Use of Critical Speed Models from World Record Data to Estimate Limits of Human Ultra-Endurance Running Performance

Abstract

Ultra-endurance running (>42.2km) has increased in popularity in recent years and elite finish times continue to improve. Models of Critical Speed (CS) - the highest speed which can be maintained without fatigue - are routinely developed from short training distances to predict limits of performance in endurance events such as the marathon. Purpose: To model critical speed (CS) from current world record data to estimate limits of ultra-endurance running performance in males (M) and females (F). Methods: Current world record running performances from distances between 100m and 42.2km were used to develop two models of CS. A linear regression model used slope and y-intercept to estimate CS and finite anaerobic running distance (D’), respectively. Additionally, a curvilinear relationship between average speed and running distance was developed to estimate average speed at a given running distance. Results: For F, CS was determined to be 5.26 m∙s^-1 and D’ was 317.86 m. For M, CS was 5.82 m∙s^-1 and D’ was 345.48 m. This model predicted average speed from current world records could be maintained for 11.5% (M, 50km) to 68.3% (F, 24h run) greater distance. The curvilinear decay relationship between average speed and running distance underestimated average speed at 42.2km for both M and F, yet predicted 1.7% (M, 50km) to 28.2% (F, 24h run) greater average speed for ultra-endurance races. Conclusions: Current world records for ultra-endurance running are likely lesser than the physiological limits of humans and the magnitude of this effect increases with greater race distance.