Empirical Cycling Podcast Watts Doc #61: Quantifying Diminishing Returns
Feb 5, 2026
A deep dive into quantifying diminishing returns from nearly 15,000 people and multi-year data. They model log-shaped growth curves, identify typical change-point timing, and show wide individual variation. The conversation compares strength and cycling progress, explores study design pitfalls, and gives practical season-planning and recovery implications without promising quick fixes.
AI Snips
Chapters
Transcript
Episode notes
Long-Term Strength Gains Follow A Log Curve
- Steele et al.'s 14,690-person gym dataset shows long-term strength gains follow a logarithmic growth curve.
- Log-transforming time turns the curved gains into a near-linear trend useful for modeling diminishing returns.
Most People Hit A Change Point Around 6–8 Months
- The study identifies a population 'change point' where week-to-week gains sharply slow.
- Most participants' change points cluster around 26–31 weeks, showing early diminishing returns.
Big Early Effects Then Long-Term Taper
- Early effect sizes are large (e.g., leg press 0.86) and weekly improvement drops quickly after a few weeks.
- Over years, group averages reach ~50–60% improvement, showing rapid early gains then tapering.