Why does my weight trend lag behind the scale?
A weight trend lags because it combines the newest reading with earlier ones. During a sustained fall, the earlier readings are often higher, so the smoothed line can remain above today’s weight. That delay is part of smoothing, not a second measurement of your body.
A small worked example
Suppose yesterday’s trend was 80 kg and today’s scale reading is 79 kg. An exponential average with a smoothing factor of 0.15 moves 15% of the way from the old trend toward the new reading.
New trend = previous trend + 0.15 × (new reading minus previous trend)
80 + 0.15 × (79 minus 80) = 79.85 kg
With a factor of 0.10, the same update gives 79.90 kg. With 0.30, it gives 79.70 kg. Those differences come entirely from the settings. None of the three numbers measures body fat.
More responsive or smoother?
A larger factor gives the newest reading more influence. It responds faster to a lasting change but also follows more of the short-term movement. A smaller factor retains more of the older trend. It looks calmer but takes longer to catch up.
There is no setting that removes every fluctuation while immediately recognizing every lasting change. The arithmetic cannot know which kind of reading just arrived. Our guide to daily weight changes explains why a single reading needs context.
Why two apps can show different trends
Apps may use a different smoothing method, a different starting value, or different treatment of days without a reading. Even two exponential averages can disagree if they assign different weights to the newest observation. Compare the same dates and history before assuming one is wrong.
Downtrend starts its trend at the first recorded weight. Its current default gives a new daily reading a weight of 0.15. After a gap, it increases the new reading’s influence according to the number of days since the previous entry. It does not create weigh-ins for missing days.
Smoothing and the projection window are separate
Smoothing controls the trend line. The projection window controls how much recent trend history is used to estimate a pace and goal date. A shorter window responds to recent changes sooner. A longer one includes more history.
Changing the projection window does not change how the trend itself is smoothed. This is why two settings that both seem to make an app more responsive can affect different parts of the display.
Sources
- NIST. Exponential smoothing. The weighted-average formula and the effect of the smoothing factor. This explains the arithmetic, not a medical interpretation of weight.
Downtrend’s behavior above was checked against its trend calculation on 24 October 2026. Its defaults and handling of gaps are product choices, not recommendations from NIST.
This article is general information, not medical advice. Talk to a doctor before a big change to what you eat or how you train, and especially if you are pregnant, under 18 or have a health condition.