving this highly complex engineering problem would allow the fabrication of optimized aerobars tailored for a specific rider and race course.

Optimization of the Foot/Pedal system through custom-molded carbon shoes for the sprinters of the French national track cycling team.

AXEL COUVEINHES

This study examines the impact of custom-molded carbon (CMC) shoes on elite track cycling sprinters by comparing them to traditional carbon shoes (CT). Through mechanical, biomechanical, and qualitative assessments, we analyzed sprint performance, force production, and athlete perception. While CMC shoes improved ankle stability and knee alignment, they did not enhance sprint performance or force output. However, their biomechanical adaptations suggest potential benefits in injury prevention. Despite perceived advantages in rigidity and efficiency, further refinements are needed to improve comfort and overall usability without compromising performance.

The Effect of Rate of W′ Utilisation and Acute Fatigue on W′ Reconstitution

Alex Welburn

The critical power (CP) and W′ concept has become more integrated within applied cycling performance science. Both parameters can be assessed in the laboratory and field and have become a useful tool for coaches and athletes (Leo et al. 2022; Moritani et al. 1981; Poole et al. 1988; Spragg, Leo, and Swart 2022). It is possible to mathematically model the depletion and recovery of W′ during intermittent exercise (Skiba et al., 2012, Skiba and Clarke, 2021). Exercising at a work rate higher than CP causes a reduction in W' whereas when exercising at a work rate below CP there is an exponential reconstitution of W' (W′_REC) that is dependent on the recovery duration and a reconstitution time constant (τ_W'). This is contingent on the difference between recovery power output and CP (D_CP) (Skiba et al., 2012; Skiba et al., 2013; Skiba et al., 2014).

There are generalised equations available for W’_BAL models that allow prediction of W′_REC(Bartram et al. 2018, 2022; Pugh et al. 2022; Skiba et al. 2012). However, for optimal use individualisation of W’_BAL is advised (Welburn et al. 2023) as W′_REC has been shown to correlate with athlete specific performance parameters such as V̇O_2peak, CP and LT₁. (Bartram et al. 2018, 2022; Caen et al. 2021; Pugh et al. 2022; Welburn et al. 2023). Recent work suggests that W′_RECis influenced by work rate (Caen et al. 2019) and slows with repeated exercise (Chorley et al. 2019). However, both factors are not currently accounted for in the W’_BALmodel.

Durability (or fatigue resistance) is considered to be a key determinant of cycling performance (Muriel et al. 2022; Spragg et al. 2022). It appears that work rate has the greatest impact on durability (Leo et al. 2024; Spragg et al. 2024), therefore understanding the impact of work rate (i.e. the rate of W′ utilisation) and acute fatigue on W′_RECcould provide an important insight into the determinants of W’_BAL which would potentially allow improvements to the accuracy of the W′_BALmodel. This study has two aims: assess the effect of (i) rate of W′ utilisation, and (ii) acute fatigue on W′_RECduring intermittent exhaustive exercise.

Methods

Currently, 6 participants (4 males, 2 females; age: 21 [2] y, height; 1.78 [0.09] m, body mass; 67.8 [6.3] kg, V̇O_2max; 59.9 [7.6] mL·min^-1·kg^-1, maximal aerobic power (MAP); 361 [83] W, CP; 260 [130] W, W′ 22.45 [10.4] kJ, mean [SD]) have been recruited for this laboratory-based investigation. All tests were performed on an electronically braked cycle ergometer (Excalibur Sport, Lode). Participants initially attended the laboratory on six separate occasions for the determination of V̇O_2max, maximal aerobic power (MAP) and CP/W′. Participants subsequently performed two experimental trials, in a counter-balanced order. In both trials, participants completed intervals consisting of 40 sec on and 20 sec off. Trial one was performed at a W′ utilisation rate of 60 J×s^-1above CP [WR_60FRESH] and trial two was performed at a W′ utilisation rate of 120 J×s^-1above CP [WR_120FRESH]. The number of repetitions were calculated to deplete ~75% of their W′. The recovery work rate between the interval efforts was at 100 W below CP (i.e. D_CPof 100 W). After the final interval participants then completed at an opened ended time to exhaustion (TTE) test at the same work rate. After a recovery period of ~60 min (involving cycling at 50% of LT₁), participants repeated the same interval protocol at the same W′ utilisation rate [WR_60FATIGUED & WR_120FATIGUED]. Total work done above CP (W′_totalTTE) was calculated for each open ended TTE. The Wʹ_BALmodel (Pugh et al. 2022) was used to calculate the depletion and reconstruction of W′ during both trials. An individual τ_W’was calculated for each exhaustive exercise under the assumption that at the point of task failure at the end of the TTE represents a W′_BAL of 0 kJ. A two-way repeated measures ANOVA was conducted assess differences in W′_totalTTE with W′ utilisation rate (WR₆₀vs WR₁₂₀) and conditions (Fresh vs Fatigued) as the primary variables. Where significant main effects or an interaction was observed, Bonferroni-corrected post hoc t-tests were used to locate differences. Statistical significance was set at P ≤ 0.05. Data are presented as mean ± standard deviation [SD] unless otherwise stated.

Results

Table 1 shows W′_totalTTE during the final open-ended TTE interval for the intermittent exercise trials. There was a main effect for condition (P = 0.008), but no main effect for W’ utilisation rate (P = 0.397) or interaction (P = 0.084). W′_totalTTE was lower (P = 0.007) in WR_60FATIGUED compared to WR_60FRESH. There was no difference (P = 0.099) in W′_totalTTE between WR_120FRESH and WR_120FATIGUED.

Table 1: W′_totalTTE during the final open-ended TTE interval for the intermittent exercise trials [n=6]. * = significant difference (P < 0.05) between Fresh and Fatigue.

WR₆₀

WR₁₂₀

W′_totalTTEFresh

[kJ]

W′_totalTTEFatigued [kJ]

W′_totalTTEFresh [kJ]

W′_totalTTEFatigued

[kJ]

11.1 ± 4.1

5.7 ± 2.5 *

8.4 ± 3.8

6.2 ± 1.8

Discussion

Preliminary data from this study suggests that acute fatigue reduces W′_totalTTE during intermittent exhaustive exercise. Moreover, there was a lower W′_totalTTE in the fatigued trials at WR₆₀than WR₁₂₀. Although an individualised τ_W'was calculated for the fresh trials, it was not possible for the fatigued trials. This is because W′_totalTTE was reduced, suggesting that W′ does not reconstitute completely up to 1 hour after exhaustive intermittent exercise. This resulted physiological implausible τ_W' values when solving the W′_BAL model for 0 kJ at task failure in the fatigued trials (e.g. fresh τ_W': 425 sec vs fatigued τ_W': 4512 sec).

Implications for applied practitioners

These observations have important implications for future refinement of the W′_BALmodel, particularly during fatiguing exercise where so-called durability becomes critical. Moreover, these data suggest the potential requirement to include a 3^rd parameter into the W′_BAL model (i.e., W′_totalTTE) to account for the deterioration in W′.

