Your weight went up 0.8 kg overnight. Almost none of it was fat.
To gain a kilogram of fat you have to eat roughly seven thousand calories more than you burn. Nobody does that between Tuesday and Wednesday. Yet the scale moves that much overnight all the time, and people quit over it.
Paste your readings in and see what is actually happening underneath them.
Weight noise calculator
Paste your daily weights, oldest first — one per line. Nothing is sent anywhere.
Use whatever unit you weigh in; the answer comes back in the same one.
What the scale is weighing
Researchers have measured the composition of short-term weight change directly, under ordinary free-living conditions rather than in a metabolic ward. Over two weeks, the average change in body weight was 0.26 kg — consistent with people simply being in energy balance — and when they tracked what that change was made of, 84% of a unit change in body weight was fat-free mass, not fat.
Fat-free mass here largely means water, glycogen and the food still in transit through you. A salty dinner, a long run, a bad night's sleep, a new medication, where you are in your menstrual cycle, and how recently you used the bathroom all move that number, and all of them move it by more than a day of eating does.
There is even a weekly shape to it. Analysing 4,657 daily weigh-ins from 80 adults found a real group-level rhythm: weight is highest on Sunday and Monday, starts falling on Tuesday, and starts rising again on Saturday. The authors' conclusion is worth quoting almost directly — variations between weekends and weekdays should be treated as normal, not as signs of weight gain.
So a Monday reading compared against a Friday reading is not a measurement of progress. It is a measurement of the weekend.
The fix is not weighing less often
The obvious response — stop weighing daily, it is too noisy — turns out to be the wrong one.
A study of 2,838 weigh-ins from 40 people found that weight loss happened during the periods when people were weighing themselves daily, while gaps longer than a month were associated with weight gain. This is an association in a small group, not proof that the scale causes anything, and the likeliest reading is that people stop weighing when things are going badly rather than the other way round. But it does mean the answer is not to look away.
The answer is to keep weighing daily and stop reading the daily number.
Why a 7-day average
Average across seven days and two useful things happen at once. A full week cancels the weekday-weekend rhythm above, because every average contains exactly one of each day. And averaging seven noisy readings shrinks the noise by roughly a factor of two and a half, while leaving any real trend untouched — because the noise is random and the trend is not.
That is what the calculator above does. It needs seven readings before it will show you an average at all, and fourteen before it will compare two averages, because a “7-day average” computed from three days is not one. It will say so rather than show you a confident number built from too little.
Then read it like this:
- The trend line is the answer. If the average is falling, you are losing weight, whatever this morning said.
- The spread is the noise floor. If your readings swung 1.2 kg this week, then a 0.4 kg rise means nothing at all — it is smaller than your own measurement error.
- Give it a fortnight before judging a change. Two averages a week apart is the minimum honest comparison.
Weighing in a way that makes this work
The average only removes noise that is random. Anything that shifts every reading in the same direction survives the averaging, so make the conditions the same each time: same time of day, first thing in the morning, after the bathroom and before eating or drinking, on the same scale, on a hard floor rather than carpet.
gura records daily weights and shows you the 7-day average instead of the raw number, for the reasons above. But the arithmetic is not the app's — the calculator on this page is the same, and it runs entirely in your browser.