This is not medical advice. General information about food tracking and weight change, not a treatment plan. If you are under 18, pregnant or breastfeeding, have ever had an eating disorder or a difficult relationship with food, or are managing a medical condition, talk to a doctor or a registered dietitian before you start counting calories or change your intake. The full list is here.

What counts as a weight loss plateau, and what is just the scale moving

There is a definition problem to get out of the way first. The scale weighs everything in you, and most of what changes in it over a fortnight is not fat.

Researchers followed 46 adults across two cohorts under ordinary free-living conditions for two weeks, measuring total body water by isotope dilution in one group and fat and fat-free mass by repeat DXA scans in the other. Average weight change was 0.26 kg with a standard deviation of 1.2 kg, which is what energy balance looks like. When they decomposed those changes, a unit change in body weight was 84% change in fat-free mass, and the energy density of the 1 to 3 kg short-term swings averaged 2,380 kcal per kg.2 Fat tissue is closer to 7,700 kcal per kg.3 So a kilo that arrives or leaves inside a fortnight is mostly water, glycogen and gut contents, and it costs about a third of what a kilo of fat costs.

That is 46 people over two weeks, and the decomposition comes from associations between measurements rather than from following one person's kilo. It still sets the scale of the noise you are reading through.

The noise also has a shape. A Finnish group analysed 4,657 daily weigh-ins from 80 adults, BMI 20.0 to 33.5, aged 25 to 62, over monitoring periods of 15 to 330 days. Weight was higher early in the week, Sunday and Monday, and fell through it. Increases began on Saturday, decreases on Tuesday. The compensation pattern was strongest in the people who lost or maintained weight and weakest in those slowly gaining.1 The authors' own conclusion is the useful sentence: weekend to weekday variation should be treated as normal rather than as a sign of weight gain.1

This is a group-level rhythm found retrospectively in 80 people, so it does not promise you a Friday low. What it does say is that comparing a Monday to a Friday, or a good week to a bad one, measures the day more than the diet.

The practical rule I use: nothing is a plateau until four weeks of seven-day averages sit flat next to each other. Three weigh-ins do not qualify, and neither does one bad week. Regular self-weighing is worth doing while you wait. A review of 17 longitudinal studies found regular self-weighing associated with more weight loss and not with adverse psychological outcomes such as depression or anxiety, although the effect sizes and the weighing frequencies varied a lot between studies.7 Those are associations in people seeking weight-loss treatment, not a randomised effect of the scale itself.

If you menstruate, there is a second layer of noise on top of this one, and it repeats on a schedule: body weight runs about 0.5 kg higher around menstruation, as extracellular water rather than fat. I went through the size of that swing and how to compare through it in the piece on calorie tracking across the menstrual cycle.

The 3,500-calorie rule set you up to be disappointed

Most people arrive at a plateau already holding a prediction, and the prediction is where the trouble starts. It comes from a 1958 paper by Max Wishnofsky, who reasoned that a pound of adipose tissue holds about 3,500 kcal and concluded that a 3,500 kcal deficit equals a pound lost. Half a century later it is in textbooks, on clinic and consumer health sites and, by the authors' count, in thousands of citations across the literature and the lay press.3 It is also the arithmetic behind any tool that tells you the date you will hit your goal weight, since dividing a deficit by 3,500 is the only way to produce one.

It has been known to be wrong for a long time, and the size of the error is not subtle. Applied to the participants of the CALERIE I calorie-restriction study, Wishnofsky's rule over-predicted weight at 12 weeks with a bias of 4.8 kg, against 0.4 kg for a modern dynamic model. By 24 weeks the over-prediction had grown to 11 kg, against 2.2 kg.3 Take the familiar web advice about dropping one can of soda a day: the 3,500 rule says 15 lb a year, while three validated dynamic models put the same change at 5.7, 8.2 and 8.4 lb for a representative 30-year-old woman of 76 kg.3

The reason is that the energy in a kilogram of lost weight is not a constant. Measured in the same CALERIE participants, it was 4,858 kcal per kg at week 4 and 6,569 by week 24, rising as glycogen ran down and the loss shifted toward fat.3 Early weight is cheap and late weight is expensive. The same deficit therefore buys fewer kilos per month as the months go by, which on a chart looks exactly like a plateau and is not one.

The authors of that review are careful about their own side of it. The dynamic models are built from energy and metabolism data that carry measurement error, they cannot capture individual biological variation, and they describe the validation work as started rather than finished.3 Nobody has a formula that tells you what next Tuesday weighs.