Optimizing Cycling Time Trial Performance in Zwift: Validation of the PACE Model for Pacing Strategy

Matthew Van Dyck

Purpose:

The application of the critical power (CP) model to intermittent exercise allows to gain valuable insights into the dynamic balance between

A Novel Energy Imbalance Approach Applied to Rolling Resistance Assessment in Cycling

Manuel SELLIER

We present a novel approach for on-field rolling resistance testing, achieving repeatability below 2% using only 100-meter test sections. This method has been successfully applied to evaluate the influence of tire pressure, compare different tire versions within the same manufacturer, assess rolling resistance variations between manufacturers, and analyze the effect of tire temperature. Significant differences were observed, including a 26% increase in rolling resistance between 6 and 3 bar tire pressures, a 33% difference between Time-Trial and non-Time-Trial tire versions, a 9% average variation across three manufacturers, and a 26% increase between warm (18°C) and cold (14°C) tires.

Effects of Different Inclines on Lower Limb Muscle Activation During Cycling

Chia-Hsiang CHEN

This study aimed to examine the effects of different inclines on lower limb muscle activation during cycling. Twelve university students (mean height: 170.6 ± 3.5 cm, mean weight: 67.8 ± 7.1 kg, mean age: 20.5 ± 0.4 years) participated in the study. Muscle activation of the rectus femoris, biceps femoris, tibialis anterior, and gastrocnemius was measured using a DELSYS electromyography (EMG) system. A smart bicycle posture adjustment system was utilized to modify seat tube angles and handlebar positions. Participants completed cycling trials at three incline conditions—flat (0°), uphill (+5°), and downhill (-5°)—while maintaining a workload of 100 W and a cadence of 90 rpm. A repeated-measures ANOVA was performed to analyze differences in muscle activation across inclines. The results indicated that rectus femoris, biceps femoris, and gastrocnemius activation was significantly higher during uphill cycling compared to flat and downhill conditions. Additionally, activation of these muscles was greater on flat terrain than during downhill cycling, whereas tibialis anterior activation did not differ significantly across inclines. These findings suggest that even a 5° change in incline alters lower limb muscle activation patterns, providing valuable insights for cycling training and biomechanics research.

Development of a Non-Invasive Screening Method for Positional Sensitivity to Blood Flow Limitations in the Iliac Arteries

Jem Arnold

Flow limitations in the iliac arteries (FLIA) or endofibrosis is a sport-related vascular condition affecting elite endurance athletes, most commonly cyclists. This condition is rare, but awareness is growing that FLIA is a potential occupational disease in professional cyclists. Non-specific presentation of local pain and weakness in the working muscles (often unilateral quadriceps) and performance limitation during high-intensity exercise poses a challenge for early detection, diagnosis, and management. Aggressive racing positions tend to aggravate existing symptoms, but whether bike position alone increases risk of developing FLIA is not established.

We are developing a non-invasive field assessment using muscle oxygenation with near-infrared spectroscopy (mNIRS) and pedal power balance to help detect and monitor FLIA, with low burden to the athlete and test administrators, and without requiring referral to a specialist centre. We have shown that mNIRS reoxygenation after provocative exercise is delayed in patients with FLIA compared to healthy athletes (1). We are currently investigating whether these methods are also sensitive to adjustments in racing position, including validation with an Intravenous dye tracer (indocyanine green; ICG) to observe relative blood flow index (BFI) in the quadriceps during high-intensity cycling.

We hypothesise that a more hip-flexed racing position will result in impaired performance and mNIRS oxygenation kinetics, associated with reduced BFI. However, we expect to see high individual variability. We will present preliminary findings on the potential for screening athletes based on positional sensitivity to altered oxygenation kinetics and BFI. mNIRS and pedal power balance may help to screen athletes who are at higher risk for FLIA as part of a bike fitting position assessment.

1. van Hooff, Arnold, et al. (2022). Diagnosing Sport-Related Flow Limitations in the Iliac Arteries Using near-Infrared Spectroscopy. Journal of Clinical Medicine. doi:10.3390/jcm11247462

Effect of Carbohydrate Ingestion on the Power-Duration Relationship Following Prolonged Endurance Exercise in the Moderate Exercise Intensity Domain

Bernardo Norte

Introduction: Physiological assessments of endurance athletes have typically occurred in a well-rested state, which albeit important do not account for the magnitude of deterioration in physiological profiling variables (e.g., lactate threshold, V̇O_2max, critical power, functional threshold power) over time during prolonged endurance exercise (also known as ‘durability’). We tested the hypotheses that the parameters of the power-duration relationship, estimated as the end-test power (EP) and work done above EP (WEP) during the 3-min all-out critical power test (3MT), would be reduced following prolonged exercise in the moderate exercise intensity domain and that carbohydrate (CHO) ingestion would (at least) attenuate the reduction in power output at the heavy-to-severe intensity transition in direct proportion to the rate of CHO supplementation during exercise.

Materials and Methods: In a repeated measures, randomised study design, eight endurance-trained cyclists and triathletes (one female) performed two characterisation trials to establish baseline physiological parameters (V̇O_2max: 55.41 ± 7.21 mL·kg^-1·min^-1; gas exchange threshold, GET: 164 ± 29 W), including a 3MT performed in a well-rested state to determine EP (‘Fresh’-EP) and WEP (‘Fresh’-WEP). Subsequently, on three separate occasions, participants completed a fatigued-3MT (Fatigued-EP and Fatigued-WEP) immediately following 180-min of moderate intensity exercise at 95% of GET while consuming a CHO supplement containing either 0 g·h^-1, 60 g·h^-1 or 120 g·h^-1 of CHO at a glucose-to-fructose ratio of 1:0.8. Each trial was commenced following the provision of a standardised high CHO diet in the 24 h prior to each lab visit (8 g·kg^-1 CHO) and 2 h after a high CHO breakfast (2 g·kg^-1) on the morning of the trial. Venous blood, gas exchange, whole-body CHO and lipid oxidation, exercise energy expenditure, subjective gastrointestinal symptoms, capillary blood lactate and rating of perceived exertion were measured throughout exercise.

Results: Power output at the heavy-to-severe transition was significantly lower in Fatigued-EP_Water (251 ± 30 W, P = 0.016) and Fatigued-EP_CHO60 (269 ± 27 W, P = 0.022) than ‘Fresh’-EP (285 ± 31 W. There were no differences between ‘Fresh’-EP and Fatigued-EP_CHO120 (276 ± 33 W, P = 0.220), nor between Fatigued-EP_CHO60 and Fatigued-EP_CHO120 (P = 0.587). There was no difference between ‘Fresh’-WEP (14.67 ± 4.02 kJ) and Fatigued-WEP for all three conditions. Mean whole-body CHO utilisation rates were significantly higher in CHO₁₂₀ (2.12 ± 0.21 g·min^-1, P = 0.010) compared to Water only trial (1.49 ± 0.49 g·min^-1), with no differences between CHO₁₂₀and CHO₆₀ trials (1.86 ± 0.37 g·min^-1, P = 0.119). Total exercise energy expenditure did not differ between conditions (P = 0.731).

