The 3,500-Calorie-Per-Pound Rule Is Wrong: The Real Math

“A pound of fat is 3,500 calories: cut 500 calories a day and you’ll lose a pound a week.” That line, or something close to it, has appeared on more diet websites than anyone could count, for decades. The arithmetic is simple — and that simplicity is exactly what makes it misleading. Kevin Hall, whose model underlies the reference framework published in The Lancet in 2011, documented the problem in a paper whose title leaves little room for doubt: “Why is the 3500 kcal per pound weight loss rule wrong?” This page walks through why, with the published numbers that make the case — not a hunch, a measured gap.

Where the number actually comes from

The rule is meant to approximate the energy in a pound of body tissue lost. Metric-system countries round the same idea differently, as 7,700 kcal per kilogram — 3,500 ÷ 0.453592 works out to roughly 7,716 kcal/kg, rounded down to 7,700 by convention. It is not a different rule depending on the country; it is the same conversion, expressed in two units.

That figure did not come from nowhere: it sits near the published energy density of pure body fat, around 9,440 kcal per kilogram. The gap between those two numbers — and the fact that a pound of weight lost is never 100% pure fat — is not, however, the rule’s real problem. The real problem lies elsewhere, and it matters more.

The hidden assumption that breaks everything: expenditure that never moves

Dividing a deficit by 3,500 (or 7,700) implicitly assumes that every pound lost costs the same effort — that the body’s energy expenditure stays identical from day one of a diet to day one thousand, no matter how much weight has already come off in between. That assumption matches no published physiological model.

In reality, expenditure keeps falling for two distinct reasons, covered in our article on metabolic adaptation: a mechanical effect — a lighter body is cheaper to maintain and cheaper to move — and a separately measured adaptive effect, adaptive thermogenesis, which gives back roughly 14% of any sustained change in intake. Add the thermic cost of digestion itself (roughly 10% of what is eaten), and nearly a quarter of any intake cut is handed back as lower expenditure before a single gram of fat is lost. The 3,500/7,700 rule ignores both mechanisms entirely: it applies a fixed exchange rate to an expenditure that, in reality, never stops changing.

The gap in numbers: what the published model predicts, versus the rule

Hall and colleagues apply this exact comparison to a published scenario: a sedentary 100 kg man sustaining a 480 kcal/day cut (about 2 MJ/day) — precisely the scenario in their Figure 2A.

That 78% gap comes from reproducing the published equations of the model directly. The study authors themselves, in their original paper, describe an even larger gap for this same scenario once the body-water compartment is included: they describe the popular rule as predicting “about 100% greater weight loss” than their own model — close to double. Both figures point the same way: the rule is not slightly optimistic, it overstates real weight loss by something close to a factor of two over a year.

Why the gap widens over time

The rule is, by construction, a straight line: the same number of pounds subtracted every month, indefinitely, as long as the stated deficit stays the same. Expenditure that never falls, no matter how much weight has already been lost, matches no published physiological data — that missing assumption is exactly why the rule never predicts a plateau, while our article on the weight-loss plateau shows the real trajectory approaching one mathematically from day one of the deficit. The longer the time horizon, the further the rule’s straight line drifts from the real curve, which keeps slowing continuously.

What the rule is still useful for

It is not a rule to throw out entirely. Over a modest deficit and a short horizon — weeks, not months — expenditure has not yet had time to fall by much, so the gap between the straight line and the real curve stays small. It is precisely on large changes and long horizons — what most people mean when they type “how long to lose 10kg” — that the gap grows large, as our article on how long weight loss really takes covers in detail.

To reason without this rule, the best alternative remains a model that lets expenditure move with weight lost — which is exactly what our own calorie calculator does for your starting expenditure, and what the weight-loss timeline simulator in development on this site will do more fully: weight-loss timeline tool.

In short: 3,500 calories per pound (or 7,700 per kilogram) is not a bad number in itself — it is a bad way of using it, as a fixed exchange rate applied to an expenditure that, in published reality, never stops moving.