A blank day is not a zero, and it is not an average day either#
When you miss a day, the move is to put a number in it rather than to leave it empty or to abandon the week. Estimate what you ate — badly is fine — and estimate it slightly above your usual day rather than at it. Then read the seven-day figure and carry on. What you must not do is let the gap sit there, because an empty cell in a log is not neutral: depending on how your tracker handles it, it either reads as zero calories or it silently vanishes from the average, and both of those quietly rewrite your week in the same direction.
The reason a missed day deserves an above-average guess rather than an average one is the part almost nobody accounts for. Missing days are not a random sample of your days. They cluster on exactly the occasions that were unusual enough to disrupt logging in the first place — the party, the trip, the long Saturday — which are also the occasions when intake runs high. A log that drops its own gaps is not merely incomplete. It is incomplete in a specific, predictable direction.
The days you skip are not the days you would like them to be#
This is measurable, and the pattern is consistent enough to plan around. In a six-month technology-supported weight-loss trial, researchers modelled within-person variation in dietary self-monitoring across time in study, day of week and month of year among 31 adults with obesity. Two things moved: recording declined as the study progressed, and "fewer foods were reported on the weekends compared with weekdays"1.
The decline itself is not unique to that trial — in a 210-person randomized comparison of logging methods, only about a third of days were still being recorded on paper by the half-year mark4, so gaps are the normal steady state of a long-running log rather than a personal failing.
That is 31 people, so treat the weekday–weekend gap as a direction rather than a coefficient. But it lines up with something better established from the intake side: the week has a repeating shape, and the weekend sits at the top of it — the subject of reviewing weekly rather than daily. Put the two together and you get the mechanism that matters here. Logging is lightest exactly when eating is heaviest. The blanks in your diary are drawn disproportionately from the upper end of your own distribution.
Which means the instinct to "just skip it and average the rest" is not the neutral, conservative choice it feels like. It is a choice to compute your average from your better days.
What a gap actually costs, three ways#
Work the arithmetic through, because the three options are not close to each other. Take a week where six days ran at your typical intake — call it X — and the seventh, the one you missed, actually ran about 10 percent above it.
| How the gap is handled | What your week reads as | Error against the real week |
|---|---|---|
| Blank treated as zero calories | 0.857X | about 15% low |
| Day dropped, six days averaged | X | about 1.4% low |
| Day imputed at your usual | X | about 1.4% low |
| Day imputed at usual + 10% | 1.014X | none |
That table is my arithmetic on those assumptions, not a published result — but the ordering is the point and it does not depend on the exact numbers. Zeroing a day is a catastrophe: it invents a fifteen-percent deficit that never happened, and it does so on the very week you most want to read honestly. Dropping the day is far better and still slightly optimistic. Imputing it, roughly, at a bit above your normal is the only one that lands.
The practical upshot is a rule you can apply in ten seconds without doing any of that arithmetic: put in a guess, and lean it high. You do not need it to be right. You need it to be present and pointing the correct way, because a stated estimate can be wrong by 20 percent and still beat a blank that is wrong by 100.
It also helps to know how little a single day was ever going to tell you. When 15 healthy women recorded food intake daily for 17 days alongside doubly labeled water, the intraindividual coefficient of variation for energy intake averaged ±25 percent, ranging from 16 to 34 percent across individuals3. Your own days already scatter by about a quarter around your own mean. One of them going missing is a smaller event than it feels like — and one of them being guessed is a smaller event still.
The lapse is cheap. The interpretation is expensive.#
The real damage from a missed day is rarely arithmetic. It is what happens next, and there is a specific psychological pattern with a name and some data behind it.
Marlatt and Gordon described the abstinence violation effect: after a period of holding a rule, a single violation produces a cascade of guilt and self-blame that makes further violation more likely — the lapse is read as evidence about you rather than about the day. It has been tested in a dieting context. Among 76 patients enrolled in a very-low-calorie-diet and behaviour-education programme, 41 lapsed within the first 11 weeks and rated their attributions for that first lapse. Those who attributed it to their own character rather than to circumstances lost a smaller percentage of their excess weight: r(39) = −.36, p < .0252.