Conclusion: These data suggest that CHO ingestion during prolonged exercise in the moderate exercise intensity domain at a rate of 120 g·h^-1, but not 60 g·h^-1, negates the reduction in power output at the heavy-to-severe boundary measured in a well-rested versus semi-fatigued state.

Women in Cycling: A Comparative Study of Speed Choices and Lane Position Preferences Across Europe

Giuseppina Pappalardo

Women generally have a lower cycling participation rate, partly due to their heightened perception of risk and safety concerns, especially related to cycling in traffic. They are more likely than men to report apprehension about cycling in traffic and perceive a higher risk of injury from collisions with motor vehicles. Women experience higher discomfort and risk perception when cycling in mixed traffic. However, these differences tend to decrease among regular cyclists, and women may even perceive greater personal benefits from cycling compared to men (Lawson et al., 2023, Part, 2019). Women tend to prefer cycling on roads with cycling-friendly features, such as dedicated bicycle lanes, and are more likely to cycle during off-peak hours and on weekends (Debnath et al., 2021). However, many of the previous studies are based on interview ad subjective evaluation of risk perception that should be complemented by in field observation (Sanjurjo-de-No et al., 2025).

Data collection and Methods

To analyze geographical and gender differences in cycling, field video recordings were conducted on various road sections in Győr (Hungary), Krakow (Poland), Valencia (Spain), and Malmö (Sweden). To consider also the effect of bike path characteristics, two different cycling facilities were chosen:

Case 1: A sidewalk bike path (two-way bicycle path without a separated sidewalk).
Case 2: A protected or buffered bike path on the roadway (two-way bicycle path with a separated sidewalk).

The selected sections were straight and flat, located at least 50 meters from the nearest intersection, to limit the influence of physical constrains in the choice of speed among cyclists. A camera was positioned on the opposite side of the road at each site to capture a broad view (approximately 25 meters) while providing enough detail to estimate the rider's gender and age and to calculate speed and lane position.

By analyzing several hours of video frame sequences, the final selected dataset comprises 2010 free flow observations. It includes 537 records from Hungary, 756 from Poland, 257 from Spain, and 460 from Sweden.

Decision trees are popular machine learning supervised algorithms in data mining given their no-parametric approach, the capacity to handle both numerical and categorical data, non-linear relationships, intelligibility and interpretability of results. For this study we used the CHAID (Chi-squared Automatic Interaction Detector) as regression tree to envisage speed variability based on variables like countries, gender, and case and CART (Classification and Regression Trees) as classification tree to analyse position categories based on the same variables.

Discussion and Conclusions

The CHAID analysis provided more insights, showing that gender significantly influenced cycling speed in all countries, with males generally cycling faster than females. In Sweden and Spain, lane type influenced speed, with the outer lane showing a gender impact and lower speeds than the inner lane. A detailed assessment of risk perception using Classification and Regression Tree (CRT) analysis indicated that country and road cross-section significantly affect cyclists' positions within the lane. However, gender did not significantly influence position choice in the outer lane. Overall, the study highlights that women generally cycle at slower speeds than men and prefer safer, less risky cycling environments. The location of the bicycle lane within the road cross-section is a crucial factor influencing speed choices, particularly for women.

Cycling speed varied by country, with Poland recording the highest speeds, about 1.5 m/s faster than other countries. Males generally rode 0.4 m/s faster than females. The inner lane, being further from motorized traffic and closer to pedestrian areas, saw higher average speeds by about 0.5 m/s compared to the outer lane.

The study employs CHAID and CART decision tree techniques to investigate more in details the factors influencing cycling speeds and positions in the four European countries.

CHAID showed that in Poland, males cycle always faster significantly than females, while in Hungary, males especially in Case 2 exhibit higher speeds than females in the same group. Similarly, in Sweden and Spain, males in the outer lane cycle faster than females in the same lane.

Infrastructure type and lane also influence cycling speeds behaviour. In Hungary, the type of bike path affects speed, with Case 1 having a lower average speed. In Sweden and Spain, lane type affects speed differently, with inner lanes having higher speeds than outer lanes. The case type also influences speed in the inner lane for Sweden and Spain, with Case 1 having a higher mean speed than Case 2.

The CART decision tree analysis revealed that country and case are the primary factors affecting cyclists' positions, with gender playing a minor but notable role, especially in Sweden for Case 1. This suggests that while gender slightly influences cyclists' positions, country and infrastructure are more significant factors.

The study's results highlight the overall gender differences in cycling speeds and preferences. Women generally cycle at slower speeds compared to men and prefer safer, less risky cycling environments. The type of bicycle infrastructure significantly impacts these differences, with women favoring bike lanes that offer greater protection from motorized traffic. Overall, women perceive risk more than men and prefer more separation from motorized traffic. Conversely, men are more likely to engage in risky behaviours, adhere less to traffic regulations, and ride at higher speeds.

In conclusion, the study emphasizes the importance of considering gender differences when designing and implementing cycling infrastructure. By understanding and addressing these differences, cities can create safer and more inclusive cycling environments that cater to the needs of all cyclists, regardless of gender.

Modeling vs. Measuring: How accurate is the calculated Maximal Lactate Steady State?

Mattice Sablain

Introduction

The maximal lactate steady state (MLSS) represents the highest sustainable workload without net lactate accumulation. In 1986 Mader & Heck proposed a model estimating MLSS by determining the oxygen consumption at the crossing point (V̇O_2ss) between the rate of lactate formation and elimination. The sole input variables in this model are the V̇O_2max and the maximal rate of blood lactate accumulation (VLa_max) which is believed to represent the maximal glycolytic flux. Subsequently, an equation is often used to transform V̇O_2ss into a power output using the resting oxygen consumption (V̇O_2rest) and cycling efficiency (K_ef). Previous research found no significant difference between the calculated power at MLSS (cPMLSS) and experimentally measured power at MLSS (mPMLSS) (Hauser et al., 2014). However, the Bland-Altman plots showed quite broad limits of agreements (LoA). It remains unclear if these broad LoA are caused by an error in the determination of V̇O_2ss or the conversion into a power output as no prior studies have directly compared the measured (mV̇O_2ss) and calculated V̇O_2ss (cV̇O_2ss). Therefore, this study wanted to compare mV̇O_2ss and mPMLSS with cV̇O_2ss and cPMLSS. Furthermore, we examined whether individualizing specific model constants improved the LoA, potentially enhancing the model’s applicability.

