The scale showed a kilogram more this morning. The number is real. A kilogram of fat overnight does not follow from it.
A scale does not separate your body into fat, water, glycogen and the contents of your digestive system — it adds them all together. So the useful question is not “what happened since yesterday?” but “where is the series of comparable measurements heading?”
This matters especially when a smart scale is already sending readings into an app automatically. Without that boundary, automation simply delivers alarming points faster, and every high morning starts prescribing a restriction for itself.
What moves the number before fat changes#
| Cause | What happens | What one morning cannot tell you |
|---|---|---|
| Water | Fluid intake and output shift through the day. | That the whole difference is fat. |
| Salt | Changes in sodium take part in regulating extracellular fluid. A particularly salty day temporarily shifts weight. | How much water this particular person retained. |
| Carbohydrate | Some carbohydrate is stored as glycogen, which binds water. Filling and using those stores changes total mass. | That a fast rise equals an equal change in fat. |
| Food in transit | Food and drink have mass while they pass through the digestive system. | That a late dinner changed your body composition. |
| The menstrual cycle | For some women, mass and fluid volume change across phases. The size and timing of the swing are individual. | That the jump came from the cycle, or that it is the same for everyone. |
| Weighing conditions | Time of day, clothing, surface and a different scale make neighbouring readings less comparable. | That the difference arose inside the body rather than in the conditions. |
In a study of 46 adults under ordinary conditions, 84% of short-term weight change over two weeks was in fat-free mass, which includes water Bhutani 2017. That is an average across two small cohorts, not a formula for one individual morning.
The body-weight model developed by researchers at the US National Institute of Diabetes and Digestive and Kidney Diseases accounts separately for glycogen-bound water and for changes in extracellular fluid when sodium changes.1
For the cycle there is no single figure either. In a study of 42 women measured twice a week across one cycle, mean mass during menstruation was about 0.45 kg higher than in the first week, accompanied by a change in extracellular water. The authors note the small sample, the single cycle observed and the limits of bioimpedance.2
These data explain the mechanism. They do not let anyone look at one number and decide “that was salt” or “that was the cycle” — not a person and not an app.
The point is real, but the cause of the difference is unknown: water, food, measuring conditions and tissue are mixed together.
Repeated readings under similar conditions show a rise, a fall or a flat line.
What “true weight” would mean#
In everyday talk it means weight with the water and the other interference removed. Home scales do not produce such data, and no app can reconstruct it from total mass. What Calk builds is a smoothed trend of total mass.
The algorithm keeps the real points and draws a calmer line through them. It accounts for the gaps between measurements and starts a new segment after a long pause, so that an old rate does not describe a new situation. Direction is judged from several points of the smoothed series; when there is not enough data, no conclusion is shown.
The points swing, the direction changes more slowly
After a salty dinner the point rises noticeably earlier than the line turns.
One high morning tugs the line slightly but does not set a new direction. If the rise repeats and holds, its influence on the trend grows steadily — which is exactly as it should be.
How weight reaches the trend#
Apple Health and Health Connect are system stores. Connecting does not give an app access to “all of your health”: permission is granted per data type, and you can withhold it or revoke it later.34
For the trend, Calk needs one type — body mass. Permitted measurements are imported to the device, the earliest reading is taken for each local day, and the series of points becomes a smoothed line. The earliest is chosen deliberately: morning readings are easier to compare with each other.
Calk does not read water-intake or menstrual-cycle data. Those factors matter for interpretation, but the app does not connect your weight jump to salt or a cycle phase — that conclusion stays your own context. What Calk processes and where is described in privacy and your data.
Comparability beats ritual#
Use the same scale on a hard surface and pick a similar moment in your routine. Do not replace yesterday’s point with a nicer repeat reading, and do not make a decision from one jump. Weighing frequency, what to do on a plateau and the detailed measuring order are covered in how to read your weight trend.
Daily weighing is optional. A few comparable points a week can be more useful than a daily series, if daily numbers cause anxiety or provoke abrupt restriction.
If the trend really has stalled, look at recurring meals and portions. That calls for a short food check, not a punishment for one morning. Salt and its sources are covered separately in salt, water and potassium.
When a trend is not enough#
Smoothing does not turn an app into a medical device. It does not explain oedema, a change in your cycle, a drug side effect or an illness.
Rapid unexplained weight change alongside marked swelling, breathlessness, pain or any other new symptom calls for medical assessment, not for waiting until the line settles. During pregnancy, with heart or kidney disease, and on treatment that affects fluid, weighing guidance is worth agreeing with your doctor.
For ordinary monitoring, a more modest rule is enough: a point tells you what you weighed at that moment; a trend tells you where mass is heading. Neither tells you why.
To see the trend in Calk, allow weight to be read or enter a few measurements by hand, and keep the conditions comparable. There is no need to open access to every health category.
Calk is available on iPhone, iPad, and Android.
Frequently asked questions#
Why did my weight go up overnight when I did not overeat?
Does a smoothed trend show a change in fat?
Does Calk work out whether a jump came from salt or the cycle?
Do you have to weigh yourself every day?
Sources and the limits of this check#
Integration and smoothing behavior was checked on August 11, 2026 against the current version of the app. The list of requested permissions has to be rechecked after every change to the integrations.
Hall KD, Sacks G, Chandramohan D, et al. Supplementary web appendix: Dynamic Mathematical Model of Body Weight Change in Adults, supplement to the article in The Lancet, 2011. The model accounts for glycogen-bound water and changes in extracellular fluid; it is population physiology, not a way to determine the cause of one weigh-in. ↩︎
Kanellakis S, Skoufas E, Simitsopoulou E, et al. Changes in body weight and body composition during the menstrual cycle. American Journal of Human Biology. 2023;35(11):e23951. ↩︎
Apple Developer Documentation: Authorizing access to health data. ↩︎
Android Developers: Read raw data from Health Connect. By default Health Connect returns other apps’ data from 30 days before permission was first granted. ↩︎