Author(s): PROTZEN, G., INOUE, A., RIESCO-VILLAR, J., DA SILVA, E., GONZÁLEZ-FERNANDÉZ, A., BOULLOSA, D., Institution: UNIVERSITY OF LEÓN, Country: SPAIN, Abstract-ID: 1680

INTRODUCTION:
Critical power (CP) is widely used to characterise endurance performance and guide training prescription in cycling. Although the traditional 3-point CP model is commonly used as a reference method, the short exhaustive trial (~3 min) is highly fatiguing and sensitive to pacing, which may limit feasibility and ecological validity in applied settings. Consequently, practitioners often adopt 2-point CP models; however, CP estimates are sensitive to the number and duration of trials, and simplified approaches may introduce bias [1]. Contemporary mountain biking (MTB) is characterised by frequent high-intensity efforts, which increase the importance of accurately representing the high-intensity end of the power–duration relationship [2]. Therefore, this study examined whether replacing the 3-min trial with a more reproducible 30-s maximal sprint within a three-point CP model yields CP estimates closely aligned with the traditional 3-point approach in trained MTB athletes.
METHODS:
Twenty-five well-trained MTB athletes (Tier 3) completed a field-based CP testing protocol consisting of multiple maximal efforts performed across two testing days. Critical power was estimated using: (i) a traditional 3-point model (3-, 7-, and 12-min trials); (ii) a sprint-based 3-point model combining the 7- and 12-min trials with a 30-s maximal sprint; and (iii) a 2-point model (7- and 12-min trials). Agreement between models was assessed using intraclass correlation coefficients (ICC), linear regression (slope and intercept), and Bland–Altman analysis.
RESULTS:
The sprint-based 3-point model showed very close agreement with the traditional 3-point model (r = 0.96; slope = 0.99; intercept = −3 W), with a small mean bias (+4 W) and narrow limits of agreement (−10 to +19 W). In contrast, the traditional 2-point model demonstrated magnitude-dependent bias relative to the traditional 3-point model (r = 0.95; slope = 0.89; intercept = +36 W), with wider limits of agreement (−26 to +14 W).
CONCLUSION:
Replacing the 3-min trial with a 30-s maximal sprint within a 3-point CP model yields CP estimates that are closely aligned with those obtained from the traditional 3-point approach, and with superior agreement compared to a 2-point model. This sprint-based 3-point model may provide an ecologically valid, less fatiguing, and more practical alternative for MTB athletes, while better reflecting the short-duration high-intensity demands characteristic of MTB competition.
REFERENCES
[1] Maturana et al., 2018; [2] Protzen et al., 2026.