Methods

7 healthy male participants (age: 31 ± 9 years; V̇O_2max: 58.1 ± 8.6 mL∙min^-1∙kg^-1) visited the lab on 4-7 occasions with a minimum of 24h in between. On the first testday V̇O_2rest was measured followed by a familiarization of a 15” isokinetic sprint test at 120 rpm (Cyclus2 ergometer; RBM Electronics, Leipzig, Germany). The second day consisted of the same 15” isokinetic sprint test with blood lactate measurements to determine VLa_max and a ramp incremental exercise test to determine V̇O_2max. On the following testdays mPMLSS was determined as per method of 30’ constant load tests. Related-samples Wilcoxon tests were used to compare measured and calculated V̇O₂_ss and PMLSS. Additionally, the same test was applied to assess whether individualization improved cPMLSS. Bland-Altman plots were used to visualize differences between measurements.

Results

No significant differences were observed between mV̇O₂_ss and cV̇O₂_ss (p = 0.176) nor between mPMLSS and cPMLSS (p = 0.917), despite broad LoA (Fig. 1). Moreover, while not statistically significant, individualizing V̇O₂_rest and K_ef led to larger biases between mPMLSS and cPMLSS, further widening the LoA.

Discussion

In conclusion, our study shows that while the Mader and Heck model is effective at the group level, it lacks accuracy for individual assessments. The errors in V̇O_2ss calculation may stem from the input variables, model constants, or the model itself. This highlights the need for caution when applying the model in training and performance settings. Further research is needed to identify the source of these inaccuracies.

A 5-minute Time Trial to Assess Durability in Professional Cyclists: A Field Study

Borja Martinez-Gonzalez

In recent years, durability has emerged as an additional factor in endurance performance models. The literature on protocols to assess durability in the field is limited. This study investigated the impact of prior accumulated work on 5-minute time trial performance. Twenty male professional cyclists performed the test in a single session, which was divided into three phases: 1) 60 minutes at 60% of the lactate turnpoint (LT2), followed by a 5-minute time trial; 2) 60 minutes at 60% of LT2, followed by 3 repetitions of 5 minutes at 105-110% of LT2 with 5 minutes of active recovery; and 3) 60 minutes at 60% of LT2, followed by a 5-minute time trial. Power output, heart rate, and cadence were measured throughout the test. The total work done before the 5-minute time trial in the fatigued state was 3432 ± 242 kJ. The results showed a significant reduction in both absolute and relative power output in the fatigued state compared to the fresh state (p < 0.001), while heart rate and cadence did not show significant differences (p > 0.05). The decline in performance was -16 ± 16 W (96.4 ± 3.5%). These findings suggest that prior accumulated work leads to a notable decline in 5-minute time trial performance. The study provides valuable insights into the development of a test to assess durability in a single training session. Future research should aim to validate and establish the reliability of this protocol, exploring different amounts of prior accumulated work to further investigate fatigue resistance in endurance athletes.

Exogenous ketosis increases D-β-hydroxybutyrate concentrations in human cerebrospinal fluid

Simon K Kjær

Background:

Over the last decade, there has been growing interest in the use of ketone supplements to

explore their effects on elite-level cycling performance, training adaptations, and, in particular,

recovery. This includes areas related to the brain, such as enhancing cognitive resilience post-

exercise, improving sleep quality and efficiency, and increasing the neuroprotective hormone

brain-derived neurotrophic factor (BDNF) and dopamine levels in circulation. We

hypothesized that ketone ester intake before a planned lumbar puncture would increase D-β-

hydroxybutyrate (βHB) and subsequently BDNF and dopamine in the cerebrospinal fluid

(CSF), a reflection of the levels in the brain.

Methods:

In a randomized, placebo-controlled, single-blind, parallel-group study, patients referred for

elective lumbar puncture received either 30 g of ketone ester 1 hour before (KET 1H), 2 hours

before (KET 2H), or a placebo (PLA) before the lumbar puncture procedure. βHB

concentrations were measured using point-of-care testing (POCT) and Liquid

Chromatography-Mass Spectrometry (LC-MS). The primary endpoint was βHB concentrations

in the CSF.

Results:

The study is ongoing; as of submission, 20 out of 24 patients have been included: 6 in PLA, 7

in KET 1H, and 7 in KET 2H. Blood βHB concentrations, measured by POCT, averaged 0.2 ±

0.09 in PLA, 3.22 ± 0.85 in KET 1H, and 2.66 ± 1.01 in KET 2H following lumbar puncture.

CSF βHB, also measured by POCT, was 0.05 ± 0.05 in PLA and increased to 0.17 ± 0.08 (P =

0.07) in KET 1H and to 0.47 ± 0.17 (P < 0.001) in KET 2H.

Conclusion:

KE treatment administered 2 hours before lumbar puncture was associated with the highest

increase in βHB levels in CSF compared to placebo. Verification by LC-MS and

neurotransmitter measurements will follow upon study completion. Our results suggest a delay

in the passage of βHB from blood into the CSF, which underlines the importance of timing

ketone supplement ingestion to maximize its effects on the central nervous system in the

context of exercise and recovery.

Clustering professional road cycling races into distinctive race profiles

Annemiek Roete

Introduction

Professional road cycling races vary in format and terrain, resulting in diverse physiological demands on riders. These variations are largely attributable to the race profiles. A better understanding of race demands resulting from different race profiles can optimize cyclists’ performance by tailoring training to race type. Additionally, it can support the decision making around a riders’ race program, training periodization, and can enhance team composition and scouting by matching riders’ strength and weaknesses to specific race profiles. Race organizers or researchers often apply their own criteria to categorize races into distinctive types. There is currently no standardized categorization for professional men’s road cycling. This study aims to develop a categorization model for professional road cycling races using clustering analysis.

2. Materials and Methods

The initial dataset comprised international races listed on the 2024 Union Cycliste International (UCI) calendar, including .WT, .Pro, .1, and .2 classifications. Features used in earlier studies with a physiological focus which categorized races into distinct types were summarized and used as features in the clustering analysis. Relevant data to quantify those features were obtained from ProcyclingStats, an open-access database. The features identified in earlier research included race distance, vertical meters, number of climbs, total uphill distance, vertical meters of all climbs, vertical meters of climbs positioned in the last 10km and, vertical meters located at the finish. To enhance the accuracy and stability of the clustering process, noise reduction was performed by identifying races that had been explicitly categorized by organizers, as these were more likely to be well-defined and distinctive. Additionally, all data was thoroughly checked. Incomplete and or incorrect data was removed. The clustering tendency of the dataset was assessed using the Hopkins statistics. Three clustering algorithms were then applied: KMeans, hierarchical clustering with genie linkage and noise detection, and the Ordering Points To Identify Cluster Structure (OPTICS) algorithm. Each algorithm’s performance was validated using silhouette score. In addition, a subset of labeled races provided by race organizers, with the Brier score employed as a validation metric with lower values indicated better performance.

