Your resting engine barely moves; the day wrapped around it does#
Your metabolism, in the sense most people mean it — the cost of simply being a body — is one of the steadiest things about you. What is not steady is the total. Pooling 25 studies that measured free-living energy expenditure twice in the same people with doubly labelled water, the within-subject coefficient of variation came to 11.8 percent, and even between two measurements taken with no time gap at all it was 8.2 percent1. On a 2,400-calorie day, 8 percent is close to 190 calories — our arithmetic on their percentage — appearing and disappearing without you doing anything you would describe as different.
Now the part that does not move. Measured in a room-sized calorimeter, the same men's basal expenditure repeated to a within-subject variation of 2.4 percent across five measurements a day apart, and 2.9 percent across five measurements one to three weeks apart2. Stretch the calendar tenfold and the baseline barely notices. So the answer to "why do my calorie needs vary" is not that your metabolism is erratic. It is that the metabolism is the quiet part, and everything wrapped around it — movement, food, the weather, the week — is the loud part. Metabolism explained covers what the quiet part is actually spent on; this is about how much each term wobbles.
Three instruments, three answers to the same question#
The most useful way to see this is to watch the number get less stable as the measurement gets closer to real life.
| How the same person was measured | What was held constant | Within-person variation |
|---|---|---|
| Basal expenditure, calorimeter, 1 day apart | Everything | CV 2.4% |
| Basal expenditure, calorimeter, 1–3 weeks apart | Everything | CV 2.9% |
| 24-hour total, calorimeter, 1 day apart | Room, meals, schedule | CV 2.7% |
| 24-hour total, calorimeter, 1–3 weeks apart | Room, meals, schedule | CV 4.6% |
| Free-living total, isotopes, no time gap | Nothing | CV 8.2% |
| Free-living total, isotopes, 52 weeks apart | Nothing | CV 15.4% |
| Between two different people | — | CV 20.6% |
Calorimeter rows from Rumpler 1990; free-living and between-person rows from Black & Cole 2000.
Read the column downward. Nothing about anyone's physiology changes between the top row and the bottom one. What changes is how much of ordinary life the instrument is allowed to see. A modern replication of the confined case makes the ceiling explicit: 35 adults spent 23 hours on each of three consecutive days in whole-room calorimeters, and 24-hour expenditure repeated with an intraclass correlation of 97.8 percent and a coefficient of variation of 1.9 percent3.
Sealed in a room, a person's daily burn repeats to within two or three percent. Let the same person out and it moves by eight, and nothing metabolic happened on the way through the door.
The last row is the one worth sitting with. Between-person variation in daily expenditure is 20.6 percent; a single person's own variation is 11.8. Your day-to-day range is over half the width of the entire population's spread. When you compare today's number to yesterday's, you are working with a signal that is not much quieter than the difference between you and a stranger — and that stranger comparison is the subject of how much resting metabolism varies between people.
One day is a poor sample of your week#
If the variation lives in movement, the obvious question is how representative any single day of movement is. Someone has measured that directly, and the answer is bleak in a useful way.
Eighty-seven adults wore accelerometers, and researchers partitioned the variance to ask how well one day predicts a person's habitual level. For overall physical activity, a single day returned an intraclass correlation of 0.31. For moderate-to-vigorous activity, 0.31. For sedentary time, 0.21. Reaching a respectable reliability of 0.80 took 7 to 10 days of monitoring for overall and moderate-to-vigorous activity, 3 to 4 days for light activity, and 13 to 15 days for absolute sedentary time4.
An ICC of 0.31 means roughly two-thirds of what you see in one day is not a property of the person. It is the day.
The same study also checked the explanation everybody reaches for first. Weekdays did carry more activity than weekends, significantly — and the weekday/weekend split accounted for less than 5 percent of the total variation. So the wobble is not a tidy pattern you could correct for with a calendar. It is mostly unstructured: a longer walk to the station, a cancelled errand, a meeting that ran and killed the gym slot. The largest movable term in your budget is also the least schedulable one, which is the argument in what NEAT is and why it matters.
The seasons move activity and leave the total alone#
The folk theory that you burn more in winter — the body fighting the cold — has been tested with the reference method, and what came back is more interesting than a yes or no.
Twenty-five healthy Dutch adults had total expenditure measured by doubly labelled water and sleeping metabolic rate measured overnight in a respiration chamber, in both summer and winter. Physical activity level was clearly higher in summer: 1.87 ± 0.22 against 1.76 ± 0.18 (p < 0.001), and the seasonal swing was much larger in the men than the women (0.20 ± 0.14 versus 0.05 ± 0.16, p = 0.04). Total energy expenditure showed no seasonal difference at all5.
Hold those two results together, because separately each one misleads. People genuinely moved more in summer, and their days did not end up bigger. That is the shape of the compensation problem — added activity buys less than the arithmetic promises, an effect that is strongest in exactly the conditions dieters create for themselves (constrained energy expenditure). It also means neither season is a metabolic season. If you burn differently in January, the reason is on your calendar rather than in your thermostat.
