Wilke et al (2023). Differences between 48h and 72h intervals on matches' load and subsequent recovery