3. Results

In total, 1,020 race days were reviewed for inclusion into the clustering dataset, with 276 race days being included. Hopkins statistics was 0.70, indicating the dataset is clusterable. KMeans clustering produced five distinct clusters: time trial, flat, hilly, semi-mountain and mountain, with a Brier score of 0.38. Hierarchical clustering also identified five clusters corresponding to time trial, flat, hilly, semi-mountain, mountain, with a Brier score of 0.48. OPTICS generated 15 clusters, which were later grouped into three time trial types (various distance), one hilly, six flat (with varying profiles), three semi-mountain types, and two mountain clusters, yielding a Brier score of 0.71.

4. Discussion

The differences in clustering results reflect the inherent variations in the algorithms’ methodologies. KMeans demonstrated the best performance compared to hierarchical clustering and OPTICS. While hierarchical clustering and OPTICS are better suited for handling complex data structures, their results suggest a tendency to overfit the dataset in this context, especially given the inherent variability in road cycling races, even within the same category. The findings of this study suggest, that based on clustering, race profiles of men professional cycling races may be divided into flat, hilly, semi-mountain, mountain, and time trial race profiles. Using such a unified categorization of race profiles allows for a comparison of different races and studies, providing a comprehensive overview of all professional cycling races. Additionally, this standardized classification enables a deeper analysis of race demands by recognizing variations in race profiles.

The Influence of Prior Accumulated Work on the Torque-Cadence Relationship in Junior Cyclists.

Andrea Giorgi

Understanding the torque-cadence-power relationship in fatigue is crucial for assessing cyclists' performance potential. The aim of the study is to investigate the impact of prior accumulated work (like a junior cycling road race) on the power duration and torque cadence relationship. Results showed that race simulation workload induces reductions in Critical Power (CP), work capacity (W'), 15 seconds, 3 minutes and 12 minutes all out efforts along with lower cadences. On the other hand, torque values did not change under fatigue across all out efforts. These findings highlight that the power output under fatigued conditions in road cyclists relies more on torque than on cadence

The Relationship between durability and cycling performance

George Evans

The relationship between durability and cycling performance

George Evans ¹^*, Daniel Muniz-Pumares ²

^*Correspondence (GE) g.evans6@herts.ac.uk

Affiliation; School of Life and Medical Science, University of Hertfordshire, Hatfield, UK

Durability has been proposed as a component of endurance performance, however, different protocols to measure durability exist, and the extent to which durability explains variation in performance is not yet fully understood. This study investigated whether durability, assessed as the decline in physiological variables following two different fatiguing protocols, could predict cycling performance assessed as simulated time trial and road race performance.

Four competitive cyclists (28 ± 8.3 years, mass: 79.1 ± 5.4 kg, VO_2max: 62.7 ± 4.6 ml•kg•min^-1) have so far completed physiological testing in a fresh state, and following two fatiguing, work-matched protocols comprised of 15 kJ·kg^-1 of continuous and intermittent cycling, as well as a 26.8 km simulated time trial and a road race which included 4 x 5 min time trials. Provisional results show that peak ramp power (baseline = 428.5 ± 22.3 W, continuous = 399.3 ± 27.2 W, intermittent = 374.3 ± 24.2 W), critical power (baseline = 340 ± 19.1 W, continuous = 327.4 ± 16.1 W, intermittent = 304.7 ± 12 W) and W-prime (baseline = 12.3 ± 2.2 kJ, continuous = 9 ± 4.86 kJ, intermittent = 8.9 ± 5.7 kJ) seem to show differences after 15 kJ·kg^-1. The simulated time trial took participants 39.9 ± 0.5 min to complete, the power output for this duration was 306.6 ± 19.8 W. The average power of the combined 5-min performance measure during the simulated road race was 330.6 ± 23.7 W. Statistics will be completed once the sample has reached full power. For the first time, we hope to show the relationship between durability and cycling performance in a controlled laboratory environment and whether the type of fatiguing protocol (continuous or intermittent) used has an influence on the effectiveness of using durability as a predictor of performance.

Keywords: Durability; Physiology; Fatigue; Performance

Training modification decision-making and the role of athlete monitoring in elite endurance sports coaching: An instrumental case study of a world-class cycling coach and a world-class development rowing coach

Wouter Piet Timmerman

The purpose of this study is to improve the understanding of the training modifications made by elite endurance coaches and explore the role athlete monitoring has in this process. Using Stake’s (1995, 2003) case study methodology, an instrumental case study design was employed to allow naturalistic generalisations to be drawn when exploring endurance sports coaching and athlete monitoring at the elite level. The study followed a world-class cycling coach and a world-class development rowing coach for six months, and data collection included four semi-structured interviews, eleven retrospective verbal reports through stimulated recall, and eight field observations of training sessions. A qualitative analysis consisting of categorical aggregation and direct interpretation generated three themes regarding coaches’ athlete monitoring and training modifications: 1) Gathering and balancing information for continuous situational assessment and training modification triggers; 2) The only constant in coaching is change: Anticipating modifications and being ready for the unexpected; 3) Empowering athlete autonomy of training modification decision-making within coach boundaries. The findings indicated that minor to moderate training modifications were predominantly reactive to subjective athlete feedback and disruptive events, while major training modifications were predominantly in response to crisis events. Quantitative monitoring data alone rarely dictated a training modification unless there were concerns about the athlete’s honesty, however, quantitative information led to gathering additional qualitative information through intensified coach-athlete communication and coach observation of the athlete. The findings also highlighted that athlete monitoring had a central role in modifying training, yet non-athlete monitoring information such as weather, logistics, and organisational decisions was frequently the decisive information. Further, coaches demonstrated adaptability to changing circumstances through anticipatory conditional decision-making. This was done by establishing multiple decision paths in advance and, depending on the athlete monitoring outcome, selecting the most appropriate course of action. Additionally, the coaches empowered their athletes and gave them the autonomy to make minor to moderate training modifications within the coach’s boundaries. Lastly, the cross-case differences reflected sport-specific and organisational factors in training modifications and showed a higher injury incidence of the under-23 rowers compared to the professional cyclists, which may be related to the experience of the athletes.

Automated Data and Video Processing in Track Cycling

Thomas Vleeschouwers

Recent advances in sports data analytics have driven the development of integrated performance monitoring systems. This abstract describes the implementation and use cases of integrating a pan-tilt-zoom (PTZ) camera into the existing Wireless Cycling Network (WCN) deployed at the Wielercentrum Eddy Merckx in Gent, Belgium. WCN combines a precise timing system, ANT+ sensor data collection and an anaerobic capacity (W’) estimation. Automated PTZ camera tracking and recording enables immediate visual analysis by coaches. AI-computer vision models are used for detection and qualitative analysis of team pursuit changeovers. In addition to elite use cases, general engagement can be improved with event-specific real-time data display and personalized videos and session reports. These integrated capabilities and the modular design streamline coaching workflows and provide a platform for future applications across various track-based sports. Future work includes further development of analysis tools to support coaches, production deployment of the complete system and implementing it at additional tracks.

