Why weight loss slows down: the 3,500-calorie rule and what replaced it

By Robert Robinson, founder of Neato Health · September 15, 2026 · 4 min read

The arithmetic says a pound a week forever. Your body says otherwise. Here is the research on why every diet slows, how much of the slowdown is expected, and how Neato keeps the target honest as you shrink.

Every calorie calculator you have ever used made the same promise: a pound of fat is 3,500 calories, so eat 500 fewer a day and lose a pound a week, indefinitely. It is true for about a month. After that the pounds come slower on the same intake, and the slowdown is not a failure of discipline. It is physics, and it has been measured.

Where 3,500 came from

A pound of adipose tissue holds roughly 3,500 calories of energy, a figure that traces to work in the 1950s. The number is not wrong. The mistake is treating the deficit as constant. A 500-calorie deficit is 500 calories below your burn, and your burn is not fixed. It falls as you lose weight, for three reasons, so the same intake that was 500 below your burn in January is 300 below it by April.

Three reasons the burn falls

You are carrying less

The resting burn equation in the resting-burn guide moves about 10 calories per kilogram of body weight. Lose 15 kilos and your resting burn is down 150 calories a day before anything else happens. Movement costs less too: walking a mile at 70 kilos burns less than at 85.

The body adapts beyond the arithmetic

In 1995, Leibel, Rosenbaum and Hirsch measured energy expenditure in people held at their usual weight, then after a 10 percent loss, then after a 10 percent gain. After the loss, expenditure fell by more than the lost tissue explained, by about 15 percent of the day’s total, and the reduction was mostly in the non-resting part: the body moved more economically. After the gain, expenditure rose more than expected. The body defends its weight in both directions, and the effect persisted as long as the new weight did.

The extreme version was documented by Fothergill and colleagues in 2016, who followed contestants from a television weight-loss competition for six years. Their resting burn was, on average, 500 calories a day below what their new weight predicted, and the gap had not closed in six years. That was after a diet far more severe than anything Neato would plan, but it makes the direction unmistakable.

Less food, less digesting

Digesting food costs about 10 percent of the calories eaten. Eat 500 fewer and you save 50 calories of digestion, which is 50 calories of burn you no longer have.

The model that replaced the rule

Hall and colleagues published the modern account in 2011: a mathematical model of body weight change, validated against controlled feeding studies, that tracks how expenditure falls as weight falls. Its headline is a rule of thumb that replaces the old one: a sustained 10-calorie-a-day change in intake leads to about a 1-pound change in weight, but it takes about a year to get halfway there and three years to arrive. A permanent 500-calorie cut does not lose a pound a week forever; it loses roughly 50 pounds total, half of it in the first year, then nothing more, because at 50 pounds lighter you burn 500 fewer and the deficit is gone.

The model also explains the shape everyone recognizes: a fast start, then a curve that flattens toward a new equilibrium. The flattening is the deficit closing, not the diet failing.

What Neato does about it

• The target is recomputed from your latest weight. The formula is rerun every time you weigh in, so the resting estimate drops as you do and the deficit stays the size the plan intends.
• The scale corrects the estimate. The adaptation Leibel measured is not in any formula, but it is in your data: if you are eating what the plan says and losing slower than the plan expects, your real burn is lower than the estimate, and the calibration described in the scale guide walks the estimate down to meet it. The plan adjusts to your body rather than the average one.
• The rate is a percentage. Neato’s cut rate is a fraction of body weight per week, so the planned weekly loss shrinks as you shrink. That is the safe direction, as the fat-loss guide explains, and it matches what the model says is possible.

Reading a plateau

A true plateau, several weeks with no change in the trend line, has three common causes and the log tells them apart. The deficit has closed because the target was never lowered: check that the target has fallen with your weight. Intake has crept up: compare this month’s logged average with the first month’s. Or the scale is hiding a real loss behind water: give it a week and read the trend, not the day.

What a plateau is not, in almost every case, is a broken metabolism. Leibel’s subjects adapted by 15 percent, not 50. The people in Lichtman’s underreporting study, described in the logging guide, believed they had slow metabolisms and turned out to be eating twice what they thought. When the log and the scale disagree with the plan, the log is usually where the answer is.

Neato’s guides are general wellness information built from the published research cited above. They are not a diagnosis, treatment, or medical advice, and they don’t know your history. For decisions about your health, talk to your doctor.