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

Abstract

The application of the critical power (CP) model to intermittent exercise allows to gain valuable insights into the dynamic balance between W' depletion and W' recovery, thereby offering a potential tool for designing and optimizing pacing strategies in cycling. This study aimed to assess whether a pacing plan based on the PACE model - a novel model for real-time monitoring of W' during exercise - could improve time trial (TT) performance on the Zwift indoor cycling platform. Additionally, we investigated the impact of prior heavy-intensity exercise on TT performance. Twelve recreationally trained cyclists (29 ± 7 years; 75.2 ± 5.2 kg; 180 ± 3 cm; 54.1 ± 5.2 mL·min^-1·kg^-1) completed 9 to 11 laboratory testing sessions on separate days. Following an initial maximal ramp test, critical power (CP) and W' were determined through 3 to 5 constant power output (PO) tests to exhaustion. Participants then completed four 23.5 km TTs on Zwift (The Muckle Yin): two TTs with self-selected pacing (TT₁ + TT₂), one TT following a 120-minute constant work rate effort with 60 g/h carbohydrate intake (TT_CWR), and one TT with model-imposed pacing (TT_PACE). Differences in finish time and PO between the TTs were analysed using Repeated Measures ANOVA. IntraClass Correlation (ICC) was calculated between TT₁ and TT₂ to evaluate day-to-day variation. Critical power and W' were 281 ± 14 W and 22.8 ± 3.8 kJ, respectively. No significant differences in finish time were observed between TT₁ and TT₂ (p = 1.000, ICC = 0.974). Finish time was 77 ± 57 s faster in TT_PACE (2315 ± 80 s) compared to TT₁ (2359 ± 91 s, p < 0.001), TT₂ (2361 ± 102 s, p = 0.031), and TT_CWR (2458 ± 146 s, p = 0.002). Average PO was significantly higher in TT_PACE (278 ± 15 W) compared to all other conditions (TT₁ = 265 ± 18 W, p < 0.001; TT₂ = 264 ± 22 W, p = 0.029, TT_CWR = 243 ± 27 W, p < 0.001). In conclusion, the present study results demonstrate the effectiveness of a novel predictive W' model to improve TT performance in recreational cyclists, by optimizing W' expenditure and recovery.