Big data reveals secrets of running
Users of wearable exercise trackers upload astonishing number of more than billion exercises each year. This data provides interesting opportunity to study and increase our understanding of exercise adaptation and athletic performance. In a new study published in Nature Communications, Polar joined forces with French physicist, Dr. Thorsten Emig. Together they analyzed data from 14,000 individuals and 1.6 million exercise sessions using Dr. Emig’s model, which determines running performance with two indices: Maximal aerobic speed (MAS) and endurance index that describes decline in sustainable power over duration.
The first aim was to predict marathon performance without any prior information of marathon finishing times. The results were spectacular: the mean error between actual and predicted race time was 2%. The results also highlighted the importance of endurance index. Similar race time can be achieved with low MAS and high endurance index or with high MAS and low endurance index. In addition, endurance index allows determination of key performance indicators often used in exercise physiology, such as anaerobic threshold. Analysis of training history also hint the existence of optimal training.
Findings of the current study pave the way for a new approach where wearable devices are used in combination with mathematical modelling, for exercise and performance testing.
The full text is available here: https://www.nature.com/articles/s41467-020-18737-6
Emig, T., Peltonen, J. Human running performance from real-world big data. Nat Commun 11, 4936 (2020). https://doi.org/10.1038/s41467-020-18737-6