Missing days: what a gap does to your numbers

Nobody logs perfectly. This page is the reference for what that actually costs — which numbers shrug it off, which quietly go wrong, which switch themselves off, and how long each takes.

Short version: a missing weight is cheap, a missing calorie entry is expensive, and both missing together freezes your dashboard rather than breaking it. The detail is below, by how many days you've missed.

The one rule that explains most of this

Your record runs from your first weigh-in to your last one — not to today.

Everything on the dashboard is measured inside that span. Nothing on the page knows what today's date is except the missing-days list on the Data page.

That single fact explains a lot of otherwise strange behaviour:

• If you stop logging entirely, nothing changes. Your dashboard doesn't grey out, doesn't warn you, doesn't age. It shows the same picture it showed on your last logged day, for as long as you leave it.
• "Recent · 6w" means the last six weeks of data. If your last entry was in March, Recent is a window that closed in March.
• Calories you log after your last weigh-in fall outside the record and reach no calculation at all. If you're only going to log one of the two on a given day, weight is the one that keeps the door open.

What the app does with a hole

Two different rules, one per number:

A missing weight is filled in. The app draws a straight line between the weigh-ins either side and uses the points along it. They're marked — on the Weight + Trend chart, logged mornings are blue candles and filled-in ones are amber, so you can always see which is which.

A missing calorie entry is never filled in. No intake statistic invents a number: your averages, your deficit, the Consistency band and the Diet Days breakdown all simply skip that day.

With one important exception, and it's the source of nearly everything below. The live TDEE estimate on the dial can't skip a day, because it works by comparing what you ate against what your weight did about it — a day with a weight and no calories still has to be explained somehow. So for those days it stands in your recent average intake, and widens its own uncertainty a little. When your eating during the gap looked nothing like your eating before it, that stand-in is wrong, and the estimate has to blame the difference on your metabolism — because that's the only variable it has left.

If only your weight is missing

Calories still logged. This is the cheap one.

• 1–5 days — nothing at all. Days younger than five are never even listed as missing; you may still be about to fill them in.
• 6–20 days — listed on the Data page. Weight is drawn between your weigh-ins and the trend line runs straight through. Your numbers move by a few tens of calories at most.
• 21–90 days — still fine. In testing, gaps of up to three months in weighing, with calories logged throughout, moved the TDEE estimate by less than the margin the dial already shows.

Weighing weekly instead of daily costs almost nothing — around 65 calories on the estimate, comfortably inside its own uncertainty. If daily weigh-ins are what's making you want to quit, weigh weekly and keep logging your food.

One caveat: a long straight line of filled-in weights is smoother than real data, and smooth data makes the app slightly more confident than it has earned. The number stays about right; the confidence behind it is mildly overstated.

If only your calories are missing

Weighing continues. This is the expensive one, and it's the pattern to watch for, because it degrades quietly with everything still lit up.

• 1–5 days — nothing. Inside the grace window.
• 6–13 days — listed on the Data page. The estimate starts to lean, typically by under 100 calories.
• 14–16 days — around 200 calories of drift. Every tile still green. Nothing on the page tells you.
• 17 days — Intake and Diet days go quiet: "too few logged days lately to judge." They need 5 of the last 21 days to carry calories, and you've just dropped below that.
• 21 days — Consistency goes quiet too. Its rolling band has run out of days to average.
• 3–8 weeks — the estimate keeps answering and keeps drifting: around 500 calories off by four weeks, 840 by eight. Pace and the dial stay lit the whole time.

Weighing daily while logging no food is the single worst pattern in the app — worse than stopping altogether, because a full stop at least freezes the picture, while this keeps feeding the estimate weight changes it has to explain using an intake you're no longer eating.

⚠️ The uncertainty does not keep up here. The ± margin behind the dial widens only slightly (roughly 135 to 185) while the real error runs into the hundreds. When calories are what's missing, the dial deserves more suspicion than it shows.

If both are missing

Nothing logged at all. Counter-intuitively, this is the safer of the two failures.

• 1–5 days — nothing.
• 6+ days — listed on the Data page, with an alert when you arrive there.
• Any length at all — every number on your dashboard freezes at your last logged day and stays there. Nothing greys out. Nothing drifts, because nothing is being recalculated. The reliability caution counts calendar span rather than logged days, so a period keeps whatever standing it had.

The estimate you left behind stays reasonable for a long time — in testing it remained inside its own stated margin through months of simulated absence. The risk isn't that the number goes wrong; it's that a months-old number is still labelled "where you are right now", and nothing on the page says otherwise. If you've been away a while, check the date of your last entry on the Data page before you trust the dial.

Coming back

The cost of a break is paid on your return, not while you're away.

The first day you log again, the whole gap enters your record at once: the missing weights are drawn in as a straight line, and every missing day has your pre-break intake assumed for it. If your eating changed during the break — and it usually did — the estimate takes the hit immediately.

• After 1 week away — the dial reads about 120 calories low; back to normal in around 5 weeks.
• After 2 weeks — about 230 low; around 5 weeks.
• After 1 month — about 600 low; around 6 weeks.
• After 2 months — about 900 low; around 6 weeks.

Two things are worth knowing. It always errs low — after a break spent eating more than usual, the app will tell you that you burn less than you do, which is exactly the wrong direction for setting your next deficit. And recovery takes about six weeks regardless of how long you were gone, because six weeks is how far back the estimate reaches. It isn't a penalty that scales with the break.

What to do about it

Ranked by what the measurements actually showed:

1. Fill the gap in with rough calorie guesses. This is worth more than everything else put together. Guesses that were 20% wrong cut the recovery from six weeks to about ten days; one flat number typed into every day of the break cut it to three. Use Fill it in on the missing-days list — you don't have to leave the page.
2. Don't bother back-filling weights. Weighing through a break, or adding the weights afterwards, changed the recovery by about a day. The weight was never the missing piece.
3. If you'd rather not guess, just read the dial with suspicion for six weeks. Your period numbers, your trend and your history are unaffected — it's specifically the live estimate that's rebuilding.

What doesn't change anything

• Deferring or ignoring a day. Both are bookkeeping for the missing-days list and nothing else; neither feeds a single calculation. Ignoring a day you genuinely have no data for is completely safe — it stops the app asking, and does nothing more.
• Your history. Nothing is deleted, degraded or overwritten by a gap. Fill a day in two years later and every number recomputes as though it had always been there.
• The charts. A missed day simply draws nothing — no bar, no marker, no invented point. A gap reads as a gap.

Where gaps show up on purpose

• Missing days (Data page) — lists every day inside your log that lacks either number, once it's more than five days old. Fill it or ignore it from the row.
• Weight + Trend chart — filled-in weights appear as amber candles beside the blue logged ones.
• Calorie charts — days without calories are skipped, so the line breaks rather than dipping to zero.
• The live TDEE dial — needs 21 days of record and 14 days carrying calories before it appears at all. Both count over your whole record, so once it has unlocked, no later gap can lock it again.
• Signal strip — Intake and Diet days need 5 of the last 21 days with calories; Consistency needs its 14-day band to fill; Pace needs two weigh-ins in the window.
• Achievements — a logging streak breaks on the first missed day, and a day counts as logged if it has either number. The comeback badge exists precisely for gaps of two weeks or more.

The figures here were measured by running the app's own calculations over records with gaps cut into them — both real logs and test records where the true answer was known in advance. They're typical rather than guaranteed: your own numbers depend on how much your eating changed during the gap. The underlying definitions live in Metrics & Formulas §1 (data pipeline), §2b (the live estimate), §3b (the signal strip) and §4 (charts).