How many calories should I eat to lose weight?

Every calorie calculator hands you a confident number. Eat 1,600 a day, it says, and the fat will melt off. Then you eat 1,600 for three weeks, the scale doesn't move, and you assume you're broken.

You're not. The number was an estimate, and estimates need correcting. Here is how to find the intake that works for your body rather than an average one.

Start with a formula, not a guess

Your body burns energy just staying alive: keeping your heart going, your brain running, your temperature steady. That baseline is your resting metabolic rate, and the most reliable way to estimate it without a lab is the Mifflin-St Jeor equation, which held up best across a systematic review of the common formulas.¹²

Multiply that resting rate by how active you are and you get your maintenance calories, the amount that keeps your weight steady. To lose fat, you eat below it.

How big a gap?

A pound of body fat stores roughly 3,500 calories; a kilogram, about 7,700. So a daily shortfall of 500 calories points to roughly half a kilo a week on paper. In practice it's messier, but the principle holds: a moderate, steady gap beats a savage one.

Aim to lose 0.5 to 1 percent of your body weight a week. When researchers compared fast and slow weight loss in trained athletes, the slower group held on to more muscle and strength.³ Crash lower and you mostly buy yourself hunger, fatigue, and muscle loss you'll regret later.

Why the calculator's number is only a starting point

Mifflin-St Jeor lands within about 10 percent for most people, but 10 percent of a 2,200-calorie maintenance is over 200 calories. That's the difference between losing steadily and not moving at all. Your real number depends on muscle mass, how much you fidget, past dieting, and how honestly you track.

So treat the calculator as a hypothesis. The scale tests it.

The two-week rule

Pick your starting intake and hold it for two weeks. Weigh yourself most mornings, after the toilet, before food, and take the weekly average rather than trusting any single day. Daily weight swings of a kilo or two are mostly water and food in transit, not fat.

After two weeks, read the trend:

• Losing 0.5–1% a week? Leave it alone. You found your number.
• Nothing moved? Cut about 150 calories a day, or add 2,000 steps, and give it another two weeks.
• Dropping too fast (more than 1% a week, and you feel wrecked)? Add 100–150 calories back. Faster isn't better.

This loop is the whole game. Most people don't need a better calculator; they need to actually run the experiment for a fortnight instead of changing plans every three days.

A worked example

Take a 34-year-old woman, 82 kg, 165 cm, on her feet part of the day. Mifflin-St Jeor puts her resting rate near 1,520 calories and maintenance near 2,090. A steady deficit lands her around 1,570 a day.

She eats that for two weeks and her average weight barely shifts. Rather than despair, she trims to about 1,420 and adds an evening walk. The next fortnight, the average drops 0.7 kg a week. That corrected number, not the first one, is her real target.

The floors you shouldn't go under

Very low intakes are hard to sustain and make it tough to get enough protein and micronutrients. As a general guide for healthy adults, women shouldn't drop below about 1,200 calories a day and men below about 1,500 without medical supervision. If your maths pushes you under that, the answer is more movement, not less food.

If you're pregnant or breastfeeding, under 18, managing a medical condition, or have a history of disordered eating, skip the do-it-yourself calorie cutting and talk to a doctor or registered dietitian first.

What actually makes the deficit stick

The number on the label matters, but so does what you eat to reach it. In a tightly controlled inpatient trial, people eating an ultra-processed diet put away about 500 calories a day more than when they ate minimally processed food, without being told to.⁴ Whole foods and plenty of protein keep you fuller on fewer calories, which makes the whole thing feel less like willpower and more like coasting.

Do you eat back exercise calories?

Short answer: mostly no. Fitness trackers are notoriously generous with "calories burned", often by 20 to 40 percent, and eating back an inflated number is a common reason people stall. If you set your intake using an activity multiplier, as most calculators do, your everyday movement is already baked in. Eating extra on top double-counts it. Leave your target where it is and let any bonus movement help the deficit along.

Three mistakes that hide your real number

• Weekend amnesia. Five tight days and two loose ones can erase a whole week's deficit. It's the weekly total that counts, not the weekday average.
• Untracked extras. Oil in the pan, milk in your coffee, a few chips off someone's plate. These are real calories, and they're exactly where a 1,600-calorie plan quietly becomes 2,000.
• Liquid calories. Juice, sweet coffee, alcohol. They don't fill you up but they count in full.

None of this means you have to weigh food forever. It means that when your number "isn't working", the number usually is; the tracking is what slipped.

Find your starting number, hold it for two weeks, read the trend, adjust once. Repeat. That's how you eat the right amount for your body instead of an average stranger's.