Empirical Cycling Podcast
Watts Doc #63: Confronting Uncertainty In Training And Data
Why It Matters
Understanding uncertainty helps athletes set realistic goals, avoid over‑reliance on one‑size‑fits‑all training prescriptions, and make smarter adjustments based on actual performance data. This episode is timely as more cyclists turn to personalized coaching and data analytics, making the discussion of probabilistic thinking and evidence‑based training highly relevant.
Key Takeaways
- •Training outcomes are probabilistic, not guaranteed algorithmic results.
- •Experience and data reduce uncertainty, improving future performance forecasts.
- •Subjective probability judgments rely on past observations, not exact math.
- •Minimizing surprise aligns training decisions with realistic expectations.
- •Shorter feedback loops accelerate learning and better training adjustments.
Pulse Analysis
In this episode Coley Moore and coach Gita Minas unpack the myth that training follows a simple algorithm. They argue that cyclists and athletes often hear "do this and you’ll get that"—a McScience shortcut that ignores the inherent uncertainty of human performance. By comparing training prescriptions to probability estimates, they highlight how most coaches rely on subjective judgments rather than exact statistical models, acknowledging that outcomes are never guaranteed. This perspective reframes fitness goals as forecasts, not certainties, and sets the stage for a data‑driven conversation about reducing guesswork.
The duo explores how experience and real‑world data act as tools to shrink that uncertainty. Drawing on concepts from neuroscience—specifically the brain’s drive to minimize surprise—they explain how repeated exposure to training blocks builds an intuitive sense of likely outcomes. While true probabilities are hard to calculate without thousands of identical trials, athletes can use past observations to assign reasonable confidence levels, much like a Monte Carlo simulation approximates risk. This subjective probability, grounded in personal history, becomes more accurate over seasons, allowing coaches to predict FTP gains or race placements with increasing precision.
Finally, the hosts translate theory into practice for both professional coaches and self‑coached riders. They advocate for tighter feedback loops: collecting performance metrics, comparing them to expected ranges, and adjusting training plans swiftly. By treating each workout as a forecast and each race as a test of that forecast, athletes can shorten the learning curve, make evidence‑based tweaks, and ultimately achieve more reliable improvements. Emphasizing data‑driven coaching, the episode equips listeners with a mindset that balances optimism with realistic probability, turning uncertainty into a strategic advantage for cycling performance.
Episode Description
We reveal our process for dealing with the uncertainty inherent in training and racing by keeping the FO in the scientific method of FAFO. After addressing whether or not probabilities are even real and the role of subjective experience, we look at some examples of uncertainty and predictions in testing, training, and racing, how reliable and accurate that data and performance models may be, and the process of improving our confidence in their ability to guide a training program. We also discuss incorporating peer-reviewed literature as well as anecdata, and the ultimate value of coaching or self-coaching experience.
Comments
Want to join the conversation?
Loading comments...