Original episode & show notes | Raw transcript
This document provides a detailed, educational breakdown of the core concepts discussed in the Empirical Cycling podcast episode concerning training metrics. The central theme is a critical examination of the Performance Management Chart (PMC) and its components, arguing that while these tools are useful, an over-reliance on them can be detrimental to an athlete’s development and perception of their own fitness.
The PMC is a tool used in training software (like TrainingPeaks and WKO5) to visualize an athlete’s fitness, fatigue, and form over time. It is built upon a single foundational metric: the Training Stress Score (TSS).
The chart typically displays three key data lines:
CTL (Chronic Training Load): Often labeled “Fitness.” This is a rolling, exponentially-weighted average of your daily TSS over a longer period (typically 42 days). It is intended to reflect your accumulated training load and, by extension, your aerobic fitness.
ATL (Acute Training Load): Often labeled “Fatigue.” This is a rolling, exponentially-weighted average of your daily TSS over a shorter period (typically 7 days). It reflects the fatigue you’ve accumulated from very recent training.
TSB (Training Stress Balance): Often labeled “Form” or “Freshness.” This is calculated by subtracting your current ATL from your current CTL (TSB = CTL - ATL).
A negative TSB suggests you are carrying a high level of fatigue (your recent training is greater than your chronic load).
A positive TSB suggests you are fresh and recovered, as your recent training load is lower than what you are chronically adapted to.
The conventional wisdom of the PMC is that peak performance occurs when an athlete has a very high CTL (high fitness) and a positive TSB (good form/freshness).
To critique the PMC, one must first understand its building blocks.
Normalized Power (NP): Unlike average power, Normalized Power is an estimate of the power an athlete could have maintained for the same physiological cost if their power output had been perfectly constant. It is calculated using a complex algorithm that involves:
Calculating a rolling 30-second average of power.
Raising these values to the fourth power.
Averaging these results.
Taking the fourth root of that average.
This process gives significantly more weight to high-intensity spikes in power. A ride with many sprints and surges will have a much higher NP than average power, which is meant to better reflect the true metabolic strain of the effort.
Intensity Factor (IF): This is a simple ratio of your Normalized Power to your Functional Threshold Power (FTP). An IF of 1.0 represents an hour-long effort at your absolute maximum steady-state pace.
Training Stress Score (TSS): TSS quantifies the training load of a single workout. It is calculated based on the duration and intensity (IF) of the session. A one-hour ride at FTP (IF = 1.0) is the benchmark for 100 TSS.
The podcast’s core argument is a critique of the over-reliance on these metrics, a behavior they compare to the Fight Club mantra, “You are not your khakis.” Athletes can become fixated on chasing a higher CTL number, believing it is a direct and infallible measure of their “fitness.”
The Common Anecdote: An athlete spends months chasing a high CTL, constantly feeling fatigued and seeing no improvement in their FTP. Frustrated, they take a break, ride for fun, and allow their CTL to drop. When they next compete, they feel fantastic and perform exceptionally well, leading to the conclusion: “There’s something wrong with this chart.”
This reveals two modes of using the PMC:
The Flawed Approach: Letting the chart dictate how you should feel. (e.g., “My TSB is +10, so I must be fresh. I will ignore my tired legs and do a hard workout.”)
The Correct Approach: Using the chart as a “second opinion” to understand your subjective feelings. (e.g., “My legs feel heavy today. I’ll check my PMC… ah, my ATL has spiked by 20 points this week. That provides context for why I feel this way.”)
The podcast highlights several critical flaws in the assumptions that underpin the PMC model.
The TSS model assumes that training stress is interchangeable. For example, it treats the following as roughly equivalent because they can both generate ~100 TSS:
A grueling one-hour workout of 4x5 minute VO2 max intervals.
A steady two-to-three-hour Zone 2 endurance ride.
From a physiological standpoint, these are not equivalent.
Fatigue: The VO2 max session induces significant neuromuscular and central nervous system fatigue. A proper endurance ride should leave an athlete feeling pleasantly tired, not exhausted.
Adaptation: While both contribute to aerobic fitness, they stimulate different pathways and adaptations. The muscular and metabolic stress is fundamentally different.
This false equivalence leads athletes to believe they can maintain their “fitness” (CTL) during a race season by swapping volume for intensity, which is not a physiologically sound one-to-one trade.
Labeling CTL as “fitness” is a significant marketing and conceptual error. It encourages the simplistic belief that “bigger is better.” The podcast suggests a more accurate term would be “work capacity” or “accumulated training load.”
An athlete can become significantly stronger and faster while maintaining the same CTL. This is because CTL is relative to your FTP.
The PMC is a closed system; it only accounts for stress measured via a power meter. It is completely blind to other significant physiological stressors:
Poor sleep
Work or relationship stress
Illness
Nutritional deficits
This is where subjective feelings (“How do my legs feel?”) and other tools like Heart Rate Variability (HRV) attempt to fill the gap, though they also come with their own limitations (e.g., questionable accuracy of consumer-grade HRV sensors).
The podcast offers a heuristic to quickly estimate someone’s CTL: CTL ≈ (Average Weekly Riding Hours) x 0.7
This works because most structured training weeks average out to an Intensity Factor (IF) of around 0.70-0.75. This trick reveals a crucial point: for many athletes, CTL becomes little more than a proxy for their weekly training volume. It doesn’t necessarily reflect the quality or effectiveness of that training.
A common question is whether to assign a TSS value to activities like strength training. The podcast provides a clear guideline:
Only track activities that provide a direct aerobic stimulus.
Strength Training: No. While it causes fatigue, it is not an aerobic workout. Including its “TSS” in your PMC will artificially inflate your CTL, making year-over-year analysis unreliable. You may have a higher CTL but be a slower cyclist because you’re not recovering properly.
Multi-Sport Athletes: Yes, but use separate PMCs for each discipline (e.g., swimming, cycling, running) to analyze the specific stresses of each sport.
Alternative Tracking Method: For non-aerobic activities, use a simple formula of RPE (Rate of Perceived Exertion, scale of 1-10) x Duration (in hours). This creates a separate, useful “training load” metric that doesn’t corrupt your cycling-specific data.
The PMC comes with “rules,” such as a recommended ramp rate (the rate at which CTL increases) of no more than +5 to +7 per week. The podcast argues that these are merely starting points. An athlete’s ability to handle an increased training load depends entirely on their training history, genetics, and recovery capacity. Elite athletes like Mathieu van der Poel can handle ramp rates that would destroy an amateur, demonstrating that individual context trumps generic rules.
The overarching message is that the Performance Management Chart is a valuable tool for providing a second opinion and for post-hoc analysis. It can help an athlete and coach understand the relationship between training load, fatigue, and performance.
However, it is not a predictive oracle. Its flaws—the false equivalence of stress, its blindness to outside factors, and the misleading terminology—mean that it should never override an athlete’s subjective feelings and intuition. The most important questions remain: “How do my legs feel?” and “Am I getting stronger?” The PMC can help provide context to those answers, but it cannot provide the answers themselves.