Short answer
Because steps only measure one kind of effort, the one you did with your legs — and crashes have many other causes that a step counter never sees. Thinking, concentrating, talking, being upright, processing noise and light, emotional strain, poor sleep and fighting off an infection all spend energy without moving the step count. And a crash can arrive a day or more after what caused it, so yesterday's number is often the wrong day to look at.
Our own data makes the point sharply. Across 433,313 tracked days from 5,061 people, a poor-recovery morning happened on 15% of days overall. After a day with fewer than half a person's usual steps, it still happened 13.6% of the time — barely lower than after a normal day (14.4%). After a very active day it rose to 19.2%. Physical activity matters, but it explains a surprisingly small slice of bad mornings.
And if you crashed after a day you barely moved, you're not imagining it, and you didn't cause it by being careless. Post-exertional malaise is well documented as being triggered by more than physical activity. In a survey of 150 people with ME/CFS, 90% reported crashes after physical exertion, cognitive exertion and emotional distress — not physical exertion alone. (doi.org) This is also where the spoon metaphor stops being enough: it assumes you know your budget in the morning. A zero-step day spent on a difficult phone call, a medical appointment, a hard conversation, or simply sitting upright at a screen is not a rest day for your nervous system.
The step counter isn't lying. It's answering a narrower question than the one you're asking. (If you're new to the idea of crashes after exertion, start with our full guide to post-exertional malaise; this article is about why step counts mislead pacing, with our own data.)
Quiet days don't buy good mornings — and big days echo for three
We looked at every tracked day for 5,061 Welltory users and asked: how often is the next morning a poor-recovery morning — morning battery at least one standard deviation below that person's own normal — depending on how active the previous days were?
The day before:
Very active (more than 1.5× their usual) — 19.2%
Normal — 14.4%
Quiet (less than half their usual) — 13.6%
A very active day raises the risk of a bad morning by about 5 percentage points. A quiet day lowers it by less than 1. Rest barely moves the needle.
Two and three days later, the effect of a big day fades but doesn't vanish:
A busy day still leaves a 2-point mark three days on. That's the pattern people with PEM describe: effort today, cost later.
Stacking matters most. After three very active days in a row (each more than 1.3× usual), a poor morning came 21.3% of the time. After three quiet days in a row, 14.1% — essentially the same as the 15.0% baseline. Three days of rest didn't make a good morning any more likely than usual.
And it's the same for everyone. We checked people who report ME/CFS, long COVID, POTS, and no conditions at all. The pattern barely changed: after a quiet day, poor mornings came 13.5% of the time for ME/CFS, 12.1% for long COVID, 12.9% for POTS, and 14.0% for people with no conditions.
Association, not causation. Morning battery is Welltory's composite recovery measure, not a clinical measure of post-exertional malaise. Observational data from people who chose to track; conditions are self-reported.
What does a step counter miss?
A step counter records one thing: how many times your wrist or phone detected a walking stride. For pacing with a chronic illness, that's a narrow slice of what actually spends energy.
In a survey of 150 people with ME/CFS, 90% said crashes followed physical exertion, cognitive exertion and emotional distress — not physical effort alone. (doi.org) Among people seeking care for long COVID, low and medium levels of both physical and cognitive exertion were enough to trigger a crash. (doi.org) None of that non-physical effort appears in a step count.
What a step counter can't see:
Sitting or standing upright — at a desk, in a queue, in the shower. For anyone with orthostatic intolerance, upright time can cost more than walking.
Mental work — screens, forms, reading, decisions, anything new.
Social and emotional load — conversations, conflict, bad news, keeping it together in front of other people.
Sensory load — noise, bright light, crowds.
Arm and upper-body effort — cooking, carrying, washing hair, typing for hours. Wrist devices may count some of it as steps, phones in a pocket won't count it at all.
What your body is doing internally — fighting an infection, recovering from a poor night, coping with heat.
So a "zero-step day" can be a day of hard work for your nervous system. The tracker isn't wrong. It's just measuring something much smaller than what you're trying to manage. For a full explanation of what counts as exertion in PEM and why, see our full guide to post-exertional malaise.
Why is yesterday's step count the wrong day to look at?
Because the cost of effort often arrives later — and our data shows exactly how much later.