Heat Acclimatization: Bridging the Gap Between High Performance and General Health

Alberto Hermo Argibay

The increase in temperature due to climate change poses significant health risks for both the general population and athletes, particularly those with low thermoregulatory capacity, who may experience increased stress responses that could lead to vascular complications like atherosclerosis. To mitigate these effects, athletes have adopted heat acclimatization protocols, which could also benefit the general population. In our study, ten physically active subjects (6 males, 4 females) underwent a heat training protocol twice a week for eight weeks, involving 30 minutes of constant-power exercise at 36°C. Subjects were tested incrementally and had blood samples taken to evaluate physical fitness and markers of atherosclerosis risk. Results showed improvements in body composition, cardiovascular function, thermoregulatory capacity and metabolic response to heat as well as an improvement in the risk of developing cardiovascular diseases.

Applications of Language Modelling for a cycling Aerodynamics’ coach

Callum Barnes

This study investigates the application of Language Modelling in cycling aerodynamics. A novel ground truth is created through recruiting a cohort of experts in cycling aerodynamics, bike fit and biomechanics and taking that ground truth to be the collective expert consensus. Within this study 9 Large Language Models and 1 Large Reasoning Model were tested with 7 of the Large Language Models being open-source models from Google, Meta, Microsoft and Alibaba and the closed source models from OpenAI. This study tested these models without a system prompt, with a system prompt, with applied Retrieval Augmented Generation, with an enthusiast level knowledge base and Retrieval Augmented Generation with a more technical knowledgebase. The best performing model in this study was OpenAI’s Chat-GPT 4o with an average mark of (90 ± 0.41)%. And the best performing opensource model was Alibaba’s Qwen2.5:32b with a system prompt and the technical knowledge base providing an average score of (88.73 ± 0.29) %. The results from this study show that it is possible to develop a model which performs to a similar level of a human expert within the domain of aerodynamics, bike fit and biomechanics in cycling. Additionally, this study proposes a method to experimentally quantify the improvements an athlete can make through the assistance of a domain specific Large Language Model.

Traditional Resistance Training versus Torque Training: A Randomized Controlled Trial

Raúl de Pablos

Resistance training (RT) has been shown to improve endurance performance in cycling. Notably, it has been recommended that cyclists perform heavy (>70% of one-repetition maximum [1RM]) RT sessions off the bike to optimize neuromuscular adaptations. On the other hand, the so-called “torque” training (i.e., low-cadence efforts performed against an allegedly high pedaling intensity) has become increasingly popular as an on-bike alternative to conventional (off-bike) RT for enhancing lower-limb muscle strength and cycling performance. However, recent evidence suggests that the force demands of this torque training on lower-limb muscles are relatively low (i.e., <50% of the cyclists’ maximal lower-limb dynamic force), and its effectiveness compared to training at similar relative intensities but with self-selected cadences remains unclear This study compared the effects of off-bike RT (squats) and torque training on endurance-related parameters in well-trained cyclists. Twenty-seven male cyclists were randomly assigned to off-bike RT (n = 9), torque training (n = 9), or a control group (n = 9) for 10 weeks. The RT group performed 5 sets of 7 squats at 70% 1RM twice per week, with a 4-minute intraset recovery, while the torque group completed 5 sets of 4-minute intervals at 70% VO₂max power output at a target cadence of 50rpm, with a 2-minute intraset recovery, twice a week. Measured outcomes included VO₂max, maximal aerobic power (MAP), ventilatory thresholds (VT, RCP), and time to exhaustion at RCP. Off-bike RT significantly improved MAP (p = 0.05, ES = 0.60), VT (p < 0.05, ES = 0.47), and RCP (p = 0.05, ES = 0.48). No significant changes were observed in the torque training group and control group. These findings indicate that off-bike RT is a superior method for improving pedaling performance, highlighting its importance in cycling training. Future research should refine torque training protocols to enhance its efficacy.

Article Traditional Resistance Training versus Torque Training: A Randomized Controlled Trial

Raúl de Pablos

Sodium Bicarbonate Ingestion for Performance Enhancement in Cycling: From Original Concept to Endurance Ergogenic

Andy Sparks

Ingestion of sodium bicarbonate (SB) to enhance extracellular buffering and exercise performance has been researched since the 1930’s. In that time there has been a considerable volume of research on SB across many sport and exercise contexts. Much of this work was conducted during the 1980’s and 1990’s (McNaughton, 1992) and laid the foundations for the ingestion practices that some athletes are still using, but over the past 10 years, there have been notable advances in SB ingestion methods. Renewed interest in SB began with the identification of highly variable inter-individual blood bicarbonate (HCO₃) responses (Jones et al., 2016), typical with traditional ingestion modes such as fluid or gelatine capsules. In response to this, the time to peak pre-exercise ingestion timing approach demonstrated that individualising ingestion time could improve performance further (Boegman et al., 2020). Meanwhile, attention also started to focus on methods to reduce the gastrointestinal symptoms (GIS) that can negatively impact performance for some athletes following SB ingestion. Pharmaceutical technologies were used to reduce the interaction of SB with stomach acid, via delayed release and enterically coated capsules, establishing that it was possible to considerably reduce some of the GIS (Hilton et al., 2019). In so doing, the pharmacokinetic responses of blood HCO₃ were also observed to be specific to the mode of SB ingestion, which has implications for both the dose and timing of each ingestion form to maximise its ergogenic potential (Leach et al., 2023). The resolution of GIS problems has come recently in the format of a commercially available SB ingestion system that combines mini-tablets with a carbohydrate hydrogel delivery system (MBS). The MBS has demonstrated very low GIS responses and elevated and prolonged changes in blood HCO₃ (Gough and Sparks 2024a). Furthermore, the MBS has now been shown to be effective at improving performance and recovery from 4 km time trials (Gough and Sparks 2024b), to enhance 40 km time trial performance (Shannon et al., 2024) and is being used extensively in the professional peloton. Interestingly, Shannon et al., (2024) have also suggested that gross economy may be improved using the MBS, along with relative reductions in perceived exertion. With the resolution of the GIS, it may now be possible to more clearly determine the ergogenic mechanisms that SB may enhance in addition to hydrogen ion buffering. These potential mechanisms include improved economy, muscle recruitment, and reduced pain and perceived exertion, and represent an exciting new avenue of investigation in future SB performance research.

