How to Track Weight Over Time: Moving Averages, Weekly Readings, and What Noise Looks Like

You step on the scale Monday morning and see a number 1.4 kg higher than Friday. That higher number is not fat gain. It is glycogen water binding: each gram of glycogen stored in your liver and muscle binds roughly 3 to 4 grams of water. A shift in carbohydrate intake can move the reading by 1 to 2 kg within 24 hours. Tracking weight with a moving average separates that noise from real change.

The normal daily weight fluctuation in an adult is 1 to 2 kg. The causes are physiological. Glycogen water binding accounts for the largest share. When you eat carbohydrates, glycogen stores refill and carry water with them. When you restrict carbohydrates, those stores deplete and the water is excreted. The scale drops, but you have not lost fat. Sodium works the same way: a high-sodium meal pulls water into the extracellular space and the scale rises for 24 to 48 hours. Menstrual cycle weight fluctuation adds another layer. Water retention peaks in the luteal phase, adding 0.5 to 1.5 kg that is not adipose tissue.

The Weighing Protocol That Removes the Worst Noise

Weigh first thing in the morning, after using the toilet, before eating or drinking. Same time, same clothing, same scale. This is the first morning weight. It minimises measurement error from food mass, clothing weight, and hydration shifts. A consumer bathroom scale has a tolerance of ±0.5 to 2.0 kg under uncontrolled conditions. The protocol keeps the noise within a known range so you can see the trend.

A single daily reading is a snapshot of glycogen stores, sodium intake, and menstrual phase. It is not a measurement of fat mass. Take seven daily readings, discard none, and calculate the average. That number reflects your weight minus the 1 to 2 kg of noise present on any single day. If the weekly average moves down over three consecutive weeks, you are seeing real change. If it stays flat, you are not losing fat even if one day was low. If it rises for a single week, check sodium and menstrual phase before assuming calorie surplus.

Weight tracking noise follows predictable patterns. A 3 kg drop over three days is not fat loss. Fat loss at a moderate deficit is roughly 0.5 to 1 kg per week. A drop larger than that is water. A rise of 1.5 kg the day after a restaurant meal is not fat gain: it is sodium and glycogen. A rise during the luteal phase of the menstrual cycle that reverses in the follicular phase is hormone-driven water retention. The signal-to-noise ratio of body weight data is poor on a daily timescale and good on a weekly or monthly timescale. Your job is to dampen the noise, not to explain every blip.

Collect seven consecutive first-morning weights. Add them. Divide by seven. That is your weekly average. Next week, drop the oldest day and add the newest, then average again. This simple moving average smooths the 1 to 2 kg fluctuation into a line that moves slowly and honestly. An exponential moving average, which gives more weight to recent readings, is faster to respond but also faster to overreact to a sodium spike. For most people, the simple seven-day average is the best method.

A weight loss plateau occurs when the weekly moving average does not decline for two to four weeks. The first cause to check is adherence decay: you are eating more than you think, or moving less. The second is measurement error: your scale tolerance of ±0.5 to 2.0 kg can hide a real loss of 0.5 kg per week for several weeks. The third is metabolic adaptation, a reduction in energy expenditure beyond what is predicted by body mass loss. Meta-analyses of 30 studies show adaptive thermogenesis after weight loss averages 100 to 250 kcal per day below predicted BMR for the new body weight. That is real. It is not the cause of most plateaus. Before you assume metabolic adaptation, verify that your weighing protocol and calorie tracking are consistent.

The National Weight Control Registry observes that people who maintain a significant weight loss tend to weigh themselves frequently. That is an association, not a prescription. The registry data show that self-monitoring frequency is one behaviour among many. Frequent weighing without the moving average method leads to overinterpretation of single readings, which can trigger unnecessary dietary changes or discouragement. The protocol matters more than the frequency. Weigh daily, average weekly, act only on the average.

When the Normal Route Does Not Work

If you cannot weigh at the same time every day, weigh three times per week on the same days (Monday, Wednesday, Friday) and average those three. This gives a less stable estimate but removes the worst noise. If you travel, accept that the data from that week is noise and skip it. If your scale breaks, buy a calibrated digital scale with a tolerance of ±0.1 kg. Consumer bathroom scales that vary by 2 kg between uses are not worth the money.

The Single Thing That Most Often Goes Wrong

A reader sees a 1.5 kg rise on the scale, assumes it is fat, and cuts calories aggressively. That cut causes glycogen depletion, which produces a 2 kg drop over the next two days. The reader interprets this as success, then eats carbohydrates, the glycogen refills, the water returns, the scale rises again, and the cycle repeats. The moving average would have shown a flat line the whole time, because no fat was lost. Trust the average, not the day.