← Thinking

Does work stress affect training recovery? How one budget is shared

16 March 2026 · Tom Wuerden

Does work stress affect training recovery? How one budget is shared

Applied Physiology #1. The meeting and the interval are paid for out of one reserve, and reading the balance in it is a skill you can pick up inside a week. Part two takes up the withdrawal that repays the most, which is strength work after 40.

Anyone who trains hard while holding down a demanding job has had the session that fell apart for no reason they could name. The plan said threshold. The legs said something else, the numbers were down, and the nearest explanation was that you had gone soft. I reached for that one for years. It is cheap and it is usually wrong.

The physiological account is duller and considerably more useful. The system regulating your heart rate, your digestion and your overnight repair draws on a single pool of capacity, and it never asks where a demand came from before it starts paying out. A difficult call and a set of intervals get charged to the same place. Once you can see the balance in that account, the internal arguments about willpower mostly stop being worth having, because they were aimed at the wrong thing.

This is an educational and strategic perspective, not personal medical advice.

The views are the author's own and not statements by Atlas Cove Lda.


Two branches, and what each of them is doing

The autonomic nervous system is the part of your nervous system that operates without you. Heart rate, blood pressure, the movement of the gut, the width of your airways, sweating, the immune traffic that steps up while you sleep. None of it is a decision, which is rather the point, because nobody could run this much accounting consciously.

It has two branches, built to pull against one another. Output is the job of the sympathetic side. Your heart rate climbs, each contraction gets stronger, blood is redirected toward the muscle about to work and away from the gut, airways open, and adrenaline and noradrenaline go out to hold the arrangement together. That is a hard interval described from the inside. It is also a call you did not want, at a smaller amplitude and often for very much longer.

The parasympathetic branch handles the other half of the job, and it is the half almost nobody talks about. Heart rate comes down, largely through the vagus nerve. Digestion gets its blood supply back, damaged tissue is repaired, glycogen goes back into muscle. Nothing you gained from a session was gained during the session, because the adaptation happens afterwards, on the restoring side, and only when the restoring side is given a turn.


One pathway, two kinds of demand

Here is the fact the rest of this rests on. The stress response has a hormonal arm as well as a neural one, and that arm is the hypothalamic-pituitary-adrenal axis, usually shortened to HPA axis. The name is only the route. The hypothalamus signals the pituitary, the pituitary signals the adrenal glands, the adrenals release cortisol, and cortisol reports back to the brain so the sequence shuts itself down once the demand passes. A loop with its own brake, which works nicely when demands arrive one at a time.

What the pathway does not have is a separate channel for physical work and another for cognitive work. Threshold intervals, a negotiation going badly, a night cut short, a low-grade infection: all of it runs through the same machinery and is governed by the same feedback (Chrousos, 2009). The system registers size and duration, never source. That is not a metaphor about stress being stress. It is a statement about which structures are involved, and everything below follows from it.

Which is where the intuitive model fails most people I know. They manage training load with real care, hours logged, the easy-to-hard split watched closely, and then treat the remaining hundred and ten hours of the week as though nothing in them costs a thing. The deadline is a withdrawal in precisely the sense the intervals were. So is the fourth short night in a row.


Allostatic load, and how it differs from being tired

Short bursts of any of this are fine, and it needs saying early, because a whole industry now teaches people to be frightened of their own cortisol. A sympathetic surge in the face of hard effort is the job this machinery was built for. Sleep properly afterwards and it returns to baseline, no bill attached. Load is not the problem. Load without the return leg is.

The concept that puts a name to that imbalance is allostatic load, which McEwen developed while working on how stress wears on a body over time (McEwen, 1998). Allostasis is the business of holding a body stable by changing things: raising output, shifting fuel, adjusting immune activity. Allostatic load is the wear those regulating systems take on when they are asked to do it again and again with too little restoration in between. Wear is the correct word. Nothing breaks, and there is no moment at which something visibly goes wrong.

Be clear about what kind of paper that is. McEwen's is a framework: it names a concept, gathers what was known at the time and gives a field a shared way of talking about cumulative cost. It is not a trial and it sets no threshold, so anybody quoting you a number of weeks at which allostatic load arrives is inventing it. What the framework earns is the shape of the thing, that repetition without recovery costs something and the cost accumulates quietly.

Undramatic is the whole character of it. In an ordinary week it reads as a morning pulse sitting a little above where yours usually sits, sleep that stops returning you to where it once did, body composition that will not shift on an unchanged routine, attention thinning out mid-afternoon. That is a body choosing to spend less because less is what it currently has, which is a reasonable decision for a body to make. It is not a judgement on your character and it is not a diagnosis of anything. If it worries you or it persists, that is a conversation to have with a doctor rather than with an article.

It is also why this goes unnoticed for so long. Blood work frequently comes back unremarkable while every one of those things is happening, because a normal panel says very little about how well a system is regulating itself, and regulation is the thing under strain. There was a period when my own account ran well into deficit, and I have described that elsewhere, so a single line does the job here.


Recovery is trainable, and almost nobody trains it

Now the encouraging half. Your capacity to restore is not a fixed quantity issued to you at birth. It answers to a stimulus in roughly the way your aerobic engine answers to one, so it can be built on purpose rather than waited for.

