Empirical Cycling Podcast Watts Doc #62: Setting Up Your n=1 Training Experiment
Feb 25, 2026
Kyle Helson, a NASA scientist and national champion sprinter, brings expertise in experimental design and physiology. He discusses how to set up n=1 training tests, control confounders like timing and travel, and use repeated testing to separate noise from real change. Practical rules for dose standardization, tracking perceived effort, and clever within-person replications round out the conversation.
AI Snips
Chapters
Books
Transcript
Episode notes
N=1 Is Anecdote Without Controls
- N=1 experiments are common but often anecdotal and lack the controls of multi-subject science.
- Kyle contrasts casual self-experiments to lab studies that predefine metrics, simulate data, and blind analyses to avoid bias.
Apparent Responders May Be Measurement Noise
- Individual differences in training response can reflect measurement noise, not true physiology.
- Bonafiglia et al. compare sprint vs endurance results and used a parallel repeatability study to define a "no change" zone around measurement variability.
Three Sources That Muddy Individual Results
- Three major noise sources obscure individual response: measurement error, acute/chronic life factors, and true biological adaptation.
- Power meters, metabolic carts, sleep, stress, and cumulative life stress all widen your apparent variability.
