It's Wednesday morning and Tuesday is empty. Maybe you were travelling, maybe you were ill, maybe the day just got away from you and you noticed at ten at night, when opening the app felt worse than not bothering.
Either way there's a gap in the record now, and before you've finished your coffee you're making a decision about it.
There are three things people do here. Rebuild Tuesday from memory. Decide the week is spoiled and pick it up properly on Monday. Or open the app, see the gap, feel vaguely bad, and gradually stop opening the app.
Only one option is actually cheap, and it's the one nobody recommends: leave Tuesday empty and log your breakfast.
Don't fill it in from memory
The reconstruction instinct is the strong one, and I understand it. An empty row looks like a failure and a filled row looks like diligence.
But estimating a meal you're looking at and estimating a meal you half-remember are not the same act. The first one is a guess with a handle on it — you can say what you assumed, and correct it in four words when the assumption turns out wrong. The second is a guess about a guess, made by someone who wasn't taking notes.
What makes it worse is the direction the error runs. Memory doesn't fail randomly, it fails by omission, and the things it omits are always the same things. You remember dinner. You do not remember the handful of something out of the bowl at four, the second pour, the two bites of your kid's food you ate standing up. Every one of those is missing from your reconstruction, and none of them is missing in the other direction.
So a day rebuilt on Wednesday isn't a fuzzy record of Tuesday. It's a confident record that leans low, every time, for the same reason — which is the exact kind of error that actually moves your numbers, as opposed to the kind everybody frets about.
An empty day is honestly empty. A reconstructed one launders a guess into the record as though you'd been standing there when it happened.
An empty day is not a light day
Here's the part an app has to get right, and where the empty row you're worried about does its real damage.
If Tuesday has nothing in it, the arithmetic can read that two ways: you ate nothing, or nothing was written down. Those are wildly different facts and they look identical in the database. Read the wrong one and a day you forgot about becomes a day you were brilliantly disciplined, and the app quietly congratulates you on a deficit you never ran.
Graham's rule is that a day counts as complete when at least two of its meal slots have something in them — breakfast, lunch, dinner, snack — and the total reaches 60% of that day's calorie target. That bar is deliberately low. A day of rough guesses, every figure carrying a tilde, clears it without trouble — fuzzy days count fully, because fuzzy data is still data. Worth knowing that it counts slots and not entries, so three things all logged as snacks is one meal by this rule, not three.
What doesn't clear it is a day with a coffee and half a sandwich in it. That isn't a light day, it's a day you stopped logging at eleven. It still shows up in your history exactly as you left it, and it stays out of the math that decides anything.
Days with nothing at all in them aren't in the ledger to begin with. They count by their absence, not as a number.
And today never counts against you. A report pulled at nine in the morning has no business reading your breakfast as a failed day.
You can miss four days in a fortnight
Graham looks at the last fourteen days, not counting today, and wants 70% of them complete — ten of fourteen — before he'll suggest changing anything about your plan.
So: four missing or half-empty days in a fortnight, and everything runs exactly as it did before. Nothing resets. Nothing is lost. Nobody mentions it.
I'd rather print that number than have you guess at it. The reason a streak counter is so good at making people quit is that the cost of one miss is never stated, so it feels total — and a cost that feels total is one you stop trying to pay. A budget you can see is a budget you can spend. You have four days. Tuesday was one of them.
What the gaps actually cost
Less than the guilt suggests, and not the thing you'd assume.
Your weight trend doesn't need your food log at all. It's built from scale readings — a week of them, averaged, against the week before. Log nothing for a fortnight and the number still arrives, because standing on the scale is the whole input. Whatever progress you made is still there, and no amount of missed logging deletes it.
What the gaps change is how much weight that number is allowed to carry.
When the trend comes in off pace, the next move is to change your calorie target. That's a much bigger claim than reading the scale — it says we know what you ate, and that the eating is what needs adjusting. Below that 70%, Graham won't make it. You still get your rate and your direction, but they arrive with the caveat attached — too many days are missing to read the trend confidently — and if you ask him about it he says the same thing, and asks for a fuller week before he reads anything into it.
That's the honest shape of the loss. A gap doesn't erase your progress — the scale says what the scale says, and being a pound and a half lighter than you were a fortnight ago is true whether or not you logged Tuesday. What the gaps leave open is why. Whether that pound and a half means the target is right, or that you drifted somewhere under it without noticing, is the question the log was supposed to answer — and it's the one that decides what to change next.
What this doesn't fix
The completeness rule can't tell a remembered meal from one logged at the table. If you reconstruct Tuesday convincingly enough, it'll pass as a complete day and feed the math like any other, and nothing in the code will catch it. What the check measures is whether a day was recorded, not whether it was true.
Which is why the first section is advice rather than a feature. The only person who can tell the difference is you, on Wednesday morning, working out whether you genuinely remember lunch or just remember having had one.
So leave it empty. It's already the most accurate thing you can put there, and it costs you one day of explanation out of a budget that holds four.
Starting over on Monday costs you the five days in between.
— Ryan