A stall is common; arrival is rare, slow, and loud#
If the scale has been flat for a month and you are close to the weight you were aiming for, the tempting reading is that you have arrived — that this is the weight your body intends to hold, and further effort is arguing with physiology. That reading is almost always wrong, and the reason is arithmetic rather than motivation. Under a genuinely constant intake, the approach to a new stable weight takes years, not months, so a flat month at the six-month mark is far more likely to be a deficit that closed than a floor you reached.
And a real floor is not quiet. When people are pushed to the actual lower limit of what their body will hold, the scale is the least of the signals: hormones, mood, blood pressure and strength all move, and they move a long way. So the discrimination you need is not subtle. Ask whether the clock supports arrival, and ask whether your body is behaving like something under pressure. If the answer to both is no — which it usually is — you have a stall of the ordinary sort, and why weight-loss plateaus happen covers the causes. Whether a defended weight exists as a fixed number at all is a separate and genuinely unsettled question, taken up in set point theory.
The arithmetic that makes the last five kilograms unreadable#
Two things degrade at once as you get lighter, and together they make the end of a diet much harder to read than the beginning.
The first is the size of the signal. Loss rates are conventionally set as a percentage of body weight, so the same percentage buys fewer grams as the percentage applies to less. Someone at 100 kg losing 0.7% a week is shedding 700 g; the same person at 65 kg is shedding 455 g. The scale's day-to-day scatter, meanwhile, does not shrink with you — it is set by fluid, gut contents and glycogen, whose magnitudes are covered in plateau versus normal fluctuation. Signal falls, noise holds, and the number of weeks you need before a trend means anything grows accordingly.
The second is the size of the deficit relative to the error in measuring it. A 500 kcal daily gap is 18% of a 2,800 kcal expenditure and 26% of a 1,900 kcal one — but the absolute error in your own logging is roughly the same in both cases. The lighter you are, the larger a share of your intended deficit a constant logging error can consume. This is why "recount before you conclude" is such a durable first move near goal: not because you are careless, but because the same slip that was noise at 100 kg is decisive at 65.
What the clock says about arriving#
Here is the part almost nobody has been told. Mathematical models of body weight predict that under a constant calorie intake, weight approaches its new steady state exponentially — quickly at first, then ever more slowly, never quite finishing. When two widely used models were tested against two years of data from the CALERIE calorie-restriction trial, the better-performing one (the NIH Body Weight Planner) had an exponential time constant of 414 days in women and 340 days in men1.
Work that constant forward, which the paper leaves to the reader: after a full year at an unchanged intake, a woman is roughly 59% of the way to her final weight and a man roughly 66%. Getting most of the rest takes another year or more.
Arrival is a multi-year event. If you have been at this six months, the flat month you are staring at is not the end of the curve — it is a section of it you cannot resolve yet.
What the calendar does reliably show is that the last stretch takes longer than a naive projection says, and that the delay has a measurable cause. Sixty-five premenopausal women ate 800 kcal a day until their BMI fell to 25 or below. They lost 12.5 ± 3.1 kg — about 16% of body weight — over 155 ± 49 days, with a range of 66 to 252 days to hit the same target. Metabolic adaptation at the level of resting metabolic rate averaged −46 ± 113 kcal/day, and predicted how long the journey took: roughly one extra day per 10 kcal/day of adaptation, so the most adapted participants needed about 70 days more than the least2.
Read the error bars there, because they are the finding as much as the mean is. An adaptation of −46 kcal/day with a standard deviation of ±113 means a large share of that sample adapted in the other direction. Slowing is real on average and absent in plenty of individuals — which is also the conclusion reached from the other end by metabolism after weight loss. Seventy days is a meaningful delay and it is not a wall.