Crashes in ME/CFS and long COVID are frequently delayed: in the 150-person survey, 11% of people had a consistent delay of at least 24 hours before a crash began, and in focus groups most people described symptoms starting within 24 hours and peaking within 72 hours of a formal exercise test. (doi.org) (doi.org)
Our lag table puts numbers on what that means for a tracker. A very active day raises the chance of a poor-recovery morning to 19.2% the next day — and it's still at 16.9% two days later and 16.7% three days later, against about 14.7% after an ordinary day. The effect fades, but it's still visible three mornings on.
In practice: if you crashed today and yesterday's step count was low, don't conclude the crash came from nowhere. Look back three days. And look at more than steps — the day that cost you may have been a low-step day full of other effort.
Why is a step count so misleading for pacing?
It measures one input out of many. As above, cognitive, emotional, orthostatic and sensory effort don't register.
It measures on the wrong day. Delayed onset means today's crash may reflect effort from two or three days ago.
It rewards the wrong behaviour. A low step count feels like safe pacing, even on a day packed with other kinds of effort. Many people describe a pattern of "I stayed in bed all day and still crashed" — often because the day in bed involved work, messages, stress or being propped upright.
It hides stacking. Our data shows the biggest jump comes from several busy days in a row, not a single one. A step count read day by day misses the accumulation.
It can't tell you about illness. An infection, hormonal changes, poor sleep or heat can all lower your threshold without any change in activity.
None of this means steps are useless. A sudden spike in steps is a genuine warning, and a very active day did raise the risk of a bad morning in our data. It means steps are one line on a dashboard, not the whole dashboard.
What does a "zero-step" day actually cost?
Take two days that look very different on a step counter.
Day one: 4,200 steps. A slow walk to the corner and back in the morning, light pottering at home, a nap after lunch, an easy evening with a familiar programme on. Nothing pressured, nothing new, lying down whenever you wanted to.
Day two: 380 steps. A video call with a new specialist that ran over, spent explaining your history from the beginning. Two hours of forms and emails afterwards. A tense phone conversation with a family member. Most of the day spent sitting upright at a desk under bright light because the paperwork had a deadline. Too wired to sleep well that night.
The step counter rates day two as the restful one by a factor of ten. For many people with post-exertional malaise, day two is the one that brings the crash — possibly not tomorrow, but the day after.
That isn't a hypothetical category of person. People with ME/CFS describe cognitive exertion triggering PEM, usually within hours rather than minutes, and people with long COVID report that even low and medium levels of cognitive exertion can do it. (doi.org) (doi.org)
Which tracker metrics actually help with pacing?
If steps are the weakest pacing signal, what's better? Here's how the common tracker metrics compare for people pacing with ME/CFS, long COVID, POTS or fibromyalgia.
| Metric | What it catches | What it misses | Use it for |
|---|---|---|---|
| Steps | Walking | Upright time, mental and emotional load, arm work, illness | Spotting unusually big walking days |
| Heart rate during the day | Walking, being upright, stress, heat, illness | Some mental effort; wrist sensors can lag | Live alerts when you cross a personal threshold |
| Resting heart rate (overnight) | Infection, poor recovery, overexertion building up | Short spikes during the day | A morning check: is today's baseline raised? |
| Heart rate variability | Stress and recovery state | Noisy from night to night | Weekly trends, not single mornings |
| Sleep | Duration and disruption | How restorative sleep felt | Context: a bad night lowers your threshold |
| Your own daily ratings | Mental, emotional and sensory load | Nothing a device measures | The part no tracker can see |
The strongest set-up most people land on combines a device and a notebook: heart rate for live warnings, overnight resting heart rate as a morning check, and a quick evening rating of mental and emotional load. Steps stay in the picture, but as one line, not the verdict.
A note on pacing apps: several apps are built specifically for pacing with chronic illness, usually around heart rate thresholds and daily energy budgets. Whatever you use, check that it lets you log non-physical effort — an app that only counts movement has the same blind spot as a step counter.
What does the push-crash cycle look like in numbers?
The push-crash cycle — sometimes called boom and bust — is the pattern of doing more on a good day, crashing, resting, feeling better, and doing more again. It's one of the most common patterns in energy-limiting conditions, and our data shows why it's so hard to break.
The push costs more than the rest earns back. After a single very active day, the risk of a poor-recovery morning rose by about 5 points. After a single quiet day, it fell by less than 1.
