Supercompensation, briefly
It is well known that improving cycling performance requires exercise and recovery. We've written a longer post on supercompensation, which is the underpinning process, but to recap briefly.
The classical model: training stress causes a temporary drop in performance capacity (fatigue), followed by a recovery period during which the body doesn't just return to baseline, but overshoots it and adapts to a slightly higher capacity than before, as a hedge against future stress.
String these cycles together with the right spacing and you get progressive adaptation. Get the timing wrong (too little recovery) and you accumulate fatigue faster than adaptation, which is the mechanistic basis of overtraining.
The many parallel processes of recovery
When thinking about recovery, it's important to think about a parallel set of processes. They don't all run on the same clock, which means it's hard to reduce recovery to a single timeframe. It's helpful to get a feel for what these processes are:
Substrate replenishment. This includes glycogen resynthesis, where muscle and liver glycogen stores, depleted during prolonged endurance work, get rebuilt via glycogen synthase activity. This is faster if carbohydrate intake is adequate (largely complete within 24 hours for moderate depletion), but can take 48 to 72+ hours after genuinely glycogen depleting long sessions, especially if carb intake is suboptimal. See our guide to fuelling cycling performance for more on this. Also relevant here is intramuscular triglyceride and fluid/electrolyte rebalancing, which is less discussed but important for endurance athletes specifically given sweat losses.
Structural repair and remodelling. This covers muscle protein synthesis, which is distinct from the inflammatory cascade but triggered partly by it, and connective tissue remodelling, as tendons and ligaments adapt on a notably slower timescale than muscle (collagen turnover is metabolically slow), which is part of why tendon issues often show up as a lagging indicator of load increases that muscles could otherwise tolerate. It also covers mitochondrial biogenesis, a genuinely endurance specific process. PGC-1α upregulation post session drives new mitochondrial density over 24 to 48 hours; this is arguably the central adaptation endurance training is trying to provoke, running in parallel with everything else.
Neural and hormonal recovery. Central and neuromuscular fatigue recovery is distinct from peripheral (muscular) fatigue. CNS recruitment capacity and neuromuscular junction function can take longer to normalise than muscle soreness would suggest, which is part of why performance can lag behind how you "feel." Sleep plays an outsized role here too. Chris Froome has spoken about this at length; see our piece on sleep and cycling performance.
Immune function. Related to but distinct from local inflammation, the "open window" hypothesis suggests transient immune suppression after prolonged strenuous exercise, which is part of why endurance athletes sometimes report higher URTI incidence after big blocks or events. In other words, hard training can temporarily weaken your immune defences, leaving a window where you're more likely to pick up an infection.
The key practical point to take away is that these processes have different time requirements. Glycogen might be back within a day, whereas connective tissue and full neuromuscular recovery can lag a week behind how recovered you feel systemically. This is why single metric recovery tracking (just HRV, or just soreness, or just resting HR) tends to miss things, and also why a single "recovery day" number is a blunt instrument for something that's really several overlapping curves.
Practical implications for cyclists
We know there is a limit to the amount of endurance training that can be sustained, and while this is higher for cyclists (typically up to about 20 hours a week for Tour de France competitors) than for runners (10 to 14 hours for Olympic competitors), that still seems to be a hard limit. If you're building your own base volume, our Zone 2 training guide is a good place to start.
Perhaps the most pertinent practical question to answer from all of the above is: how can you better understand the signals to read the recovery you need, rather than trying to guess?
Subjective measures (cheap, surprisingly good)
- Session RPE and morning readiness ratings. A simple 1 to 10 self-report, done consistently, correlates surprisingly well with more expensive objective measures in the research. The value is less in any single day's number and more in the trend.
- Sleep quality and duration. Probably the single most load bearing subjective (and increasingly objective, via wearables) signal. Poor sleep blunts glycogen resynthesis, growth hormone release (which is largely sleep dependent), and immune function simultaneously.
- Mood state and questionnaires (e.g. the old POMS based approach, or simpler DALDA style daily logs). Persistent negative shifts, especially "staleness" or irritability, are one of the better validated early markers of non-functional overreaching, and they often show up before performance does.
Autonomic and objective measures
- HRV (heart rate variability), usually taken first thing on waking, reflects parasympathetic reactivation. The key is trend over 7 day rolling averages, not single day readings, which are noisy. A sustained downward trend suggests accumulating systemic fatigue; a single low day is often just noise (alcohol, poor sleep, stress) rather than training specific signal.
- Resting heart rate trends. Cruder than HRV but easier to track long term; a sustained upward drift (a few beats over baseline) is a classic overreaching flag.
- Heart rate variability during exercise and HR power decoupling. For cyclists specifically, tracking whether heart rate drifts upward relative to power output at a fixed effort over a session, or whether your normal power at given HR relationship shifts session to session, is a good "clean legs but tired system" indicator.
Performance based measures
- Power duration curve tracking (cycling) or pace at effort (running). Is your output at a given perceived effort or heart rate tracking normally, or dropping? This is a lagging but very concrete indicator.
- Countermovement jump height or similar neuromuscular tests. Used a lot in team sports, less common for endurance athletes training solo, but a simple jump test can flag neuromuscular fatigue that doesn't show up in HR data.
- Training stress balance models (TSB/CTL-ATL, or "form"). These are essentially a mathematical proxy for the supercompensation curve itself, modelling accumulated fitness versus fatigue. Useful as a planning tool, but they're a model, not a direct physiological reading. They can mislead if your actual recovery capacity that week (poor sleep, life stress) diverges from the model's assumptions.
The practical synthesis
No single metric is reliable alone. The useful approach is a small composite: morning HRV trend, subjective wellness score, and a performance check (does normal effort produce normal output). When multiple signals move in the same direction, HRV trending down, subjective wellness declining, and performance flat or dropping at matched effort, that's a far stronger signal than any one of them individually. Divergence (e.g. HRV is fine but you feel awful) is itself informative, and often points to something outside training load, such as sleep debt, illness incubating, or life stress, rather than training stress per se.
We will continue to research these and other, related topics over the coming months.