This is why the app I build gives you a daily calorie target and no date. A target is arithmetic on your current body. A projected goal date would be the 3,500 rule wearing a nicer interface, and it would be wrong by kilograms within six months.

Weight responds slowly, and the half-time is about a year

The other half of the expectation problem is time. A modelling paper in the Lancet that simulated energy expenditure adaptations during weight loss found that the bodyweight response to a change in energy intake is slow, with half-times of about one year, and that people with greater adiposity have a larger expected loss for the same intake change but take longer to reach their new steady state.4

Half-time means that after about a year of holding a new intake you are roughly halfway to the weight that intake eventually settles at, not all the way to it. The approach is asymptotic, so the weekly loss shrinks continuously and the last part of it takes the longest.

That paper is population-level modelling built for a question about the obesity epidemic, not a prediction for one person. But the shape it describes is the one to hold in your head when month four is slower than month one: a slowing rate is what a working deficit does, and a real plateau, the point where loss stops altogether, sits years out rather than months.

What the plateau models actually blame

Clinical weight loss in individuals typically stabilises at around six months, while validated models of dynamic energy balance consistently put the plateau between one and two years.5 Something makes real people stop losing roughly three times sooner than the physics says they should.

A group led from Montclair State and Pennington Biomedical built two models to find out which explanation could produce the six-month plateau, setting and validating them against NHANES, the CALERIE weight-loss study, the Bouchard twin overfeeding study and the Minnesota Starvation Experiment. The first model asked how much metabolic adaptation would be needed. It increased the final weight, but it did not move the predicted plateau to six months. The second model added a randomly fluctuating intake term standing for intermittent lapses. It produced the oscillating weight curves that show up all over the weight-loss literature, and it reached a six-month plateau at what the authors describe as levels that can be considered high diet adherence.5

Their conclusion is that an intermittent lack of diet adherence, not metabolic adaptation, is a major contributor to the frequently observed early weight-loss plateau.5

That sentence gets read cruelly, and the cruel reading is wrong. Nobody in the model is lying about their food. A level of slippage most of us would call doing well, a few days a month that go sideways, turns out to be arithmetically enough to flatten the curve at six months on its own. And it is a modelling result: it shows that lapses are a sufficient explanation, not that adaptation never matters to anybody.

The slippage does not have to be large. Pooled results from five validation studies that used recovery biomarkers as a reference put average under-reporting of energy intake at 28% with a food frequency questionnaire and 15% with a single 24-hour recall, with BMI, education and age predicting who under-reports.6 Those are research instruments rather than phone apps, so do not transplant the percentages onto your log. The direction transplants fine, and I went through where it comes from in how accurate calorie tracking apps actually are.

Your target moved while you were not looking

There is one more subtraction, and it is the least mysterious thing in this article. The deficit you set is the gap between your intake and what your body costs to run, and the second number falls as you get smaller.

Run it through MyPlate's own arithmetic. A 35-year-old woman, 165 cm, low activity, losing weight: at 82 kg the app puts maintenance at 2,083 calories and her target at 1,770, a gap of 313. Six kilos later, at 76 kg, maintenance is 2,000. If she is still eating 1,770 her deficit is now 230 calories, about a quarter smaller than the one she started with, and she changed nothing to make that happen.

Now the part where I criticise my own app. MyPlate does not recompute your calorie target when you log a weight. The target only changes when you go back through the setup questions from Profile, and until you do, it is a plan for the body you had when you last answered them. That is true of most trackers and it is not obvious from the screen, so a stall at month three is worth checking against a freshly calculated number before it is treated as a mystery. You can run the same equations on the calorie calculator without opening the app.

Measured expenditure can also fall by more than body size alone predicts, which is a real effect with real evidence behind it. I went through what it does and does not explain in the piece on maintenance calories and regain, including the six-year Biggest Loser follow-up that gets quoted for it. For a plateau specifically, the relevant finding is the one above: modelled adaptation changes where you end up, not when you appear to stop.5

What to check before you cut your calories again

Cutting is the instinct, and it is the last item on this list rather than the first, because five cheaper things come before it, in the order I would work through them.

