Correlating weight change with your calorie balance
A single weigh-in means nothing: water, salt and digestion move the scale by a kilo from one day to the next. What matters is the trend over several weeks, compared with the calorie deficit or surplus accumulated over the same period. The Trend tab in Happy Assistant does exactly that, and derives a measured energy expenditure — your body's, not a formula's.
The principle: two curves, one comparison
The body stores or releases roughly 7,700 kcal per kilo of body fat. Eat 500 kcal below your expenditure for three weeks and theory predicts about 1.4 kg lost. The app compares that expected change with the change actually measured across your weigh-ins.
- Measured change: the slope of a linear regression over all your weigh-ins, to smooth out daily noise.
- Expected change: the day-by-day sum of (intake − estimated expenditure), divided by 7,700.
- Estimated expenditure: basal metabolism adjusted for your activity level, plus the calories of logged workouts.
- The gap between the two curves is the useful signal, not today's number.
Logging your weigh-ins
- Weigh yourself under the same conditions: morning, fasted, after the bathroom.
- Just tell the assistant your weight ("I'm 74.2 kg this morning") or use the "Add a weigh-in" button.
- Repeat at least two or three times a week: the regression needs data points.
- Open Profile → My nutrition → Trend to read the curve and the verdict.
Daily weighing isn't obsessive if you only look at the trend: the more points, the more reliable the slope.
Measured expenditure beats a formula
Formulas (Mifflin-St Jeor, activity factors) estimate an average for a population, not your real expenditure. Knowing your average intake and your real weight change over the period, the calculation can be reversed: measured expenditure = average intake − (weight change × 7,700 / number of days).
| Situation | Reading |
|---|
| Loss matches expectation | Your reported intake and estimated expenditure are consistent. |
| Loss slower than expected | Intake likely under-reported, or expenditure overestimated. |
| Loss faster than expected | Real expenditure higher than estimated, or meals left unlogged. |
| Weight stable while in deficit | The deficit probably isn't real: recalibrate with measured expenditure. |
Conditions for a meaningful result
- A period of at least three to four weeks: below that, noise beats signal.
- At least four or five weigh-ins spread across the period.
- Enough meal coverage: missing days distort the intake total.
- Logged workouts, otherwise expenditure is underestimated.
Until those conditions are met, the app shows what's missing instead of a misleading verdict. That's deliberate: a figure computed from three weigh-ins in ten days means nothing.
A real example: seven weeks of combined tracking
Yves, creator and first user of Happy Assistant, tracked his meals for seven weeks (22 August – 9 October 2026) — mostly photographed, with portions estimated by the assistant — his workouts (estimates or Garmin / TrainerRoad screenshots, with no automatic sync) and his morning weigh-ins under the same conditions. This is the product creator's own account, not an independent review.
| Intake and needs | Daily average |
|---|
| Logged food intake | 2,409 kcal |
| Estimated needs excluding sport (rest + daily activity) | 2,161 kcal |
| Logged sport | 482 kcal |
| Estimated total needs | 2,643 kcal |
Comparing food intake with needs excluding sport alone, the balance looks like a surplus of about 247 kcal per day. Including sport, it becomes a deficit of about 235 kcal per day: sport changes how the food diary reads. This comparison does not predict what would have happened without sport — appetite and habits might have changed too.
Calories are not the whole story: with training that mixes endurance and strength work, sports guidelines put protein needs around 1.6 to 2.2 g per kilo of body weight. Over the period, protein stayed above target (147 g, ≈ 2.1 g/kg) while the balance moved into deficit — the loss happened without sacrificing muscle. Tracking also flagged one area to improve: fat intake about 30% above target.
Over seven weeks, weight went from 71.2 to 69.7 kg, a trend of about 0.25 kg per week — in two phases: a deficit close to 350 kcal per day at first, then almost balance from 22 September. The theoretical curve computed from the balance meets the last weigh-in, but this is a retrospective analysis, not a forecast announced in advance.
The value is not in collecting data, but in connecting it: meals, sport, estimated needs and weigh-ins on the same dates.
Honest limits
- Weight is not body fat: muscle gain can mask fat loss.
- The 7,700 kcal per kilo coefficient is an average and varies with body composition.
- Intake estimated from photos remains an estimate: ±10 % is normal.
- This is not a medical tool: see a health professional for unexplained weight change.
Interpret without overclaiming
The chart smooths weigh-ins with linear regression. It measures a weight trend, not body-fat change. The estimate becomes more credible with several weeks of comparable weigh-ins and a sufficiently complete food log.
- Measurement: weight entered under comparable conditions.
- Estimate: trend slope and inferred energy expenditure.
- Reference: 7,700 kcal per kilogram of body fat as an average order of magnitude.
Frequently asked questions
How long before the result is usable?
Allow three to four weeks of regular weigh-ins and logged meals. The trend stabilises beyond that, and the measured expenditure estimate gets sharper every week.
Can I log my weight by voice?
Yes. Tell the assistant "I'm 74.2 kg this morning": the weigh-in is dated and added to your history, and your current weight is updated for your daily targets.
Does measured expenditure replace my computed targets?
Not yet — it is informative: it shows the gap between theory and reality. Daily targets are still computed from your profile and your logged workouts.