why did i gain weight overnight.

Published August 19, 2026·Measurement·12 min read

An overnight rise of one to three kilograms is water, stored carbohydrate, and food still in transit — not fat. Here is what each trigger actually weighs, how long it lasts, and the weighing protocol that makes the number readable again.

you already know the exact sequence. you wake up. you go to the bathroom. you take everything off, because obviously, and you step on with your feet in the same two spots they are always in. and the number is two kilos higher than it was yesterday.

and then the day reorganizes itself around that number. you eat differently. you think about it in the car. you mentally cancel something. somewhere in there you do the math on what you ate last night and it does not add up, but the scale said what it said, so you assume the scale knows something you don't.

it doesn't. the scale is a very accurate instrument pointed at the wrong question. it weighs everything — bone, muscle, organs, fat, the four hundred grams of dinner still working its way through you, and above all water, which is the single most changeable thing in your body and the least related to whether anything is going well.

so let us go through what actually moved. not vaguely, "oh it's just water" — that is what people say when they don't know either. i mean the specific mechanisms, the measured numbers, how many days each one lasts, and one of them is going to surprise you because the most repeated explanation on the internet is the one with the weakest evidence behind it.

why fat is not on the list of suspects

start here, because it settles the panic before we get to mechanisms. fat is stored energy. to add it you have to eat energy your body did not spend, and the conventional figure used to convert between the two is about 7,700 calories per kilogram of fat tissue, or 3,500 per pound. that is an approximation and a contested one — Kevin Hall's modelling work in the Lancet exists partly to show how badly that simple rule misleads people over months. but as a floor it is fine for our purpose. even generously, a two-kilogram overnight gain of actual fat needs somewhere near 15,000 calories of surplus in one evening. you would remember.

Hall's paper adds the more useful point. modelling adult metabolism against real energy-balance data, he found that the body weight response to a change in energy intake is slow, with a half-time of roughly one year. a year. bodies are not ledgers that settle overnight; they are slow systems with enormous short-term noise on top. which means every number you see between two weigh-ins is noise, by definition. the signal is underneath it, moving at a completely different speed.

so if it is not fat, it is one of four things, and they are all things that legitimately weigh something.

the glycogen answer, which is the big one

your muscles and liver store carbohydrate as glycogen — a branched chain of glucose molecules kept on hand as fast fuel. and here is the part that makes it matter for the scale: glycogen is not stored dry. it is stored hydrated, holding water alongside it in the muscle cell.

the ratio has been quoted as 1:3 for decades, and in 2015 a group at the University of Castilla-La Mancha actually went and checked it with muscle biopsies. nine trained cyclists rode for 150 minutes in the heat until they were 4.6 percent dehydrated, dropping muscle glycogen by 44 percent. then they were refed carbohydrate and biopsied again during recovery. the finding: per gram of glycogen restored, at least 3 grams of water was stored in the muscle alongside it. when fluid losses were fully replaced as well, the ratio went higher still, up to 1:17 — though the authors are careful to say that extra water was probably not bound to glycogen, just present.

now apply it. you have a rest day, or a bigger carbohydrate meal than usual, or you eat normally after a few days of eating less. your glycogen stores refill. and every gram of carbohydrate you park in a muscle arrives with roughly three grams of water attached. that is real mass on a scale and it is also exactly what you want to be happening, because a full glycogen store is what makes the next session feel possible. the review of post-exercise glycogen resynthesis by Burke, van Loon and Hawley describes the refill as a two-phase process — a rapid window in the first four hours where depletion itself drives storage, then a slower phase over the next twenty. the water lands on the scale before the benefit shows up in the gym. tragic ordering. nobody consulted us.

the salt answer, and why it is weaker than you have been told

every article on this topic says the same thing: you ate salt, salt holds water, the scale went up. it is repeated so uniformly that it reads as settled. it is not, and the controlled evidence is genuinely awkward for it.

in 2000, a group at the German Aerospace Center's Institute of Aerospace Medicine ran 32 healthy men through a metabolic ward on fixed sodium intakes: 50, then 200, 400 and 550 milliequivalents a day. that top figure is an enormous amount of salt. plasma volume rose in a dose-dependent way, by 315 millilitres at the highest intake. but here is the result nobody quotes. total body water did not increase. and body mass did not increase. the salt moved fluid between compartments — out of the interstitial space and into the bloodstream — rather than adding any to the body.

a decade later the picture got stranger. Rakova and colleagues, publishing in Cell Metabolism, held men on fixed salt intakes for months during space-flight simulation and found total body sodium swinging by 200 to 400 millimoles on weekly and monthly rhythms — driven by aldosterone and cortisol, and, critically, without parallel changes in body weight or extracellular water. the body stores and releases sodium somewhat independently of how much water it carries.