  • Confirm it is actually flat. Four weeks of seven-day averages, compared against each other, not three weigh-ins on three mornings: a fortnight's swing is mostly fat-free mass and water.2 Weigh at the same time of day, and expect the week to have its own rhythm.1
  • Recalculate your target for the body you have now. Six kilos is worth roughly 70 calories a day off a target in the example above, and no app does this for you unless you ask it to.
  • Log an ordinary week, weekends included. A log that covers Monday to Thursday and stops describes the four days that were never the problem.
  • Count what does not feel like a meal. Cooking oil, drinks, the spoonful while cooking, what is left on someone else's plate. The modelling says the gap that flattens a curve is small, which means it hides in things this size.5
  • Look at the trend, not the week. If your average is falling at all, even 200 g a week, the deficit is working and the only thing wrong is the rate you expected: the response to an intake change is slow, with half-times of about a year.4
  • Only then adjust, and adjust once. A 100 to 150 calorie change is plenty. Whatever you do, do not go below 1,200 calories a day if you are female or 1,500 if you are male: the NIH clinical guidelines put the floor there, and eating under it belongs with a clinician rather than an app.9

Change one thing at a time and give it four weeks. Cutting calories, adding cardio and starting a new logging habit in the same week means that whatever happens next, you will not know which of the three did it. Four weeks is also roughly the shortest window in which a change in the trend can be told apart from the noise described at the top of this article.

What the evidence does not settle: diet breaks

The popular answer to a plateau is a planned break at maintenance. There is one well-known trial behind that, and it is worth reading properly rather than as a headline.

The MATADOR study randomised 51 men with obesity to 16 weeks of continuous energy restriction, or to intermittent restriction delivered as eight two-week blocks of restriction alternating with seven two-week blocks of energy balance, 30 weeks in total. Intake during restriction was 67% of maintenance requirements in both arms. Among the participants who completed per protocol, 19 continuous and 17 intermittent, the intermittent group lost 14.1 kg against 9.1 kg, and 12.3 kg of fat against 8.0 kg, while fat-free mass loss did not differ. The fall in absolute resting energy expenditure was no different between the groups, but after adjusting for changes in body composition it was significantly smaller in the intermittent arm.8

Now the limits, which are not small. Fifty-one men were randomised and 36 finished per protocol, so the headline comparison rests on two groups of under 20. Everyone was male and everyone had obesity. And the intermittent arm spent 30 weeks getting its result against the continuous arm's 16, so what was shown is more loss per week of restriction, which is not the same as faster.

I am not going to tell you a two-week break fixes a plateau, because one trial in 36 men does not support that, and nothing in the plateau modelling above predicts that it should. What a planned spell at maintenance definitely does give you is a measurement: hold your weight steady for a fortnight and you have found roughly what your body currently costs to run, which is a number worth having whatever you do next. That is the subject of the maintenance calories piece.

Common questions

How long does a weight loss plateau last?

Long enough that most of the short ones were never plateaus. Body weight over two weeks in free-living adults changes mostly as fat-free mass, and the 1 to 3 kg swings carry an energy density of about 2,380 kcal per kg against roughly 7,700 for fat tissue, so a fortnight of flat or rising weight is inside the normal range of water and gut contents.2 Weight also runs higher early in the week and falls through it.1 Treat four weeks of flat seven-day averages as the threshold for calling it a plateau at all, and expect a genuine one to need a change to your intake or your expenditure to end it.

Why am I not losing weight in a calorie deficit?

Usually because one of three things is true, and only one of them is about your body. Your deficit may be smaller than it was, since maintenance falls as you get lighter and most apps do not recalculate the target when you log a weight. Your intake may be higher than your log says, which is the ordinary case rather than the shameful one: a model that adds intermittent lapses at what its authors call high levels of adherence produces a six-month plateau on its own, while a model of metabolic adaptation did not move the plateau's timing at all.5 Or you are losing weight and cannot see it yet, because the response to an intake change is slow, with half-times of about a year.4

Should I eat less when my weight loss stalls?

Not first. Check that it is really flat over four weeks, recalculate your target for your current weight, log an ordinary week including the weekend, and count the things that do not feel like meals. If you still need to adjust after that, a change of 100 to 150 calories a day is enough, and there is a hard floor underneath it: the NIH clinical guidelines put the lowest reasonable unsupervised intake at 1,200 calories a day for women and 1,500 for men.9 Going below that is a decision for a clinician, not for an app or an article.

Has my metabolism slowed down permanently?

Measured expenditure does fall during and after weight loss, by more than body size alone predicts in some studies, but it is a poor explanation for a plateau at six months. When researchers modelled how much metabolic adaptation would be required to produce that early plateau, adaptation raised the final weight without shifting when the plateau arrived, while intermittent lapses in intake reproduced it directly.5 Both of those are mathematical models validated against trial data rather than measurements of you, so read them as the best current account rather than the last word.

