Yes. Energy cannot appear from nowhere, and a body that takes in less than it spends has to make up the difference from storage. The physics is not in dispute.
What people are usually arguing about is something else: whether the equation is useful in practice. And there the answer gets more interesting, because both sides of it move around far more than the phrase suggests.
What the model actually claims
Eat more energy than you spend and you store the surplus. Spend more than you eat and you draw on reserves. That is it.
The mistake is reading this as a promise that the two sides are fixed, knowable numbers you can control independently. They are not. Change one and the other changes in response, which is where most of the frustration comes from.
Calories in is harder to pin down than it looks
The figure you log is not the figure your body receives.
Food labels carry a tolerance and are built from averages, and rounding on small items adds up faster than people expect. Our piece on why label counts are often wrong goes through the mechanics.
Then there is absorption. Fibre passes through you. Protein costs energy to break down, so you keep less of it than the label implies. Whole nuts give up less energy than nut butter does. Cooking changes the accessible energy in food.
And underneath all that is estimation error. People consistently under-report what they eat, and not out of dishonesty. Oil in the pan, a few chips off someone else's plate, the milk in four coffees.
Calories out moves when you push on it
This is the half that surprises people. Your expenditure is not a constant you can look up.
Cut calories and several things happen at once. Your body is smaller, so it costs less to run. Your metabolism drifts down a little beyond what the size change explains. And critically, you move less without deciding to: less pacing, less fidgeting, more sitting still. That unconscious activity varies enormously between people and it is the first thing to fall when energy is scarce.
Hormones are part of the same picture. Leptin falls as fat mass falls, which turns up hunger and turns down spontaneous movement. Ghrelin pushes appetite up. Thyroid output can shift. None of this breaks the equation. It changes the numbers inside it, in the direction that makes the deficit smaller than you planned.
So are all calories the same
For the arithmetic, yes. For your behaviour, absolutely not, and that is the distinction worth holding.
200 calories of chicken breast and 200 calories of biscuits do the same thing to the ledger. They do very different things to how full you feel an hour later, how much you want to eat next, and how much of your weight loss comes from muscle rather than fat.
Protein is the clearest case: it is the most filling macronutrient and the most expensive to digest, so it wins twice over. Fibre and water content do similar work through volume.
Which is why food quality matters even though calories are what determine the direction. Quality decides whether the calorie target is something you can actually hit for six months.
Where calorie counting genuinely falls short
It is imprecise, and it presents itself as precise. That mismatch causes real harm when people trust the number over the scale.
It says nothing about nutrition. You can hit a target on junk and be malnourished.
It ignores why people eat, which is often stress or boredom rather than hunger, and no arithmetic addresses that.
And for some people, tracking every mouthful turns into an unhealthy preoccupation. That is a genuine cost, not a soft one.
How to use it well
Use the equation as a compass, not a scale. Estimate your maintenance calories, eat there for two or three weeks, and let the trend on the scale correct the estimate. Your own data beats any formula.
Track consistently rather than perfectly. A steady error is harmless because it cancels out of the comparison.
Prioritise protein and whole foods, so that hitting the number takes less willpower.
And accept that the target moves as you do. That is not the model failing. That is the model working exactly as it should, with inputs that refuse to sit still.