Why your calorie deficit stops working, and what to do about it

The rule everyone uses predicts something impossible

Cut 500 calories a day, lose a pound a week. It is on the NHS website, it was in the American guidance, and it is the arithmetic behind the calculator on this page. It is also, taken literally, a prediction that you will keep losing a pound a week forever until you disappear.

A group led by Kevin Hall modelled what actually happens. For a 100 kg sedentary man cutting 480 calories a day, the static rule predicts 22 kg lost in the first year. Their dynamic model predicts about half that, with weight settling near 75 kg only after several years: half the eventual change arrives in roughly a year, 95 percent of it after about three.

Their words for it are unusually blunt for a Lancet paper. Official recommendations, they wrote, erroneously state that reducing intake by about 2 megajoules a day will produce slow steady loss of about half a kilo a week, and the widespread use of this rule has led to drastically overestimated expectations.

Real trials agree. Across seven studies where intake was actually supervised, people lost 20.1 pounds against the 27.6 the rule predicted.

The error is not the number 3,500

This is where most retellings go wrong, and Hall published a follow-up specifically to correct them. He was explicit that readers might wrongly conclude the deficiency of the rule is the numerical value 3,500.

It is not. A pound of fat tissue does store roughly that much energy, and he notes the figure is about right for modest weight changes in people with overweight or obesity. The problem is that using a more accurate energy value does not repair the rule, because the rule's real assumption is that your deficit stays 500 calories a day while you get lighter.

It does not. A smaller body burns less at rest and less to move. Your 500-calorie deficit against your old maintenance becomes a 300-calorie deficit against your new one, then 100, then nothing. That is the plateau, and it is arithmetic rather than willpower.

Hall offers a replacement worth remembering: every 10 calories a day of sustained change produces about a pound of eventual weight change, with half of it arriving in a year.

Your metabolism does adapt, but less than the internet says

Alongside the mass effect there is a genuine additional drop, called adaptive thermogenesis: resting expenditure falls by more than the change in body composition alone predicts. The direction has been replicated for four decades. The size is contested, and anyone quoting you one number is picking a study.

The foundational work found that holding weight 10 percent below starting weight cut total energy expenditure by 6 to 8 calories per kilogram of fat-free mass per day, which is already adjusted for what you lost.

The famous large figure, roughly 500 calories a day still suppressed six years on, comes from 14 contestants in a televised competition who lost 58 kg each under extreme diet and exercise. It is real and it is not your situation.

In better-controlled ordinary conditions it is much smaller. A study of 171 women found about 54 calories a day, and crucially it did not predict who regained weight, and it was absent at one and two year follow-up. A systematic review of 33 studies and 2,528 people concluded that well-designed studies report lower or non-significant values, and that it seems attenuated or non-existent after periods of weight stability.

So: real during active restriction, probably tens rather than hundreds of calories for most people, and not a proven explanation for regain.

The uncomfortable part: the deficit you think you have

Before blaming physiology, there is a much larger and better-evidenced effect. People who believe they cannot lose weight on a low intake are, on measurement, eating considerably more than they report.

In the study everyone cites, participants with a history of diet resistance underreported their food intake by an average of 47 percent and overestimated their physical activity by 51 percent. Their measured energy expenditure was within 5 percent of what their body composition predicted. Their metabolisms were normal.

Two honest caveats. That specific finding came from 10 people, not the 224 who were screened, and it is quoted far more confidently than a sample of ten deserves. But the broader phenomenon is well replicated: reviews using doubly labelled water find intake is generally underreported across adults, children, athletes and military personnel alike.

The strongest demonstration is not about dishonesty at all. In a two-year trial with intensive support and monitoring, participants prescribed a 25 percent calorie restriction achieved 11.9 percent. Under research conditions, with people trying hard, the prescribed deficit was less than half the real one.

Bigger deficits do not scale the way you would hope

The evidence points three ways at once, and all three cut against aggression.

The loss is sub-proportional, because expenditure falls as mass does. The prescribed deficit is not the achieved deficit, and the gap widens the more you ask for. And a larger share of what you lose is lean tissue: a review found the degree of restriction positively associated with the fat-free share of loss, and a trial matching total weight loss found 8.8 percent of it was lean on a very low calorie diet against 1.3 percent on a moderate one.

What aggression does buy is short-term speed and, sometimes, better adherence to a target. In a trial of rapid against gradual loss, 81 percent of the rapid group hit the goal against 50 percent of the gradual group. So slower is not automatically more sustainable, which is the opposite of the usual advice.

But it bought nothing at the end. At 144 weeks, regain was 70.5 percent in the rapid group and 71.2 in the gradual one. The authors noted this is not consistent with dietary guidelines that recommend gradual over rapid loss.

Where the floor is, and why this one matters

US guidance puts a practical daily minimum at 1,200 to 1,500 calories for women and 1,500 to 1,800 for men, and states that diets under 800 calories should be used only in limited circumstances in a medical setting with supervision.

UK guidance is stricter about the structure. NICE says diets of 800 to 1,200 calories should be considered only within a specialist service with long-term support, and under 800 only where there is a clinically assessed need to lose weight rapidly. Either should last no more than 12 weeks, be nutritionally complete, and come with clinical supervision, an eating-disorder assessment and a medicines review.

This is the part a calculator can get badly wrong. Subtracting a fixed 550 calories from a small, older, sedentary person's maintenance can land under 1,200 without anything flagging it. If the number you are working toward is below these floors, the answer is more activity or a slower rate, not less food.

Does any of it work long term? Yes, partly

The honest answer is neither of the two you usually hear. "Diets always fail" is false and so is "just be consistent".

Most weight lost is regained. Around 70 percent by two to three years is a fair expectation from the trial data.

But a real minority sustain a meaningful loss. Roughly 20 percent of people keep off at least 10 percent of their body weight for at least a year, and maintenance appears to get easier: after two to five years of holding it, the chance of continued success rises substantially. In an eight-year trial, 27 percent of the intensive-intervention group were still 10 percent or more down.

And the difference between doing something and doing nothing persists. Across 249 trials and 59,081 people, regain was faster in the intervention groups, but the gap between them and the controls was still there five years later. Greater initial loss meant faster regain, and still meant lower weight five years on.

How to set a target that survives contact with this

Treat the number as a starting hypothesis and let two weeks of your own weight trend correct it. Your measured response is worth more than any equation, and it is the only thing that accounts for you specifically.

Expect the rate to fall. Plan for the plateau at the start so it does not read as failure when it arrives, because it will arrive whatever you do.

Pick a rate you would still choose in month four rather than the fastest you can tolerate in week one. The evidence gives no long-term advantage to aggression, and it costs more lean tissue.

Do not go below the floors. If the arithmetic points there, add activity or accept a slower rate.

And weigh the possibility that you are eating more than you think before concluding your metabolism is broken. That is not an accusation; it is the most replicated finding in this entire literature, it happens to researchers too, and it is the one thing on this list you can actually check.