Do diet breaks help when weight loss stalls?

The evidence is one trial and it does not stretch as far as the advice built on it. MATADOR randomised 51 men with obesity to continuous or intermittent energy restriction; among the 36 who completed per protocol, the intermittent group lost 14.1 kg against 9.1 kg with a smaller body-composition-adjusted fall in resting energy expenditure, but it took 30 weeks against 16 to get there.8 That is more loss per week of restriction rather than faster loss, in under 20 men per group, all of them male and all with obesity. A planned fortnight at maintenance is still useful for a different reason: it measures what your body currently costs to run.

What is the best diet tracker app for getting through a plateau?

Judge a food tracker for weight loss on three things rather than on features. It should let you log a whole ordinary week without the logging itself becoming the reason you stop, since a log that skips weekends misses exactly where the modelled lapses live.5 It should chart weight over time so you can look at a trend instead of a morning. And it should make recalculating your calorie target easy, because that number goes stale as you get lighter. No app measures your expenditure, so treat any app that claims a goal date with suspicion: applied to real trial data, the 3,500-calorie rule those dates come from over-predicted weight loss by 11 kg at 24 weeks.3

Measure for four more weeks

The scale is a noisy instrument attached to a slow process, and the four explanations above account for most of what a stalled scale turns out to be: a fortnight of water mistaken for a month of stalling, a target calculated for a heavier body, a week whose weekend never got logged, or a rate of loss that was only ever disappointing next to a 1958 rule of thumb that over-predicts by 11 kg at six months.3

So the move is almost never to eat less this week. It is to measure better for four more weeks: same scale, same time of day, seven-day averages, every day logged including the ones you would rather not. If the average is still falling, nothing is wrong. If it is genuinely flat, you now know that from data rather than from a bad Tuesday, and a 100 calorie adjustment made once, on purpose, is a different act from cutting in frustration.

One thing outranks all of it. If the flat scale has turned into something you check three times a day, or the log has started deciding whether the day went well, stop and talk to your doctor or a registered dietitian. A plateau is a measurement problem. What that feeling turns into is not, and no amount of better tracking fixes it.

Sources

Every figure above traces to one of these. If you find a number that doesn't match the source it claims, tell me and I'll correct it.

  1. Orsama AL, Mattila E, Ermes M, van Gils M, Wansink B, Korhonen I (2014). Weight rhythms: weight increases during weekends and decreases during weekdays. Obesity Facts, 7(1):36-47. Read on PubMed →
  2. Bhutani S, Kahn E, Tasali E, Schoeller DA (2017). Composition of two-week change in body weight under unrestricted free-living conditions. Physiological Reports, 5(13):e13336. Read on PubMed →
  3. Thomas DM, Gonzalez MC, Pereira AZ, Redman LM, Heymsfield SB (2014). Time to correctly predict the amount of weight loss with dieting. Journal of the Academy of Nutrition and Dietetics, 114(6):857-861. Read on PubMed →
  4. Hall KD, Sacks G, Chandramohan D, et al. (2011). Quantification of the effect of energy imbalance on bodyweight. The Lancet, 378(9793):826-837. Read on PubMed →
  5. Thomas DM, Martin CK, Redman LM, et al. (2014). Effect of dietary adherence on the body weight plateau: a mathematical model incorporating intermittent compliance with energy intake prescription. American Journal of Clinical Nutrition, 100(3):787-795. Read on PubMed →
  6. Freedman LS, Commins JM, Moler JE, et al. (2014). Pooled results from 5 validation studies of dietary self-report instruments using recovery biomarkers for energy and protein intake. American Journal of Epidemiology, 180(2):172-188. Read on PubMed →
  7. Zheng Y, Klem ML, Sereika SM, et al. (2015). Self-weighing in weight management: a systematic literature review. Obesity, 23(2):256-265. Read on PubMed →
  8. Byrne NM, Sainsbury A, King NA, Hills AP, Wood RE (2018). Intermittent energy restriction improves weight loss efficiency in obese men: the MATADOR study. International Journal of Obesity, 42(2):129-138. Read on PubMed →
  9. National Heart, Lung, and Blood Institute, NIH (1998). Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults. NIH Publication 98-4083. Read the guidelines →

The formulas and limits behind MyPlate's own numbers, with their citations, are on the health sources page.