Two other rhythms deserve a mention and not a paragraph, because each has its own page. The menstrual cycle adds a small luteal-phase rise to resting rate whose size is genuinely contested — our piece on why TDEE differs for women and men sets the meta-analysis against the newer reviews. And short sleep moves daily expenditure in the direction opposite to the folklore — up, not down — while doing considerably more damage on the intake side.
Even your own long-run average is only mostly yours#
There is one more layer, and it is the least intuitive. If a week of your life is a noisy sample, is a year of it a clean one?
The IAEA doubly-labelled-water consortium pulled 696 repeat measurements from 348 adults, spaced on average 1.9 years apart, and asked how repeatable an individual's total expenditure is. It came out at R = 0.64 (95% CI 0.578–0.703) — which the authors correctly call high, and which also means that a bit over a third of the variance in any one measurement of your daily burn belongs to the occasion rather than to you6. In children the same analysis returned R = 0.00.
So "my TDEE" is not a personal constant that a better instrument would eventually pin down. It is a distribution with a stable centre and a genuinely wide body, and the width is not measurement failure — it is the quantity. That is a different claim from the one in why every TDEE calculator is an estimate, which is about how badly an equation locates your centre. This is about the fact that even a perfect equation would be aiming at something moving.
What to do with a number that will not sit still#
The practical consequences are short, and mostly consist of not overreacting.
- Judge the deficit on 7 to 14 days, never on one. A single day's expenditure carries the same information about your habitual burn as a single day's weigh-in carries about your fat mass, and for a related reason — the noise is larger than the daily signal. The accounting version of this argument is in does your body count calories daily or weekly.
- Do not adjust your target by less than the noise. Free-living expenditure moves by roughly 8 percent between measurement periods with no time gap. On most people's days that is 150 to 200 calories, which is larger than almost any adjustment you would be tempted to make on a single bad week.
- Match the window to the question. Light daily activity settles within 3 to 4 days; hard training and sedentary time need one to two weeks before the average means anything. If you want to know whether a change in routine worked, give it a fortnight.
- Stop diagnosing the baseline. The part of you that could plausibly be "slow" varies by 2 to 3 percent, measured. The part that swung 400 calories last Thursday was your afternoon.
The reason this matters is not accuracy for its own sake. It is that a metabolism framed as unstable invites the wrong repairs — supplements, resets, another recalculation — for a fluctuation that was never metabolic. Your engine idles at a very consistent rate. The day is what keeps changing around it.
FAQ#
How much does one person's daily calorie burn actually vary?#
Measured with doubly labelled water across 25 studies with repeat measurements, within-subject variation in free-living energy expenditure averaged 11.8 percent, and 8.2 percent between measurements with no time gap between them. That is over half the size of the 20.6 percent variation between different people. Under confinement in a metabolic chamber, where behaviour is constrained, the same quantity varies by only 2.7 to 4.6 percent — the difference is life, not physiology.
Do you burn more calories in winter than in summer?#
Not in the one study that measured both seasons in the same people with the reference method. Twenty-five Dutch adults showed a clearly higher physical activity level in summer (1.87 versus 1.76, p < 0.001), and no difference in total energy expenditure between the seasons. Being more active in summer did not make the summer days bigger, and the cold did not make the winter days bigger either.
Should I eat more on days I move more?#
Matching intake to a single day's activity is chasing a number you cannot measure well enough to chase. One day of accelerometry predicts habitual overall activity at an intraclass correlation of 0.31, so most of what you would be responding to is the day rather than a real change in your needs. Setting a steady intake and reading the outcome over one to two weeks handles the same variation without asking you to estimate it daily.
Sources#
- Black AE, Cole TJ. Within- and between-subject variation in energy expenditure measured by the doubly-labelled water technique: implications for validating reported dietary energy intake. Eur J Clin Nutr. 2000;54(5):386-394.
- Rumpler WV, Seale JL, Conway JM, Moe PW. Repeatability of 24-h energy expenditure measurements in humans by indirect calorimetry. Am J Clin Nutr. 1990;51(2):147-152.
- Allerton TD, Carnero EA, Bock C, et al. Reliability of measurements of energy expenditure and substrate oxidation using whole-room indirect calorimetry. Obesity (Silver Spring). 2021;29(9):1508-1515.
- Aadland E, Ylvisåker E. Reliability of objectively measured sedentary time and physical activity in adults. PLoS One. 2015;10(7):e0133296.
- Plasqui G, Westerterp KR. Seasonal variation in total energy expenditure and physical activity in Dutch young adults. Obes Res. 2004;12(4):688-694.
- Rimbach R, Yamada Y, Sagayama H, et al. Total energy expenditure is repeatable in adults but not associated with short-term changes in body composition. Nat Commun. 2022;13(1):99.