Performance Level Influences Racing Profiles of 1-km Time Trial in Male High School Cyclists

Tomohiro Nakamura

Analyses of 1-km time trial (TT) racing profiles specifically for high school cyclists are lacking. This study aimed to clarify the racing profiles of male high school cyclists in a 1-km TT. The 1-km TT performance was analyzed in 50 male high school cyclists who participated in competitions held on the same track (333.3 m). The cyclists were divided into the High (n = 25, 72.9 ± 2.0 seconds) and Low (n = 25, 79.9 ± 3.1 seconds) groups based on their 1-km TT performance. We obtained panning shots of cyclists during the 1-km TT race and calculated the section velocity at every 83 m, average velocity, peak velocity, normalized velocity, and fatigue index from lap times. Analysis of variance showed significant main effects for section velocities and group, as well as a significant interaction in the 1-km TT velocity curve. The peak velocity was significantly higher in the High group than in the Low group (p = 0.000). However, the fatigue index did not differ between the groups. Significant negative correlations were found between 1-km TT record and velocity in each section and peak velocity in both groups (p < 0.05 or p < 0.01). In conclusion, the racing profiles of 1-km TT related to better performance are determined by higher peak velocity but are not influenced by the rate of velocity decline after reaching the peak velocity in male high school cyclists.

The Impact of rider bicycle weight, tyre pressure and surface condition on rolling resistance of bicycle tyres using a laboratory treadmill (p.a.v.e.)

Jens Buder

Rolling resistance of tyres on bicycles are commonly considered as performance determining parameter in competitive cycling. There are several constraints to consider in setting up a bicycle for a given scenario for racing in order to provide maximum material performance to a rider, by means here a low rolling resistance, such as: rider bicycle weight, tyre filling pressure, the tyres itself and surface conditions.

In this study, a set of 5 different tyre widths, each of the same model, have been examined on their rolling resistance in context of varying conditions, by means: 4 different rider bicycle weights, two different surface conditions each at 6 different tyre filling pressures. The study used a laboratory treadmill (p.a.v.e.) to determine the rolling resistance in a coast out style test design, introduced by Buder et.al. in 2024 (1).

The findings support the idea of setting up bicycles application-, condition- and rider specific. A strong correlation between rider weight, tyre pressure, tyre width as well the presence of surface induced vibrations could be established. Each of the parameters itself but but also combinations of these impacted the rolling resistance and therefore appear to be valid as performance setting values. Based on the findings of this study, it is suggested to introduce rider specific indices and impact factors for each of the parameters that prioritise bike set up measurements in order to achieve a maximum support for that specific athlete. These indices are suggested to be examined in yet to be established physiological test designs, with subjects, representing a valid target audience (national level or higher level competing cyclists), during several cycling seasons in order to establish a general or even specific guideline for bike setups in professional bicycle racing.

Influence of seating posture and gender on pelvic movement and saddle pressure during cycling

Frank Michel

The aim of the present study was to compare female and male cyclists riding in an inclined and an upright seating posture based on saddle pressure distribution and pelvic movement. The results show differences between females and males as well as between the two seating postures. In general, females show less mean saddle pressure and less pelvic movement in the frontal plane. However, females reveal a higher anterior forward pelvic tilt which contributes to higher mean pressure around the anterior saddle region. The results of the study confirm the importance of selecting the appropriate saddle and seat pad respectively depending on gender and cycling specific activities to improve seating comfort and to reduce the risk of injury.

Pelvic motion and saddle pressure of female and male cyclists during a cycling specific graded exercise test

Frank Michel

The aim of the present study was to analyse the movement of the buttock on the saddle based on pelvic motion and saddle contact pressure during a cycling specific graded exercise test. Nine female and ten male recreational cyclists took part in the study. Overall, both the saddle pressure distribution and the pelvic motion show that with increasing intensity/fatigue there is more movement between the buttock and the saddle to observe. In particular, pelvic movement in the frontal plane (rocking) should be considered for bike fitting interventions as well as for the design of saddles and seat pads at high load intensities/fatigue. Substantial pelvic movement can lead to critical shear forces between the buttock and saddle. These shear forces, combined with moisture development can therefore lead to a higher risk of seating discomfort and pain.

The Impact of drivetrain configuration on overall cycling efficiency

Bart Blockmans

Despite frequent displays of undisputed dominance in cycling’s biggest road races, time trialing remains a test of precision, where victory is often determined by razor-thin margins. In all of the major time trials of the 2024 season, champions were separated from runner-ups by time differences of less than 0.5% (or 9 seconds per 30 minutes). As a consequence, the pursuit of cycling efficiency has never been as relentless as today, with countless hours of wind tunnel testing being used to scrape off the last fraction of a watt from a cyclist’s jersey. Besides aerodynamic efficiency, bicycle drivetrain efficiency has received increased interest in recent years. On the verge of what could be a paradigm shift in bicycle drivetrain technology, the traditional front derailleur-based 2x systems are increasingly being challenged by “derailleur-killing” technologies such as 1x configurations and drivetrains equipped with hub gears. Whether or not said technology shift will actually happen, and if so in which of cycling’s many disciplines, depends to a large extend on the proven efficiency of these new drivetrain concepts. In the current work, a comparative analysis of various drivetrain configurations is presented, based on a combined simulation-based and experimental approach, with case studies from World Tour time trialing in the 2024 and 2025 seasons.

Double Threshold Training in Cycling: A Case Study on Physiological Adaptations in an Amateur Female Cyclist

Mollie Brewer

Introduction: Double training days, where two structured sessions are performed in a single day, are widely used in endurance sports such as swimming, running, or cross-country skiing (Haugen et al., 2022; Mikkelsen, 2023; Torvik et al., 2021). Traditional threshold training protocols usually prescribe single-session efforts at or near lactate threshold to improve endurance performance. However, emerging research in endurance sports, particularly in elite distance running, has explored the physiological benefits of performing two sessions at threshold intensity within a single day (Casado et al., 2023; Kelemen et al., 2023). This approach aims to increase total time at threshold while minimizing muscular strain and optimizing metabolic adaptations (Kjøsen Talsnes et al., 2024). Despite its increasing adoption in other endurance sports, its application to cycling remains underexplored.

This study was motivated by challenges with traditional threshold training, which did not yield the desired physiological adaptations. Specifically, achieving sufficient time at threshold in a single session proved difficult without excessive fatigue. This led to an exploration of whether increasing threshold frequency by performing two separate sessions in a single day (double threshold training), spaced 5-7 hours apart, could provide sufficient adaptive stimulus while better managing recovery demands. To assess these training adaptations, this study utilizes lactate profile shifts to evaluate physiological changes associated with both a traditional threshold approach and a double threshold training approach in a female amateur cyclist using a two-year comparative case study design.

Method: This single-subject case study (the author, a female amateur cyclist) investigates the physiological effects of different threshold training structures in cycling. Over two consecutive years, the author completed two 12-week training interventions during the same time period each year within a structured block periodization model. In 2022, training included one weekly threshold session, while in 2023, greater time at threshold was achieved by performing two threshold interval sessions within the same day. Both interventions used a periodized, progressive approach to increase the duration of the threshold intervals. In both training interventions, each week also included an additional high-intensity session, typically a cyclocross race or group ride, while the remaining training consisted of low-intensity, high-volume endurance riding. Threshold interval training sessions alternated between on-road cycling and an ergometer. Low intensity training comprised 70-80% total training volume in both years (150.6 hours in 2022, 167.5 hours in 2023), ensuring endurance adaptations were maintained while allowing for differences in threshold training structure. While every effort was made to maintain rigorous, comparable conditions, this study was conducted in a real-world training environment rather than a controlled laboratory setting.