Better cardiac parasympathetic regulation goes along with easy aerobic training (Buchheit and Gindre, 2006). Read the design carefully, because it decides what you may take away: that work compares fitness and training load against measures of parasympathetic activity in people who already train. It tells you the slow hours travel alongside a stronger restoring branch. It cannot tell you how many hours, at what intensity, or how long any of it takes. Direction is what I take from it, and the direction is convenient, since most endurance athletes already do those hours for another reason entirely.

Slow breathing is the other lever, and it costs nothing at all. At about six breaths a minute, which works out as a five-second inhale and a five-second exhale, your breathing falls into step with the natural rhythm at which blood pressure oscillates, and the swing in heart rate across each breath gets much larger. Vagal traffic to the heart rises with it. A review of the work in healthy adults found the shift toward the parasympathetic side happening within minutes, and reported that the effect appears to compound when practice is regular (Russo, Santarelli and O'Rourke, 2017). Small studies, healthy volunteers, a review rather than a trial. I would stand behind the direction and treat any dose as guesswork.

Deep sleep is where parasympathetic traffic runs highest, which is the entire reason protecting your bedtime and your evening light returns more than nearly anything you might swallow instead. I know of no trial that measures recovery capacity itself and follows people for a year. I would like one to exist.

The part I find slightly embarrassing is how long it took me to notice the obvious. The bulk of my training week was easy aerobic work, put there purely for the engine, or so I assumed. Those hours turned out to be the one thing holding the restoring branch up, and I had arrived there by accident. Training the restoring branch on purpose is among the very few moves that make the budget larger instead of rationing whatever is left in it.


Which numbers tell you what you can carry today

There is a distinction worth holding on to between measuring what your body has produced over months and measuring what it can carry right now. Most people own only the first kind of number and then try to make daily decisions with it.

Outcome measures describe where you have been. Cholesterol, VO2 max, body composition, whatever the last panel said. All worth having, and I have mine done. They move slowly by design, which makes them good for judging a training block and close to useless for judging a Tuesday. VO2 max is the clearest example of the type, and I have argued elsewhere that it settles far less about a long race than its reputation implies.

State measures get much closer to the question you are actually asking. Your morning heart rate against your own baseline. The trend in your heart rate variability, meaning the beat-to-beat variation in the gaps between heartbeats. An honest reckoning of how you have slept lately. How heavy the opening ten minutes of easy work feel under you.

HRV deserves its own paragraph, because it gets explained badly nearly everywhere. A healthy heart does not tick like a metronome. The intervals between beats vary, and much of that variation comes from the vagus nerve adjusting the heart from one beat to the next, so more variation generally indicates more parasympathetic influence. Thayer and colleagues pooled neuroimaging and HRV studies and found the measure tracking activity in the brain regions that regulate autonomic output (Thayer et al., 2012). Notice how much narrower that is than the marketing on the box. It supports HRV as an index of central regulation, which is why it tracks stress and health more closely than fitness. It is not a stress meter, and there is no threshold above which you are permitted to train.

The version of all this you can run this week is small. Take four readings at once, before a session you have already committed to, and read them as a set:

  1. Morning heart rate, compared against your own baseline and nobody else's.

  2. Whatever recovery trend your watch reports, read across seven days rather than as this morning's verdict.

  3. An honest account of the past two nights, including the one you would prefer to round up.

  4. How heavy the opening ten minutes of easy work feel. No device required, and this is the reading I have come to trust above the other three.

A single one of them heading downward is noise. Once three or four point the same way, you have a week that will not lose an argument with you, and the sensible response is to put something back rather than draw more out, whatever your plan happens to say. When all four read fine, spend the money. That is the lens in its entirety.

Precision is not on offer here. What one morning's figure buys you is close to nothing, and two devices on the same wrist will contradict each other. Whatever signal exists lives in the direction of travel over a week. Read where you are, then choose.


My training spreadsheet had one column in it

For years I ran something close to that lens, and I ran it well, on precisely one half of the ledger. Each week of my last build was written out properly, the hours and the balance between easy and hard, and I knew in advance which sessions would come off if a week went sideways.

Then consider everything else those months contained. My sister and I were starting a company, in a country whose administrative machinery I was still working out, with money decisions turning up most weeks. Every so often a problem woke me at four and had to be solved again properly, awake, at nine.

None of that was recorded anywhere. Training hours had a column and nothing else did, so everything outside that column was treated as free. At work, a model that quietly ignored an entire class of input would have earned a short comment from me.

Beyond the mechanism there is human evidence, and I would like to be exact about what sort. Reviewing the literature on stress and exercise, Stults-Kolehmainen and Sinha (2014) found that people carrying more psychological stress recover less well, adapt less to the same training and get hurt more often. Found together is the whole of what a design like that delivers. The work pooled there is observational, and no ethics board is ever going to randomise volunteers into a difficult year, which is the difficulty in one sentence. People carrying a heavy year sleep badly and eat differently too, which means no observational design can pull the strain cleanly out of that bundle. What survives is narrower and still worth having. A training week that looks entirely reasonable written down can push your total past what you can absorb, because it was written in ignorance of everything else you were carrying.