so what should you take from a salty dinner. the honest version is less exciting and more useful. the mass that shows up the next morning is mostly the food itself and the water you drank with it — a large restaurant meal plus drinks is genuinely a kilogram or more of physical material sitting inside you until it finishes moving through. that is not water retention. that is groceries, in transit, standing on the scale with you.

the caveat that keeps this honest: both studies used men, on sustained intakes under controlled conditions. neither one tested a single salty dinner in a person who does not normally eat that way, and short-term shifts around an abrupt change are a different question from the steady state. what the evidence does not support is the confident claim that salt durably adds body water and body mass. at the doses tested, it did neither.

what your cycle is actually doing, according to the tracking data

if you have a menstrual cycle, you have been told the bloat comes in the week before your period. that is the folklore, and it is repeated everywhere, and the largest prospective dataset on the question does not say that.

the Prospective Ovulation Cohort collected daily diary entries across 765 menstrual cycles in 62 healthy women over a year, with ovulation confirmed by quantitative basal temperature analysis. self-reported fluid retention was scored daily on a 0 to 4 scale. the pattern: scores were lowest in the mid-follicular phase — the stretch after your period ends and before you ovulate — climbed gradually from 0.22 to 0.50 across the eleven days around ovulation, and peaked on the first day of menstrual flow, at 0.9. not the week before. day one.

two more findings worth carrying. neither estradiol nor progesterone levels were significantly associated with the fluid retention scores — the hormones everyone names as the cause did not track with the symptom. and there was no significant difference in perceived fluid retention between cycles where ovulation happened and cycles where it did not.

the necessary honesty here, and it is a real limitation: this study measured what women reported feeling, on a 0 to 4 scale — it did not weigh anyone or measure body water. so it tells you when the sensation peaks, not how many grams moved. treat it as a correction to the timing folklore, not as a measurement of your body weight. the practical use is simple: if you know the sensation peaks around day one rather than the week before, you stop reading a mid-cycle scale jump as a cycle event and go look at what you actually ate and trained.

the new-workout answer, which lasts longer than you would think

you start a program, or come back after two weeks off, or your coach adds a lift you have never done. the next morning the scale is up and you have not eaten anything unusual. this one is muscle damage and the swelling that follows it, and the timing is the useful part.

in a controlled arm-to-arm study, ten untrained people performed ten sets of six maximal eccentric elbow flexions — the lowering half of a curl, done hard — and were then measured across fourteen days. upper arm circumference increased, and the changes in swelling and creatine kinase, a marker of muscle damage, were still being measured out at three and four days post-exercise. soreness and swelling do not peak the next morning. they build.

so the scale reading you take the morning after a hard novel session is not even the top of that curve. if you started something new on Monday and panicked on Tuesday, Thursday was going to be worse, and none of it was fat. it is inflammation and fluid moving into damaged tissue so it can be repaired, which is the actual process you signed up for. it resolves as the repair finishes.

the same logic covers creatine, if you started taking it. Powers and colleagues loaded 32 resistance-training men and women with 25 g a day for a week and measured body water directly with deuterium oxide and sodium bromide dilution. the creatine group gained body mass and total body water, with fluid distribution unchanged — the water went into the cells along with the creatine. Hultman's classic loading work established the protocol that produces this: roughly 20 g a day for six days raises muscle total creatine by about 20 percent. the scale moves in week one and it is supposed to. that is the supplement arriving.

the weekly rhythm nobody tells you about

here is the finding that reframes the whole thing. researchers at VTT in Finland took 4,657 daily weight measurements from 80 adults tracked for anywhere from 15 to 330 days, and modelled them for a weekly pattern. There is one.

body weight rises starting Saturday, peaks on Sunday and Monday, and declines from Tuesday onward. it is a group-level rhythm, it is statistically real, and it means a Monday weigh-in compared against a Friday weigh-in is measuring the day of the week as much as it is measuring you. the authors' own conclusion is worth quoting plainly: weekend-to-weekday variation "should be considered as normal instead of signs of weight gain."

and then the part that genuinely surprised me. the compensation pattern was strongest in the people who lost or maintained weight, and weakest in those who slowly gained. the fluctuation was not the problem. the people whose weight swung up on the weekend and came back down during the week were the ones succeeding. a flat, unmoving line was not the marker of control. the return was.

how much noise is normal, in actual numbers

you deserve a number for this rather than reassurance. researchers in Brisbane weighed ten well-hydrated, medically stable adults every hour for nine hours a day, across three days, recording every input and output. these were people with nothing wrong with their fluid status.

their weight fluctuated across the three days by 1.1 to 3.6 percent. four of the ten swung more than 2 percent, and two swung more than 3 percent. for a 70 kg person, 3.6 percent is 2.5 kilograms of completely unremarkable variation.

and then the finding that hands you the protocol: the time of the weigh-in accounted for 99.8 percent of the variation. when weights were compared at the same hour each day, the difference collapsed to 0.4 kilograms or less. almost the entire signal you have been reading as your body was the clock.

the caveat, stated plainly: this was ten older hospitalised adults, not a large sample of training women. it is a small pilot study and the exact percentages should not be treated as a population norm. what transfers is the mechanism and the ranking — timing dominates, and same-time comparison shrinks the noise by most of its size.

the protocol: how to weigh yourself so the number means something

this is the part to keep. it takes no extra effort — it is the same weigh-in, made readable.