Lactate profiles were used as the primary metric for evaluating changes in training adaptations. Pre- and post-intervention graded exercise tests to exhaustion were conducted on an electronically controlled cycling ergometer (Wahoo Kickr, Wahoo Fitness, Atlanta, GA, USA). The test followed a modified step-test protocol (Millán et al., 2009), with capillary blood lactate (Lactate Plus, Nova Biomedical, Waltham, MA, USA), power output (Wahoo Kickr, Wahoo Fitness, Atlanta, GA, USA) and heart rate (Polar F10, Polar Electro Oy, Kempele, Finland) measured at 5-minute intervals during incremental workloads of 0.5 watts/kg to establish a lactate profile. A 10-minute stage at 3.5 W/kg was included to assess sustainable threshold intensity, defined as the highest workload at which lactate accumulation remained stable (<1.0 mmol/L increase between mid- and end-stage measurements). Throughout both training interventions, threshold training intensities were prescribed based on pre-test results. Heart rate and power were used to regulate intensity, with frequent lactate sampling during threshold intervals to ensure precision in training intensity, targeting lactate values between 3.5-4.5 mmol/L.

Results: Lactate profile analyses revealed clear differences between the two training interventions, with lower lactate accumulation at equivalent workloads following double threshold training.

After 12 weeks of traditional threshold training (2022) the lactate profile showed minimal changes except that the athlete was able to complete an additional workload stage, increasing from 3.5 W/kg to 4.0 W/kg. At the highest completed stage, peak lactate levels increased from 6.3 to 12.6 mmol/L, reflecting an increased tolerance for high-intensity efforts rather than sustainable efficiency.

In contrast, following 12 weeks of double threshold training (2023), the lactate profile exhibited a lower and rightward shift, reflecting an improved ability to sustain higher workloads with reduced lactate accumulation. The athlete completed an additional workload stage, progressing from 4.0 W/kg to 4.5 W/kg. At 3.0 W/kg, levels dropped from 3.3 to 1.8 mmol/L, while at 3.5 W/kg, they fell from 5.8 to 3.8 mmol/L. This suggests the body became more efficient at clearing lactate, allowing for harder efforts with less fatigue.

Post-test comparisons between traditional threshold training (2022) and double threshold training (2023) revealed clear differences in lactate response for equivalent workloads. While there was some variation in pre-test values, they were not drastically different-6.4 mmol/L for traditional threshold and 5.8 mmol/L for double threshold at 3.5 W/kg. After 12 weeks, lactate at 3.5 W/kg barely changed with traditional training, decreasing only slightly from 6.4 to 6.3 mmol/L. In contrast, double threshold training led to a much greater reduction, with lactate dropping from 5.8 to 3.8 mmol/L. Post-test results show greater adaptation after 12 weeks of double threshold training, with a larger reduction in lactate levels compared to the traditional approach.

Discussion/Conclusion: These findings suggest that increasing training volume at threshold intensity through two sessions in a single day may offer a potentially promising training strategy in cycling. While the single-subject design limits generalizability, the results provide insights into how training frequency and structure may influence physiological responses to threshold work, highlighting a need for further research to validate these results in a broader cycling population.

Physiological Predictors of a super-sprint Triathlon Performance in junior and U23 Triathletes

Chiel Poffé

The mixed-team super-sprint triathlon relay (~300m swimming, ~6.6km cycling and ~1.5km running) performed in teams of four athletes, is the newest Olympic triathlon discipline. Until today, no scientific data is available on the specific physiological attributes. Our study aims to predict a super-sprint triathlon performance from physiological exercise testing. Fourteen national level triathletes performed physiological tests in swimming and running and competed in a super-sprint triathlon. The physiological profile in swimming and running was calculated and applied to predict triathlon performance. The speed at VO_{2max_swim} corresponded well to the swimming speed during the triathlon (0.013 m·s^-1, n.s.), whereas the running speed was similar to the running speed at metabolic steady state (MMSS_run) (0.05 m·s^-1, n.s.). Stepwise multiple regression analysis selected MMSS_run as the primary predictor of triathlon performance (r² = 0.66, p<0.05). The swimming speed at metabolic steady state (MMSS_swim) (r² = 0.84, p<0.05) and the amount of work that can be performed above MMSS_run (W’_run) were also included in the prediction model (r² = 0.91, p<0.05). Our data indicate that MMSS_run, MSS_swim, and W’_runallow for a precise prediction of a super-sprint triathlon performance. This information can be used to optimize training and pacing strategies.

Optimal mass scaling of power output for complex courses and its relationship to time-trial performance in cycling

Marton Horvath

The allometric scaling of mean power output is a long-standing debate in cycling research as its pivotal role in performance prediction is commonly acknowledged. In contrast to previously used testing-based approaches, in this study, a numerical method was used to determine optimal mass exponents for scaling of power output. Time-trial performances of typical elite-level road cyclists were estimated utilizing the power-duration relationship and normative data for both simplified hypothetical and complex real-life courses. Optimal mass exponents were determined based on the estimated average speed of typical cyclists over the respective course sections, their morphological characteristics, and external factors such as incline and wind velocity. The magnitude of derived optimally scaled power outputs, namely W/kg^0.6068and W/kg^0.4891, showed a near-perfect negative correlation with the estimated performance of five typical elite-level cyclists on two recent Grand Tour individual time-trial courses. Additionally, complex analyses revealed that optimal scaling of power output relies on the interplay of internal and external factors. The findings of this study indicate that optimal mass exponents are course-specific and characterized by the magnitude of headwind velocity relative to the cyclist. Further in-field research is required to validate the presented method for determining optimal mass exponents.

Calculating Maximal Oxygen Consumption from Outdoor Cycling Training Data

Tzyy-Yuang SHIANG

This study aims to develop a novel Power to Heart Rate (P/H) index to estimate maximal oxygen consumption (VO₂max) for cyclists, a critical indicator of aerobic capacity and endurance performance. Traditional methods for measuring VO₂max can be expensive and time-consuming, which limits their applicability among non-elite athletes. The P/H index leverages readily available heart rate and power output data from common cycling equipment, offering a more accessible alternative. The methodology was applied to a cross-sectional sample of 15 male cyclists, both experienced and non-experienced. The result demonstrated a strong correlation between the indoor P/H index and measured VO₂max, as well as a moderate correlation for the outdoor P/H index. Despite some limitations, the P/H index presents a promising approach to estimate VO₂max, especially for recreational or inexperienced cyclists. Future research should focus on larger, diverse samples and explore the applicability of the P/H index to other endurance activities.