HRV (heart rate variability) is a non-invasive maker of cardiac autonomic modulation utilized in many hundreds of scientific studies each year. The reliability of heart rate variability has been frequently investigated yet remains poorly quantified. Assessing the reliability of a measure that assesses dynamic physiological processes and shows large between- and within-subject variation is a complex task. In this issue of Clinical Science, Pinna and co-workers provide excellent insight into the test-retest reliability of commonly used HRV indices and put the values obtained into context by comparing them with levels of between-subject variation and by producing sample size estimates.