Decoupling of Internal and External Workload During a Marathon: An Analysis of Durability in 82,303 Recreational Runners

Figures

• 

  Fig. 1. Time-course of the decoupling of internal-to-external workload for athletes with low, moderate, and high decoupling. Low, moderate and high decoupling was defined as athletes with a decoupling < 1.1, between 1.1 and 1.2, and > 1.2 in the 35–40 km segments. Decoupling is expressed relative to the 5–10 km segment of the marathon

• 

  Fig. 2. Estimated onset of decoupling during a marathon and decoupling type (low, moderate and high), for male (M) and female (F) runners. The filled circles in the high decoupling indicate a male—female difference (p < 0.01)

• 

  Fig. 3. The onset (distance and time) and the magnitude of the decoupling of internal-to-external workload ratio relative to marathon performance, where marathon performance is calculated: a relative CS, and b in absolute units (min). Estimated onset of the decoupling of internal-to-external workload relative to marathon performance, where marathon performance is calculated: c relative CS, and d in absolute units (min). Filled markers indicate a significant difference between male and female runners (p < 0.01) and a solid line between two makers indicates a statistically significant difference between consecutive pace bins (p < 0.01)

• 

  Fig. 4. Error associated with predictions of marathon performance derived from a CS and D’ only, b CS and D’ plus the magnitude of the decoupling, and c CS and D’ plus the decoupling degree and time to decoupling onset. The error is calculated as the mean absolute difference between the predicted finish-time and the actual finish-time as a fraction of actual finish-time for each finish-time group and the dotted lines show the mean error for male and females for all finish-times. In (a) a filled marker indicates a difference between the corresponding male and female means (p < 0.01), and a solid line between two makers indicates a difference between relative pace segments (p < 0.01). The overall R2 value for each finish-time is also shown

• 

  Fig. 5. Overall performance of different models based exclusively on CS and D′, as well as parameters related to the decoupling of the internal-to-external workload ratio