Weigh yourself daily if you comfortably can, and at least once a week if you cannot. Daily readings give the trend line the most to work with; once a week is the floor MacroDiet's expenditure estimate needs to keep updating. In between, three mornings a week is a sensible compromise. What follows is why those numbers are what they are, and what happens to the maths when you miss days.

illustrative synthetic data76 kg77 kg78 kg79 kg80 kg81 kg82 kgday 1day 8day 15day 22day 29day 36scale readingtrend weight (EMA, 7-day half-life)start 81.1 → trend 78.7 kg
Fig. Daily scale readings (dots) against the trend weight (line) — an exponential moving average with a 7-day half-life, α = 1 − exp(−ln 2 / 7) ≈ 0.0943, the same smoothing Zyra Workouts applies. Each new reading moves the trend only about 9% of the way toward it, so a single heavy or light morning barely registers while a real drift shows within a week or two. Illustrative synthetic data, not a person's log.

What the algorithm needs

Two different things consume your weigh-ins, and they have different appetites.

The trend line is an exponential moving average with a 7-day half-life (the formula is in trend weight vs scale weight). It works with whatever readings exist. More readings mean more noise averaged out and a faster response to real change; fewer readings mean each one carries more weight. It never refuses to run.

The expenditure estimate in MacroDiet is different. It compares how much you ate with how your trend weight moved, and it cannot do that arithmetic without both halves. The app's own rules, from its tutorial cards:

RequirementNumberWhat happens otherwise
To start updating3 consecutive logged daysThe starting estimate from the formula stands
Food logging, per 7-day periodat least 3 daysEstimate pauses ("holding")
Weigh-ins, per 7-day periodat least 1Estimate pauses ("holding")
To feel dialled in20–30 days of consistent trackingEstimate is still ramping

The asymmetry — three food days but only one weigh-in — is deliberate. A missing food day cannot be guessed with any accuracy, because intake varies a lot from day to day. A missing weigh-in can be guessed quite well, because body weight rarely moves far between two readings a few days apart. So the weight floor is low, and the nutrition floor is higher.

In plain terms: the trend line likes daily data but tolerates anything; the expenditure estimate needs one weigh-in a week to stay awake.

Daily is easiest for the trend, not required

Daily weigh-ins are the ideal input for a smoothing algorithm, and also the easiest habit, because a habit tied to "every morning" survives better than one tied to "Tuesdays". With seven readings a week, an odd reading is one of seven and the trend barely notices it (a 2 kg blip moves it about 0.19 kg). With one reading a week, the same blip is the only information the week contributes, so the trend moves the same 0.19 kg on that reading but has nothing to correct it against until next week.

CadenceReadings per 28 daysWhat you getThe cost
Daily28Smoothest line; real change visible within 2–3 weeksSeeing the noisy number every day
3 × per week12Nearly as smooth; still fastRemembering which days
Weekly4Meets MacroDiet's floor; coarse lineOne odd morning colours the whole week
Less than weekly< 4Trend still computesMacroDiet's estimate holds between readings

The choice is about you, not the maths. If a daily number on the scale pulls your attention toward the wrong thing, weigh three mornings a week and look only at the trend. If you can treat the reading as data, daily is simplest. This is arithmetic, not a trial result, and there is no right answer beyond "consistently".

Gaps and how they are filled

Missing days are ordinary and handled. Two approaches exist, and Zyra uses both depending on the app.

Interpolation. Trend algorithms of the kind MacroDiet's method descends from fill a gap by drawing a straight line between the readings either side. If you weighed 81.0 kg on Monday and 80.6 kg on Wednesday, Tuesday is filled as 80.8 kg before the smoothing runs. Over a week-long gap the same rule spreads the difference evenly across the missing days. The filled days carry no new information, but they stop the smoothing from treating the first reading after a gap as a sudden jump.

Continue from the next reading. Zyra Workouts merges one weight per day from your entries (and from Apple Health where you have connected it; a value you typed wins over an imported one) and runs the smoothing over the days that exist. After a gap, the next reading moves the trend by the usual 9.4% of the difference — no more, no less — so a long gap simply means the line picks up where the data resumes.

Either way, nothing is lost by missing days. In MacroDiet the expenditure estimate pauses while the weight floor is unmet and resumes with the next entry. In Zyra Workouts the consistency screen shows the last 42 days as a grid, one cell per day, today last, so you can see your own cadence rather than guess it.

The short version

Daily weigh-ins give the trend line its best input; one weigh-in per 7-day period (with three logged food days) is the minimum MacroDiet needs to keep its expenditure estimate updating; three mornings a week is a good middle. Gaps are filled by interpolation or simply continued from the next reading, and a paused estimate resumes with the next entry. Pick the cadence you will actually keep, under the same morning conditions, and let nutrition and training decisions follow the trend. Educational overview only — not medical advice.