Where the model stops being true

An account you can read a balance from is a model, and models have edges past which they describe nothing. Nothing inside you is adding up your week. No organ computes a figure, and my liking for tidy arithmetic says more about the engineer in me than it does about physiology. The pooling is the part the model gets right, one reserve with a queue of claims against it. Precision is the part it gets wrong. Start putting a point score on your Tuesdays and you have swapped one comforting fiction for another wearing a better suit.

A hard week is not damage, either. Bodies are built for load and get better at carrying it by carrying it, and the requirement is only that the second half of the cycle actually takes place. That half is the first thing negotiated away, every time.

There is a larger claim I hold and cannot support at the strength I would like. My honest suspicion is that a meaningful share of what gets handled as purely a problem in the head is sitting on top of a body that has been under-recovered for years. That is a belief and I am labelling it as one. What the literature offers me is associations, and I would rather hand you the claim with its label still attached.


Sources

  1. McEwen, B. S. (1998). Protective and damaging effects of stress mediators. New England Journal of Medicine, 338(3), 171-179. DOI: 10.1056/NEJM199801153380307

  2. Chrousos, G. P. (2009). Stress and disorders of the stress system. Nature Reviews Endocrinology, 5(7), 374-381. DOI: 10.1038/nrendo.2009.106

  3. Stults-Kolehmainen, M. A., and Sinha, R. (2014). The effects of stress on physical activity and exercise. Sports Medicine, 44(1), 81-121. DOI: 10.1007/s40279-013-0090-5

  4. Buchheit, M., and Gindre, C. (2006). Cardiac parasympathetic regulation: respective associations with cardiorespiratory fitness and training load. American Journal of Physiology-Heart and Circulatory Physiology, 291(1), H451-H458. DOI: 10.1152/ajpheart.00008.2006

  5. Russo, M. A., Santarelli, D. M., and O'Rourke, D. (2017). The physiological effects of slow breathing in the healthy human. Breathe, 13(4), 298-309. DOI: 10.1183/20734735.009817

  6. Thayer, J. F., Åhs, F., Fredrikson, M., Sollers, J. J., and Wager, T. D. (2012). A meta-analysis of heart rate variability and neuroimaging studies: implications for heart rate variability as a marker of stress and health. Neuroscience and Biobehavioral Reviews, 36(2), 747-756. DOI: 10.1016/j.neubiorev.2011.11.009


Questions people ask about this

Does stress at work really affect training recovery?

Mechanistically yes, and the reason is structural. The hormonal arm of the stress response runs through one pathway, the HPA axis, and that pathway has no separate setting for a hard week at a desk and a hard session on the road (Chrousos, 2009). Both draw down the same reserve.

The human evidence is weaker and points the same way. More psychological stress goes with worse recovery, less adaptation and more injury at comparable loads (Stults-Kolehmainen and Sinha, 2014). Observational work, so it is a strong hint rather than a proven cause.

How do I tell ordinary tiredness from a system that is overdrawn?

Ordinary tiredness clears. A decent night or two and you are back where you were. The overdrawn version survives a normal weekend, and it tends to arrive as a cluster rather than as one symptom: a raised resting pulse, sleep that no longer restores, afternoons where concentration costs visibly more.

Read it as information rather than as a verdict. If it persists, or if anything about it worries you, take it to a doctor. An article cannot examine you.

Should I skip a session when my HRV reading is low?

Not on one reading, no. A single morning figure carries very little, and the devices are noisy enough to contradict themselves inside a week. What carries information is the direction over seven days and whether anything else agrees with it.

Use the four readings above together. One of them pointing down is noise. Three or four pointing the same way is worth acting on, and the action is to put something back rather than take more out.

How long does it take to rebuild the recovery side?

Nobody can give you an honest number, and I would be suspicious of anyone who tries. The easy aerobic evidence is an association in people who already train (Buchheit and Gindre, 2006), the breathing evidence is a review of small studies in healthy adults (Russo, Santarelli and O'Rourke, 2017), and neither design was built to establish a timeline.

What the two support is the direction. Keep easy hours easy, protect sleep with the seriousness you would give a session, and practise slow breathing often enough that it stops being an event. The first things I would expect to move are your morning readings and how heavy easy work feels.


Where Atlas Cove fits

State before load is the order the Atlas Cove week is built around. What a guest can currently carry is established before anyone decides what to ask of them, and read again at the end so the direction of travel is visible rather than assumed. If that order makes sense to you, the method behind the week is written out in full.

Tom Wuerden

Tom Wuerden · Co-Founder

Engineer turned Ironman

we don’t take everyone.

One cohort a month, ten rooms behind one gate. Applying commits you to nothing but a conversation.

Next cohort: Soft Opening · 1–6 September 2026 · from €1,950 per guest, all-in.

Apply

every guest completes a health screening questionnaire before the work begins.

prefer to talk to a person first? ask us anything on WhatsApp →

next cohort: Soft Opening · 1–6 September 2026from €1,950 all-in · ten rooms, one cohort a month
Apply