Several caveats are load-bearing here. That is a correlation in 41 people from 1992, in a very-low-calorie-diet population that does not resemble ordinary tracking, and the same paper found that first-lapse attributions did not predict dropout overall — the one dropout-related correlation it reports rests on 14 people. So this is a suggestive result about how lapses get read, not a demonstration that a thought pattern causes weight regain.
What survives the caveats is the framing, and the framing is useful. "I missed Saturday" and "I can't do this" are two different sentences about the same blank cell, and only one of them ends the project. A tracking gap is a data problem with a ten-second fix. Treating it as a verdict is what turns one missing day into ten, and it is the same all-or-nothing posture that separates tracking that stays useful from tracking that stops being.
Filling a day you did not log#
A workable procedure, in descending order of effort — use the first one you can actually manage.
- Reconstruct it from anchors. Bank card entries, photos on your phone, the empty packaging in the bin, who you were with. Most days leave more evidence than you expect, and you are only aiming for the big items.
- Copy a comparable day. If last Saturday and this Saturday were structurally similar, duplicate the entry and adjust the one or two things you remember being different. This is much better than it sounds, because the repeating parts of your diet are the parts you are least likely to have gotten wrong.
- Use a typical-day default, adjusted up. Have a saved "ordinary day" figure and a saved "big day" figure. Missed days get the big-day figure unless you have a reason to think otherwise.
- Mark it as estimated. Not for accuracy — for reading. When you look back at a month, you want to know which days were measured and which were reconstructed, so a flat trend built on three imputed days gets the confidence it deserves and not more.
None of this recovers the information you lost. It just stops the gap from lying. And the goal was never an unbroken record — logging half your days has been enough to produce weight loss in trials, which is the argument the counting workflow makes at length and the reason starting small beats starting perfectly. A week with one guessed day is a week you can still read. A week you abandoned because Saturday was blank is not.
FAQ#
Does one missed day ruin a week of tracking?#
No, and the arithmetic is reassuring in one direction and alarming in the other. If you estimate the missing day, even roughly, your weekly figure lands within a couple of percent of the truth. If you leave the cell empty and your tracker averages it as zero, the same week reads about 15 percent lower than it was — my arithmetic, on a week of six typical days plus one slightly bigger one. The missed day is not the problem. The blank is.
Is it better to guess at a missed day or leave it blank?#
Guess, and lean the guess high. Days go unlogged disproportionately when something disrupted the routine, and disrupted days tend to be bigger ones — self-monitoring in one trial was measurably lower on weekends than weekdays1, which is the opposite of when intake peaks. An estimate that is 20% wrong is a far better input than an absence, and it stops your average being computed from your easiest days.
Why do I always stop logging on weekends?#
Because logging is anchored to routine and weekends are where routine goes. In a six-month trial, fewer foods were recorded on weekends than on weekdays, and recording declined steadily as the months passed1. The fix is not more resolve on Saturday; it is a lower-effort weekend version — one photo per meal, or a saved "typical Saturday" entry you adjust — plus a standing rule that Sunday evening fills in whatever is missing.
Sources#
- Pellegrini CA, Conroy DE, Phillips SM, Pfammatter AF, McFadden HG, Spring B. Daily and Seasonal Influences on Dietary Self-monitoring Using a Smartphone Application. J Nutr Educ Behav. 2018;50(1):56-61.e1.
- Mooney JP, Burling TA, Hartman WM, Brenner-Liss D. The abstinence violation effect and very low calorie diet success. Addict Behav. 1992;17(4):319-324.
- Champagne CM, Han H, Bajpeyi S, et al. Day-to-day variation in food intake and energy expenditure in healthy women: the Dietitian II Study. J Acad Nutr Diet. 2013;113(11):1532-1538.
- Burke LE, Conroy MB, Sereika SM, et al. The effect of electronic self-monitoring on weight loss and dietary intake: a randomized behavioral weight loss trial. Obesity (Silver Spring). 2011;19(2):338-344.