Pushes stack. After three very active days in a row, a poor morning came 21.3% of the time — the highest risk in our data.
Rest doesn't bank credit. After three quiet days in a row, a poor morning came 14.1% of the time — about the same as the 15.0% baseline. Resting didn't make the next morning more likely to be good; it just stopped making it worse.
So the practical lesson isn't "rest more". It's avoid stacking. One demanding day followed by a lighter one is very different from three demanding days in a row, even if the total effort is similar. Spreading effort out is what breaks the cycle — and because effort includes much more than steps, spreading it means planning the calls, appointments and screen-heavy work too, not just the walks.
How do you pace when steps can't be trusted?
Pacing — keeping total effort within the amount your body can handle without a crash — is the approach most people with PEM rely on, and in surveys it is one of the most common strategies people report for preventing crashes, alongside avoiding specific triggers. (doi.org) Rest is the most commonly reported way to recover once one has started. (doi.org) The difficulty is measuring "total effort" when the obvious metric misses most of it.
A few approaches that work better than steps:
Count all effort, not just movement. At the end of each day, give physical, cognitive, emotional and upright effort a quick score from 0 to 3. The total is a rough but honest measure of the day's cost.
Budget across days, not within one. Because crashes are delayed and stacking matters, think in three-day blocks. A demanding day should be followed by a deliberately lighter one — and ideally not by another demanding one. Our data shows three busy days in a row carried the highest risk of all.
Plan recovery before the demanding day, not after. If you know an appointment or an event is coming, lighten the days either side in advance rather than waiting to see how you feel.
Break up cognitive work. Short blocks with lying-down breaks between them are usually easier than one long stretch, in the same way that splitting a walk into pieces is easier than doing it all at once.
Reduce upright time where you can. Working reclined, sitting to shower, and lying down between tasks all cut orthostatic load without cutting what you get done.
Use heart rate as a live signal. Because heart rate rises with being upright, stress and illness as well as walking, a wearable alert when you cross a personal threshold can catch effort a step counter misses.
What to track instead of — or alongside — steps
Upright time. How many hours you spent sitting up or standing, not just walking. For people with orthostatic intolerance this may matter more than steps.
Cognitive load. A simple 0–3 rating each evening: none, light, moderate, heavy mental effort.
Emotional load. The same 0–3 rating for stress, conflict, difficult news or social demand.
Sensory load. Noise, light, crowds — a quick yes or no.
Heart rate during the day. Many people with ME/CFS and long COVID pace with heart rate because it responds to upright posture, stress and illness, not just to walking. Time spent above a personal threshold captures more types of effort than steps do.
Sleep and resting heart rate. A raised resting heart rate or a bad night is a signal your threshold may be lower before you've done anything.
A three-day window. When you crash, look back across the previous three days, not just yesterday. Our data shows the effect of a big day is still visible three days later.
Within two or three weeks of this kind of logging, most people start to see their own pattern — and it often has very little to do with step counts.
What does the evidence not show?
That our morning measure equals PEM. Morning battery is a composite recovery estimate. It will capture some crashes and miss others, and it will flag some bad mornings that have nothing to do with exertion. We are not measuring post-exertional malaise directly.
That activity doesn't matter. It does. Very active days raised the risk of a bad morning by about 5 percentage points, and three busy days in a row raised it by about 6. What our data shows is that low activity doesn't protect you nearly as much as people expect.
Why the pattern was the same in ME/CFS as in everyone else. People with ME/CFS may already pace their activity carefully, so their "very active" days may be much smaller in absolute terms than other people's. Our measure is relative to each person's own normal, which is what makes the comparison fair — but it also means we can't say how much absolute activity triggers a crash.
When should you see a doctor?
See a clinician if crashes are frequent, severe or getting worse, if you have never had an assessment for ME/CFS or long COVID and suspect PEM, if orthostatic symptoms are part of the picture, or if exertion intolerance is stopping you working, studying or looking after yourself. Also see someone about new symptoms that don't fit your usual pattern.
Seek urgent care for chest pain, fainting, severe breathlessness, new weakness or numbness, confusion, or high fever.
How to bring this up with your doctor — and what to ask for
Bring a log that includes non-physical effort. Two weeks with upright time, cognitive and emotional load ratings, sleep, and crashes marked. That shows a pattern steps alone never could.
