Prediction of Half-Marathon Power Target using the 9/3-Minute Running Critical Power Test

Abstract

Running power output allows for controlling variables that have been previously overlooked by relying solely on speed, such as surface, gradient and weight. The ability to measure this external load variable now enables the analysis of concepts that have predominantly been studied in cycling, such as the Critical Power (CP), in the context of running. This study aims to predict the CP target at which trained athletes run a half-marathon and determine whether races of this distance can serve as a valid alternative to update the CP record. A group of nine trained athletes performed the 9/3-minute Stryd CP test and participated in a half-marathon race in two separate testing sessions conducted in the field. The average power during a half-marathon race is a valid alternative method for determining the CP in trained athletes, as evidenced by the agreement (95% CI: -0.11 to 0.37 W/kg) and trivial systematic bias (0.13 W/kg) between methods. The linear regression model half-marathon power = 0.97 + 0.75·CP (W/kg) showed low standard error of estimate (0.29 W/kg) and significant large association between methods (r = 0.88; p = 0.002). Coaches and athletes should be aware that the CP target for a half-marathon race is 97.3% of the CP determined by the 9/3-minute Stryd CP test.

Keywords: IMU; Running power; endurance performance; exercise testing; field.

Figure 1.

Difference between the CP and the average power of the HM race.

Figure 2.

Agreement between the power output of the 9/3-minute Stryd CP and the half-marathon: (a) Linear regression analysis. Solid central line represents the fitted equation; upper and lower dashed lines show 95% confidence intervals; dotted line represents line of equality y = x ; (b) Bland Altman plot. Solid central line represents the mean between methods (systematic bias); upper and lower dashed lines show systematic mean ± 1.96 SD (random error); dotted line depicts linear regression (proportional bias).