The scale is noisy. The trend is not.

What is actually on the scale

A scale measures everything: tissue, bone, the water in and around your cells, the glycogen in your muscles and liver, the food and fluid currently in your digestive tract, and whatever you are wearing. Fat is one component and it is the slowest-moving one.

Over a single day the fast-moving components can swing by a few pounds in either direction. Bhutani and colleagues, weighing adults at baseline and again two weeks later in Physiological Reports (2017), found that about 84% of a two-week change in body weight was fat-free mass rather than fat, and note that total body water alone can fluctuate by up to about 5% daily, with short-term swings of one to two kilograms being ordinary. The slow component changes by a small fraction of that. This is why a daily reading tells you very little on its own.

The usual culprits

Sodium. A change in salt intake moves fluid around the body, and the balance takes a few days to settle: Bie, reviewing sodium and water regulation in the American Journal of Physiology (2018), puts the half-time of the shift to a new steady state at roughly 21 hours. It resolves on its own.

Carbohydrate and glycogen. Glycogen is stored in hydrated form, with three to four parts water, as Kreitzman and colleagues describe in the American Journal of Clinical Nutrition (1992). Eat noticeably more or fewer carbohydrates than usual and stored water follows, which is most of what people see in the first week of any low-carbohydrate approach: Yang and Van Itallie, comparing two 800-calorie diets at the same calorie level in the Journal of Clinical Investigation (1976), found 61% of the weight lost on the ketogenic diet was water, against 37% on the mixed diet.

Training. Unaccustomed or hard training causes short-term fluid retention in the muscles worked. Nosaka and Clarkson, in Medicine & Science in Sports & Exercise (1996), measured swelling in untrained arms that began a day after the session and was still visible on MRI weeks later. Perversely, a good week of training can flatten the scale.

Hormonal cycles. For people who menstruate, water shifts across the cycle are real and modest. Kanellakis and colleagues, measuring 42 women twice weekly in the American Journal of Human Biology (2023), found body weight about half a kilogram (roughly a pound) higher during menstruation than in the first week of the cycle, almost all of it extracellular water. Comparing a reading to one taken at a different point in the cycle compares two different things.

Digestive contents. Food in transit has weight. A large or high-fiber meal is still physically present the next morning.

Everything else. Total body water moves by several percent across an ordinary day, and not every shift has a cause you can name.

Why a rolling average works

Each of those causes pushes the number around in a way that is large but short-lived and roughly as likely to go up as down. Averaging across seven days lets most of that cancel, leaving the slower change underneath visible.

It is the same reason nobody judges a climate from one afternoon. The daily figure is real, it is not a measurement error, but it is answering a different question from the one you are asking.

Weighing in a way that compares

You cannot remove the noise, but you can stop adding to it. Weigh under the same conditions each time. Typically first thing, after using the bathroom, before eating or drinking, wearing about the same thing. Consistency matters far more than which conditions you pick.

Weighing daily and reading the average is generally more informative than weighing weekly, because one weekly reading might land on an unusually high or low day and you would have no way to know.

When a flat trend means something

A genuinely rolling average that is flat across several weeks (not several days) is signal rather than noise. That is worth looking at intake, expenditure, or whether the original estimate was simply wrong. Estimates being wrong is common and expected.

If weight changes sharply without a change in habits, or you have symptoms alongside it, that is a reason to speak to a clinician rather than to adjust a target.

How Loopa handles this

Loopa charts your weigh-ins with a seven-day average drawn over them, so the trend is the thing you see first and the daily dots are context rather than the headline. Logging weight and reading that trend is free, over a rolling thirty days; Premium opens the full history, the forecast that projects from the trend rather than the last reading, and What If, which lets you change the inputs and watch the projected date move before you commit to anything.