1. same time, every time. first thing, after the bathroom, before eating or drinking. this single change removes most of the noise, because timing explained 99.8 percent of the variation in the study above.
2. same clothing state. naked or the same light layer. clothes were a recorded variable in that study for a reason.
3. weigh daily, but never react to a day. daily data is good. daily interpretation is what breaks people.
4. read the 7-day average against last week's 7-day average. a single reading contains all the noise. a week of readings mostly cancels it.
5. compare same weekday to same weekday if you are looking at individual numbers at all. Monday against Monday. the weekly rhythm is real, and comparing across it is measuring the calendar.
6. give it four weeks before drawing a conclusion. the weight response to changed intake has a half-time of around a year — four weeks is already an aggressive read on a slow system.
7. keep one measurement that is not weight. the load and reps in your training log, or how a specific pair of trousers fits. when the scale is being noisy, that is the record that still reads true.

one more thing about the morning ritual

i want to name the actual problem, because it is not really a measurement problem.

the reason a number moving two kilos ruins a day is that it feels like a verdict. like the body filed a report overnight and the report says you failed. and once that lands, the day gets built around it — the eating gets weird, the training gets punitive or gets skipped, and by the following week the thing that actually damaged your progress was not the number. it was the eleven days of chaos the number caused.

this is why the protocol above matters more than any single fact in this article. not because the number is meaningless, but because it only means something at a scale of weeks, and reading it daily is like judging a film by one frame and being surprised the plot makes no sense.

the body you had last night is the body you have this morning. it is the same amount of muscle, doing the same things, holding a slightly different amount of water. nothing was undone while you slept. water is just heavy, and it moves.

---

quick answers

why did i gain weight overnight?

almost certainly not fat. an overnight rise of one to three kilograms is water, stored carbohydrate, and the physical mass of food and fluid still moving through you. the mechanism that does most of the work is glycogen: muscle stores carbohydrate bound to water, and biopsy measurements found at least three grams of water stored per gram of glycogen. refill a depleted store after a rest day or a big carbohydrate meal and the water arrives with it. body weight also follows a weekly rhythm, rising over the weekend and falling from Tuesday, which researchers who tracked 4,657 daily weigh-ins concluded should be treated as normal rather than as weight gain.

is it possible to gain fat overnight?

not in any amount the scale would show as a jump. storing body fat requires eating energy you did not use, and the conventional figure is roughly 7,700 calories per kilogram of fat tissue. a two-kilogram overnight gain would mean eating around 15,000 calories beyond your needs in one evening. modelling published in the Lancet also shows the body weight response to a change in energy intake is slow, with a half-time of roughly one year, so real tissue change appears over months, not between two weigh-ins.

does salt make the scale go up?

less than the internet claims, and the evidence points somewhere unexpected. in a controlled metabolic ward study, sodium intake was raised in stages from 50 to 550 milliequivalents per day and plasma volume rose by 315 millilitres, but total body water did not increase and body mass did not increase. the salt shifted fluid between compartments rather than adding it. what a salty meal does move on the scale is real but simpler: the mass of the food and the water you drank with it, both of which pass through within a day or two. that study was conducted in 32 healthy men on sustained intakes, not on a single salty dinner.

how should i weigh myself so the number stops lying?

weigh at the same time every day, after the bathroom and before eating or drinking, in the same amount of clothing, then ignore the individual number and read the seven-day average against last week's seven-day average. in a study of well-hydrated adults weighed hourly across three days, the time of the weigh-in accounted for 99.8 percent of the variation in weight, and same-time-of-day weights differed by no more than 0.4 kilograms. one weigh-in is a data point with the noise still in it. a weekly average is a trend.

keep reading

• why don't i look like i lift yet. — the honest timeline, and why the mirror lags the work by months
• how much protein do i actually need. — the intake numbers, since the scale will not tell you what you are built of
• when am i supposed to add weight. — the log-based progress metric that replaces the morning weigh-in
• why can't i stay consistent. — what actually breaks a run of good weeks, and it is usually this
• what does toned actually mean. — why body composition and body weight are two different conversations

---

references

educational content, not medical advice. the numbers above describe normal fluctuation in healthy people — rapid or persistent swelling, weight gain with shortness of breath, or swelling in one limb are clinician questions, not scale questions. sample sizes and populations are stated in the table above for a reason: the hourly-weighing study used ten older hospitalised adults, the sodium work used men, and the cycle data is self-reported sensation rather than measured body water.

- MoMo