What a real floor looks like from the inside#
The clearest published picture of a body at its actual limit is a single drug-free male bodybuilder, tracked for six months of contest preparation and six months of recovery. Body fat went from 14.8% to 4.5%. Testosterone fell from 9.22 to 2.27 ng/mL. Resting heart rate dropped from 53 to 27 beats per minute and blood pressure from 132/69 to 104/56 mmHg. Total mood disturbance rose from 6 units to 43. Strength fell and had not fully recovered six months later, though body fat, testosterone and mood all had3.
That is one person, and a case study proves nothing about frequencies. But it is the shape of the thing, and the shape is the useful part: a genuine physiological floor is a systemic event with a symptom list, not a quiet week on the scale. Nothing in that profile is subtle or deniable.
There is also an objective line for the input side. Luteinizing hormone pulsatility in regularly menstruating women is undisturbed at an energy availability of 30 kcal per kilogram of lean mass per day and breaks down below it4 — energy availability being intake minus exercise expenditure, divided by lean mass. If you are comfortably above that line, whatever is stalling your scale is not your body running out of room.
Deciding without a scan#
Put the two questions side by side and the answer usually falls out.
| Signal | Points to an ordinary stall | Points to a floor |
|---|---|---|
| Time in deficit | Under a year | Multiple years at a genuinely stable intake |
| Body composition | Still carrying fat you can pinch | At or near the lean end for your sex |
| Energy availability | Comfortably above 30 kcal/kg lean mass | At or below it |
| Sleep, libido, mood, training | Unremarkable | Several visibly degraded at once |
| Recount of intake | Not done recently, or shows drift | Weighed and verified for a fortnight |
Most people reading this will land in the left column on four rows out of five, and the honest conclusion is that the deficit closed rather than the body objected. If you land in the right column, the useful move is not a bigger cut — it is to decide whether the weight below you is worth what the profile above costs, and to set a target you can hold. That decision is really a goal-setting problem, handled in setting a realistic weight goal, and what comes after it is a maintenance problem, which has its own evidence and its own habits. Arriving somewhere you can stay is a better outcome than briefly touching a number you cannot.
FAQ#
How do I know if I've reached my body's lowest sustainable weight?#
By symptoms, not by the scale. A body at its actual limit shows a systemic picture — in one tracked bodybuilder, testosterone fell from 9.22 to 2.27 ng/mL, resting heart rate from 53 to 27 bpm, and mood disturbance rose sevenfold at 4.5% body fat3. If your sleep, mood, libido and training are all fine and only the scale is flat, you are looking at a stall, not a floor.
Why does the last 5 kg take so much longer than the first 5 kg?#
Three effects stack. Loss rates scale with body weight, so the same percentage buys fewer grams. Metabolic adaptation adds time in proportion to its size — about one extra day per 10 kcal/day, up to roughly 70 days in the most adapted2. And weight approaches a new steady state exponentially, so progress is slowest exactly at the end: a time constant of 414 days in women means a year of unchanged intake covers only about 59% of the distance1.
Should I cut calories further to finish the last few pounds?#
Usually not first. Near goal the deficit is already a large fraction of a smaller expenditure, so a further cut buys little gap and costs disproportionately in adherence and lean mass. Verify the intake you think you are eating before changing it — a constant logging error consumes a much bigger share of a small deficit than a large one. If a verified deficit still produces nothing over a full year, the question has stopped being about the number and started being about whether this weight is one you want to defend.
Sources#
- Guo J, Brager DC, Hall KD. Simulating long-term human weight-loss dynamics in response to calorie restriction. Am J Clin Nutr. 2018;107(4):558-565.
- Martins C, Gower BA, Hunter GR. Metabolic adaptation delays time to reach weight loss goals. Obesity (Silver Spring). 2022;30(2):400-406.
- Rossow LM, Fukuda DH, Fahs CA, Loenneke JP, Stout JR. Natural bodybuilding competition preparation and recovery: a 12-month case study. Int J Sports Physiol Perform. 2013;8(5):582-92.
- Loucks AB, Thuma JR. Luteinizing hormone pulsatility is disrupted at a threshold of energy availability in regularly menstruating women. J Clin Endocrinol Metab. 2003;88(1):297-311.