Use the term. "I get worse after exertion, and it's often delayed by a day or two. The exertion isn't only physical — mental and emotional effort trigger it too. Could this be post-exertional malaise?" Naming PEM precisely helps it get assessed properly.
Ask these specifically. Could this be ME/CFS or long COVID? Should we assess for orthostatic intolerance with a stand test? What pacing approach do you recommend? Is there a heart rate threshold I should use? Should I avoid graded exercise programmes that increase activity on a fixed schedule?
Why this conversation matters. In a Norwegian survey of specialist care for ME/CFS, failing to address PEM roughly doubled the risk of health deterioration after rehabilitation — 63.2% deteriorated when PEM wasn't addressed against 40.1% when it was. (doi.org)
If you are dismissed. "Can you note in my record that I reported delayed worsening after physical, mental and emotional exertion?" A documented symptom is much harder to overlook at the next appointment.
How Welltory helps
Welltory measures steps, but it also measures heart rate, heart rate variability, sleep and daily stress load — signals that respond to upright time, emotional strain, poor sleep and illness, not just walking. Seen together, against your own baseline, they give you a much fuller picture of what a "zero-step" day actually cost.
The most useful habit the data supports is the three-day look-back. When a bad morning arrives, check the previous three days — heart rate, sleep, stress load, and your own notes on cognitive and emotional effort — rather than yesterday's step count alone. That's where the real pattern usually is.
How we made it
The data section uses Welltory's daily panel: 5,061 users and 433,313 days with enough data to compare each morning against that person's own baseline. A poor-recovery morning was defined as morning battery at least one standard deviation below the person's own median. Activity was classified relative to each person's own median step count. We looked at one-, two- and three-day lags, at three-day runs, and separately for people reporting ME/CFS, long COVID, POTS and no conditions.
The clinical content draws on studies of post-exertional malaise triggers and timing (Chu et al. 2018; Hartle et al. 2021; Stussman et al. 2020), PEM in long COVID (Vernon et al. 2023), a meta-analysis of pain after exercise in ME/CFS and fibromyalgia (Barhorst et al. 2021), and a survey of specialist ME/CFS care (Wormgoor & Rodenburg 2023).
Discounts for blog readers: up to 36% off
See what affects your energy, stress, sleep, and daily state with Welltory
This article is for educational purposes only and is not a substitute for medical advice, diagnosis, or treatment from a qualified clinician.
Was this helpful?
Ask AI for a summary of page
Written by Jane Smorodnikova
The founder and CEO of Welltory. A recognized tech leader with two Master's degrees and experience at MIT, she has scaled Welltory to over 17 million users.
Written by Tatsiana Yashyna
Deputy COO at Welltory. With a background in medicine and years of working with health data, she translates research and real physiological signals — sleep, stress, heart rate, and hormones — into clear, evidence-based explanations that help people understand what their bodies are telling them.
References
- Chu L, Valencia IJ, Garvert DW, Montoya JG. Deconstructing post-exertional malaise in myalgic encephalomyelitis/chronic fatigue syndrome: a patient-centered, cross-sectional survey. PLOS ONE 2018;13(6):e0197811. https://doi.org/10.1371/journal.pone.0197811
- Vernon SD, Hartle M, Sullivan K, et al. Post-exertional malaise among people with long COVID compared to myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). Work 2023;74(4):1179-1186. https://doi.org/10.3233/wor-220581
- Stussman B, Williams A, Snow J, et al. Characterization of post-exertional malaise in patients with myalgic encephalomyelitis/chronic fatigue syndrome. Frontiers in Neurology 2020;11:1025. https://doi.org/10.3389/fneur.2020.01025
- Hartle M, Bateman L, Vernon SD. Dissecting the nature of post-exertional malaise. Fatigue: Biomedicine, Health & Behavior 2021;9(1):33-44. https://doi.org/10.1080/21641846.2021.1905415
- Barhorst EE, Andrae WE, Rayne TJ, et al. Pain-related post-exertional malaise in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) and fibromyalgia: a systematic review and three-level meta-analysis. Pain Medicine 2022;23(6):1144-1157. https://doi.org/10.1093/pm/pnab308
- Wormgoor ME, Rodenburg SC. Focus on post-exertional malaise when approaching ME/CFS in specialist healthcare improves satisfaction and reduces deterioration. Frontiers in Neurology 2023;14:1247698. https://doi.org/10.3389/fneur.2023.1247698